JD» ,
Jannfi

. 3:...
O ..Q ‘3

z .
5.33%qu
y 3. I. P» l. A , .. .. n
33.9.53 n V fir?
. . . 3r
y. «m btmhnufi .
an... lung. mi

.323183‘3
xv. 51‘ I

. hr in”!
5d“... E
1 angina!
.I... .
.p.
2 Q
.oh. {and
Jana-mm.

. § .
1550.51 I. £
5...? 93.3.5...»

1.3....
31);).

11:! >6: z...
[.1:!.?:...{ .
:55)! ’lxi.‘:.

i). .9. 1. .tv
I\a

’91.. Mug.
.1 .
y. ) '1‘)

3.. ‘
rp‘. 2
t I
a,\
o, l

I

 

THESIS

                     

’HIIIUNHHIHIHIJWI’HIHIIIHllllJUIllll

31293 01417 2252

LIBRARY
Michigan State
University

This is to certify that the

dissertation entitled
Evidence for an Association Between

Splicing Components: Galectin-3
and Polypeptide“) of snRNA

presented by

SING-YUAN WANG

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in Biochemistry

W7wfl

Major professor

Date 575 HI

 

MS U is an Affirmatiw Action/Equal Opportunity Institution 0-12771

PLACE IN RETURN BOX to remove thle checkout from your record.
TO AVOID FINES return on or More dete due.

DATE DUE DATE DUE ‘ DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

[—T—l l

MSU le An Afflnnetlve Action/Emu Opportunity lnetltution
more» i

EVIDENCE FOR AN ASSOCIATION BETWEEN SPLICING COMPONENTS:
GALECTIN-3 AND POLYPEPTIDE(S) OF snRNP

By

Sung Yuan Wang

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Biochemistry

l 995

ABSTRACT
EVIDENCE FOR AN ASSOCIATION BETWEEN SPLICING COMPONENTS:
GALECTIN-3 AND POLYPEPTIDE(S) OF snRNP
By

Sung Yuan Wang

Galectin-3 is a galactose/lactose-binding protein found in the nucleus and cytoplasm
of a variety of cell types. Using a cell-free assay for pre-mRNA splicing, previous studies
have demonstrated that galectin-3 is a required factor for spliceosome formation and
splicing activity. Depletion of the lectin from a splicing competent extract resulted in a loss
of both activities, which could be restored upon reconstitution with purified recombinant

galectin-B.

The goal of this thesis was to demonstrate an association of galectin-3 with
components of the splicing machinery; specifically, the Sm epitopes found on polypeptides
of the small nuclear ribonucleoproteins (snRNPs). This was accomplished at two levels: (a)
at the level of single cells, the immunofluorescence patterns of galectin-3 were Shown to be
similar to that of the snRNP Sm polypeptides under a variety of conditions and
colocalization could be demonstrated in double labeling experiments; and (b) at the level of
protein-protein interactions, it was shown that when anti-Mac 2, a monoclonal antibody
specific for galectin-B, immunoprecipitates galectin-3 from nuclear extracts of HeLa cells

and mouse 3T3 fibroblasts, the Sm B polypeptides are coprecipitated.

The immunofluorescence experiments were carried out with mouse 3T3 fibroblasts.
The staining patterns for galectin-3 and snRNP Sm polypeptides were monitored in parallel
and were found to behave in the same fashion: (a) both yielded diffuse distribution of
fluorescence covering the entire nucleus, except for several nucleoli, in fixed and
permeabilized cells; (b) using conditions of penneabilization and extraction that preserved a
RNP—containing nuclear matrix, both yielded speckled staining patterns; (c) the fluorescence
staining was lost for both galectin-3 and snRNP Sm polypeptides when permeabilized
nuclear residues were treated with ribonuclease A, whereas neither staining was sensitive to
deoxyribonuclease Itreatment; and ((1) during the various stages of mitosis, both galectin-3
and snRNP Sm polypeptides were found to be excluded from the area of chromosomal

localization, condensation and separation.

Nuclear splicing extracts were subjected to immunoprecipitation with anti-Mac 2.
Immunoblotting of the anti-Mac 2 precipitate with polyclonal rabbit anti-galectin-3 revealed
the presence of the lectin. Immunoblotting of the same anti-Mac 2 precipitate with an
autoimmune serum reactive with the snRNP B polypeptides showed that a fraction of the
Sm B polypeptides in the nuclear extract was coprecipitated with the lectin by the
monoclonal antibody. The specificity of this coimmunoprecipitation was analyzed by
comparing the polypeptides precipitated by anti-Mac 2 versus those found in the
corresponding precipitate fractions obtained with anti-Sm (positive control) and with anti-
transferrin receptor, an isotype-matched monoclonal antibody (negative control). The
coprecipitation of Sm B with galectin-3 by anti-Mac 2 was dependent on the presence of the

lectin. Nuclear extracts depleted of galectin-3 by prior adsorption on a lactose affinity resin

failed to yield Sm B in the anti-Mac 2 precipitate. The coprecipitation was not affected by
the addition of lactose; nor was it perturbed by prior treatment of the nuclear extract with
ribonuclease. These results suggest that at least a fraction of the Sm B polypeptides in the

splicing extract can interact with galectin-3 through protein-protein interactions.

In the course of these studies on the polypeptides precipitated by anti-Mac 2, we
also analyzed the composition and identities of polypeptides bound to the affinity resin
lactose-agarose under conditions used to deplete the nuclear extract of splicing activity. We
found that, in addition to galectin-3, the bound fraction of the lactose-agarose contained at
least three other polypeptides. Of particular interest was the identification of one of the
other polypeptides to be galectin-1 (the prototype member of the galectin family), whose
presence in the nucleus of HeLa cells was confirmed by confocal fluorescence microscopy.
These findings provide an explanation for previous observations that while lactose-agarose
can deplete nuclear extracts of splicing activity, anti-Mac 2 precipitation failed to yield the
same effect. Together with the observation that galectin-1, as well as galectin-3, can alone
reconstitute splicing activity in a lactose-agarose depleted extract, these results suggest that

the activities of galectin-1 and galectin-3 are redundant in the cell nucleus.

To Yi Yu and Catherine
and
To My Parents

for their love, support and faith in me

 

 

{would iil
muggemem d:
science and life. i
encouragemenl. S
in: scientific an.

love a s;
lit he provided ‘
when of my g
hmmmm;

Finally. I
colleagues Anan
ET Amoys. Sh.
Emmml
ltelugan State.

ACKNOWLEDGMENT

I would like to express my sincere appreciation to Dr. John L. Wang for his advice and
encouragement during my graduate study in his laboratory. I thank him for teaching me about
science and life. I would also like to extend my gratitude to Mrs. Patty Voss for her
encouragement, support and tolerance. I thank her for providing a conductive environment for
both scientific and social growth.

I owe a special note of thanks to Dr. Ronald Patterson for his valuable advice and support
that he provided me throughout the years. I would also like to express my appreciation to
members of my guidance committee, Dr. Richard Anderson, Dr. Zachary Burton and Dr. Arnold
Revzin for their advice and criticisms.

Finally, I wish to thank Dr. Melvin Schindler, Dr. Siu-Cheong Ho and my friends and
colleagues Anandita, Mark Kadrofske, Yeou Guang Tsay, Dr. Jamie Laing, Dr. Neera Agrwal,
Eric Amoys, Sharon Grabski who shared both the ups and downs and who made my time in the
lab very special. Knowing them has been a privilege and worth by itself the time I have spent at
Michigan State.

vi

 

LIST Oil
LlSl OF I
Clltl’lIl
LllTRj
A l C

BIC
CIG

Slim

MATE
Cell

TABLE OF CONTENTS

 

 

 

 

 

 

 

 

 

 

 

 

Page

LIST OF TABLES x
LIST OF FIGURES xi
CHAPTER I: Literature Review ' 1
LITERATURE REVIEW ON GALECTINS 2
A) Carbohydrate Binding Proteins of Animal Cells 2

B) C-Type Lectins 3

C) Galectins 5

1) Structure of Galectins 5

2) Galectin-3: Chimera of Two Domains lO

3) Carbohydrate-Binding Properties of Galectins l2

4) Subcellular Localization of Galectins l4

5) Functions of Galectins l7

LITERATURE REVIEW ON snRNP PARTICLES AND PRE-mRNA SPLICING - l9

 

 

 

 

A) Structure and composition of snRNP 19
B) Nuclear Localization of the snRNP 29
C) Cytoplasmic Assembly and Nuclear Import of the snRN P 31
D) Non-snRNP Splicing Factors 34
LITERATURE CITED 38

 

CHAPTER II: Similarities in the Nuclear Matrix Localization of Galectin-3 and
Small Nuclear Ribonucleoproteins (snRN P): Evidence from a

 

 

 

 

Comparative Immunofluorescence Analysis 48

SUMMARY 49
INTRODUCTION 5 1
MATERIALS AND METHODS 53
Cell Cultures - 53

 

Antibodies 53

 

vii

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Irnmunofluorescence Microscopy 53
EffeCts of Perrneabilization, Extraction and Nuclease Treatment 54
V RESULTS 56
Irnmunofluorescence Analysis of Galectin-3 S6
Immunofluorescence of Perrneabilized Cells after Salt Extraction or Nuclease
Digestion 59
Immlmofluorescence Analysis as a Function of the Cell Cycle 65
DISCUSSION 76
REFERENCES 79
CHAPTER III: Coimmunoprecipitation of Sm Polypeptides of snRN P with
Galectin-3 by a Monoclonal Antibodies Directed against the
Lectin 82
SUMMARY 83
INTRODUCTION 85
MATERIALS AND METHODS 87
Cell Cultures 87
Antibodies and Affinity Columns 87
Buffers 88
Nuclear Extract Preparation 89
Immunoprecipitation and Lac-depletion of Nuclear Extract 90
Analysis of RNA in Bound Fraction of Immunoprecipitation and Lac-
Adsorption 92
Cesium Sulfate Gradient Fractionation and Immunoprecipitation 93
RESULTS 95
Immunoprecipitation of Nuclear Extract of 3T3 Cells by Autoimmune Anti-Sm
and by Monoclonal Anti-Mac 2 95
Immunoprecipitation of HeLa Cell Splicing Extracts 102
Effects of Galectin-Depletion on Coimmunoprecipitation 105
Coimmunoprecipitation Experiments in the presence of Lactose and at High
Ionic Strength 106

 

viii

DEC

flilPl

Sl'l

ill

 

 

CH

 

Analysis of the Anti-Mac-2 Precipitate for RNA 1 14

 

 

 

 

 

 

 

 

 

 

 

 

 

 

DISCUSSION 121
REFERENCES 126
CHAPTER IV: Nuclear and Cytoplasmic Localization of Galectins: Evidence
from Laser Scanning Confocal Microscopy 128
SUMMARY -- 129
INTRODUCTION 130
MATERIALS AND METHODS 133
Antibodies and Affinity Columns 133
Lac-agarose Affinity Adsorption 133
Immunofluorescence Staining and Laser Scanning Confocal Microscopy --------- 134
RESULTS 136
Polypeptides Bound by Lac-agarose 136
Nuclear and Cytoplasmic Localization of Galectin-l: Evidence from Laser
Scanning Confocal Microscopy 139
DISCUSSION 144
REFERENCES 147
CHAPTER V: Concluding Statement 149

ix

 

 

 

 

 

 

Cllll’ll

Tile 1. (

Table I. l

Mei 3

Table 6. F

LIST OF TABLES

Page

CHAPTER I:

Table 1. Comparison of C-type Lectins and Galectins ....................................................... 4
Table 2. Members of the Galectin Family .......................................................................... 7
Table 3. Previous Names for Galectin-3 ........................................................................... 11
Table 4. Major Mammalian snRNAs ................................................................................ 21
Table 5. Major mammalian snRNPs ................................................................................. 25
Table 6. Proteins of Mammalian snRNPs ......................................................................... 26

LIST OF FIGURES

Page

CHAPTER I:
Figure 1: Schematic diagram illustrating the domain context and organization of

the various galectins ............................................................................................ 8
Figure 2: Schematic diagram illustrating the two major steps in pre-mRN A

splicing. ............................................................................................................. 20
Figure 3: Schematic diagram delineating the number and order of assembly of the

intermediates in spliceosome formation. .......................................................... 23
Figure 4: Schematic diagram illustrating the nuclearcytoplasmic shuttle in the

assembly of the RNA and polypeptide components of U1 and U2

snRNPs .............................................................................................................. 33
Figure 5: Schematic diagram summarizing the context and organization of specific

domains in SR family of proteins. .................................................................... 35
CHAPTER II:

Figure l: Immunofluorescence staining of 3T3 cells after fixation with

paraformaldehyde (4%) and permeabilization with Triton X-100 (0.5%). ....... 57
Figure 2: Comparison of the immunofluorescence staining pattern of 3T3 cells. ............ 60
Figure 3: Effect of enzyme treatments on the intranuclear staining pattern for

galectin-3 and the Sm antigen of snRNPs ......................................................... 63
Figure 4: Double immunofluorescence staining for galectin-3 and Sm antigen of

snRNPs in the same cells. ................................................................................. 66
Figure 5: Immunofluorescence staining patterns of galectin-3 in 3T3 cells as a

function of time following serum stimulation. ................................................. 69
Figure 6: Analysis of the immunofluorescence staining pattern of galectin-3 as 3T3

Cells undergo mitosis ......................................................................................... 71
Figure 7: Analysis of the immunofluorescence staining pattern of the Sm antigens

as 3T3 cells undergo mitosis. ............................................................................ 74
CHAPTER III:

xi

 

 

Figure 1: Immunoprecipitation of nuclear extracts of 3T3 cells by autoimmune
anti-Sm and by monoclonal anti-Mac 2. ........................................................... 96

Figure 2: Profiles of the density and immunoprecipitation patterns of the individual
fractions derived from cesium sulfate gradient sedimentation of
nucleoplasm. ..................................................................................................... 99

Figure 3: Immunoprecipitation of HeLa cell splicing extracts by autoimmune anti-
Sm and by monoclonal anti-Mac 2. ................................................................ 103

Figure 4: The effect of galectin-3 depletion on the immunoprecipitation of Sm B
by anti-Mac 2. ................................................................................................. 107

Figure 5: Immunoprecipitation in the presence of lactose (Lac) of HeLa cell
splicing extract by autoimmune anti-Sm and by monoclonal anti-Mac 2 ....... 109

Figure 6: Immunoprecipitation under high ionic strength of HeLa cell splicing
extract by autoimmune anti-Sm and by monoclonal anti-Mac 2. ................... 1 12

Figure 7: Analysis of the RNA components when HeLa cell splicing extracts are
immunoprecipitated by autoimmune anti—Sm and by monoclonal anti-

Mac 2. ............................................................................................................. l 15
Figure 8: The effect of RNase treatment on the immunoprecipitation of Sm B by

anti-Mac 2. ...................................................................................................... l 19
CHAPTER IV:

Figure 1: The composition and identities of polypeptides in the bound fraction of
Lac-agarose adsorption. ................................................................................... 137

Figure 2: The subcellular localization of galectin-1 as revealed by laser scanning
confocal fluorescence microscopy. ................................................................. 141

xii

CHAPTER I

Literature Review

LITERATURE REVIEW ON GALECTINS

A) Carbohydrate Binding Proteins of Animal Cells

There are three main classes of proteins involved in protein-carbohydrate
interactions (1): a) enzymes that use carbohydrates and glycoconjugates as substrates,
such as glycosidases and glycosyl transferases; b) sugar-binding antibodies; and c) the
carbohydrate binding protein (CBP) group. Carbohydrate binding proteins are defined as
non-enzymatic and non-immune proteins which can selectively bind specific
carbohydrate structures and are now referred to as lectins. Ricin, a toxic protein, was
documented as the first lectin from a plant source (2). Since the initial identification,
lectins have been found in a variety of organisms, in various tissues and cell types. More
recently, much attention has been paid to animal lectins. Since carbohydrate structures
such as glycoproteins and glycosaminoglycans have been found in the cytoplasm and the
nucleus (3), there has been, in turn, greater interest in the intracellular localization of

lectins.

Animal lectins have been classified based on the nature of their carbohydrate
ligands, their involvement in biological processes, their subcellular localization, and their
dependence on divalent cations. However, since the primary structures of many lectins
have been determined, the shared sequence characteristic is now one of the most useful
way to classify lectins. Based on their amino acid sequences, especially those of
characteristic carbohydrate-recognition domains (CRD), most animal lectins can be

Classified into two major distinct families (4-6). One is the family of C-type lectins

which is found extracellularly or associated with plasma membrane and which requires
calcium ions for their sugar-binding activity. The other one is the galectin family (7),
consisting of soluble, metal-independent, B-galactoside-binding proteins formerly known

as S-type or S-Lac lectins (Table 1).

B) C-Tyng Lectins

The C-type CRDs derive their name from the fact that they require calcium ions
for sugar binding activity. The carbohydrates that their CRDs recognize are diverse,
including simple carbohydrates such as galactose, N-acetylglucosamine, and complex
glycoconjugates such as sialyl Lewis antigens (5). Although calcium ions are required for
CRD binding to its ligand, the other domain(s) of C-type lectins is involved in many of
the functions of the different proteins. All of the C-type lectins have primary structures
typical of extracellular proteins or transmembrane proteins in that they exhibit a signal
sequence for entry into the endoplasmic reticulum. C-type animal lectins are found in
serum, extracellular matrix and membranes; thus they can be divided into insoluble and

soluble groups (6).

The insoluble C-type lectin group includes those transmembrane receptors, such
as a type-I receptor whose N-terminal region is extracellular and C-terminal region is
cytoplasmic. This group includes selectins, mannose receptor and thrombomodulin. The
selectins mediate the initial phase of adhesion between leukocytes and endothelia in a

weak transient adhesion. A larger number of membrane-integrated C-type lectins are

Table 1. Comparison of C-type Lectins and Galectins

   

 

 

 

 

 

 

 

Pro ierties, 3 .. . r. r a ” ..   i. _-_; Galectins  .
Calcium requirement Yes No
Solubility Variable Soluble
Cys residues Disulfides Free thiols
Cellular localization Extracellular Intracellular
Transmembrane Extracellular
Carbohydrate specificity Various B-galactosides

 

 

 

 

Ill/:1"

categorized into type-H receptors such as hepatocyte receptor and macrophage receptor.
A type-II receptor has a cytoplasmic N-terminal region and an extracellular C-terminal
domain. Some of the functions ascribed to these receptors include the removal of
glycoproteins by endocytosis and phagocytosis. Soluble C-type lectins include collagen-
like proteins such as the mannose-binding protein. This type of lectin is composed of an
N-terminal collagenous domain and a C-terminal CRD. Other soluble C-type lectins

include proteoglycan core proteins, snake venom lectins, etc.

C) alectins
1) Structure of Galectins

S-type lectins which possess affinity for B-galactosides and show a significant
sequence similarity in the carbohydrate-binding site are now renamed as galectins (7).
The general designation of the genes encoding galectins is LGALS (lectin, galactoside-
binding, soluble), and gene numbering is being kept consistent with the numbering of the
proteins; so that LGALSI encodes galectin-1, and LGALSZ encodes galectin-2, etc (7).
The CRD of the galectin family is clearly distinguishable from the corresponding CRD of
the calcium-dependent C-type lectin (4, 8). Unlike C-type lectins, galectins have been
found both inside and outside of cells. They do not depend on cations for carbohydrate

binding activity and are isolated usually as soluble proteins.

At present, seven galectins in this family have been studied (Table 2). Galectin-l

is isolated as a homodimer with monomer molecular weight of ~ 14 kDa. It is abundant

in smooth and skeletal muscle, but is also found in many other cell types. Galectin-Z,
originally found in human hepatoma, is isolated as a homodimer (9). Galectin-3,
previously known as CBP35 was first isolated from mouse 3T3 fibroblasts (10, 11). It
consists of a single polypeptide of ~ 33 kDa. Galectin-4 was found as an abundant rat
intestinal 36 kDa lectin ( 12). It is homologous to a 32 kDa B-galactoside-binding protein
from Caenorhabditis elegans ( l3). Galectin-S was originally isolated from rat lung as a
monomer of ~ 18 kDa (14). Galectin-7 is a 15 kDa monomeric protein found in human
keratinocytes (15). Galectin-8 was originally cloned from rat liver (16) with a molecular
weight of 34 kDa. Unlike galectin-4, which is abundant in the intestine galectin-8 was
found in liver, kidney, cardiac muscle, lung and brain. In addition, there are invertebrate
lectins whose CRDs fit the consensus sequence of the galectin CRDs but whose
subfamily designations has not been decided. These include the sponge (17) and

nematode (l3) S-type lectins.

As implied by their nomenclature, the structures of all the galectins have at least
one carbohydrate recognition domain (CRD), with conserved amino acid sequence
between members of the family and homologues of the same member across different
species (Figure 1). Thus, galectin-l galectin-2, galectin-5 and galectin-7 each has one
CRD, representative of the prototype of the family. Galectin—4 and galectin-8 represent
tandem-repeat type with two CRDs in the structure. Each of the CRDs in galectin-4
shows about 25% sequence identity to vertebrate galectin-1, and has carbohydrate-

binding activity. Sequence analysis also showed that galectin-8 exhibits about 34%

Table 2. Members of the Galectin Family

 

 

 

 

 

 

 

 

 

 

 

 

:‘7 ’GaIeCtin H " Subunit StruCture POIypeptideMr Source ChrOmOsome i

l Homodimer 14,000 Various species, Human 22
tissues/cell types

2 Homodimer 14,000 Human Human 22 I

3 Monomer 33,000 Various species, Human 13 l
tissues/cell types

4 Monomer 36,000 Rat - 1

(C. elegan)
5 Monomer 18,000 Rat Mouse 11
Mouse 4

6 ? ? ?

7 Monomer 15,000 Human Human 19 1

8 Monomer 34,000 Rat -

 

Figure 1: Schematic diagram illustrating the domain context and
organization of the various galectins.
The carbohydrate recognition domain(s) (CRD) in each galectin is highlighted
by the shaded rectangle. Amino acid residues that are conserved among the
various galectins are highlighted in the top. The sequence of a nine residue
motif that is tandemly repeated (n times, depending on species) giving rise to a
proline- and glycine-rich domain in galectin-3 is indicated.

 

 

 

 

 

' e
e o 0.
....................
...............
0.... ..............

a 1 135:} 59' 1' F 7. L I; v ’
. I 1 Galectin-1
‘ ' 0‘ , . . . . _ .3, .
0

xi f’fif Q {Ie*’ ! ‘”,sz.efx }
4: 140.72.! >3." .me .o;.-.t . :-t-.‘

 

 

 

 

Galectin-Z

 

 

... . . .‘. (y. . .1. ._. . ..............................

 

 

 

 

 

(pcAvaxxx). Galectin-s

 

 

7 VIVV v v’ve ‘vvv ,I. I . I, v‘ .
.'.' ~ ,‘,' ' . . i - . .i . ' '.'~‘ ..'.- . 223.7 v'-_ 2‘77 .‘..’ 3.. ,7,‘ .I v. '4
I ' ‘ i A ‘ A ' '. " :5..- l. .......................
37...: ‘.Va'-.' 7'1 l'.-'.'>' ~~ v’.‘ '4v-ivn. ..... - ...... ‘reel‘g .--‘ ...... A
.......... We

 

.xxg.xs '~.q';_ C.n““"“ui§‘ .
-------------------------- V... e .'-'.'-‘e'u‘e n‘e'e‘
.......... .................
Dine-enve'v-v' ‘ugfletinjuj

 

zaiééséaizizéaé;E;;;?;i;i:z:i:i::;i§;:;:.i.:1:;..:i::si;é: l Galectin-7

 

 

 

 

..............................................

J GalectIn-B

...............................................
.....................

 

 

10

sequence identity to galectin-4. Galectin-6 has been mentioned (7), but its

characterization has yet to be published.

2) Galectin-3: Chimera of Two Domains

The galectin—3 group includes proteins originally isolated from various sources
which exhibit B-galactoside binding activity and which have been designated different
names (Table 3). Genomic Southern blot analysis suggested that there is only one gene
coding for the polypeptides of the galectin-3 group. The C-domain of galectin-3
polypeptide contains both hydrophilic and hydrophobic regions, as is characteristic of
many globular proteins. This region is about 35% identical with several members of the
galectin-1 group, and 14 amino acid residues in this region are highly conserved among
galectin-3 and other members of galectin family (Figure 1). Consequently, the C-domain
of galectin-3 is assumed to harbor the CRD. In contrast, the N-domain of galectin-3
polypeptide contains neither a highly hydrophilic nor a hydrophobic region. This domain
includes a stretch of eight contiguous 9-residue repeat units having the sequence Pro-Gly-
Ala-Tyr-Pro-Gly followed by three other amino acid residues. As a result, this
stretch of the sequence is characterized by a high proportion of proline and glycine

residues (18). Thus, galectin-3 is classified as a chimera—type galectin .

Study on the N-terminal domain of galectin-3 showed that it exhibits no apparent
B-galactoside-sugar binding activity; in contrast, the C-terminal domain is sufficient for
two independent transitions representing the thermal denaturations of the C-domain

sugar-binding (19). Differential scanning calorimetry analysis of galectin-3 has yielded

Table 3. Previous Names for Galectin-3

 

 

Rat, Human,

ll

. .9} ‘ “.35...

 

Lung, Brain, Kidney

 

Non-classical

 

 

 

 

 

Dog secretory pathway
88? 31,000 Rat, Human Basophilic leukemia cell IgE binding
actixity
Mac-2 32,000 Mouse, Human Macrophage Cell surface
antigen
L-34 34,000 Mouse Fibrosarcoma Cell surface lectin
Melanoma metastatic marker
CBP35 35,000 Mouse, Human Lung, Fibroblast Intracellular lectin
LBP 35,000 Mouse Macrophage Laminin binding
activit

 

 

 

 

 

_-

 

It;

"—3

5m

12

(~ 40 0C) and C-domain (~ 55 0C). The binding of lactose to the intact galectin-3
polypeptide and to the C-domain half molecule shifted the 55 °C transition to ~ 65 °C.
These results suggest that ligand binding by galectin-3 is accompanied by a
conformational change that significantly stabilizes the polypeptide against thermal

denaturation.

Analysis of the isoelectric point (pI) of E. coli expressed recombinant galectin-3
showed that the pI value of unmodified polypeptide is 8.7, as determined by calculation
from the deduced amino acid sequence and by isoelectric focusing gel electrophoresis
analysis (20). However, when mouse 3T3 cell extract was subjected to two-dimensional
gel electrophoresis and immunoblotting analysis for galectin-3, two spots corresponding
to pI values of 8.7 and 8.2 were observed. The pI 8.2 species represents a post-

translational modification of the pI 8.7 polypeptide by the addition of a single phosphate

group.
3) Carbohydrate-Binding Promrties of Galectins

All mammalian galectins recognize the same structural determinant on lactose and
related B-galactosides (10, 12, 21-24); however, they share a higher affinity of binding
for lactose than that for galactose (25—27). The affinity of galectin-3 for lactose binding
was found to be 50-100 times greater than that for galactose. The critical positions of the
lactose molecule include the hydroxyls at positions 4 and 6 of Gal and position 3 of Glc,

since substitutions at any of these positions resulted in reduced binding activity.

13

Modification of lactose by incorporating an acetamido group at position 2 of glucose

moiety to yield N-acetyllactosamine enhanced the binding of galectins.

The sequences responsible for carbohydrate binding are mainly encoded by three
exons (28-30), for example exons II, III and IV for galectin-l and IV, V and V1 for
galectin-3. Most of the residues directly interacting with the carbohydrate ligand are
conserved among the galectins and encoded by a single exon. Mutation of some of these
residues in exon HI of galectin-1 resulted in impaired carbohydrate binding activity (31,
32). Site-directed mutagenesis on the CRD of galectin-1 showed that the conserved
hydrophilic residues, such as His44, Asn46, Arg48, Asn61, Glu71, and Arg73, are
important in carbohydrate recognition (32). Deletion of the N-terminal and C-terminal
regions of galectin-1 also decreased the carbohydrate-binding affinity indicating that the

other two exons may also be important for carbohydrate binding activity (31).

Recently, crystallographic studies of galectin-1 have shown that the lectin appears
to form a dimer structure. Each of the monomer associates with one N-acetyllactosamine
molecule (33, 34). The hydroxyl groups of the galactose molecule in the disaccharide
play major roles in its interactions with the protein. The main determinant of the
specificity for galectin-1 binding to carbohydrate resides on the hydroxyl at position 4 of
the galactose, which interacts with the side chain of Arg48 and His44 through hydrogen
bonds. Other residues involved in the carbohydrate binding include Asn46, which
interacts with hydroxyls at positions 3 and 4 through a water molecule, and Asn6l with
hydroxyl at position 6. The interaction between the side chains of Lys63 and Trp68 also

aligns the two residues to interact with the galactose ring. The hydroxyl group of

l4

galactose at position 2 shows no interaction with the proteins. The GlcNAc moiety of the
disaccharide also interacts with the protein, but the interaction is less extensive than that
of the galactose portion. The GlcNAc-protein interactions are mediated through the
contact of the hydroxyl at position 3 with Arg73, Arg48 and Glu71. In addition, the N2
position of the N-acetyl group is involved in the interaction with Arg73, Asp54 and His52.
This may explain the enhanced binding of galectin to N—acetyllactosamine over that of
lactose. A similar result has been found in the study of the crystal structure of N-

acetyllactosamine and recombinant galectin-2 complex (35).
4) Subcellular Localization of Galectins

Most of the C-type lectins have been found to be localized either in a defined
cellular compartment such as membrane-integrated receptors or exported extracellularly
as soluble proteins. In contrast, the cellular localization of galectins is somewhat less
clear cut. Galectin-l polypeptides have been shown to be intracellular, though reports
vary regarding the nuclear and cytoplasmic distribution. Immunofluorescence staining
for galectin-1 on cryostat sections from adult chicken kidney (36), calf pancreas (37) have
shown that the lectin is localized in cell nuclei as well as cytoplasm. Similarly,
immunocytochemical studies in dorsal root ganglion neurons have shown that galectin-1
could be detected both in the nucleus and cytoplasm of the neurons (38).
Immunoelectron microscopy has also localized galectin-l in the nucleus of the epidermal
cells from the intermediate layer of chick embryonic skin (39). However,
immunofluorescence and ultrastructural studies have also led to explicit statements that

anti-galectin-l antibodies failed to label the cell nucleus. In a recent study,

15

immunofluorescence staining of Chinese hamster ovary cells by anti-galectin-l
antibodies, detected galectin—1 both intracellularly and extracellularly. However galectin-

] was absent from the nucleus (40).

Despite the intracellular localization, there is considerable evidence for the
extracellular localization of galectin-1, though no typical secretory signal sequence has
been found (8). Studies of export of muscle galectin-1 have shown that during myoblast
differentiation the lectin becomes concentrated in evaginations of plasma membrane
which pinch off or bud off to form labile lectin-rich extracellular vesicles (41, 42). The
exported lectin can then interact with carbohydrate on laminin (43) and other extracellular
glycoproteins (44). These observations suggest a possible mechanism for lectin export
from the cytosol to the extracellular matrix. Galectin-l is also found to be extemalized
upon injection of epinephrine without the involvement of secretion vesicle in Xenopus

laevis skin cells (45).

Similarly, galectin-3 has been found on the cell surface and inside the cell. Most
of the galectin-3 was localized intracellularly, especially in the nucleus as shown by
immunofluorescence staining, though a small amount of the lectin was detected on the
cell surface (46). Subcellular fractionation studies on rat basophilic leukemia cells
showed that the majority of the rat galectin-3 is intracellular including in the nucleus (47).
Other studies have also shown the intracellular and extracellular localization of galectin-3
in dorsal root ganglion neurons (38, 48), in murine fibrosarcoma cells and human
carcinoma HeLa-S3 (49), and mouse macrophages (50). In Baby Hamster Kidney (BHK)

cells, the majority of galectin-3 is found in the cytoplasm and small amounts are

 

l6

deposited on the cell surface and substratum (51). In Madin-Darby canine kidney cells
the lectin is expressed and secreted at the apical domain of the polarized cells (52),

suggesting that apical secretion may be the possible way for galectin—3 extemalization.

A wide difference in the level of galectin-3 in various tissues of mice during
developmental stages was noted (26). The induction of galectin-3 in mouse macrophages
was found to be dependent on the addition of thioglycolate (50). The regulation of
expression and subcellular localization of galectin-3 was studied in detail in mouse 3T3
fibroblasts. Galectin-3 was found primarily in the cytoplasm of quiescent cells; however,
in proliferating cells, the expression level increased and is predominantly localized in the
nucleus (53). In quiescent cells, the cytoplasmic galectin-3 was found to be
phosphorylated (pI 8.2). Upon serum stimulation, the synchronized cells showed an
increased level of the phosphorylated form (pI 8.2), both in the cytosol and the nucleus.
The level of unmodified form (pI 8.7) also increased but could only be found in the
nucleus of the stimulated cells (20). The significance and mechanism of the differential
expression and localization of the two isoelectric variants of galectin-3 remain unclear.
Nevertheless, one study on human colonic tissue specimens has shown that galectin-3 is
concentrated in the nuclei of differentiated epithelial cells (54). The expression and
nuclear localization changed significantly during the progression from normal mucosa to
adenoma to carcinoma; the lectin is absent from the nuclei of carcinoma cells but still
localized in the cytoplasm. These observations suggest that the exclusion of galectin-3

from the nucleus may be related to the neoplastic progression of colon cancer.

 

l7

5) Functions of Galectins

It has long been suggested that animal lectins might function in modulating cell-
cell and cell-extracellular matrix interactions. For example, galectin-l could be involved
in modulating transitory adhesions during embryonic cell movement (55), and may also
play a role in the elaboration or organization of extracellular matrix components (56). In
skeletal muscle, galectin-1 has been shown to either promote or inhibit cell adhesion.
One study showed that galectin-1 binds to laminin and inhibits cell-matrix interactions
(43); presumably, binding of galectin-1 to polylactosarnine on laminin in myoblasts
interferes with laminin recognition by the major laminin receptor integrin 0431. In
contrast, a recent report has shown that that galectin-1 binds to laminin and promotes cell
adhesion to laminin (57). Nevertheless, these results suggest that galectin-l may have a
role in muscle development. Other studies suggested that galectin-l might participate in

regulating cell proliferation (58, 59).

On the basis of the identification as a laminin-binding protein (see Table 3), it has
been proposed that galectin-3, like galectin-l, might also play a role in cell adhesion (60).
Some studies suggested that binding of galectin-3 for both IgE and IgE receptor can
trigger activation of mast cells and basophils and play a role in inflammation (61). Two
glycoproteins named M2BP-l and M2BP-2 were identified from certain colon cancer cell
lines by coprecipitation with galectin-3 (62). The interaction is mediated by CRD of
galectin-3 and sugar moieties of the glycoproteins. The localization of MZBP-l and

M2BP-2 is not clear, but these glycoproteins may play a role in mediating galectin-3

18

function. The exclusion of galectin-3 from the nuclei of carcinoma cells may be related

to the neoplastic progression of colon cancer (54).

Several lines of evidence suggest that galectin-3 may play a role in pre-mRNA
splicing (63, 64). First, immunofluorescence staining and immunoblotting analyses have
shown that ribonuclease A (RN ase) treatment of perrneablized cells can remove nuclear
galectin-3 while DNase I failed to do so. In addition, fractionation of nucleoplasm
derived from Swiss 3T3 cells on a cesium sulfate density gradient revealed galectin-3 in
fractions with densities of 1.30 -1.32 g/ml , corresponding to the range of densities
reported for hnRNP and snRNP. Finally, the most direct evidence for a role of galectin-3
in pre-mRNA splicing comes from experiments using a cell free assay. The nuclear
extracts used for this assay contain galectin-3, as detected by immunoblotting. Addition
of saccharides that bind to galectin-3 with high affinity inhibits the cell free splicing.
Most persuasively, depletion of galectin-3 by affinity adsorption on a lactose column
resulted in a concomitant loss of splicing activity, which could be restored by

reconstitution with recombinant galectin-3.

LITERATURE REVIEW ON snRNP PARTICLES AND PRE-mRNA SPLICING

A) Structure and commsition of snRN P

During and after transcription, the pre-mRNA molecules are processed and
transported to the cytoplasm where translation occurs. For most of the RNA polymerase
II transcripts, this process includes the addition of m7G cap at the 5’-end,
polyadenylation at the 3’-end, assembly with proteins to form hnRNP, splicing to remove
noncoding introns and ligation of exons, and the exchange of hnRNP proteins for mRN A
proteins. The catalytic elements required for pre-mRN A splicing reaction include
snRNPs, non-snRNP splicing factors and probably many other factors. The steps in
nuclear pre-mRN A splicing have been elucidated and are illustrated in Figure 2 (65).
Pre-mRNA splicing proceeds through a two step mechanism involving two separate
transesterification reactions. The first step is cleavage at the 5’ splicing site generating
free exon 1 by activating the 2’-OH at the branch site producing a lariat RNA; the second
step involves another transesterification reaction at the 3’ splice site to join the exons and

remove the intron.

The catalytic elements, snRNPs, contain uridine-rich small nuclear RNA, a
common set of core polypeptides and one or more snRNP-specific proteins. In
mammalian cells, there are six major snRNAs (Table 4), named U1 through U6, ranging
in size from 107 to 216 nucleotides (66). The major snRNPs are found located in the
nucleoplasm, except for U3 which is localized in the nucleolus. Other minor snRNAs

have also been characterized; however, this review will focus on the major snRNPs. The

19

 

20

 

 

 

Figure 2: Schematic diagram illustrating the two major steps in pre-mRNA
splicing.
Step 1 is a transesterification reaction leading to the formation of a lariat
intermediate. Step 2 is another transesterification reaction leading to intron
removal and exon joining. This scheme was modified from reference 65.

21

Table 4. Major Mammalian snRNAs

AAUUUG
GAUUUUUGG

AAUUUUUG

 

22

major snRNAs are transcribed by RNA polymerase II, except for U6 which is transcribed
by RNA polymerase H1 (66, 67). Analysis of pre-mRNA intron regions has revealed
some consensus sequences to be complementary to conserved internal sequences in U2
snRNA and at the 5’-end of U1 snRNA (68-70). U1 and U2 snRNPs recognize the
consensus sequences at the 5’ splice site and branch site in the initial steps in the splicing
reaction (Figure 3). In the meantime U4 and U6 snRNAs bind to each other through an
extended complementary region to form the U4/U 6 snRNP complex (71) and then the
U4/U6 snRNP particle associates with U5 snRNP (72). This tri-snRNP is then bound to
the pre-mRNA/U 1/U 2 complex along with other components for intron removed. The

snRNPs are either recycled for other splicing events or degraded.

Patients with systemic lupus erythematosus (SLE) often possess antibodies against
two nuclear antigens called Sm and RNP. Anti-Sm antibodies can immunoprecipitate
five major snRNAs, U1, U2, U4, U6 and U5 (Table 5), and minor snRNAs U7, U8, U9,
U10, U11 and U12 (73). This was the first indication that different snRNP particles share
identical or at least similar proteins. The major epitopes recognized by the anti-Sm class
autoantibodies are antigenic determinants on the 28 kDa B and 16 kDa D polypeptides of
the snRNP core proteins (Table 6). In addition to these shared proteins, each snRNA is
f011nd to be associated with one or more specific polypeptides to form a specific snRNP
Particle. The anti-(U1)RNP autoantibodies found commonly in SLE patients recognize
determinants on the Ul-specific proteins (73, 74); this antibody reacts mainly with a U1-

SPeCific 70 kDa protein, a 33 kDa A and a 22 kDa C protein (75-77) (Table 6).

 

23

Figure 3: Schematic diagram delineating the number and order of assembly
of the intermediates in spliceosome formation.
The pre-mRN A substrate contains two exons (El and E2), a 5’-splicing site
(S’SS), a 3’ splicing site (3’88), and a branch point (BP). The association of
pre-mRNA with hnRNP leads to the formation of the ATP-independent H
complex. In the presence of ATP, addition of U1 and U2 snRNPs results in
complex A. The addition of a preassembled U4, U5, U6 snRNP complex to
complex A results in complex B formation. An ATP-dependent reaction
converts complex B to the spliceosome (complex C). After the product
mRNA (E1 and E2 joined) is released, the lariat form of the intron is still
associated with the snRNPs in a complex. While the snRNPs are recycled, the
lariat intron is degraded. This scheme was modified from reference 65.

25

Table 5. Major mammalian snRN P

gun-1'.) ._.,. J ~, ,-.~ ,\1 q“. ..............................

 
    
 

          

               

 

 

 

 

U1 Core Polypeptides Nucleoplasm Anti-Sm
U1 70K Anti-m3G

A Anti-(U1)RNP

C

U2 Core Polypeptides Nucleoplasm Anti-Sm
A’ Anti-m3G

B” Anti-(U2)RNP

U3 ? Nucleolus Anti-(U3)RNP
34 K Anti-m3G
' U4/U6 Core Polypeptides Nucleoplasm Anti-Sm
Anti-m3G
U5 Core Polypeptides Nucleoplasm Anti-Sm
100 K Anti-m3G

 

 

 

 

 

 

26

Table 6. Proteins of Mammalian snRNPs

Anti-Sm
Anti-Sm
Anti-Sm
Anti—Sm

 

27

With the exception of U6 and the nucleolar snRNP U3, the major snRNPs share a
common core of polypeptides, which include B (28 kDa), D’ (18 kDa), D (16 kDa), E (13
kDa), F (12 kDa) and G ( 11 kDa) proteins (Tables 5 and 6). Some of these polypeptides
recognize a conserved sequence motif on snRNAs called the Sm binding site (Table 4)
and are assembled to form a BzD’zDzEFG-containing snRNP particles (78, 79). In human
cells, one B protein is replaced by a B’ protein (80), but the B’ protein is not found in
non-primate mammals. Since U6 snRNA is base-paired with U4 snRNA, it could be
isolated by antibodies against the core proteins, though it has no Sm binding site (71). At
high ionic strength, only the D, E, F and G polypeptides remain associated with the
snRNA. These polypeptides are tightly associated and some of the proteins bind to RNA
to protect the Sm consensus sequences from RNase digestion (81). A similar RNase
protection pattern was also found with particles containing the B protein. These
observations suggest that B protein is associated with the snRNP particle through contact
with other polypeptides, not by direct binding to snRNA. Consistent with this notion,
biochemical studies of the assembly of snRNP particles have demonstrated that D, E, F
and G proteins initially assemble to form a 6S particle; the 6S intermediate binds to
snRNA followed by the association of two copies of B protein or one copy each of B and
B’ proteins (78, 79, 82). Binding of the core proteins to snRN A is directed by a sequence
motif of AUnG (n = 2 - 6) (Table 4). This sequence motif is found in the single-stranded
region of the 3’ half of all the anti—Sm precipitable snRNAs U1, U2, U4, and US (81).

Insertion of this sequence into a single-stranded heterologous RNA has been shown to

 

 

in

ma
if S!
M

l
“19’:
a.“

in. \

v“ ‘1‘

SF'!‘

.‘Ml

28

induce the assembly of the snRNP core proteins suggesting that this sequence is necessary

and sufficient for core protein binding (81).

Besides the core polypeptides, Ul snRNP also contains the Ul-specific 70 kDa
protein, the 33 kDa A protein, and the 22 kDa C protein (Tables 5 and 6) (83-85).
Sequence elements in the stem-loop structure at the 5’-end of U1 RNA are essential for
binding of Ul-specific proteins, U1 70 kDa and 33 kDa A (86-88). Purification of U2
snRNP has identified two unique proteins, a 32 kDa methionine-poor A’ protein and a 28
kDa B” protein (84, 85). The U2-specific B” protein has been found to be homologous to
the Ul-specific A protein (89, 90). Both A’ and B” proteins have been found to associate
with the 3’-end stem-loop region of U2 RNA (91). Although the shared core proteins
remain associated with the snRNA under rigorous isolation conditions, the association of
the snRNP-specific proteins is sensitive to ionic strength during isolation. For example,
there is no specific protein other than core proteins that could be isolated with the U4/U 6
and U5 snRNPs from HeLa cell extracts prepared in 0.5 M NaCl (85). This suggested
that the polypeptide composition of snRNP identified may depend on the isolation
condition. Protein cross-linking experiments have shown that B and D proteins are
directly associated with E, F and G core proteins but not with each other, and the U1
snRNP-specific U1 70 kDa and 33 kDa A polypeptides can be cross-linked to each other,
but not to the core proteins (92, 93). These observations suggested that the U1 snRNP-

Specific proteins and the core proteins occupy different domains on U] snRN A.

29

B) Nuclear Localization of the snRN P

Immunofluorescence localization studies using human autoantibodies (anti-Sm
and anti-UlRNP) have shown that the major snRNPs are localized in the nucleoplasm,
while U3 is confined to the nucleolus (94, 95). The snRNPs appear to be concentrated in
20-50 nuclear speckles with a diffuse distribution throughout the nucleoplasm, except for
nucleoli (96). The diffuse nuclear staining may represent an excess population of
snRNPs, or those snRNPs in transit to or from nascent transcripts, or snRNPs in transit to
speckles from a cytoplasmic assembly site, and possibly those Sm antigens dissociated
from snRNPs. Non-snRNP splicing factors such as SC-35 also have been localized to
nuclear speckled structures (97). However, a splicing factor does not necessarily have to
be localized to the speckled structure. The protein U2AF (U2 snRNP auxiliary factor)
which facilitates U2 snRNP binding to the intron branch site was shown to be distributed
throughout the nucleoplasm (98, 99). Fluorescence in situ hybridization and immuno-
electron microscopy have revealed a close association between the newly transcribed c-
fos RNA and the nuclear speckled structures (100). This result suggested that the nuclear
speckled structures observed by immunofluorescence staining for certain snRNPs and

non-snRNP splicing factors are indeed the sites for pre-mRNA splicing.

Immunoelectron microscopy studies of the distribution of splicing factor SC35
have shown that the speckled staining pattern corresponds to nuclear structures enriched
in interchromatin granules and perichromatin fibrils (101, 102). The interchromatin
granules with a diameter of 20-25 nm are linked together by thin fibrils (103). It has been

shown that after inhibition of RNA polymerase II transcription, a stable population of

30

poly(A)+RNA reorganizes into fewer large interchromatin granule clusters. Along with
thereorganization of interchromatin granules, splicing factor SC35 could also be found in
these regions (104). Perichromatin fibrils with a diameter of 3-5 mm are found at the
peripheral regions of condensed chromatin and dispersed throughout the interchromatin
spaces including on the surface of interchromatin granule clusters (103). In contrast to
interchromatin granules, perichromatin fibrils could be rapidly labeled with [3H]uridine,
suggesting that these fibrils correspond to the structure containing newly transcribed
RNA (103). Various components involved in splicing, snRNPs, hnRNP antigens and a

non-snRNP splicing factor SC-35, have all been localized to these fibrils (103).

Detergent extraction of cells followed by fluorescence in situ hybridization studies
have shown that nuclear poly(A)+RNAs are concentrated primarily within several discrete
transcript domains which often surround nucleoli in interphase nuclei. Moreover,
poly(A)+RNA localization in the cell nucleus is coincident with the Sm containing snRNP
(105). Detergent extraction of cells has previously been shown to retain the nuclear
matrix and most of the matrix-associated hnRNP (106). Comparison of nuclei with and
without detergent extraction demonstrated that specific nuclear RNAs are unambiguously
retained upon nuclear matrix preparation (107). Taken together, these results indicate
that upon detergent extraction, these RNP molecules remain in the nucleus and suggest
that poly(A)+RNA and snRNP may associate tightly with subnuclear structures, and
possibly, interchromatin granules and perichromatin fibrils may correspond to these
subnuclear structures. These results provide support that these subnuclear structures are

involved in pre-mRNA splicing.

31

Previous studies on the distribution of snRNPs have demonstrated that during
mitosis, along with the breakdown of the nuclear envelope, the speckled structures
observed in the interphase nucleus disappeared and the snRNPs are distributed diffusely
throughout the cell except the region for condensed chromatin (108, 109).
Immunofluorescence and immunoprecipitation analyses showed that during mitosis the
snRNP particles still retain the structure and protein composition of the particles seen
during interphase (108). The snRNP particles begin returning to the daughter nuclei after
the chromatin begins decondensing in telophase, and they are returned quantitatively to
the daughter nuclei during early G1 phase, at which time the speckled structures are

reformed.

C) anlasmig Assembly and Nuclear Immrt of the snRNP

The snRNAs are transcribed as precursors, with several nucleotides longer than
the mature nuclear snRNA. These extra nucleotides of newly transcribed snRNA are
removed and a 2,2,7-trimethyl-guanosine cap (m3G cap) is added to the snRNA during
maturation and snRNP assembly in the cytoplasm (110). The U6 snRNA is transcribed
by polymerase III and is capped at the 5’ end by a methyl group (Table 4) (1 l l). The cap

structure provides resistance to 5’ exoribonuclease activity ( l 12).

Kinetic studies of assembly of snRNP have shown that newly transcribed snRNA
appears transiently in the cytoplasm. Analysis of cytoplasm generated from mouse
fibroblasts by sucrose gradient centrifugation followed by immunoprecipitation with anti-

Sm antibodies has identified a 68 snRNA-free protein intermediate which consists of the

32

D protein and the E, F, and G proteins. This 6S protein intermediate assembles with free
snRNA and then with two copies of the B protein to form the snRNP particle. The
mature snRNP is then transported back into the nucleus (Figure 4) (79, 80, 82). In
contrast to the core polypeptides, some of the U1 and U2 specific proteins are restricted

to the nucleus (Figure 4).

The binding of core proteins to the snRNA appears necessary for transport of the
snRNA into the nucleus. Mutation of the U2 snRNA Sm consensus sequences,
preventing the core protein assembly with snRNA, resulted in the accumulation of the
snRNA in the cytoplasm; however, removal of sequence responsible for binding of U2
specific protein does not prevent snRNA core protein assembly or nuclear localization of
U2 snRNP. These results suggest that the snRNP nuclear localization signal may be
generated by the association of core proteins and snRN A. However, nuclear localization
of snRNP may also be affected by the snRNA structure. Both mutant U2 snRNA with 3’
extension sequences and mature U2 snRNA could be immunoprecipitated by anti-Sm
antibodies; however, mutant snRNA with 3’-end extension could only be found in the
cytoplasm (113). Analysis of the 3’-end extension sequence has predicted an interaction
between the 3’ extension and the 5’-end stem-loop structure, but this does not prevent
binding of core proteins. Similar results have also been obtained for mutant U1 snRN A
carrying 3’—end extension sequence (114). These results indicated that core protein
binding alone is not the only factor involved in snRNP nuclear import. Proper snRN A

structure is also required for snRNP transport into the nucleus.

33

 

 

 

 

 

UZSnRNP U1snRNP

Figure 4: Schematic diagram illustrating the nuclearcytoplasmic shuttle in
the assembly of the RNA and polypeptide components of U1 and
U2 snRNPs.
The snRNAs transcribed in the nucleus are exported to the cytoplasm, where
the core polypeptides B, D, and E-G are added to form snRNPs. Certain U1-
specific and U2-specific polypeptides are also assembled on the respective
RNPs. These are then translocated into the nucleus, where addition U-specific
and U2-specific polypeptides are added. This scheme are modified from
reference 79.

34

D) Non-snRNP Splicing Factors

A number of proteins essential for splicing activity are not associated with
snRNPs. These proteins contain either a serine/arginine-rich (SR) domain or a RNA
recognition motif (RRM). In mammalian cells, splicing factors 8035 (115), ASF/SF2
(116), U2AF (117) and U1 70 kDa polypeptide (l 16) contain the RRM and a SR-rich
domain (Figure 5). In Drosophila, the products of the transformer (tra) and transformer
2 (tra2) genes (118), which regulate doublesex (dsx) pre-mRN A splicing, also contain the
SR-rich domain (119). It has been shown that tra, tra2, and members of the SR family
are sufficient to commit dsx pre—mRNA to female-specific splicing, though individual SR
proteins differ significantly in their ability to participate in commitment of splicing
complex formation. More importantly, the localization signal for nuclear speckled
structures may reside in the SR-rich domain (119). A monoclonal antibody which labels
mammalian cells in a speckled pattern (120) has been used to identify a family of nuclear
phosphoproteins that contain an SR-rich carboxyl terminal domain (121). SDS-PAGE
showed that this SR protein family consists of at least five different proteins (20, 30, 40,
55, and 75 kDa). Except for the 20 kDa polypeptide, a repeated sequence that
encompasses an RRM is found in all four SR proteins, and each of the four SR proteins
individually can complement a splicing-deficient cytoplasmic S-100 extract. The 30 kDa
band contains two distinct proteins, SRp30a and SRp30b which have been shown to be
ASF/SF2 and SC35, respectively (122). Each individual SR protein initiated splicing of a

different pre-mRNA with substrate specificity (123). These observations suggest that

35

 

 

 

 

_@ RRM @_ Tra 2

 

 

 

 

 

 

 

 

 

RRM ————-- G @—— ASFISF2

 

 

 

 

 

 

 

 

 

—4- RRM PG ——.— SC35
——@———— RRM1 RRMz RRM3~—— U2AF

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 5: Schematic diagram summarizing the context and organization of
specific domains in SR family of proteins.
The RRM box indicates the RNA recognition domain. The relative location
of the serine/glycine rich domain (SR) is shown for each protein. PG and G
indicate proline/glycine and glycine rich region, respectively.

36

specific SR proteins have distinct and essential roles in pre-mRNA splicing. One
possible role is that SR proteins have different specificities for subclasses of pre-mRNAs,
like tra and tra2 gene products for Drosophila sex determination, and regulation of the

levels of SR proteins in different cell types contributes to the regulation of cell-specific

splice choices.

The non-snRNP splicing factor U2AF is required for the binding of U2 snRNP to
the pre-mRNA intron branch site. Purified U2AF consists of two polypeptides of 35 kDa
and 65 kDa. An in vitro splicing assay using cytoplasmic S-100 extract showed that
U2AF is an essential splicing factor and all U2AF activity resides in the 65 kDa
polypeptide (99). SC35 has been shown to be required for the splicing reaction,
especially for the ATP-dependent complex formation (124). In the presence of ATP,
SC35 is required for the interaction of U1 snRNP with both the 5’ and 3’ splice sites
along with the U2 snRNP binding to the branch site. ASF/SF2 displays two activities in
an in vitro splicing assay. It is required for the assembly of the first detectable ATP—
dependent splicing complex (116), and is able to switch utilization of alternative 5’ splice
sites in a concentration-dependent manner (125). The SR domain in SF2/ASF is not
required for alternative site selection, but is necessary for constitutive splicing, and the
RRM motif is essential for alternative splicing (126). The alternative splicing activity of
ASF/SF2 can be antagonized by hnRNP A] protein (127). Both ASF/SF2 and hnRNP

A] have strand annealing activities, which may promote base pairing of snRNAs to the

alternative splice sites in the intron.

37

Recently, the SR proteins have been shown to be able to complement splicing
activity in reactions depleted of U1 snRNP, but can not restore splicing activity to either
U2 snRNP or U4/U6 snRNP-depleted reactions (128). However, the concentration of SR
proteins required to complement the U1 snRNP-depleted splicing reaction is about 10-
fold greater than that of the endogenous concentration, and the activity of SR proteins to
complement the U1 snRNP-depleted splicing reaction is dependent on the particular pre-
mRN A. This observation suggests that U1 snRNA is not essential for catalysis of either
of the transesterification steps in splicing, and since SR proteins are thought to bind to
RNA in a sequence-specific manner, different pre-mRNAs might have differential

activity for recognition by SR proteins.

10.

11.

12.

Literature Cited

Goldstein, I. J. 1979. Carbohydrate-Protein Interaction. American Chemical
Society. ACS Symposium Series; 88.

Sharon, N ., and Lis, H. ed. 1989. Lectins, Chapman and Hall Ltd., London

Hart, G. W. Haltiwanger, T. S., Holt, G. D., and Kelly, W. G. 1989. Ann. Rev.
Biochem. 58, 841-874.

Drickamer, K. 1988. J. Biol. Chem. 2Q, 9557-9560.

Drickamer, K., and Taylor, M. E. 1993. Annu. Rev. Cell Biol. _9_, 237-264.
Hirabayashi, J. 1993. Trend in Glycosci. Glycotech. 5, 251-270.

Barondes, S. H., Castronovo, V., Cooper, D. N. W., Cummings, R. D., Drickamer,
K., Feizi, T., Gitt, M. A., Hirabayashi, J ., Hughes, R. C., Kasai, K., Leffler, H.,
Liu, F. -T., Lotan, R., Mercurio, A. M., Monsigny, M., Pillai, S., Poiries, F., Raz,
A., Rigby, P., Rini, J. M., and Wang, J. L. 1994. Cell 7_6_, 597-598.

Hirabayash, J ., and Kasai, K. 1993 Glycobiolgy 3, 297-304.

Gitt, M. A., Massa, S., Leffler, H., and Barondes, S. H. 1992. J. Biol. Chem. 261,
10601-10606.

Roff, C. F., Rosevear, P. R., Wang, J. L., and Barker, T. 1983. Biochem. J. 2L1,
625-629.

Roff, C. F., and Wang, J. L. 1983. J. Biol. Chem. 258, 10657-10663.

 

Oda, Y., Herrmann, J ., Gitt, M. A., Turck, C. W., Burlingame, A. L., Barondes,

S. H., and Leffler, H. 1993. J. Biol. Chem. 26_8, 5929-5939.

38

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

39

Hirabayashi, J., Satoh, M., and Kasai, K. 1992. J. Biol. Chem. _Zfl, 15485-15490.
Gitt, M. A., Wiser, M. F., Leffler, H., Herrmann, J. Xia, Y. R., Massa, S. M.,
Cooper, D. N. W., Lusis, A. J., and Barondes, S. H. 1995. J. Biol. Chem. fl,
5032-5038.

Madsen, P., Rasmussen, H. H., Flint, T., Gromov, P., Kruse, T. A., Honore, B.,
Vorum, H., and Celis, J. E. 1995. J. Biol. Chem. 2E, 5823-5829.

Hadari, Y. R., Paz, K., Dekel, R., Mestrovie, T., Accili, D., and Zick, Y. 1995. J.
Biol. Chem. ;7_0, 3447-3453.

Pfeifer, K., Haasermann, M., Gamulin, V., Bretting, H., Fahrenholz, F., and
Muller, W. E. G. 1993. Glycobiology 3, 179-184.

Jia, S and Wang, J. L. 1988. J. Biol. Chem. 2_63, 6009-6011.

Agrwal, N., Sun, Q., Wang, S. Y., and Wang, J. L. 1993. J. Biol. Chem. 268,
14932-14939.

Cowles, E. A., Agrwal, N ., Anderson, R. L., and Wang, J. L. 1990. J. Biol. Chem.
265, 17706-17712.

Lee, R, T., Ichikawa, Y., Allen, H. J., and Lee, Y. C. 1990. J. Biol. Chem. 3,
7864-7871.

Abbott, W. M., Hounsell, E. F., and Feizi, T. 1988. Biochem. J. _25_2, 283-287.
Sato, S., and Hughes, R. C. 1992. J. Biol. Chem. 261, 6983-6990.

Knibbs, T. N., Agrwal, N., Wang, J. L., and Goldstein, I. J. 1993. J. Biol. Chem.
_fi, 14940-14947.

Leffler, H., and Barondes, S. H. 1986. J. Biol. Chem. 261, 101 19-10126.

26.

27.

28..

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

4O

Crittenden, L. L., Roff, C. F., and Wang, J. L. 1984. Mol. Cell. Biol. 4, 1252-
1259.

Sparrow, C. P., Leffler, H., and Barondes, S. H. 1987. J. Biol. Chem. _2_6_5, 7383-
7390.

Gitt, M, A., and Barondes, S. H. 1991. Biochemistry 30, 82-89.

Gritzmacher, C. A., Mehl, V. S., and Liu, F. T. 1992. Biochemistry 3_1_, 9533-
9538.

Rosenberg, I. M., Iyer, R., Cherayil, B., Chiodino, C., and Pillai, S. 1993. J. Biol.
Chem. 2_68, 12393-12400.

Abbott, W. M., and Feizi, T. 1991, J. Biol. Chem. ;6_6, 5552-5557.

Hirabayashi, J., and Kasai, K. 1991. J. Boil. Chem. 2Q, 23648-23653.

Liao, D. I., Kapadia, G., Ahmed, H., Vasta, G. R., and Herzberg, O. 1994. Proc.
Natl. Acad. Sci. USA 21, 1428-1432.

Boune, Y., Bolgiano, B., Liao, D. L., Stecker, G., Cantau, P. Herzberg, 0., Feizi,
T., and Cambillau, C. 1994. Str. Biol. 1, 863-870.

Lobsanov, Y. D., Gitt, M. A., Leffler, H., Barondes, S. H., and Rini, J. M. 1993. J.
Biol. Chem. _2_6_8, 27034-27038.

Beyer, E. C., and Barondes, S. H. 1980. J. Supramol. Struct. 13, 219-227.
Childs, T. A., and Feizi, T. 1980. Cell Biol. Int. Rep. 4, 775.

Hynes, M. A., Gitt, M., Barondes, S. H., Jessell, T. M., and Buck, L. B. 1990. J.

Neurosci. _1_Q, 1004- 1013.

 

41

Akimoto, Y, Kawakami, H., Oda, Y., Obinata, A., Endo, H., Kasai, K., and

39.
Hirano, H. 1992. Exp. Cell Res. 1%, 293-304.
40. Cho, M., and Cummings, R. D. 1995. J. Biol. Chem. 23, 5207-5212.
41.. Cooper, K. N. W., and Barondes, S. H. 1990. J. Cell Biol. 1_10, 1681-1691.
42. Harrison, F. L., and Wilson, T. J. G. 1992. J. Cell Sci. 201, 635-646.
43. Cooper, K. N. W., Massa, S. M., and Barondes, S. H. 1991. J. Biol. Chem. _1_1_5,
1437-1448.
44. Cu, M. J., Wang, W. W., Song, W. K., Cooper, D. N. W., and Kaufman, S. J.
1994. J. Cell Sci. M, 175-181.
45. B013, N. C., Roberson, M., Haywood-Reid, P. L., Cerra, R. F., and Barondes, S.
H. 1986. J. Cell Biol. E2. 492—499.
46. Moutsatsos, I. K., Davis, J. M., and Wang, J. L. 1986. J. Cell Biol. 192, 447-483.
47. Gritzmacher, C. A., Robertson, M. W., and Liu, F. T. 1988 J. Immunol. 1_1,
2801-2806.
48. Regan, L. J ., Dodd, J ., Barondes, S. H., and Jessell, T. M. 1986. Proc. Natl. Acad.

Sci. U. S. A. _8_3, 2248—2252.
Raz, A., and Lotan, R. 1987. Cancer Metastasis Rev. 6, 433-452.

49.

50. Cherayil, B. J., Weiner, S. J ., and Pillai. S. 1989. J. Exp. Med. fl, 1959-1972.
51. Sato, 8., Burdett, I., and Hughes, R. C. 1993. Exp. Cell Res. 201, 8-18.
12. Lindstedt, R., Apodaca, G., Barondes, S. H., Mostov, K. E., and Leffler, H. 1992

J. Biol. Chem.l6_8. 11750-11757.

 

42

53. Moutsatsos, I. K., Wade, M., Schindler, M., and Wang. J. L. 1987. Proc. Natl.

Acad. Sci. USA _83, 6452-6456.

54. Lotz, M. M., Andrews, C. W., Jr., Korzelius, C. A., Lee, E. C., Steele, G. D.,

Jr.,Clarke, A., and Mercurio, A. M. 1993. Proc. Natl. Acad. Sci. USA. 9_(), 3466-

3470.

55. Zalike, 8.15., Thomson, L. W., and Ledsham, I. M. 1987. J. Cell Sci. _8_8,, 483-493.

56. Catt, J. W., and Harrison, F. L. 1985. J. Cell Sci. 3, 347-359.

57. Zhou, Q., and Cummings, R. D. 1993. Arch. Biochem. Biophys. M, 6-17.

58. Yamaoka, L., Ohno, 8., Kawasaki, H., and Suzuki, K. 1991. Biochem. Biophys.

Res. Commun. m, 272-279.

59. Wells, V., and Mallucci, L. 1991. Cell 64, 91-97.
Woo, H. J., Lotz, M. M., Jung, J. U, and Mercurio, A. M. 1991, J. Biol Chem.

60.
266, 18419-18422.
61. - Frigeri, L. G., Zuberi, T. I., and Liu, F. T. 1993. Biochemistry fl, 7644-7649.
62. Rosenberg, I., Cherayil, B. J ., Isselbacher, K. J., and Pilai, S. 1991. J. Biol. Chem.

266, 18731-18736.
Laing, J. G., and Wang. J. L. 1988. Biochemistry 27_, 5329-5334.

$4, Dagher, S. R, Wang, J. L., and Patterson, R. J. 1995. Proc. Natl. Acad. Sci. USA

22, 1213-1217.
Sharp, P. A. 1994. Cell fl, 805-815.
Guthrie, C., and Patterson, B. 1988. Ann. Rev. Genet. 22, 387-419.

 

43

Kunkel, G. R., Maser, T. L., Calvet, J. P., and Pederson, T. 1986. Proc. Natl.

67.
Acad. Sci. USA Q, 8575-8579.
68. Moore, M. J., and Sharp, P. A. 1993. Nature 366, 364-368.
69. lemer, M. R., Boyle, J. A., Mount, S. M., Wolin, S. L., and Steitz, J. A. 1980.
Nature E, 220-224.
70. Parker, R., Sillciano, P. G., and Guthrie, C. 1987. Cell 42, 229-239.
71. Hashimoto, C., and Steitz, J. A., 1984 Nucl. Acids Res. fl, 3283-3293.
72. Konarska, M. M., and Sharp, P. A. 1987.Cell Q, 763-774.
73. Lerner, M. R., and Steitz, J. A. 1979. Proc. Natl. Acad. Sci. USA 76, 5495-5499.
74. Williams, D, G. Stocks, M. R., Smith, P. R., and Maini, R. N. 1986. Immunol.
68, 495-500
75. White, P. J., and Hoch, S. O. 1981. Biochem. Biophys. Res. commun. _192, 365-

371.
White, P. J., Billings, P. B., and Hoch, S. O. 1982. J. Immunol. 22_8, 2751-2756.

76.
Pettersson, 1., Hinterbergwer, M., Mimori, T., Gottlieb, E., and Steitz, J. A. 1984

77.
J. Biol. Chem. 2Q. 5907-5914.

78. Feeney, R. J., Sauterer, R, A., Feeney, J. L., and Zieve, G. W. 1989. J. Biol.

Chem. 264, 5776-5783.
Fisher, D. E., Conner, G. E., Reeves, N. H., Wisniewolski, R., and Blobel, G.

 

1985. Cell .42. 751-758.
Rokeach, L. A., Jannatipour, M., Haselby, J. A., and Hoch, S. O. 1989. J. Biol

Chem. 264, 5024-5030.

 

 

81.

82.

83.

84.

85.

86.

87.

88.

89.

90.

91.

92.

44

Liautard, J. P., Sri-Widada, J ., Brunel, C., and Jeanteur, P., 1982. J. Mol. Biol.
1Q, 623-643.

Sauterer, R. A., Goyal, A., and Zieve, G. W. 1990. J. Biol. Chem. 265, 1048-
1058.

Hinterberger, M., Pettersson, I., and Steitz, J. A. 1983 J. Biol. Chem. 2_58_, 2604-
2613.

Kinlaw, C. S., Roberson, B. L., and Berget, S. M. 1983. 23. 7181—7189.
Bringmann, P., and Luhnnann, R. 1986. EMBO J. 6, 3509-3516.

Hamm, K., van Santen, R. A., Spritz, R. A., and Mattaj, I. W. 1988. Mol. Cell.
Biol. _8_, 4781-4791.

Patton, J. R., and Pederson, T., 1988. Proc. Natl. Acad. Sci. USA 8_5_, 747-751.
Patton, J. R., Habets, W., van Venrooij, W., and Pederson, T. 1989. Mol. Cell.
Biol. _8_, 3360-3368.

Habet, W. J ., Sillekens, P. T. G., Hoet, M. H., Schalken, J. A., Boebroek, A. J .,
Leunissen, J. A., van de Ven, W. J. M., and van Venrooij, W. I. 1987. Proc. Natl.
Acad. Sci. USA M., 2421-2425.

Sillekens, P. T., Habets, W. J ., Beijer, R. P., and van Venrooij, W. I. 1987.
EMBO J. _6_, 3841-3848.

Mattaj, I. W., and De Robertis, M. 1985. Cell fl, 111-118.

Harris, S. G., Hoch, S. 0., and Smith, H. C. 1988. Biochemistry 21, 4595-4600.

93.

94.

95.

96.

97.

98.

99.

100.

101.

102.

103.

104.

105.

106.

45

Luhrmann, R. snRNP Proteins. 1988. in Structure and Function of Major and
Minor Small Nuclear Ribonucleoprotein Particles. Bimstiel, M. L. ed. Springer-
Verlag, New York.

Nyman, U., Halman, H., Hadlaczly, G., Petterson, 1., Sharp, C., and Ringertz, N.
R. 1986. J. Cell Biol. fl, 137-144.

Spector, D. L., 1984. Biol. Cell. _5_1, 109-112.

Spector, D.L. 1990. Proc. Natl. Acad. Sci. USA 22, 147-151.

Fu, X. D., Maniatis, T. 1990. Nature 346, 437-441,

Carmo-Fonseca, M., Tollervey, D., Pepperkok, R., Barabinok S. M. L., Merdes,
A., Brunner, C., Zamore, P. D., Green, M. R., Hurt, E., and Lamond, A. I.
1991.EMBO J. _1_Q, 195-206,

Zamore, P. D., and Green, M. R. 1991. EMBO J. m, 207-214,

Huang, 8., and Spector, D. L. 1991. Genes Develop. 6, 2288-2302,

Spector,D. L., Fu, X. D., and Maniatis, T. EMBO J. 1991.1_Q, 3467-3481,
Carmo-Fonseca, M., Pepperkok, T., Carvalho, M. T., and Lamond, A. I. 1992 J.
Cell Biol. M. 1—14,

Spector, D. L. 1993. Ann. Rev. Cell Biol. 2, 265-313,

Huang, 8., Deerinck, T. J ., Ellisman, M. H., and Spector S. L. 1994. J. Cell Biol.
1%, 877-899,

Carter, K. C., Taneja, K. L., and Lawrence, J. B. 1991. J. Cell Biol. 115, 1191-
1202.

Fey, E. G., Krochmalnic, G., and Penman S. 1986. J. Cell Biol. 202, 1654-1665.

107.

108.

109.
110.
111.

112.

113.

114.

115.
116.
117.
118.
119.
120.

121.

122.

123.

46

Xing, Y., and Lawrence, J. B. 1991. J. Cell Biol. Q2, 1055-1063.

Renter, R., Appel, B., Rinke, J ., and Luhnnann, R. 1985. Exp. Cell Res. 15_9, 63-
79.

Spector, D. L., and Smith, H. C. 1986. Exp. Cell Res. E3, 87-94.

Mattaj, I. W. 1986. Cell 46, 905-911.

Singh, R., and Reddy, T., 1989. Proc. Natl. Acad. Sci. USA _86, 8280-8283.
Murthy, K. G. K., Park, P., and Manley, J. L. 1991. Nucl. Acids. Res. 12, 2685-
2692.

Konings, D. A. M. and Mattaj, I. W. 1987. Exp. Cell Res. 112, 329-339.
Neuman De Vegvar, H. E., and Dahlberg, J. E. 1990. Mol. Cell. Biol. m, 3365-
3375.

Fu, X. D., and Maniatis, T. 1992. Science A6, 535-538.

Krainer, A. R., Mayeda, A., Kozak, K., and Binns, G. 1991. Cell _66, 383-394.
Zamore, P. D., Patton, J. G., and Green, M. R. 1992. Nature 355, 609-614.
Tian M., and Maniatis, T. 1993. Cell E, 105-114.

Li, H., and Bingham, P. M. 1991. Cell 67, 335-342.

Roth, M. B., Murphy, G., and Gall, J. G. 1990. J. Cell Biol. 1_11, 2217-2223.
Zahler, A. M., Lane, W. S., Stolk, J. A., and Roth, M. B. 1992 Genes Develop. 6,
837-847.

Zahler, A. M., Neugebauer, K. M., Lane, W. S., and Roth M. B. 1993. Science
@, 219-222.

Fu, X. D. 1993. Nature 366, 82-85.

124.

125.

126.

127.

128.

47

Fu, X. D., and Maniatis, T. 1992. Proc. Natl. Acad. Sci. USA 32, 1725-1729.
Zuo, P., and Manley, J. L. 1993. EMBO J. _12, 4727-4737.

Caceres, J. F., and Krainer, A. R. 1993. EMBO J. 1_2, 4715-4726.

Mayeda, A., and Krainer, A. R. 1992. Cell fl, 365-375.

Crispino, J. D., Blencowe, B. J ., and Sharp, P. A. 1994. Science E5, 1866-1869.

CHAPTER II

Similarities in the Nuclear Matrix Localization of Galectin-3 and Small Nuclear

Ribonucleoproteins (snRNP): Evidence from a Comparative Immunofluorescence

Analysis

48

SUMNIARY

Galectin-3 is a galactose/lactose-specific lectin identified as a required factor in pre-
mRNA splicing assayed in a cell-free system. In the present study, a comparative analysis
of the immunofluorescence staining patterns was carried out using an antiserum specific for
galectin-3 and human autoimmune sera reactive against the Sm antigens of small nuclear
ribonucleoproteins (snRNP). Using mouse 3T3 fibroblasts, fixed with paraformaldehyde
and permeabilized with Triton X-100, both anti-galectin-3 and anti-Sm yielded intense
staining covering the entire nucleus, with the exception of ~5 circles devoid of fluorescence.
These "black holes" most probably represent the nucleoli within the nucleus.
Permeabilization without fixation, followed by extraction with 0.25 M ammonium sulfate
prior to fixation and staining yielded speckled patterns for both anti-galectin-3 and anti-Sm.
Thus, differences in description of staining patterns, such as diffuse versus speckled, reflect,
for the most part, quantitative differences in either the antigen, the antibody or both rather
than intrinsic differences in the localization of the two antigens. The nuclear staining was
lost for both anti-galectin-3 and anti-Sm when permeabilized 3T3 cells were treated with

ribonuclease A while parallel treatment with deoxyribonuclease I failed to yield the same

effect.

During the mitotic phase of the cell cycle, when chromosomal condensation and
nuclear envelope breakdown begins, galectin-3 can be found predominantly around the area
of the nucleus but a band along the axis of the chromosomes is totally devoid of galectin-3.

Thus, it appears that galectin-3 is found throughout the cell during mitosis except where

49

50

chromosomal DNA is located. Essentially the same results are obtained with anti-Sm
staining. These immunofluorescence data suggest that both galectin-3 and the snRNPs are
colocalized on the RNP—containing nuclear matrix. This notion is supported by double
iminunofluorescence staining, which indicate that there is a direct one-to-one
correspondence between the speckled patterns obtained with anti-galectin-3 and with anti-

Sm.

INTRODUCTION

Galectin-3 is a galactose/lactose-specific carbohydrate-binding protein found in both
the cytoplasm and nucleus of a wide variety of cells (1). The amino acid sequence of the
polypeptide, deduced from the nucleotide sequence of a cDNA clone (2), exhibited
structural features, such as tandem repeats of a sequence motif rich in proline and glycine
residues, similar to proteins found in the core polypeptides of heterogeneous nuclear
ribonucleoprotein complex (hnRNP) and in spliceosomes (3, 4). Using a cell-free assay for
pre-mRNA splicing, it was recently demonstrated that galectin-3 constituted a required

factor in spliceosome formation and mRNA processing (5).

The subcellular localization of several splicing factors has been studied by
immunocytochemistry and by in situ hybridization. Immunofluorescence studies using
antibodies specific for the small nuclear ribonucleoproteins (snRNP) and non-snRNP
splicing factors, such as SC35, have shown that these components are distributed in a
speckled pattern in interphase nuclei (see reference 6 for review). In situ hybridization with
oligonucleotide probes complementary to the major spliceosomal snRNAs resulted in a

similar observation (7, 8).

The demonstration of a role for galectin-3 in cell-free pre-mRN A splicing called to
question the subnuclear distribution of the protein, relative to the other known components
of RNA splicing. Our initial immunofluorescence studies on galectin-3 showed intense
staining of the nucleus, covering the entire organelle, and some staining of the cytoplasm

(9). Subsequent studies have found varied labeling patterns, ranging from diffuse

51

52

distribution of fluorescence to distinct punctate staining (10, 11). Thus, the present study
was undertaken with three main goals in mind. First, we wanted to clarify the issue of the
apparent discrepancy in the labeling patterns of splicing factors by a direct comparison
between the immunofluorescence yielded by anti-galectin-3 versus an autoimmune antibody
reactive against polypeptides of the snRNPs. We found that qualitative differences in the
description of the labeling patterns (diffuse versus speckled) may reflect quantitative
differences in the antigen, antibody or both. Second, although we had previously reported
the sensitivity of the nuclear staining of anti-galectin-3 to ribonuclease treatment (11),
implicating anchorage of the protein to a RNA-containing matrix, no direct comparison
with a known splicing factor was available. We now report the results of such a study.
Finally, we wished to follow the subcellular localization of galectin-3, along with a known
splicing factor, through the entire cell cycle. In the present study, we found that both
galectin—3 and snRNPs were excluded from the segregating chromosomes during mitosis.
Combined with the biochemical data to be reported in Chapter III, the various lines of
evidence strongly suggest that galectin-3 is associated with the storage/assembly sites of

splicing factors.

MATERIALS AND METHODS

Q11 Culture. Mouse 3T3 fibroblasts were obtained from American Type Culture
Collection (Rockville, MD). The cells were grown as monolayers in Dulbecco modified
Eagle's Medium containing 10% calf serum, 100 U/ml penicillin and 100 ug/ml
streptomycin at 37 0C in a humidified atmosphere of 10% C02. Cells grown at a density
less than 5 x 104 cells/cm2 were proliferative (sparse culture); above this density, the cells
were confluent. In some experiments, cells at low density were arrested by removal of
serum and maintenance in medium containing 0.2% calf serum for 48 hours (quiescent

cells). Upon readdition of serum (10%), the cells were reactivated (10).

Antibodies. In the present study, antiserum against galectin-3 was derived from
rabbit #33. This Flemish Giant rabbit was immunized with recombinant galectin-3, purified
from an E. coli expression system. The details for the production of recombinant galectin-
3, the generation of the antisera, and the characterization of its specificity have been
previously reported (12). Antiserum directed against RAP30 ( 13) was a gift of Dr. Zachary
Burton (Michigan State University). Human autoimmune serum reactive with the Sm

antigens of snRNPs (ENA anti-Sm) was purchased from The Binding Site (San Diego, CA).

Immunofluorescence Microscopy. Mouse 3T3 fibroblasts were seeded onto
coverslips (22 x 22 cm), which were then placed in 6 -well (8 cmzlwell) cluster dishes. The
cells were washed twice (4 mee11) with phosphate-buffered saline (PBS, 140 mM NaCl,
2.68 mM KCl, 10 mM NazHPO4, 1.47 mM KH2P04, pH 7.4). Cells were fixed by

incubation (20 minutes at room temperature) in 4% paraformaldehyde in PBS (2 ml/well),

53

54

followed by washing twice with PBS (10 minutes each, on a rotating platform). The
residual aldehyde groups were blocked by 0.1 M glycine in PBS (4 ml/well; 20 minutes at
roOrn temperature). After removal of the glycine-PBS solution, cells were permeabilized
with 0.5% Triton X-100 in PBS (2 mee11; 4 minutes at room temperature). The cells were
again washed twice with PBS. The cells were then incubated with antiserum, with
appropriate dilution, for one hour at room temperature. After washing three times in T-TBS
(10 mM Tris pH 7.5, 500 mM NaCl, 0.05% Tween 20) (15 minutes each), the cells were
incubated with the secondary antibody, at appropriate dilution. Fluorescein-conjugated or
rhodamine-conjugated goat anti-rabbit immunoglobulin (BMB, Indianapolis, IN) was used
to detect rabbit anti-galectin-3 and rabbit anti-RAP30 binding and fluorescein-conjugated or
rhodamine-conjugated goat anti-human immunoglobulin was used to detect human
autoimmune ENA anti-Sm binding. After immunostaining, cells were counterstained for
DNA with 4',6-diamidino-2-phenylindole-2-HC1 (DAPI) at a concentration of 1 ug/ml (14).
Finally, the coverslips were washed three times in T-TBS (15 minutes each) before

mounting with Penna-Fluor on glass microscope slides.

Effects of Permeabilization Extraction and Nuclease Treatment. In some

 

experiments, 3T3 cells were permeabilized (0.5% Triton X-100 in 300 mM sucrose, 100
mM NaCl, 3 mM MgC12, 10 mM 1,4-piperazinediethanesulfonic acid, pH 7.4; 10 minutes
at 4 oC) without prior fixation. The permeabilized cells were washed with PBS twice (30
seconds each) and were either extracted with 0.25 M ammonium sulfate for 10 min at room
temperature (15) or were digested with ribonuclease A (RNase; Sigma; 100 ug/ml) or

deoxyribonuclease I (DNase; BMB; 100 U/ml) at 37 °C for 1 hour. The treated cells were

55

washed with PBS twice (30 seconds each), and cells were fixed by incubation (20 minutes
at room temperature) in 4% paraformaldehyde in PBS (2 ml/well), followed by washing
twice with PBS (10 minutes each, on a rotating platform). The residual aldehyde groups
were blocked by 0.1 M glycine in PBS (4 mee11; 20 minutes at room temperature). After

removal of the glycine-PBS solution, the cells were then stained with antiserum as

described above.

RESULTS

Immunofluorescence Analysis of Galectin-3. We had previously reported the

development of polyclonal rabbit antisera against recombinant galectin-3, purified from an
E. coli expression system (12). One antiserum (#33) and its preimmune control serum were
used in experiments throughout this entire study. This antiserum immunoblots a single
polypeptide (Mr 33,000) in extracts of mouse 3T3 fibroblasts. Moreover, immunoblotting
analysis of purified NHz- and COOH-tenninal domains of galectin-3 indicated that the
principal epitopes recognized by anti-serum #33 were localized within the NHz-terminal
half of the polypeptide (12). These results obviate any complication in the interpretation of
the staining patterns, particularly in terms of the possibility that the antibody reagent might
cross-react with the carbohydrate recognition domain of the members of the galectin family,

which exhibit sequence homology with each other (1).

Mouse 3T3 fibroblasts, fixed with paraformaldehyde and permeabilized with Triton
X- 100, yielded both nuclear and cytoplasmic staining with anti-galectin-3. This was
observed both in low magnification micrographs showing a field containing several cells
(Figure 1A) as well as high magnification micrographs showing a single cell (Figure 1B).
Parallel analysis with the corresponding preimmune serum showed negligible staining
(Figure 1E and Figure 1F). In general, the staining appeared to be diffuse, covering the
entire nucleus with the exception of ~5 circles devoid of fluorescence (Figure 1B). These

"black holes" were also observed when the same cells are counterstained with the DNA-

56

57

Figure 1: Immunofluorescence staining of 3T3 cells after fixation with
paraformaldehyde (4%) and permeabilization with Triton X-100
(0.5%).
A, E: low magnification showing fields containing several cells; B-D, F-H: high
magnification showing single cells. A and B, rabbit anti-galectin-3 (1:150
dilution of antiserum #33); E and F, preimmune serum from the same rabbit #33
(1:150 dilution). The binding of the primary antibody was detected by
rhodamine-conjugated goat anti-rabbit immunoglobulin. The cells in B and F
are shown in DAPI counterstaining and phase contrast micrographs in C, G and
D, H, respectively. (Bar, 50 um)

58

 

59

specific dye DAPI (Figure 1C and Figure 1G). In phase contrast micrographs of the same
cells, these "black holes" correspond to phase dense structures in the nucleus (Figure 1D
and Figure 1H). Moreover, the "black holes" were seen in cells treated with antibodies
directed against two other nuclear proteins:(a) RAP30, a 30 kDa component of the general
transcription factor, TFIIF; and (b) Sm, an epitope defined on proteins B and D of snRNPs
(Figure 2, column A). On the basis of these results and comparison with previously
published staining patterns of snRNPs and the non-snRNP spliceosome component SC35
(14, 16), we believe the "black holes" correspond to nucleoli. Thus, the nuclear distribution

of galectin-3 appears to be diffuse throughout the nucleoplasm, excluding the nucleoli.

Immunofluorescence of Permeabilized Cells after Salt Extraction or Nuclease
Digestion. When 3T3 cells were permeabilized with Triton X-100 (without fixation), then
fixed and stained with antibodies against galectin-3, Sm, and RAP30, the labeling patterns
(Figure 2, column B) were qualitatively different from those obtained with cells that were
initially fixed prior to permeabilization (Figure 2, column A). First, cytoplasmic staining
was lost from the galectin-3 and RAP30 samples. Second, there was reduction in the
staining intensity in all three samples. Finally, there appeared to be a general smearing of
stained structures, yielding a diffuse distribution of fluorescence and accentuating the
nucleolar "black holes" (galectin-3 and Sm). The most striking labeling patterns were
obtained, however, when the permeabilized cells were extracted with 0.25 M ammonium
sulfate prior to fixation and staining. This procedure extracts the majority of the nonhistone

nuclear proteins, leaving chromatin, nuclear matrix, and associated RNAs (15, 17). In all

A81 On .38 832030888 888-83 “mow 339.806.8888.... .3 co 8.3238288

838-85 “new 33w3809088meofi .3 388% 83 .3088“ 38.5 05 .8 9583 05. 888% can” 3

838a (95 888 0888685 58:: new .8826 On _ u C 83.85 838 Anne 88883 co noes—6

o2 ” C Wazoo—«mean :38 5:5 3:338 :05 83 cocoa—U8 some 898 0:28.. 2F .82.: :05 28 63:8

83:68.5 2 mmd 55 3885 .38388& 803 8:8 CV 696 :05 was 85 38.538ch 803

2.8 EC 5888888 came 87x :28. 2 38 see 8283888 53 use 2»; £00 93
8:3 nah he F538. 9883» 3:888:=:==EE_ 05 .8 85.—38:0 "N 953m

is”

61

 

Sm

62

cases except RAP30 (which has lost all of its staining), distinct speckled patterns were

observed (Figure 2, column C).

It should be noted that differences between the diffuse staining pattern seen in
column A of Figure 2 and the speckled pattern seen in column C of the same figure (for
galectin-3 and Sm) probably reflect, at least in part, quantitative differences in either the
antigen, the antibody, or both. With high titer antiserum and high levels of antigen (e.g. in
fixed and permeabilized cells of column A), the staining intensity is so strong that it covers
the entire nucleus, obscuring even the nucleolar "black holes" (Figure 2, column A, anti-
galectin-3 staining). With loss of antigen from the nuclei during the permeabilization and
extraction procedures, the staining of discrete structures/regions becomes more distinct,
giving rise to a speckled pattern (Figure 2, column C, anti-galectin-3 staining).
Alternatively, the same staining protocol could yield, with two different antisera against the
same antigen, either diffuse or speckled patterns. In this connection, we have also observed
variations in the DAPI staining of DNA, ranging from a completely fluorescent nucleus,
diffuse fluorescence except for nucleolar "black holes", to punctate patterns (see, for

example, DAPI staining in Figure 3 below).

The speckled staining pattern observed for galectin-3 and Sm (Figure 2, column C)
suggests that these antigens are associated with subnuclear structures. Therefore, the effects
of RNase and DNase on the immunofluorescence patterns of Triton X-100 permeabilized
cells (Figure 2, column B) were studied. RNase digestion resulted in drastic reduction in

the intensity of galectin-3 staining (Figure 3, columns A and B). There appeared to be some

63

A81 0m .38 823030888 888-88 “mow vouawsmcoerEaEE 80 8:50—38:88
8598-88" “mow Bemw=809088m00€ .3 080200 83 because 3.8.8 05 .80 9583 2;
#75 8_3 808882800 08* 80:28 cowH 3 8m85 <Zm 53» 80 828:0 ow _ ”S T802888“ 858
.23 083308 05 03c :05 803 8:00 0038-08.35 2E. .8880m82 808 8 80888 om .8 :85
8: 8020 55 00308 8:00 008—53888 87X coach. 6v ”888882 800. .0 8888 so 80% 28E:
8: 082d 55 00:88 8:00 008.5388 02-x coach. Ev 8:00 008—53888 @8de 87X couch $5
89728 H0 swung

8m 05 88 0.8325» 8.. 50:8. «8858 awe—0:858 05 :0 35:53.: 0838. he 30am an 0.89%

 

K<o

Em

3:18

65

rearrangement of nuclear organization, as manifested by the DAPI staining. On the other
hand, DNase digestion completely removed the DAPI fluorescence, while the galectin-3
staining was retained (Figure 3, column C). Essentially the same results were obtained with
anti-Sm staining: drastic reduction of intensity in RNase-treated samples and little or no
difference in staining in the DNase-treated samples. We conclude from these results that
nuclear staining of galectin-3 (and Sm) was sensitive to RNase but not to DNase,

implicating the association of the proteins with RNP structures.

Permeabilized 3T3 cells were extracted with 0.25 M ammonium sulfate, fixed, and
then sequentially labeled with anti-Sm and anti-galectin-B. Figure 4A shows the speckled
immunofluorescence pattern due to anti-galectin-3, detected with fluorescein-conjugated
goat anti-rabbit immunoglobulin. Figure 4B shows the immunofluorescence pattern, over
the same two cells as those shown in Figure 4A, due to anti-Sm, detected with rhodamine-
conjugated goat anti-human immunoglobulin. There was a direct (one-to-one)
correspondence between the two speckled patterns in each cell, suggesting the co-

localization of the two polypeptides.

Immunofluorescence Analysis as a Function of the Cell Cycle. We had

previously shown that the addition of serum growth factors to quiescent cultures of 3T3
fibroblasts resulted in an elevation of the expression of the galectin-3 gene. This
stimulation was observed: (a) at the level of transcription rate of the gene as assayed by

nuclear run-off experiments (18); (b) at the level of accumulation of the mRNA as

66

2800308888:

00882280 Sow 08088006280008 3 02032 .230 002800: 05 ..o>0 8m-:80 .«0 8000: 8:203

05 0305. m _oc0m 280030888: 2000380 80» 02088092000800: .3 00.032 6-2828280
.0 Essa 0550; 20 $2... < .28 .32 588:5 .o .5020 02 8 0-5.8.0055 fine 3 832.8
828:0 com” S 8m-0:0 <Zm 883. 2888680 008:: 523 338808 080802 003 8202 05.

.008 55 05 .383 5285.5 2 00.0 as, 0280: .50; :28. 0.0.9 03:82.50 2...; £8 0,:

6:00
080m 05 2 mmzmfi H0 cows—.0 8m 0:0 9280—0» .8: mam—:50 350800388088: 03009 "0 0.5»:—

 

67

 

68

determined on Northern blots (18); and (c) at the level of the protein as measured on
immunoblots (19). At the level of single cells, this elevation of galectin-3 expression is
manifested by an increase in the general level of fluorescence in both the cytoplasm and the
nucleus, when assayed in fixed and permeabilized cells. Serum-starved cultures of 3T3
cells show weak labeling with anti-galectin-3 (Figure 5). Upon serum stimulation, there is
intense nuclear labeling, as well as a noticeable increase in cytoplasmic fluorescence. The
first wave of DNA synthesis was observed between 16 and 24 hours following serum
addition to the quiescent cultures (10). Thus, the increase in the expression and nuclear
localization of galectin-3, as revealed by immunofluorescence, can be observed well before
theionset of the S-phase of the first cell cycle. In fact, the most intense nuclear staining was
observed at 16 hours, followed by a decrease in the staining intensity and partitioning of the
protein between the nuclear and cytoplasmic compartments (Figure 5). This prompted us to
analyze the staining patterns of galectin-3 during the later portion of the cell cycle in some

detail.

Cells from cultures between 20 and 24 hours post serum stimulation were fixed,
permeabilized, and stained with anti-galectin-3 (Figure 6). The position of chromosomal
DNA was simultaneously followed by DAPI staining while the presence of the nuclear
membrane was observed by phase contrast microscopy. In prophase, when chromosomal
condensation and nuclear envelope breakdown begins, galectin-3 can be found
predominantly around the area of the nucleus but there appears to be a band along the axis
of the chromosomes where galectin-3 is excluded (Figure 6A). This band, devoid of

galectin-3, is accentuated in metaphase; chromosomes aligned at the metaphase plate, as

69

.0=0no_w0000000 0500-000
0% 08030090560005 ‘3 088.00 .903 33000 E0003 0.0 .8 wE00B 05. Anna 60.08000 he 023:0
Om _ u S m-008_0w-000 0500.. 003 08000000 000 005—53530 .030... 803 0:8 2F 5908002 002000
000.08 0200 0050 00 000 .Nm .2 .w .c 00 030030 00.8000 m0_0.§m 008208 m-< 0300a .3053 0:08
:00 00 002000 20 .3 080—2008.. 000 A0023. :00 03.0 00 300: Mg 0003003. 000.50 8 038.38 203 2.00
0050355» 8:08

”5.30:8 20: be 03302 0 m0 0:8 mph 5 mauve—paw be 000030: 90580 oou3$ue==e§EE~ “m 95!..—

70

 

71

.E& 30880:: 80.0000 8000 00000 03000 20 0:00

0.08 2:. 8:00 0003000 000:8 .2300 030 pt 000 8000030050 Amv "8.00022 Ae "8000000 n! ”8000000:—

Amv ”80000.00 A3 “03008 80 £800. .00 mowsm 000t0> €00,823. 0:00 .5405 .23 w0_0m08020000 .3

0032.8 003 <20 3080000000 .«0 00280 05. 020003000005 0300-000 80m 0203.88.05.00005 .3

080800 003 3000000 b00000 05 .8 m0m003 05. .390 8008000 «0 025:0 om _ H: 9000208000 03000 :03

00000005 000 0002302500 .0050 803 0000—0050 05.8 800 0000: en 000 cm 0003003 0003—00 05.0 0:00
600300 000000: 0:00 mph 00 Wage—0w *0 000300 05550 00000000033055“ 2: no $920000 "0 0.53..—

 

 

 

‘?
E
33
a
U

72

DNA
H

73

revealed by DAPI staining, appears as a dark band in the immunofluorescence photograph
(Figure 6B). As the chromosomes separate in anaphase (Figure 6C), and decondense in
telophase (Figure 6D), galectin-3 is found throughout the cell except where chromosomal
DNA is located. As the nuclear envelope reforms around the decondensing chromosomes
and the cell undergoes cytokinesis (Figure 6E), galectin-3 appears to be excluded from the
nucleus. This cytoplasmic localization of galectin-3 persists in the newly formed daughter
cells (Figure 6F). Thus, galectin-3 must be retranslocated into the nucleus as the cells enter
the 61 phase of the next cell cycle and cells from cultures 32 hours post serum stimulation

yielded staining patterns similar to those from cultures 8 hours after serum addition (Figure

5).

Essentially the same results are observed when the Sm antigen of snRNPs are
stained with autoimmune serum (Figure 7). The prophase through cytokinesis stages of
mitosis are again discerned from DAPI staining of DNA. Like galectin-3, the staining of
Sm was excluded from the position of chromosomes. After separation of the daughter cells,

Sm was found throughout the cytoplasm and must be retransported into the nucleus.

74

.50: 300000200 0000000 000:: .0000 03000 0.00 0:00 0.000
000. 0:00 0003000 00:00.: .2300 030 Pb 000 ”00000—003 Amy ”000000.00 ADV ”00000000 6V ”0000000000 av
0000:0000 $3 ”03000 80 080000 00 00w000 00000> 300000.500 0:00 .720 0:3 330000000000 3 0032—8
003 <20 00000080000 .00 0000000 05. 00000—300005 000000-000 000w 0000w0.n000.00_E000€ .3 00000000
003 3000000 b00000 0.0 00 w0_005 00,—. 300.50 com” S Em-000 <Zm 0:08 0000:5030 00.00: :03
00000005 000 0000500000: .0000 0003 0000—0800 50000 0000 0000: Va 000 om 0003000 00.00.00 800.. 0:00
0000::—

0w00000 0:00 mph 00 0000000 Em 05 .00 E030: 2:580 0000000000c0§E§ 05 «0 0:00—0:00 ”h 0.53.:

75

 

 

DISCUSSION

The experiments documented in this study lead to two main conclusions. At the
technical/language level, it calls for caution in interpreting the qualitative description of
immunofluorescence patterns, such as diffuse, speckled, and punctate, derived from
separate studies, even from the same laboratory. Thus, the initial reports of galectin-3
localization, with intense fluorescence over the entire nucleus (10), and snRNP proteins,
with a speckled pattern (6, 20) were taken to imply that the two sets of proteins localized to
different subcompartments of the nucleus. We have shown in the present study, however,
that with high levels of endogenous protein antigen and a high titer antibody, one can
observe intense labeling of the nucleus, diffuse over the entire organelle (except for "black
holes") for both anti-galectin-3 and anti-Sm. By lowering the amount of antigenic target,
accomplished in the present study by extraction with ammonium sulfate, one can obtain a
speckled pattern for both anti—galectin—3 and anti-Sm. It appears, therefore, that rigorous
conclusions regarding similar versus different localization of specific protein antigens can
be best made by direct comparisons of their immunofluorescence patterns in parallel assays,

or more ideally, by double labeling experiments.

When such a double immunofluorescence was carried out with anti-galectin-3 and
anti-Sm, there was a one-to-one correspondence between the two speckled patterns yielded
by the two antibodies. Therefore, the second major conclusion, at the conceptual/scientific
level, is that galectin-3 and Sm are colocalized in the nuclear matrix. Studies at the light
microscopy level (21), as well as at the ultrastructural level (22, 23), have identified an RN P

network which contains components of the nuclear matrix. Moreover, the use of RN ase

76

77

inhibitors and ammonium sulfate extraction have significantly improved the preservation of
the RNP network in nuclear matrix preparations (17, 24). The nuclear matrix has been
proposed to play an important role in many activities, such as DNA replication,
transcription, and RNA processing (23). In particular, the interaction of newly synthesized
RNA with the nuclear matrix has been demonstrated by Xing and Lawrence (15), who
showed that Epstein-Barr virus primary transcripts were found in the nuclear matrix, and by
Huang et al. (25), who demonstrated that the majority of poly (A)+RNA is closely

associated with the matrix.

The speckled patterns observed for immunofluorescence of snRNP antigens and
non-snRNP splicing factors have been found to correspond, at the ultrastructural level, to
interchromatin granule clusters and perichromatin fibrils (14, 26). Interchromatin granule
clusters are not labeled with short pulses of [3H]uridine. Perichromatin fibrils, however, do
label with [3H]uridine (27) and anti-RNA polymerase II antibodies (28), suggesting that
they represent nascent transcripts at the sites of mRN A synthesis (29) and early events of

pre-mRNA processing (30).

Galectin-3 has recently been identified as a factor required for the splicing of pre-
mRNA in a cell-free assay (5). Indeed, electron microscopic analysis, using the same anti-
galectin-3 antibody preparation of this study (antiserum #33), yielded immunogold labeling
of interchromatic spaces and at the borders of condensed chromatin (perichromatic fibrils)
(31). This localization of galectin-3 to interchromatic spaces and perichromatic fibrils is

similar to the reported distribution of the non-snRNP splicing factors U2AF (30) and SC35

78

(14, 20), respectively. In addition, immunoelectron microscopy has also localized hnRNP

proteins mostly to the perichromatic fibrils (26).

Finally, it should be noted that our ultrastructural studies have also shown that there
was little or no anti-galectin-3 labeling of chromatin. Thus, galectin-3 does not appear to be
a DNA-associated protein. This conclusion is consistent with our present
immunofluorescence evidence that galectin-3 is excluded from the location of chromosome

during mitosis.

10.

11.

12.

13.

REFERENCES

Barondes, S. H., Cooper, D. N. W., Gitt, M. A., and Leffler, H. 1994, J. Biol. Chem.

_2_6_9_, 20807-20810.

Jia, S., and Wang, J. L. 1988. J. Biol. Chem. 26;, 6009-601 1.

Dreyfuss, G., Matunis, M. J ., Pinol-Roma, S., and Brud, G. C. 1993. Ann. Rev.
Biochem. 62, 289-321.

Bennett,M., and Reed, T. 1993. Science LCZ, 105-108.

Dagher, S. R, Wang, J. L., and Patterson, R. J. 1995. Proc. Natl. Acad. Sci. USA 2,
1213-1217.

Spector, D. L. 1993. Ann. Rev. Cell Biol. 2, 265—313.

4 Carmo-Fonesca, M., Pepperkok, R., Sproat, B. S., Ansorge, W., Swanson, M. S., and

Lamond, A. I. 1991. EMBO J. _IQ, 1863-1873..

Huang, 8., and Spector, D. L. 1992. Proc. Natl. Acad. Sci. USA 89, 305-308.
Moutsatsos, I. K., Davis, J. M., and Wang, J. L. 1986. J. Cell Biol. 1_02, 447-483.
Moutsatsos, I. K., Wade, M., Shindler, M., and Wang. J. L. 1987. Proc. Natl. Acad.
Sci. USA &, 6452-6456.

Laing, J. G., and Wang. J. L. 1988. Biochemistry 22, 5329-5334.

Agrwal, N., Sun, Q., Wang, S. Y., and Wang, J. L. 1993. J. Biol. Chem. £8, 14932—
14939.

Burton, Z. R, Killeen, M., Sopta, M., Ortolan, L. G., and Greenblatt, J. 1988. Mol. Cell

Biol. _8_, 1602-1613.

79

14.
15.
16.
17.
18.

19.

20.
21.

22.

23.
24.

25.

26.
27.

28.

29.

30.

80

Spector, D. L., Fu, X.-D., and Maniatis, T. 1991. EMBO J. 1Q, 3467-3481.

Xing, Y., and Lawrence, J. B. 1991. J. Cell Biol. £2, 1055- 1063.

Spector, D. L. 1990. Proc. Natl. Acad. Sci. USA 81, 147-151.

Fey, E. G., Krochmalnic, G., and Penman, S. 1986. J. Cell $01.19;, 1654-1665.
Agrwal, N., Wang, J. L., and Voss, P. G. 1989. J. Biol. Chem. 2_6_4_, 17236-17242.
Cowles, E. A., Agrwal, N., Anderson, R. L., and Wang, J. L. 1990. J. Biol. Chem.
_2___5_, 17706-17712.

Jimenez-Garcia, L. F., and Spector, D. L. 1993. Cell 23, 47—59.

Smetana, K., Steele, W. J ., and Busch, H. 1963. Exp. Cell Res. _3_l, 198-201.
Berezney, R., and Coffey, D.S. 1974. Biochem. Biophys. Res. Commun. 6_Q, 1410-
1419.

Berezney, R. 1991. J. Cell. Biochem. 4_7, 109-123.

Belgrader, P., Siege], A. J ., and Berezney, R. 1991. J. Cell Sci. 3, 281-291.

Huang, S, Deerinck, T. J ., Ellisman, M. H. and Spector, D. L. 1994. J. Cell Biol. _12_6,

877-899.

Fakan, S., Leser, G., and Martin, T. E. 1984. J. Cell Biol. %, 358-363.

Fakan, S., and Puvion, E. 1980. Int. Rev. Cytol. fl, 255-299.

Spector, D. L., O'Keefe, R. T., and J imenez-Garcia, L. F. 1993. Cold Spring Harbor
Symp. Quant. Biol. _5_8, 799-805.

Fakan, S., Puvion, E., and Spohr, G. 1976. Exp. Cell Res. 2, 155-164.

Zhang, G., Taneja, K. L., Singer, R. H., and Green, M. R. 1994. Nature £2, 809-812.

81

31. Hubert, M., Wang, S.-Y., Wang, J. L., Seve, A.-P., and Hubert, J. 1995. Exp. Cell Res.,

' submitted for publication.

CHAPTER III

Coimmunoprecipitation of Sm Polypeptides of snRNP with Galectin-3 by a

Monoclonal Antibody Directed against the Lectin

82

SUMMARY

In previous studies, it had been shown that the galactose/lactose-binding protein,
designated galectin-3, was a required factor in the splicing of pre-mRNA as assayed in a
cell-free system and that its nuclear localization, as detected by immunofluorescence
microscopy, was similar to the distribution of the small nuclear ribonucleoproteins
(snRNPs). In the present study, nuclear splicing extracts derived from HeLa cells were
subjected to immunoprecipitation with anti-Mac 2, a monoclonal antibody specifically
directed against galectin-3. Irnmunoblotting of the anti-Mac 2 precipitate with polyclonal
rabbit anti-galectin-3 revealed the presence of the lectin. Irnmunoblotting of the same
precipitate with an autoimmune serum reactive with the snRNP B (Sm B) and D (Sm D)
polypeptides revealed that a fraction of the Sm B in the nuclear extract was coprecipitated
with the lectin by the monoclonal antibody. Finally, RNA components, one of which
corresponded in mobility to U2 snRNA, could also be identified in the anti-Mac 2
precipitate. The specificity of this coimmunoprecipitation was ascertained by comparing
the results obtained with anti-Mac 2 versus those observed with precipitation by
autoimmune anti-Sm (positive control) and by an isotype-matched monoclonal antibody
against the transferrin receptor (negative control). Moreover, the coprecipitation of Sm B
with galectin-3 by anti-Mac 2 was dependent on the presence of the lectin; nuclear extracts
depleted of galectin-3 by prior adsorption on a lactose affinity resin failed to yield Sm B in
the immunoprecipitate of anti-Mac 2. The coprecipitation of Sm B by anti-Mac 2 was not

perturbed by the addition of lactose or by prior treatment of the nuclear extract with

83

84
ribonuclease. These results suggest that at least a fraction of the Sm B polypeptides,

possibly on U2 snRNP, may interact with galectin-3 through protein-protein interactions.

INTRODUCTION

Galectin-3 is a galactose/lactose (Lac)-binding protein found in the nucleus and
cytoplasm of a variety of cell types (1). It has been implicated as a required factor for pre-
mRNA splicing in a cell-free assay (2). Nuclear extracts capable of carrying out in vitro
splicing contain galectin-3, as revealed by immunoblotting. When these splicing competent
extracts are subjected to affinity adsorption on Lac-containing beads, the unbound fraction
is depleted of the lectin. Concomitant with the depletion of galectin-3, the unbound fraction
also loses the in vitro splicing activity. Reconstitution of the depleted extract with
recombinant galectin-3, purified from an E. coli expression system, restores the splicing

activity in a dose-dependent fashion.

In the course of these studies, it was also observed that in a gel mobility shift assay,
splicing competent nuclear extracts yielded both ATP-independent H complex, as well as
ATP-dependent, higher order complexes (designated A, A', and B) intermediate in the
formation of the spliceosome. In contrast, nuclear extracts depleted of galectin-3 and
splicing activity yielded only the ATP-independent H complex (2). These results suggested
that galectin-3 may interact with components that constitute the spliceosome. Coupled with
the demonstration that the association of galectin-3 with the cell nucleus was sensitive to
ribonuclease (RNase) treatment (reference 3 and chapter II), the possibility is raised that the
lectin may interact with a RNA-containing component. However, the sequence of galectin-

3 exhibits no obvious RNA recognition motif(s) (4). Nor have attempts to demonstrate an

85

86
association of galectin-3 with RNA, using cross-linking by irradiation with ultraviolet light,

yielded any evidence for direct galectin-3-RNA interactions (S. F. Dagher and R. J.

Patterson, unpublished observations).

One approach to discern if galectin-3 interacts with any components of the
splicesome is to determine if immunoprecipitation of galectin-3 using a highly specific
monoclonal antibody can simultaneously coprecipitate known spliceosomal components.
We now report that, indeed, the Sm B (and perhaps Sm D as well) protein of nuclear small
ribonucleoprotein (snRNP) particles is precipitated by a galectin-3 specific monoclonal
antibody. This coimmunoprecipitation is dependent on the presence of galectin-3 in the

mixture subjected to the immunoprecipitation protocol.

MATERIALS AND METHODS

Cell Cultures. Mouse 3T3 fibroblasts were obtained from American Type
Culture Collection (Rockville, MD). The cells were grown as monolayers in Dulbecco
modified Eagle’s Medium containing 10% calf serum, 100 U/ml penicillin and 100 U/ml
streptomycin at 37 0C in a humidified atmosphere of 10% C02. Cells grown at a density
less than 5 x 104 cells /cm2 were proliferative and were harvested to generate nuclear
extracts. In some experiments, cells at low density were arrested by removal of serum
and maintenance in medium containing 0.2% calf serum for 48 hr (quiescent cells). Upon
readdition of serum (10%), the cells were reactivated and were cultured for another 16

hours (5).

HeLa S3 cells were cultured in spinner flasks in Minimum Essential Medium
containing 5% fetal bovine serum, 100 U/ml penicillin and 100 U/ml streptomycin at 37
°C in a humidified atmosphere of 5% C02. They were grown to 4 to 6 x 105 cells per ml

prior to harvesting for nuclear extract preparation.

Antibodies and Affinity Columns. A rat monoclonal antibody was developed
against the Mac 2 antigen (6), which has been shown to be galectin-3 (7). The hybridoma
line producing this monoclonal antibody (M3/38.1.2.8.HL.2) was obtained from the
American Type Culture Collection (TIB 66) (Rockville, MD). The hybridoma cells were
cultured in serum—free medium (RPMI 1640 containing Nutridoma SP (BMB, IN)). After
centrifugation to pellet the cells, supematants from the cultures were pooled, subjected to

87

88
ammonium sulfate precipitation (45% of saturation), dialyzed against phosphate-buffered

saline (PBS; 0.13 M NaCl, 5 mM sodium phosphate, pH 7.5) exhaustively, and stored in
aliquots at a concentration of 25 ug/ml. This antibody preparation, hereafter designated as

anti-Mac 2, consists of rat IgGZaK.

An isotype matched rat monoclonal antibody (IgGZaK), directed against the
transferrin receptor (TfR), was used as a control for anti-Mac 2. The hybridoma (R17
217.1.3) producing the rat anti-TfR was also obtained from American Type Culture
Collection (TIB 219) and cultured in serum-free medium. Anti-TfR was isolated in the

same fashion as described for anti-Mac 2.

Polyclonal rabbit anti-galectin-3 antiserum was obtained from rabbit #32. This New
Zealand White rabbit was immunized with recombinant galectin-3. The characterization of
the specificity of this antiserum had previously been documented (8). Human autoimmune
serum reactive with the Sm antigens of snRNPs (ENA anti—Sm) was purchased from The
Binding Site (San Diego, CA). Lactose immobilized on 6% beaded agarose (Lac-agarose)
was purchased from Sigma (St. Louis, Mo), as was agarose derivatized with cellobiose
(Cello-agarose) used as a control. Sepharose CL4B and protein G-agarose beads were

purchased from Pierce (Rockford, IL)

Buffers. Buffers used for nuclear extract preparation are designated as follows:

 

(i) buffer A contains 10 mM I-IEPES pH 7.9, 1.5 mM MgClz, 10 mM NaCl, 10 mM KCl,

0.5 mM phenylmethylsulfonyl fluoride (PMSF), and 0.5 mM dithiothreitol (DTT): (ii)

89
buffer C contains 20 mM HEPES pH 7.9, 25% (v/v) glycerol, 0.42 M NaCl, 1.5 mM

MgC12, 0.2 mM EDTA, 0.5 mM PMSF, 0.5 mM DTT; and (iii) buffer D contains 20 mM
HEPES pH 7.9, 20% (v/v) glycerol, 0.1 M KCl, 0.2 mM EDTA, 0.5 mM PMSF, and 0.5

mM DTT. Both DTI‘ and PMSF were added fresh to the buffers just before use.

Buffers used for nucleoplasm preparation and cesium sulfate gradient
sedimentation are designated as follows: (i) RSB contains 10 mM Tris , pH 7.5, 10 mM
NaCl, 1.5 mM MgC12; and (ii) cesium sulfate stock solutions which contain cesium

sulfate in RSB with final densities of 1.20 and 1.75 g/ml.

Washing buffers for immunoprecipitation and Lac depletion are designated as: (i)
buffer 1 containing 10 mM Tris, pH 7.5, 150 mM NaCl, 0.5% NP—40, 0.2 mM EDTA;
and (ii) buffer 2 containing 10 mM Tris, pH 7.5, 500 mM NaCl, 0.5% NP—40, 0.2 mM

EDTA; and (iii) buffer 3 containing 10 mM Tris, pH 7.5.

Nuclear Extract Preparation. Nuclear extract was prepared according to the
procedure of Dignam et a1. (9). Briefly, mouse 3T3 cells cultured as monolayers were
washed once with PBS containing 0.2% EDTA at 37 0C. The cultures were then
incubated for 3 minutes with PBS containing 0.1% trypsin and 0.2% EDTA. Cells
detached from the flasks were then harvested and washed twice with ice-cold PBS.
Human HeLa cells grown in spinner flasks were harvested from cell culture media and
washed twice with ice-cold PBS. In some experiments, 3T3 cells and HeLa cells were

metabolically labeled with 100 ltCi/ml of [35S]methionine (NEN, MA) for 12-16 hours.

90
The volumes of the packed cell pellets were determined. The harvested cells were

then suspended in five volumes of ice-cold buffer A and allowed to stand on ice for 10
minutes. The cells were collected by centrifugation and suspended in two volumes of
buffer A and lysed by 10 strokes in a Dounce homogenizer at 4 °C. The homogenate was
checked for cell lysis under a microscope and then centrifuged at 4 °C for 10 minutes at
1,500 x g to pellet nuclei. The supernatant was carefully decanted and the pellet was
subjected to a second centrifugation for 20 minutes at 25,000 x g to remove residual
cytoplasmic material and this pellet was designated as crude nuclei. The crude nuclei
were resuspended in 3 ml of buffer C per 109 cells and then homogenized with 10 strokes
in a Dounce homogenizer. The nuclei suspension was stirred gently with a magnetic
stirring bar for 30 minutes and then centrifuged for 30 minutes at 25,000 x g. The
resulting clear supernatant was designated as high salt nuclear extract and frozen at -70
°C. The high salt nuclear extract was dialyzed against 50 volumes of buffer D twice prior
to the subjected experiments. In some experiments, the high salt nuclear extract was

dialyzed against buffer D containing 0.3 M lactose.

Immunoprecipitation and Lac-depletion of Nuclear Extract. Nuclear extract

was precleared by incubating with Sepharose beads for 1 hour at 4 °C. The unbound
fraction of this preclearing step was subjected to immunoprecipitation or Lac-agarose
adsorption. ENA anti-Sm antibody (10 pl) and 5.0 pg of anti-Mac 2 antibody were
preabsorbed with a 50 [.1] suspension of protein G-agarose beads at room temperature for

30 minutes, and the antibody bound beads were washed three times with the buffer D. In

91
a typical precipitation, 50 [1] of the precleared nuclear extract was incubated with the anti-

Sm or anti-Mac 2 bound protein G beads for 1 hour at 4 °C. In Lac-agarose depletion
experiment, 50 ul of Lac-agarose beads was used instead. The bound fractions were
washed three time with buffer D. The material remaining bound to the beads after the
washing steps was eluted with SDS-PAGE sample buffer (10) and subjected to SDS-
PAGE (12.5% acrylamide; 200 volts (V) for 30 minutes, followed by 300 V for 2 hours),

followed by immunoblotting analysis.

For immunoblotting analysis, ENA anti-Sm was used at l:10,000 dilution; rabbit
anti-galectin-3 was used at 1:10,000 dilution; and rat anti-Mac 2 was used at 1:5,000
dilution. All three primary antibodies were diluted in TBS (10 mM Tris, 0.5 M NaCl, pH
7.5). Horseradish peroxidase conjugated goat anti-human, sheep anti-rabbit, and goat
anti-rat antibodies (BMB) were used at 1:5,000 dilution in T—TBS (TBS containing 0.05%
Tween 20) as secondary antibody. Proteins resolved on SDS-PAGE were
electrophoretically transferred to Immobilon-P (Millipore) membrane in a transfer buffer
containing 0.025 M Tris, 0.19 M glycine and 20% methanol for 3 hour at 50 milliamp
(mA), 9 hours at 100 mA and 3 hour at 200 mA sequentially; the membrane was briefly
rinsed twice with T-TBS and blocked with 5% dry milk in T-TBS for at least 1 hour at
room temperature. After briefly being washed with T-TBS twice, the membrane was then
incubated with primary antibody for 1 hour at room temperature. The membrane was
then rinsed with T-TBS twice and washed extensively with T-TBS (15 minutes once, then

10 minutes 4 times). The washed membrane was incubated with horseradish peroxidase

92
conjugated secondary antibody at room temperature for 1 hour. The blot was rinsed and

washed as described above. Proteins on the membrane were detected by ECL Western
blotting detection reagents (Amersham, England) and revealed by exposing the

luminescence to a sheet of autoradiography fllm (Hyperfllm-ECL (Amersham)).

Immunoprecipitation and Lac-agarose adsorption were also carried out on
[3SS]methionine labeled nuclear extracts at high ionic strength (in buffer C). The same
procedure was used, as described above, except the bound fractions were subjected to the
following washing steps: (i) twice with buffer 1; (ii) twice with buffer 2; and (iii) twice
with buffer 3. The material remaining bound to the beads were then analyzed by SDS-

PAGE as detailed above and the radioactive polypeptides were revealed by fluorography.

Analysis of RNA Commnents in Bound Fractions of Immunoprecipitate and

Lac-Adsomtion. HeLa cells were cultured as described above and labeled with 50
uCi/ml of 32P04 (NEN, MA) for 12 hours. Nuclear extract generated from these 32P-
labeled cells was subjected to immunoprecipitation and Lac-agarose adsorption. After
the washing steps, the bound fraction was resuspended in 100 1.11 of TE (10 mM Tris, pH
8.0, 10 mM EDTA) containing 0.3 M sodium acetate (NaOAc, pH 5.2) and 1 jig/ml
glycogen. RNA was extracted with 100 u] of phenol. The aqueous phase was recovered
and the residual phenol was extracted by adding 100 ul of chlorofonn/isoamyl alcohol
(24: 1, v/v). RNAs in the aqueous phase were precipitated with 600 u] of ethanol at - 70

°C overnight. The RNA pellet was rinsed once with ice-cold ethanol (70%) and air dried.

93
The extracted RNAs were subjected to electrophoresis in a polyacrylamide—urea gel (8%

acrylamide (acrylamide/bisacrylamide 38: 1; 8.3 M urea)). The gel was dried and the

resolved RN As were revealed by autoradiography.

Cesium Sulfate Gradient Fractionation and Immunoprecipitation.

Nucleoplasm was prepared from 3T3 cells 16 hours post serum stimulation of quiescent
cultures (48 hours in 0.2% calf serum, followed by 16 hours in 10% calf serum). The
harvested cells were resuspended in RSB and kept on ice for 15 minutes to allow the cells
to swell. The cells were lysed through 20 strokes in a Dounce homogenizer and the
nuclei were collected by centrifugation at 1,500 x g for 5 minutes at 4 °C. The nuclear
pellet was resuspended in RSB containing 2 mM PMSF and 2 mM vanadyl-
ribonucleoside complex (BRUGIBCO, MD) and ruptured by sonication for 15 seconds
four times. The material was then layered on a 30% sucrose/RSB and centrifuged at
4,500 x g for 15 minutes at 4 °C in a Sorvall HB-4 rotor (DuPont, DE) to remove
chromatin and other insoluble material (11, 12). The resulting opaque interface of the
30% sucrose cushion was collected as nucleoplasmic material and used for cesium sulfate

gradient fractionation.

Preformed linear cesium sulfate gradients (4.5 ml) were prepared in cellulose
nitrate tubes (Beckman, CA) from stock solutions having initial densities of 1.20 g/ml
and 1.75 g/ml in RSB. Samples of nucleoplasm (0.5 ml) were layered and the gradients
were centrifuged at 112,000 x g for 60 hours at 15 °C in a SW50.1 rotor (Beckman, CA).

These gradients were then fractionated with a Beckman fraction recovery system (0.5 ml

94
per fraction) and the profile of densities of the gradient was determined by weighing 100

it] portions of individual fraction at room temperature (11, 12). The individual fractions
were diluted with two volumes of RSB and then subjected to immunoprecipitation by
ENA anti-Sm antibody or anti-Mac 2 antibody. The immunoprecipitates were washed
three times with RSB, subjected to SDS-PAGE, and western blotting by anti-Sm or rabbit

anti-galectin-3.

RESULTS

Immunoprecipitation of Nuclear Extracts of 3T3 Cells by Autoimmune Anti-
Sm and by Monoclonal Anti-Mac 2. Nuclear extracts were prepared from mouse 3T3
fibroblasts that had been metabolically labeled with [3531methionine The extract was
passed over a Sepharose column and the unbound fraction was then subjected to immuno-
precipitation with ENA anti-Sm. The bound fraction of both the pre-clearing column
(Sepharose) and the anti-Sm column were subjected to SDS-PAGE. Fluorographic analysis
of the [3SS]methionine labeled polypeptides revealed that the pre-clearing column bound
one predominant polypeptide of M, 43,000 (Figure 1a, lane 1); this polypeptide has been
identified to be actin, on the basis of immunoblotting with an antiserum raised against calf
thymus actin (data not shown). The immunoprecipitate of the anti-Sm-protein G-agarose
beads contained many radiolabeled bands, including polypeptides that correspond in
molecular weight to the snRNP polypeptides A (33 kDa), B (28 kDa), C ( 22 kDa), D (18
kDa), and E/F/G (1 1-13 kDa) (Figure 1a, lane 2). When the same material (bound fraction
of the anti-Sm precipitate) was subjected to immunoblotting with ENA anti-Sm, prominent
immunoreactive bands were observed for Sm B and Sm D, as well as a conglomerate of
high molecular weight bands (Figure 1c, lane 2). These results are consistent with the
reported properties of this autoimmune serum (13, 14), whose principal epitopes are found
on core polypeptides B and D, with some reactivity against U1 snRNP, which contains its

specific polypeptides, A and C. In any case, the results illustrate that although anti-Sm

95

96

.2»: be
.5 83323.. a 28%» ab 8:235 do 828a 2: e5 :2 2: co 3..»sz 2e eta. waged Became
-mzmfi 05 e6 558%.: .«o 2028““ of. Seawaé 589:. N 82-25 Gov Ea ”caesium A8 ”80.3w.“
-0 £265 Eméfiézm AS ”omgzmom CV :23 vegetoq 89$ mcoufiafioaocsfifim .313 22.8 E .Em
-35 <Zm 53, 832.255 3 EB ”Wazoo—«mica :38 .23 83055.5 3v ”monnaoax—oa 3:82—68 05
w==3>2 883.52”. A3 .wfizofiocsEE .8 Pap—mega .3 332—8 .AoEEEcCoa bend: m0<m-mDm E
3388 263 m0§§02a0=38§ 05 E moc_&o§_om 05 was :o_§_&oemo:=EE_ 2 3.8.33 203 3255
2:. 6555065an 5; 3.32 bag—390E mama—nob: n.5,. 8:08 :5: taxman 6.83 88.58 30.022
.N 32.3.3 3:22:35

.3 EB Emég SSE—£35 .3 3.3 mph we $3.55 .322:— ue :esfimnmuoaegEE— "a 0.53%

97

1

1

 

 

. 98
immunoblots Sm B and Sm D, the entire complement of snRNP polypeptides, A through G,

is immunoprecipitated by the autoimmune serum as an RNP complex.

The same [3SS]methionine labeled nuclear extract was subjected to pre-clearing and
immunoprecipitation with anti-Mac 2, a monoclonal antibody reactive against galectin-3
(7). The bound fraction of the pre-clearing column (Sepharose) again showed actin to be
the predominant polypeptide (Figure 1a, lane 3). On the other hand, the anti-Mac 2
immunoprecipitate yielded many radiolabeled polypeptides on SDS-PAGE and
fluorography (Figure 1a, lane 4). Two of these polypeptides, at the low molecular weight
end, can be identified as galectin-3 (33 kDa) (Figure 1b, lane 4) and Sm B (Figure 1c, lane
4), on the basis of immunoblotting with a rabbit anti-galectin-3 antiserum (Figure 1, panel
b) and ENA anti-Sm (Figure 1, panel c), respectively. These results are provocative in two
respects: (a) when anti-Mac 2 immunoprecipitates galectin-3 (Figure 1a and 1b, lane 4), it
coimmunoprecipitates a small fraction of the nucleoplasmic Sm B molecules (Figure 1a and
1c, lane 4); and (b) in contrast, when anti—Sm immunoprecipitates the snRNPs, no galectin-

3 is apparently coprecipitated Gigure 1a and 1b, lane 2).

These conclusions are corroborated by experiments in which nuclear RNPs are first
enriched by cesium sulfate gradient sedimentation. The gradient used ranged in density
from 1.18 g/ml at the top to 1.75 g/ml at the bottom. Naked protein, such as recombinant
galectin-3 (p 51.20 g/ml) collected at the top of the gradient. Nucleoplasm from 3T3 cells

was fractionated over such a gradient and the individual gradient fractions were subjected to

99

Figure 2: Profiles of the density and immunoprecipitation patterns of the

individual fractions derived from cesium sulfate gradient

sedimentation of nucleoplasm.

Nucleoplasm was prepared from serum-stimulated (16 hours after addition of
10% calf serum to quiescent cultures) 3T3 fibroblasts and subjected to
fractionation on a cesium sulfate gradient. The density was determined by
weighing 100 [.11 aliquot of each fraction. The individual fractions were diluted
with RSB and subjected to immunoprecipitation with: (A) anti-Mac 2; and (B)
ENA anti-Sm. For both (A) and (B), the polypeptides in the immunoprecipitates
were resolved by SDS-PAGE (12.5% acrylamide) and then subjected to
immunoblotting, in sequence, with anti-galectin-3 and ENA anti-Sm. The
positions of migration of galectin-3, as well as Sm B and Sm D polypeptides of
snRNP, are revealed by immunoblotting of nucleoplasm with both antibodies
and are highlighted on the right. The letter X on the left designates a
polypeptide that is observed in anti-Sm immunoprecipitates immunoblotted by
anti-galectin-3. Its position of migration is distinct from that of galectin-3.

—.__ _m. l-L—n

)f

ltcd

llales

s of
as

100

 

 

1234.07.01.

 

 

 

 

immunoprecipi
with anti-galec
indeed. the ma;
However, thert
fraction 4 (~1.I
anti-Mac 2 hat

in the anti-M 31

When
anti-Sm (Fl gu
galectin-3. Sn'
from 1.20 gm:
prominent Sir
with anti-gale

Pm’iously (_Fj

In the
anti-Sm and i
23» fraction 4
distinct’ fTOm
pOIypePIlde i5
imm”"0blom

HeLa Cel] $3.8

. 101
immunoprecipitation with anti-Mac 2 (Figure 2A), followed by sequential immunoblotting

with anti-galectin-3 and anti-Sm. Irnmunoblotting with anti-galectin-3 revealed that,
indeed, the majority of the galectin-3 was found in fractions 1 and 2 (~1.20 g/ml).
However, there was a population of galectin—3 found in the anti-Mac 2 precipitate of
fraction 4 (~1.34 g/ml). This fraction, upon immunoblotting with anti—Sm, revealed that
anti-Mac 2 had, again, coprecipitated Sm B (Figure 2A). In this case, Sm D was also found

in the anti-Mac 2 immunoprecipitate.

When the individual gradient fractions were subjected to immunoprecipitation with
anti-Sm (Figure 2B), followed by sequential immunoblotting with anti-Sm and anti-
galectin-3, Sm B was found spread over many fractions of the gradient, ranging in density
from 1.20 g/ml to 1.50 g/ml. The fraction with the majority of Sm was fraction 4, where
prominent Sm B and Sm D were revealed by immunoblotting (Figure 2B). Irnmunoblotting
with anti-galectin-3 failed to reveal the lectin in the anti-Sm precipitate, as was found

previously (Figure 1a and 1b, lane 2).

In the course of these studies, we observed a band that was immunoprecipitated by
anti-Sm and immunoblotted by anti-galectin-3. This band is labeled as band X in Figure
2B, fraction 4. The mobility of this band on SDS-PAGE was slightly slower and therefore,
distinct, from that of galectin-3 (33 kDa) (Figure 2B). Although the identity of this
polypeptide is not known, the phenomenon of its being immunoprecipitated by anti-Sm and
immunoblotted by anti-galectin-3 has been reproducibly observed, in both the 3T3 and the

HeLa cell systems (see Figure 3 and Figure 4).

Immll'
carried out on
extracts were
extracts used
splicing extra
was observec
Westem blot
(Figure 3b. 1
Sm B is actL
extract was
immunopre.
32L lane 5").

revealed th,

Sex
served as tl
revealed th
anti-sm in
galectins
howevfl, ;
Sm immm

is“Filed n

102
Immunoprecipitation of HeLa Cell Splicing Extracts. A similar analysis was

carried out on nuclear extracts derived from human HeLa cells. In this case, the nuclear
extracts were prepared according to the procedure of Dignam et al (9), equivalent to the
extracts used to demonstrate a role for galectin-3 in a cell—free splicing assay (2). When
splicing extracts were Western blotted with anti-galectin-3, a single polypeptide (29 kDa)
was observed (Figure 3a, lane 1), corresponding to human galectin-3. When subjected to
Western blotting with ENA anti-Sm, this nuclear extract yielded two sets of doublets
(Figure 3b, lane 1), corresponding to Sm B (28 kDa) and Sm D (18 kDa). In human cells,
Sm B is actually a doublet, Sm B/B'; Sm D is also a doublet, Sm D/D' (15). The splicing
extract was immunoprecipitated with anti-Mac 2. Western blotting of this
immunoprecipitate with anti-galectin-3 yielded the expected galectin-3 polypeptide (Figure
3a, lane 5). Western blotting of the same anti-Mac 2 immunoprecipitate with anti-Sm

revealed the Sm B doublet prominently, as well as traces of Sm D (Figure 3b, lane 5).

Several crucial controls were performed in parallel. First, anti-Sm precipitation
served as the positive control. Western blotting of the anti-Sm precipitate with anti-Sm
revealed the expected Sm B and Sm D bands (Figure 3b, lane 4). Western blotting of the
anti-Sm immunoprecipitate with anti-galectin-3 failed to reveal a band that matched human
galectin-3 exactly (Figure 3a, lane 4). As noted previously in Figure 2B (fraction #4),
however, a polypeptide designated as band X was observed in anti-galectin-3 blots of anti-
Sm immunoprecipitates (Figure 3a, lane 4). Second, the monoclonal antibody anti-TrR was

isotyped matched to anti-Mac 2 and served as a negative control. The immunoprecipitates

103

 

63:»:sz 2a deuce—aw mm :03 ma .0 Em .m Em me 5:83.: we 302.com
2:. 63:98:“ N 82.25 .8 Gv 28 ”233603 Eméfi SE5: A3 6538er ”Eaten “8 Amv 688mm-u3
65 Co c0393 9502 Amy 3855 9:213 =8 30E :v ”803 052 some 5 BEES 05 A013 388 £ .N 032
.93 :2 .23 wE:oEo==E§ 8v c5 55.93 <Zm .23 wage—90:35.5 3V ”Ann’s mascofiwéca :38 £5
wE:oBo==EE~ A3 .mEzoEoSEg .3 330:8 .AoEEflboa bend C ”#3??QO .3 3388 803 20:25
953 05 5 0230938 65 23 Q datum E =2§E62qo=58§ no 5:883 .025? 9 688.33 203
36838 2:. .5 .3 B Enema e6 226889 05 9 @68an £8 303 Sec 3335 203 88.30 memozmm
.N 32-5.3 3:22:58

.3 can Emtuca 0:58.533" .3 $2228 wag—am =3 «40: we :eflflimooEegE—E an charm

104

 

of anti-T3 f'
(Figure 3a. 1:
precipitate g
anti-galectin
antigens we!
samples. (i)
anti-Sm pre:
anti-Mac 2.
polypeptide
was due to i
failed to rec

or in the an

Eff:
anti~Mac 2
COITObOFate
CODditiOng 1
0’ WEre [he
galectin fen
adSOrpthn ‘
agarose Unt

105
of anti-TfR failed to yield any prominent bands when immunoblotted with anti—galectin-3

(Figure 3a, lane 3) and anti-Sm (Figure 3b, lane 3). Third, Lac-agarose was used to
precipitate galectin—3 through saccharide-binding. Western blotting of this precipitate with
anti-galectin-3 revealed the 29 kDa polypeptide of the lectin (Figure 3a, lane 2) but no Sm
antigens were observed upon blotting with anti-Sm (Figure 3b, lane 2). Finally, all of the
samples, (i) splicing extract; (ii) Lac-agarose bound fraction; (iii) anti-TfR precipitate; (iv)
anti-Sm precipitate; and (v) anti-Mac 2 precipitate, were subjected to Western blotting with
anti-Mac 2. This was done to test whether anti-Mac 2 interacted directly with the Sm B
polypeptide or whether the observed coprecipitation by anti-Mac 2 of Sm B with galectin-3
was due to interactions between the two polypeptides. The results showed that anti-Mac 2
failed to recognize the Sm B (and Sm D) polypeptide in splicing extracts (Figure 3c, lane 1)

or in the anti-Sm precipitate where Sm B and Sm D are enriched (Figure 3c, lane 4).

Effect of Galectin-depletion on Coimmunoprecipitation. The conclusion that
anti-Mac 2 can precipitate Sm B only through the latter's interaction with galectin-3 was
corroborated by experiments in which the lectin was depleted from the splicing extract. The
conditions used for depletion, adsorption on Lac—agarose under high ionic strength (Buffer
C), were the same as those used to show the concomitant loss of splicing activity with
galectin removal from the splicing extract (2). The nuclear extract was subjected to
adsorption on cellobiose-agarose in parallel as a control. Both Lac-agarose and cellobiose-
agarose unbound fractions were dialyzed against buffer D. Irnmunoblotting with anti-

galectin-3 showed that the unbound fraction of the cellobiose adsorption contained galectin-

3mam4
In contrast.

4, lanes 3 a:
Sm. The UI
doublets (F
derived fr01
lane 2). In

found in the
to the cone]
precipitatio
Thus, all of

bflWCCn g3

Coi
Ionic Strer
msmyn
binding of 1
beads revea
galectin-3 (g
lime or no é
5’“ B and s

106
3 (Figure 43, lane 1), which can be immunoprecipitated by anti-Mac 2 (Figure 4a, lane 2).

In contrast, the corresponding fractions of the Lac adsorption showed no galectin-3 (Figure
4, lanes 3 and 4). The same fractions were subjected to Western blotting with EN A anti-
Sm. The unbound fraction of both the cellobiose and Lac adsorptions yielded the Sm B
doublets (Figure 4b, lanes 1 and 3). However, only the Sm B in the unbound fraction
derived from the cellobiose beads could be immunoprecipitated by anti-Mac 2 (Figure 4b,
lane 2). In contrast, the Sm B in the unbound fraction derived from the Lac beads was not
found in the anti-Mac 2 precipitate (Figure 4b, lane 4). These results provide strong support
to the conclusion that anti—Mac 2 does not recognize Sm B directly and that the observed
precipitation of Sm B by the monoclonal antibody depended on the presence of galectin-3.
Thus, all of the coimmunoprecipitation data are interpreted to implicate an interaction

between galectin-3 and the Sm B polypeptide(s).

Coimmunoprecipitation Experiments in the Presence of Lactose and at High
Ionic Strength. The experiment shown in Figure 3 was repeated in the presence of Lac
(0.3 M). The presence of the Lac in the HeLa nuclear extract completely inhibited the
binding of galectin-3 to the Lac-agarose beads. The bound fraction of the Lac-agarose
beads revealed neither galectin-3 nor Sm polypeptides upon Western blotting with anti-
galectin-3 (Figure 5a, lane 2) or with anti-Sm (Figure 5b, lane 2). The presence of Lac had
little or no effect on the immunoprecipitation by anti-Sm. Western blot by anti-Sm revealed
Sm B and Sm D (Figure 5b, lane 4), just as was observed in the absence of Lac (Figure 3b,

lane 4). Most importantly, Western blotting of the anti-Mac 2 precipitate showed galectin-3

107

1.1‘ III.

caupcaup

62:3sz 2a m Em 9a 338% a0 SEEE Mo 822:8 2F .03 25 so: we as are?
2: mo EOESEBEQ N ”ca—2&5 mo c0308.. UESQE Eocfioém 03 05 we caveat EEBEAQ ”8052—8
9.5 8: Ba 35 _«thE 2: ho cocsmafien N 82-55 me .5325 959:8 ”noun—onus 8052—3 05 .8
.8:ch 253:2 C ”225 052 some E moEEam .SV 28 A3 283 E .Em-_Em <Zm 53, wage—nocsEEHEV
”2:83-25 fine 55 3:532:55? .wszozogeé 3 832.8 .3282?“ and: mo<a-mom
,3 3288 203 sonnet canon 05 E onEomboa 05 can N 022-35 53, co_§E_ooao==EE_ 9 c8833
:05 203 222983 035 mo 2288.2 253:: 25. .omobwwabj .5 8033683230 FEB coup—08¢
bEEn 2 880.33. 20>» $8.58 05 98 U Stan E 2.8 30: E0: 3835 203 38.88 wEBEm

.N $2.35 an m Em he :o:8_&uo.ae:=EE_ 2: :e Ease—now naive—aw ue «coho 2; 3 ensure

108

 

109

.Uofiwznwfi 9a fiasco—mm mm :03 an .O Em .m Em mo :23qu no mcoEmom 2F .BSEBBQ N 82-35 “8
AB 28 63:98:“ Eméqm 5:5: Avv U§§8a Ehéca “8 GS "omoawaéj mo 839$ wagon NV 39253
mag—am =8 30: CV “803 0:3 nomo E 8.95% A6 new AS £258 5 .Emé—a <Zm 5m? 390505.55 3V
”Wazoo—«wig E38 55 wage—Dogfish E .wcfioEocsEg 3 330:8 Antwan—boa §mNC m0<m-mnm
3 3288 803 30:23 canon on» E moanmbom 05 98 Q Bram E :o_§_a_oemo::EE_ 8 5398?
385a 8 380.33 803 30838 05 EB mum—«E macaw D Stan E E md we cougcoocoo _acc a 9 vows—2:
83 23 .av .E 8 Eacwa ho c.5385 05 8 @5508 230 Sam 6on Banana 203 3855 wag—mm
.N 32.55 3:23—8:— E can Ewéca 255.532"

3 8323 9:21? :8 «do: go find $32: no 358.5 2: E 553E323=555~ um PS»:—

110

 

menup

(figure 521.13!
respectively.
immunopreci

w'nh gzflecm

In 0
grdect'm 3 t
expen'men‘
the galecti
expefimet
135$]metlr

Specific 2

\
fraction
(data nc
Similar
The b0
Polype
(_14 k[
Chaptc

Polyp

(FiOU

é

1 11
(Figure 5a, lane 5) and Sm B and Sm D (Figure 5b, lane 5) with anti-galectin—3 and anti-Sm,

respectively. These results indicate that Lac-binding did not perturb the
immunoprecipitation of galectin-3 by anti-Mac 2 and the coprecipitation of Sm B and Sm D

with galectin-3 by the monoclonal antibody.

In contrast to this lack of an effect by Lac, the coimmunoprecipitation of Sm B with
galectin 3 by anti-Mac 2 appears to be sensitive to high salt. The immunoprecipitation
experiments were carried out in buffer C, corresponding to the conditions used to deplete
the galectins from the splicing extract by Lac-affinity chromatography (2). In this
experiment, the nuclear extracts were prepared from HeLa cells metabolically labeled with
[3SS]methionine, such that polypeptides can be analyzed independent of reactivity with

specific antibodies.

When the labeled nuclear extract was pre-cleared with Sepharose beads, the bound
fraction, upon SDS-PAGE and fluorography, yielded a polypeptide corresponding to actin
(data not shown), as well as polypeptides of higher molecular weight (Figure 6, lane 1).
Similar results were obtained with precipitation by anti-TrR control beads (Figure 6, lane 3).
The bound fraction of the Lac-agarose precipitation revealed three predominant
polypeptides (Figure 6, lane 2): (a) p50 (50 kDa); (b) galectin-3 (29 kDa); and (c) galectin-1
(14 kDa). The identities and analyses of these polypeptides will be addressed in the next
chapter. The bound fraction of the anti-Sm beads yielded many bands, including
polypeptides that correspond in molecular weight to the A-G polypeptides of snRNPs

(Figure 6, lane 4). Finally, the anti-Mac 2 precipitate yielded a prominent polypeptide

112

Figure 6: Immunoprecipitation under high ionic strength of HeLa cell
splicing extract by autoimmune anti-Sm and by monoclonal anti-
Mac2.
Splicing extracts were prepared from HeLa cells metabolically labeled with
[3SS]methionine. The extracts were subjected to affinity adsorption or
immunoprecipitation in buffer C and the polypeptides in the bound fractions
were resolved by SDS-PAGE (12.5% acrylamide), followed by fluorography.
The samples in each lane were: (1) bound fraction of "pre-clearing" Sepharose;
(2) bound fraction of Lac-agarose; (3) rat anti-TfR precipitate; (4) human anti-
Sm precipitate; and (5) rat anti-Mac 2 precipitate. The positions of migration of
the snRNP-associated proteins A-G are highlighted on the left and the positions
of migration of galectin-1 and galectin-3 are highlighted on the right.

113

 

correspondmil

 

bandS. also oh.
importantly. tl‘.
D) (Figure 6. l:

of Sm B with :

Analy

Sm B. nonnal

 

components c;
prepared from
P01)"acrylamic'
revealed by at
[RVAS and L‘.
probably Con
(labeled as N
COITESPondin
the basis 0ft
lane 4). The
decrease in ‘
band M, ant

adSOrptiOn (

1 14
corresponding to galectin-3 (Figure 6, lane 5), as well as certain higher molecular weight

bands, also observed as non-specific adsorptions to control beads (Figure 6, lane 3). More
importantly, there was little intensity corresponding to a radioactive band for Sm B (or Sm
D) (Figure 6, lane 5). These results indicate that the coimmunoprecipitation by anti-Mac 2

of Sm B with galectin-3 was sensitive to high ionic strength.

Analysis of the Anti-Mac 2 Precipitate for RNA. The coimmunoprecipitation of
Sm B, normally found as snRNPs, by anti-Mac 2 prompted the question of whether RNA
components can also be identified in the immunoprecipitates. Thus, nuclear extracts were
prepared from HeLa cells that had been labeled with 32P04. When resolved on
polyacrylamide-urea gels, the following low molecular weight RNA components were
revealed by autoradiography (Figure 7, lane 1): (a) two prominent bands corresponding to
tRNAs and U1 snRNA; (b) three bands of intermediate abundance, two of which most
probably correspond to SS rRNA and U2 snRNA while the third remains unidentified
(labeled as M); (c) a band toward the top of the gel, labeled L; and ((1) minor species
corresponding to U4, U5, and U6 snRNAs. Some of these components were identified on
the basis of the 32P-labeled RNA components observed in the anti-Sm precipitate (Figure 7,
lane 4). There were intense bands for U2 and U1 snRNAs, while there was a significant
decrease in the intensity (relative to unfractionated nuclear extracts of band L, SS rRNA,
band M, and tRNAs. None of the RNA bands are enhanced in the bound fraction of the Lac

adsorption (Figure 7, lane 2). This was not surprising inasmuch as Lac-agarose bound

115

Figure 7: Analysis of the RNA components when HeLa cell splicing extracts
are immunoprecipitated by autoimmune anti-Sm and by
monoclonal anti-Mac 2.
Splicing extracts were prepared from HeLa cells labeled with 32P0... The
extracts were subjected to affinity adsorption or immunoprecipitation in buffer
D and the RNA components in the bound fractions were resolved by
polyacrylamide-urea gel (8% acrylamide, 8.3 M urea), followed by
autoradiography. (1) HeLa cell splicing extract; (2) bound fraction of the Lac-
agarose; (3) rat anti-TrR precipitate; (4) human anti—Sm precipitate; and (5) rat
anti-Mac 2 precipitate. The positions of migration of U1, U2, U3, U4, and U6
snRNAs, SS rRNA, and tRNAs are indicated on the left. In addition, two bands
corresponding unidentified RNA species are also highlighted as band L and
band M.

 

 

116

than

let»
(.51 rat
mdU6
to hint

 

NE Lac (1133 0.8!!! m2

117
galectin-3 but did not coprecipitate Sm components (Figure 3b, lane 2). Similar results

were also observed in the bound fraction of the control anti-TrR precipitate (Figure 7, lane
3). Thus, the faint bands observed in the bound fractions of the Lac-agarose and anti-TfR
precipitates most probably reflect nonspecific adsorptions of the snRNAs components on
the respective beads. Note that the nonspecific binding was most pronounced for the U1
snRNA and SS rRNA; there were drastic reductions in the intensities of band L, U2 snRNA,

band M, and, in particular, tRNAs

When the bound fraction of the anti-Mac 2 precipitate (Figure 7, lane 5) was
compared to the negative control anti-TrR (Figure 7, lane 3) and to the positive control anti-
Sm (Figure 7, lane 4), two bands were strikingly noteworthy: (a) The band labeled L was
almost as intense as that seen in the original nuclear extract, certainly enhanced over the
background seen in the other lanes (Figure 7, lane 5 versus lanes 2 and 3); and (b) The band
corresponding to U2 snRNA was also apparently enriched, relative to background controls
and comparable to the positive control in anti-Sm (Figure 7, lane 5 versus lane 4). The
intensities of the U1 snRNA, SS rRNA, and tRNAs were comparable to those found in the
other precipitates. Thus, it appears that the anti-Mac 2 precipitates preferentially enhanced
U2 snRNA and band L. The identity of band L remains to be determined; on the basis of
size markers, it is estimated that the position of migration of band L corresponds to a

nucleotide of approximately 300 base pairs.

To test whether the coprecipitation of Sm B by anti-Mac 2 was dependent on the

presence of RNA, nuclear extracts were treated with RNase (100 ttg/ml, 1 hour at 37 °C)

118
and then subjected to immunoprecipitation by anti-Sm and anti-Mac 2. RN ase—treated

nuclear extract immunoprecipitated by anti-Sm yielded Sm B and Sm D on anti-Sm blots
(Figure 8b, lane 2), but no observable galectin-3 (Figure 3a, lane 2) as described previously.
More importantly, the results showed that even after RNase treatment, anti-Mac 2 can

immunoprecipitate both galectin-3 (Figure 8a, lane 3), as well as Sm B (Figure 8b, lane 3).

119

.uofiwzcwfi 8a m Em can Wazoo—aw mo 59¢me .«o 20:63 2F .vouombxo «.83.:
.0372 us 28562“. N 82-35 :2 AC was combos Basebomazm co BEEooE Emrzca 5E2: ANV 38.—axe
wEozam 322:. A: ”8o? 2:: some E moEEam .BV use A3 £28m E .Emtzcm <Zm E3 szoEocaEExa
$5822.35 fies .23 wceosoeaeé 3 .wseosocoeé 3 8326.. .Aoeeane s3:
m0<mm0m .3 3382 203 28:an canon on. E 823253 05 98 N 837:5 “E E3 8 Emécm <Zm
E3 cosmEEooEocsEE 2 c8838 55 825 $8.38 Baowfi 032% 05. .Go hm .moEEE Om ._E\w1
8: 0373 2E3 330.: 0.53 322:8 2: EH 0 coats E m=oo So: Eat Egan 0.83 38.98 mEozmm

.N 32-5.8 3 m Em he :EHEEBEQEEEE 05 :e EoEth amalg— uc «auto 2E. ”m usury

120

Qb.‘ '

  

‘ ' . ".5623
l

D.
-2!
1'
u

n N p

DISCUSSION

The experiments presented in this study document the following key points:(a)
Wh'en nuclear extracts are subjected to immunoprecipitation with anti—Mac 2, a monoclonal
antibody specific for galectin-3, a fraction of the snRNP B proteins are specifically
coprecipitated with the lectin; (b) RNA components, one tentatively identified as U2
snRN A and another of approximately 300 nucleotides, are also found in the anti-Mac 2
precipitate; and (c) The coimmunoprecipitation is dependent on the presence of galectin-3
but is perturbed neither by Lac/ligand-binding nor by prior RNase treatment of the extract.
There are several peculiar features of the observed coimmunoprecipitation that require

discussion.

First, the coimmunoprecipitation does not appear to be reciprocal. Although anti-
Mac 2 always coprecipitates some Sm B with galectin-3 from nuclear extracts of both
human HeLa cells and mouse 3T3 fibroblasts, the lectin is not detected in the anti-Sm
precipitates of the same extracts. One possible explanation is that anti-Mac 2 precipitates
only those snRNPs bearing galectin-3 and this population represents a very minor fraction
of the total nuclear snRNPs. Thus, when anti-Sm precipitates snRNPs, the lectin
coprecipitated would constitute a barely detectable portion of the polypeptides present in the
precipitate. An alternative explanation is that the binding of galectin-3 to Sm B and the
binding of the particular autoimmune serum (ENA anti-Sm) are mutually exclusive. On

this basis, the immunoprecipitation of Sm B by anti-Sm releases galectin-3 from the

121

122
complex and thus precludes the lectin from being detected in the precipitate fraction.

Finally, we note that there is a minor polypeptide that is reproducibly immunoprecipitated
by anti-Sm and subsequently immunoblotted by polyclonal anti-galectin-B. Although the
mobility (molecular weight) of this band X is close to that of galectin-3, it is nevertheless
distinct from the major population of galectin-3 molecules. It is possible that this represents
a post-translationally modified form of galectin-3 and only this form of the lectin can

associate with the snRNP complexes.

Second, it is interesting to note that although the presence of Lac in the nuclear
extract does not inhibit the coprecipitation by anti-Mac 2 of Sm B with galectin-3, passing
the nuclear extract over a Lac-agarose column resulted in the adsorption of only of galectin-
3. No Sm B could be detected in the bound fraction of the Lac affinity beads. One possible
explanation of these observations may be the different effects of univalent ligands versus
multivalent ligands on the conformation of galectin-3. The binding of saccharides such as
Lac to galectin-3 results in a conformational change, as detected by an increase in the
transition temperature of thermal denaturation (8) and by alterations in the intrinsic
fluorescence spectrum (16). This structural change does not result, however, in self-
association (17,18). On the other hand, when galectin-3 is presented with a multivalent
ligand (e.g. polylactosamine on laminin (19) or oligosaccharides on immunoglobulin E
(17)), the resulting conformational change is sufficient to expose sites for protein-protein
interactions such that there is self-association of the galectin-3 molecules and cooperative

binding to the multivalent carbohydrates (17, 19). In the present context, the binding of

123
galectin-3 to the multivalent Lac-agarose beads would result in self-association to the

exclusion of protein-protein interactions such as with Sm B. An alternative explanation is
that a minor fraction of posttranslationally modified galectin-3 molecules (represented by
band X discussed above) is bound to Sm B and this complex, at low ionic strength
conditions of buffer D, binds neither free Lac nor Lac-agarose. At high ionic strength
(buffer C), however, the complex is dissociated and thus Lac-agarose can still deplete all the
galectin-3 molecules, including band X. Thedepleted extract no longer exhibits

coimmunoprecipitation of Sm B by anti-Mac 2 (this work) or splicing activity (2).

The analysis of these various possibilities would be greatly facilitated by the
rigorous identification and characterization of the RNA species in the anti-Mac 2
precipitate. For example, we could attempt to isolate the particular RNP fraction containing
both galectin-3 and Sm B using antisense oligonucleotides rather than by anti-Sm
immunoprecipitation, thus obviating difficulties associated with dealing with a majority of
RNP complexes that do not bear galectin-3. Unfortunately, the identification of the RNA
species in the immunoprecipitated fraction has been less than definitive. Anti-Mac 2
reproducibly enriches band L and a band matching in mobility to that of U2 snRN A. For
example, U1 snRNA and tRNA are the two major RNA species found in unfractionated
nuclear extract. In the anti-Mac 2 precipitate, however, the intensities of band L and U2
snRNA are equal or greater than these RNA species. The identity of band L is not known.
Its position of migration indicates that it contains approximately 300 nucleotides, thus

precluding it from being one of the minor snRNAs. (The minor snRNAs, U7 through U14,

124
range in size from 60 to 150 nucleotides (20)). The estimate of 300 nucleotides for band L

puts it in the range of the 78 small RNAs, 7SL (303 nucleotides) and 78K (330 nucleotides)
(21, 22). The 7SL RNA has been identified as the RNA component of signal recognition
particle, which functions in protein translocation across and integration into the membrane
of the endoplasmic reticulum (23). Although this function associates the RNP with
polysomes or microsomal membranes of the cytoplasm, it should be noted that the 7SL
RNA has been detected in nuclear fractions ( 13, 24). In fact, in the original identification of
snRNAs by Lerner and Steitz (l3), autoimmune sera precipitation of U1, U2, U4, U5 and
U6 snRNAs also included a RNA species with a mobility of 7S RNA. The significance of
the presence of 7S/78L RNA in the nuclear fraction is not known; it may simply represent

newly transcribed RNA species on their way out of the nucleus.

In any case, the immunoprecipitation of Sm by anti-Mac 2, while stringently
dependent on the presence of galectin-3, does not appear to be affected by RNase treatment
of the nuclear extract. The results suggest that the interaction of Sm B with galectin-3
involves direct or indirect protein-protein interactions. This would be consistent with the

notion that neither Sm B (25) nor galectin—3 (26) binds to RNA directly.

In previous studies (2), it was shown that when nuclear extracts were depleted of
galectins by Lac-agarose adsorption, the formation of virtually all of the ATP-dependent
splicing complexes was inhibited. The depleted extracts, when reconstituted with
recombinant galectin-3, regained the ability to form higher order complexes by the addition

of the snRNPs. The present identification of an association between galectin-3 and Sm B

125
(and possibly U2 snRNP) suggests that the role of the complex may be to organize the ATP-

independent H complex for accepting the snRNPs.

 

10.

11.

12.

13.

14.

126

REFERENCES

Barondes, S.H., Cooper, D.N.W., Gitt, M.A., and Leffler, H. 1994. J. Biol. Chem. 269,

20807-208 10.

. Dagher, S.F., Wang, J .L., and Patterson, R.J. 1995. Proc. Natl. Acad. Sci. USA 92,

1213-1217.

Laing, J .G. and Wang, J .L. 1988. Biochemistry 27, 5329-5334.

Jia, S. and Wang, J.L. 1988. J. Biol. Chem. 263, 6009-6011.

Moutsatsos, I.K., Wade, M., Schindler, M. and Wang, J .L. 1987. Proc. Natl. Acad. Sci.
USA 84, 6452-6456.

Ho, MK. and Springer, T.A. 1982. J. Immunol. 128, 1221-1228.

Cherayil, B.J., Weiner, SJ. and Pillai, S. 1989. J. Exp. Med. 170, 1959-1972.
Agrwal, N, Sun, Q., Wang, S.-Y., and Wang, J .L. 1993. J. Biol. Chem. 268, 14932-
14939.

Dignam, J .D., Lebovitz, RM. and Roeder, R.G. 1981. Nucleic Acids Res. 11, 1475-
1489.

Laemmli, UK. 1970. Nature 227, 680-685.

Calvet, J. P., and Pederson, T. 1979. J. Mol. Biol. 122, 361-378.

Mayrand, S., and Pederson, T. 1983. Mol. Cell. Biol. 3, 161-171.

Lerner, M. R., and Steitz, J. A. 1979. Proc. Natl. Acad. Sci. USA 16, 5495-5499.
Williams, D, G. Stocks, M. R., Smith, P. R., and Maini, R. N. 1986. Immunol. fl,

495-500

 

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

127

Rokeach, L. A., Jannatipour, M., Haselby, J. A., and Hoch, S. O. J. Biol. Chem.
2%, 5024-5030.

Mehul, B., Bawumia, S., Martin, S. R., and Hughes, R. C. 1994. J. Biol. Chem. 262,

' 18250-18258.

Hsu, D.K., Zuberi, RI. and Lin, ET. 1992. J. Biol. Chem. 267, 14167-14174.
Wang, J .L., Anandita, Wang, S.-Y., and Agrwal, N. 1993. in Lectins: Biology,
Biochemistry and Clinical Biochemistry, Vol. 9 (P. Chakrabarti, J. Basu, and M.

Kundu, eds.) M/S Wiley Eastern, Ltd., New Delhi, pp. 89-99.

.13

Massa, S.M., Cooper, D.N.W., Leffler, H. and Barondes, SH. 1993. Biochemistry 32,
260-267.

Shanab, G. M., Maxwell, E. S. 1992. Eur. J. Biochem. &, 391-400.

Ullu, E., Murphy, 8. and Melli, M. 1982. Cell 29, 195-201.

Ullu, E. and Melli, M. 1982. Nucleic Acids Res. 10, 2209-2223.

Walter, P. and Blobel, G. 1982. Nature 299, 691-698.

Reddy, R., Li, W.-Y., Henning, D., Choi, Y.C., Nohga, K., and Busch, H. 1981. J.
Biol. Chem. 256, 8452-8457.

Liautard, J .P., Sri-Widala, J ., Brunel, C. and Jenaeur, P. 1982. J. Mol. Biol. 162, 623-
643.

Dagher, SF. 1994. Ph.D. thesis, Michigan State University

CHAPTER IV

Nuclear and Cytoplasmic Localization of Galectins: Evidence from Laser Scanning

Confocal Microscopy

128

SUMMARY

We have analyzed the composition and identities of the polypeptides bound to
lactose agarose when splicing extracts of HeLa cells are subjected to affinity adsorption
under conditions used to deplete splicing activity. We found that, in addition to galectin-3,
the bound fraction of the lactose-agarose contained at least three other polypeptides. On the
basis of immunoblotting, two of these were identified to be actin and galectin-l. The
presence of galectin-1 in the nucleus of HeLa cells was confirmed by laser scanning

confocal fluorescence microscopy.

129

INTRODUCTION

The galectins comprise of a family of galactose (Gab/lactose (Lac)-specific
carbohydrate-binding proteins (1). At present, this family consists of eight members,
designated galectin-l through galectin-8 (see Table 2, Chapter I). As implied by their
nomenclature, the structures of all the galectins have at least one carbohydrate recognition
domain (CRD), with conserved amino acid sequence between members of the family and
between homologues of the same member across different species. Thus, galectin-l, -2, -5,
and -7 each has one CRD, the prototype of the family (2; see also Figure 1, Chapter I).
Galectin-4 and galectin-8 have tandem repeats of two CRDs (see Figure 1, Chapter 1).
Finally, galectin-3 has two domains: the amino tenninal domain is rich in proline and
glycine residues, with limited homology to certain polypeptides of the heterogeneous
nuclear ribonucleoprotein complex (hnRNP) while the carboxyl terminal domain contains
the CRD. Thus, galectin-3 exhibits a chimeric fusion of two distinct domains (2; see also

Figure 1, Chapter 1).

Previous studies had shown that galectin-3 was localized in the nucleus and
cytoplasm (3) and that its level of expression (4) and nuclear localization (5) was dependent
on the proliferative state of the cell. On the basis of the limited sequence homology to
proteins of the hnRNP (6) and on the basis of experiments showing that the nuclear
localization of the protein was ribonuclease sensitive (7), experiments were performed to
test for a role of galectin-3 in nuclear pre-mRNA processing. Several key findings were

made (8): (a) nuclear extracts capable of carrying out splicing in a cell-free assay contain

130

131

galectin-3; (b) nuclear extracts depleted of galectin-3 by affinity adsorption on Lac-agarose
become deficient in splicing; (c) the activity of the Lac-agarose depleted extract can be
reconstituted by addition of purified recombinant galectin-3; and (d) saccharides that bind
galectin-3 with high affinity inhibit the splicing reaction. These results strongly suggest that

galectin-3 may be a splicing factor/regulator.

During the course of these studies, however, Sue Dagher and Ron Patterson made
another striking observation (9). While Lac-agarose can deplete the splicing activity, a
monoclonal antibody specific for galectin-3 (the anti-Mac 2 antibody (10, 11)) failed to do
so, despite the fact that >95% of the galectin-3 polypeptide had been removed from the
extract. Moreover, the splicing activity of the galectin-3 depleted extract, derived from the
anti-Mac 2 adsorption, was still sensitive to inhibition by Lac. This suggests that other Lac-
binding proteins may be present in the splicing extract and they can be depleted, along with
galectin-3, by Lac-agarose but not by anti-Mac 2. Thus, it seemed important to define the
components that are bound by the Lac-agarose column and to test the role of each or

combination of these components in terms of the reconstitution of splicing activity.

In the meantime, I had been attempting to test the hypothesis, by
coimmunoprecipitation studies reported in Chapter III, that galectin-3 interacted with
components of the spliceosome machinery. In comparing the polypeptides found in the
bound fractions of anti-Mac 2, anti-TfR, anti-Sm, and Lac-agarose precipitations (see for
example, Figure 6, Chapter III), it became apparent that at least two other polypeptides,
besides galectin-3, was adsorbed by the Lac affinity resin. Thus, it seemed possible to

provide an explanation to Sue Dagher's observations (9) if we could definitively identify

 

132

these polypeptides as Lac-binding proteins and document their nuclear localization. The

results of these studies are now reported in this chapter.

 

133

MATERIALS AND METHODS

Antibodies and Aff'mity columns. Polyclonal rabbit anti-galectin-3 antiserum
was obtained from rabbit #33 (10). Polyclonal rabbit anti-galectin-l antibody was a gift of
Dr. S. H. Barondes (University of California, San Francisco). Rabbit antiserum against
lactose dehydrogenase (LDH) was a gift from Dr. John Wilson (Michigan State University).
The derivatization and characterization of rabbit anti-calf thymus actin have been reported
(11, 12). ENA anti-Sm antibody was purchased from The Binding Site (San Diego, CA).
Lactose immobilized on 6% beaded agarose (Lac-agarose) was purchased from Sigma

(St. Louis, Mo). Sepharose CL4B beads was obtained from Pierce (Rockford, IL)

Lac-agarose Affinity Adsorption. Nuclear extract was prepared according to
Dignam et. al.(15) as described in Chapter 3. The [358]methionine labeled nuclear extract
(in buffer C, 20 mM HEPES pH 7.9, 25% (v/v) glycerol, 0.42 M NaCl, 1.5 mM MgC12,
0.2 mM EDTA, 0.5 mM PMSF, 0.5 mM DTT) was precleared by incubating with
Sepharose beads for 1 hour at 4 0C. The unbound fraction of this preclearing step was
subjected to Lac-agarose adsorption. Precleared nuclear extract (50 111) was incubated
with 50 pl of Lac-agarose beads (pre-washed with buffer C twice) for 1 hour at 4 0C. The
Lac-agarose were washed: (i) twice with buffer 1 (10 mM Tris, pH 7.5, 150 mM NaCl,
0.5% NP-40, 0.2 mM EDTA); (ii) twice with buffer 2 (10 mM Tris, pH 7.5, 500 mM
NaCl, 0.5% NP-40, 0.2 mM EDTA); and (iii) twice with buffer 3 (10 mM Tris, pH 7.5).

The material remaining bound to the beads were subjected to SDS-PAGE (12.5%

134

acrylamide) as detailed in Chapter 3 and the proteins were revealed by fluorography and

by immunoblotting.

For immunoblotting analysis, rabbit anti-galectin-3, anti—galectin-l and anti-actin
were used at 1:500 dilution. All three primary antibodies were diluted in TBS (10 mM
Tris, 0.5 M NaCl, pH 7.5) containing 0.2% gelatin. Horseradish peroxidase conjugated
sheep anti-rabbit antibodies (BMB) were used at 1:2,000 dilution in T-TBS (TBS
containing 0.05% Tween 20) as secondary antibody. Proteins resolved on a SDS-PAGE
were electrophoretically transferred to Immobilon-P (Millipore) membrane as described
in Chapter III. The membrane was briefly rinsed twice with T-TBS and blocked in 2%
gelatin/TBS for at least 1 hour at room temperature. After briefly washing with T-TBS
twice, the membrane was incubated with primary antibody for 1 hour at room
temperature. The membrane was then rinsed with T-TBS twice and washed extensively
with T-TBS (15 minutes once, then 10 minutes 4 times). The washed membrane was
incubated with horseradish peroxidase conjugated secondary antibody at room
temperature for 1 hour. The blot was rinsed and washed as described above, and then

developed with 4-chloro-1-naphthol.

Immunofluorescence Staining and Laser Scanning Confocal Microscopy.

HeLa cells grown as monolayers were obtained from American Type Culture Collection
(Rockville, MD). They were seeded onto coverslips (22 x 22 cm) in 6-well (8 cm2/well)
cluster dishes. The cells were washed twice (4 ml/well) with phosphate-buffered saline

(PBS, 0.14 M NaCl, 2.68 mM KCl, 10 mM NazHPO4, 1.47 mM KH2P04, pH 7.4). Cells

were fixed by incubation (20 minutes at room temperature) in 4% paraformaldehyde in PBS

135

(2 ml/well), followed by washing twice with PBS (10 minutes each, on a rotating platform).
The residual aldehyde groups were blocked by 0.1 M glycine in PBS (4 ml/well; 20 minutes
at room temperature). After removal of the glycine—PBS solution, cells were permeabilized
with 0.5% Triton X-100 in PBS (2 mee11; 4 minutes at room temperature). The cells were
again washed twice with PBS. The cells were then incubated with antiserum (anti-galectin-
3, 1:150; anti-galectin-l, 1:150; anti-LDH, 1:150; and ENA anti—Sm, 1:500) for one hour at
room temperature. After washing three times in T-TBS (10 mM Tris pH 7.5, 0.5 M NaCl,
0.05% Tween 20) (15 minutes each), the cells were incubated with the secondary antibody
for one hour. Fluorescein-conjugated goat anti-rabbit immunoglobulin (BMB) (1 :50) was
used to detect rabbit anti-galectin-3, rabbit anti-galectin-l , and rabbit anti-LDH binding and
fluorescein-conjugated goat anti-human immunoglobulin (BMB) (1:100) was used to detect
human autoimmune ENA anti-Sm binding. Finally, the coverslips were washed three times

in T-TBS (15 minutes each) before mounting with Penna-Fluor on glass microscope slides.

A Meridian Instruments (Okemos, MI) Insight bilateral laser scanning confocal
microscope was used with an argon ion laser as the excitation source. An 100 x objective
lens and laser power of 50 mWatts were used for scanning the image. One photomultiplier
was used to detect fluorescence emission of fluorescein (530 nm). Images (360 x 360
pixels) were collected from 9 consecutive focal planes as with an increment of 0.5 pm for

each step in the x and y directions (16).

RESULTS

Polypeptides Bound by Lac-agarose. Nuclear extracts were prepared from HeLa
cells metabolically labeled with [3SS]methionine. The extract was "precleared" by
sepharose beads to remove components that would bind nonspecifically to the beads. The
unbound fraction of the "preclearing" step was then subjected to binding with Lac-agarose.
The conditions for these experiments (buffer C) were identical to those used to deplete
splicing extracts of galectins and splicing activity (8, 9). The material bound to the beads
was subjected to SDS-PAGE, followed by fluorography and immunoblotting. Five
polypeptides were found in the bound fraction of the Lac-agarose adsorption with molecular
weights of approximately: (a) 70,000; (b) 50,000; (c) 43,000; (d) 30,000; and (e) 15,000

(Figure 1, lane 1).

Three of these bands have been identified on the basis of immunoblotting analysis.
First, a rabbit anti-calf thymus actin antiserum (reactive predominantly with non-muscle
actins (13, 14)) immunoblotted the 43 kDa band (Figure 1, lane 4), suggesting it to be actin.
This would be consistent with the observation that this polypeptide was also bound to the
preclearing sepharose beads in a nonspecific fashion (see, for example, Figure 6, Chapter
111). Second, polyclonal rabbit anti-galectin-3 identified the 29 kDa band as galectin-3
(Figure 1, lane 3), as was reported previously (8) that under these conditions (high ionic
strength of buffer C), Lac-agarose depleted splicing extracts of galectin-3 with concomitant

loss of splicing activity. Finally, the 14 kDa band was identified as galectin-1, on the basis

136

137

Figure 1: The composition and identities of polypeptides in the bound fraction
of Lac-agarose adsorption.
Nuclear extracts were prepared, following the procedures of Dignam et a1. ( 15),
from HeLa cells metabolically labeled with [35$]methionine. The nuclear
extracts, in buffer C, were subjected to affinity adsorption on Lac-agarose beads.
The bound fraction was then subjected to SDS-PAGE ( 12.5% acrylamide),
followed by fluorography or by immunoblotting with various antibodies. Lane
1: Radioactive polypeptides revealed by fluorography. Lane 2: Irnmunoblot
with anti-galectin-l. Lane 3: Irnmunoblot with anti-galectin-B. Lane 4:

Irnmunoblot with anti-actin. The positions of migration of authentic protein
markers are highlighted on the left.

tlmrllncfin

es of Dignam cl .
mine. 111cm:
on on lac-aging:
125% acryltrai
arious mimic)
Lane 2: Irrrmrttt1
actin-3. land:
1 of authenticgm

138

 

 

139

of immunoblotting with affinity purified anti-galectin-l antibodies (Figure 1, lane 2). Thus,
it appears that Lac-agarose was indeed depleting more than one Lac-binding protein from

the nuclear extract (i.e. galectin-1, in addition to galectin-3).

The identities of the other prominent band (Mr 50,000) and minor band (Mr
70,000) are not known. That the prominent band represents a bona fide Lac-binding protein
is suggested by preliminary experiments showing its specific elution from the Lac-agarose
beads by the addition of Lac. Neither of these two bands reacted with any of the anti-

galectin antibodies available to us, including an antiserum against rat intestinal galectin-4.

Nuclear and Cytoplasmic Localization of Galectin-l: Evidence from Laser
Scanning Confocal Microscopy. The identification of galectin-1 in nuclear extracts of
HeLa cells and its adsorption onto Lac affinity resins provided an explanation to the
observation that while Lac-agarose can deplete the splicing activity, anti-Mac 2, a
monoclonal antibody specifically reactive with galectin-3, failed to do so. Coupled with the
demonstration that either galectin-l or galectin-3 alone is sufficient to reconstitute splicing
activity in a Lac-agarose depleted extract (9), the results would suggest that the activities of
galectin-1 and galectin-3 in the nucleus may be redundant. On the other hand, however, the
notion that galectin-1 can be found in the nucleus is controversial, particularly since, in the
original publication reporting the purification of the chicken homologue, there was an
explicit statement indicating that anti-chicken heart galectin-1 failed to label the cell nucleus
of chicken embryo fibroblasts under immunofluorescence (17). The main difficulty in the
analysis of galectin-1 is that, unlike galectin-3, the predominant portion of the former lectin

is found in the cytoplasm. As a result, a strong fluorescent antibody probe staining the

140

cytbplasm surrounding the nucleus sometimes results in an image containing fluorescent
flare that exceeds the resolution of the light microscope. To circumvent this difficulty, we
took advantage of the availability of a laser scanning confocal microscope to analyze the

staining pattern obtained with anti-galectin-l.

HeLa cells cultured on coverslips were fixed with paraformaldehyde, permeabilized
with Triton X-100 and stained with various antibodies plus fluorescein—conjugated second
antibody. The fluorescence staining was visualized by laser confocal microscopy, collecting
images through 9 consecutive focal planes. In this study, ENA anti-Sm was used as a
positive control for nuclear staining (Figure 2). In a plane above the cell nucleus, there was
little fluorescence. Through the middle sections, the plane of focus cuts through the nucleus
and yielded intense staining (Figure 2, column B). One of the middle sections is enlarged in
column A (Figure 2), showing diffuse distribution of fluorescence throughout the nucleus,
except for the "black holes" (see Figure 2, Chapter 2). Similarly, anti-galectin-3 yielded,
through the middle sections, bright nuclear staining with the exception of the "black holes."
Thus, the subcellular distribution of galectin-3 and the Sm epitopes of snRNPs are very
similar, a conclusion documented in Chapter II. Staining for LDH, serving as a negative
control, showed that the enzyme was predominantly cytoplasmic. With the enlarged middle
section (Figure 2, column A), the anti-LDH yielded a distinct pattern, showing that the

majority of fluorescence in the cytoplasm.

The staining for galectin-1 yielded sections that resembled neither those obtained

with anti-Sm or anti-galectin-3 nor that seen with anti-LDH. Due to high levels of the lectin

 

 

141

Figure 2: The subcellular localization of galectin-1 as revealed by laser
scanning confocal fluorescence microscopy.
The staining of snRNP Sm polypeptides and galectin-3 and the staining of LDH
provided reference pattems for nuclear and cytoplasmic localizations,
respectively. For each staining, images (360 x 360 pixels) were collected from 9
consecutive sections focal planes, with the increment of 0.5 pm for each step in
the x and y directions. They are displayed serially in column B; column A
provides an enlarged view of one middle section (#5).

 

 

 

142

Sm

Galectln-S

Galectin-1

 

 

143

in the cytoplasm, the middle sections yielded generally diffuse fluorescence throughout both
the cytoplasm and the nucleus. The areas devoid of fluorescence are not nuclei; rather, they

correspond to the "black holes" observed with anti-Sm or anti—galectin-3 staining.

 

 

 

144

DISCUSSION

The previous observations on the difference in the ability of Lac-agarose versus anti-
Mac 2 to deplete splicing activity from nuclear extracts of HeLa cells suggested: (a) in
addition to galectin-3, there must be other Lac-binding proteins in splicing competent
extracts; (b) depletion of a single lectin from such an extract cannot remove the splicing
activity; and (c) the splicing activity is lost only when all the Lac-binding proteins are
adsorbed on the Lac-agarose resin. Coupled with the demonstration that either galectin-1 or
galectin-3 can reconstitute the splicing activity in a depleted extract (9), these results
suggest that there may be multiple members of the galectin family in the nucleus and that
their activities may be redundant. Indeed, the results of this series of experiments have
identified at least galectin-1 to be present, in addition to galectin-3, in the bound fraction of
the Lac-agarose resin when nuclear extracts are depleted of splicing activity. The
localization of galectin-1 to the nucleus was also confirmed by confocal

immunofluorescence microscopy.

Previous comparisons of the immunofluorescence staining of live cells versus cells
fixed with formaldehyde followed by permeabilization indicate that galectin-1 is found
predominantly in the intracellular compartment. Within the cells, the staining is mostly
cytoplasmic. Observations and interpretations pertaining to the nuclear localization of

galectin-1 are somewhat more difficult to establish.

First, immunofluorescence and ultrastructural studies have led to explicit statements

that anti-galectin-l antibodies failed to label the cell nucleus. In chicken embryo

145

fibroblasts, anti-chicken heart lectin (chicken galectin-1) staining was described to be
localized only to the cytoplasm (17). When 17-day-old chicken embryonic keratinized
epidermis was stained with gold-labeled anti-galectin-l under electron microscopy, gold
particles were found around desmosomes, tonafilament bundles, and the intercellular space,

while the cell nucleus was free of the particles (18).

Second, in a number of studies originally performed for other objectives (e. g. to
show overlap in cells expressing lectins and lactoseries glycoconjugates), the investigators
provide no specific conclusion regarding the intracellular distribution of galectin-1. Dorsal
root ganglion neurons were subjected to immunofluorescence with anti-galectin-l and with
a monoclonal antibody directed against a lactoseries glycoconjugate (19). Although the
intracellular distribution of the lectin was not discussed in that particular report, the same
investigators have more recently concluded, on the basis of the previously published data,
that galectin-1 could be detected in both the nucleus and cytoplasm of the neurons (20).
Similarly, reassessment of previously published immunocytochemical studies in non-

neuronal cells has suggested the presence of rat galectin-1 in nuclei and cytoplasm (21).

Finally, there are studies that specifically show the localization of galectin-1 in the
nucleus as well as the cytoplasm. Cryostat sections of tissues subjected to
immunofluorescence showed labeling of both nuclei and cytoplasm in anti-chicken galectin-
l staining of adult chicken kidney (22) and anti-bovine galectin-1 staining of calf pancreas
(23). A monoclonal antibody, designated 36/8, was generated against bovine heart galectin-
] and stained the nuclei and cytoplasm of lymphoblastoid and leukemic cells (24). This

monoclonal antibody immunoblotted polypeptides of apparent molecular mass 13, 36, 65,

 

 

146

80, and 130 kDa in extracts of lymphoid cells. More recent studies have shown that the
36/8 antibody recognizes the tetrapeptide sequence Trp—Gly-Ala/Ser—Glu/Asp (25) and,
therefore, it is not clear whether the nuclear staining component is a galectin or some other
polypeptide bearing this epitope sequence. At the ultrastructural level, it has been reported
that anti-galectin-l labels the nucleus of epidermal cells of the intermediate layer of a

stratified epithelium, the chick embryonic skin (26).

On the basis of these previous reports, it seems apparent that the localization of
galectin-1 has been somewhat controversial. It was important to establish, therefore, that
galectin-1 can indeed be found in the cell nucleus. The criteria used in the present study
consisted of parallel labeling studies with antibody reagents directed against known nuclear
components (e. g. anti-Sm) and an enzyme generally accepted as a cytosolic marker (e. g.
anti-LDH). The staining patterns obtained with anti—galectin-l under confocal
immunofluorescence analysis resembled neither anti-Sm nor anti-LDH. Rather, the anti-
galectin-l staining represented a composite of the labeling patterns of nuclear and

cytoplasmic controls, suggesting that the lectin is found in both compartments of the cell.

10.

11.

12.

13.

14.

147

REFERENCES

. Barondes, S. H., Cooper, D. N. W., Gitt, M. A., and Leffler, H. 1994. J. Biol. Chem.

262, 20807-20810.

Hirabayashi, J. and Kasai, K. 1993. Glycobiology 3, 297-304.

Moutsatsos, I. K., Davis, J. M. and Wang, J. L. 1986. J. Cell Biol. 102, 477-483.
Agrwal, N., Wang, J. L. and Voss, P. G. 1989. J. Biol. Chem. _246, 17236-17242.
Moutsatsos, I. K., Wade, M., Schindler, M. and Wang, J. L. 1987. Proc. Natl. Acad.
Sci. USA 84, 6452-6456.

Jia, S. and Wang, J. L. 1988. J. Biol. Chem. 263, 6009-6011.

Laing, J. G. and Wang, J. L. 1988. Biochemistry _2_7, 5329-5334.

Dagher, S. R, Wang, J. L. and Patterson, R. J. 1995. Proc. Natl. Acad. Sci. USA 22,
1213-1217.

Dagher, S. F. 1994. Ph.D. thesis, Michigan State University.

Ho, M. K. and Springer, T. A. 1982. J. Immunol. £8, 1221-1228.

Cherayil, B. J ., Weiner, S. J ., and Pillai, S. 1989. J. Exp. Med. fl), 1959-1972.
Agrwal, N., Sun, Q., Wang, S. Y., and Wang, J. L., 1993. J. Biol. Chem. _2_6_8, 14932-
14939.

Metcalf, T. N., Szabo, L. J ., Schubert, K. R., and Wang, J. L. 1984. Protoplasma 1_2§,
91-99.

Villanueva, M. A., Ho, S. -C. and Wang, J. L. 1990. Arch. Biochem. Biophys. _2_71, 35-

41.

15.

16.

17.

18.

19.

20.:

21.
22.
23.

24.

25.

26.

148

Dignam, J .D., Lebovitz, RM. and Roeder, R.G. 1981. Nucleic Acids Res. 11, 1475-
1489.

Wade, M. H., Defeijter, A. W., Frame, M. K., and Schindler, M. 1993. Bioanlyt.

Instrumen. 31, 117-141.

Barak-Briles, E., Gregory, W., Fletcher, P. and Komfeld, S. 1979. J. Cell Biol. 8_1,
528-537.

Hirano, H., Akimoto, Y., Kawakami, H., Oda, Y. and Kasai, K. 1988. in Proceedings
of the Kagoshima International Symposium of Glycoconjugates in Medicine,
Professional Postgraduate Service International, Tokyo, pp. 8-13.

Regan, L. J ., Dodd, J., Barondes, S. H., and Jessell, T. M. 1986. Proc. Natl. Acad. Sci.
USA 83, 2248-2252.

Hynes, M. A., Gitt, M., Barondes, S. H., Jessell, T. M. and Buck, L. B. 1990. J.
Neurosci._1_Q, 1004-1013.

Barondes, S. H. 1988. Trends Biochem. Sci. l_3, 480-482.

Beyer, E. C. and Barondes, S. H. 1980. J. Supramol. Struct. 13, 219-227.

Childs, R. A. and Feizi, T. 1980. Cell Biol. Int. Rep. 4, 775.

Carding, S. R., Thorpe, S. J ., Thorpe, R. and Feizi, T. 1985. Biochem. Biophys. Res.
Commun. l_27, 680-686.

Abbott, W. M., Mellor, A., Edwards, Y. and Feizi, T. 1989. Biochem. J. _2_52, 283-290.
Akimoto, Y., Kawakami, H., Oda, Y., Obinata, A., Endo, H., Kasai, K., and Hirano, H.

1992. Exp. Cell Res. 1%, 297-304.

CHAPTER V

Concluding Statement

Removal of intron and ligation of exons is one of the most important steps for pre-
mRN A processing. Many catalytic elements have been elucidated for their involvement
in splicing machinery. Previous studies have shown that galectin-3 is a required factor for
spliceosome formation and splicing activity using a cell-free assay for pre-mRNA splicing.
However, the subnuclear localization and the RNP components which galectin-3 is

involved in pre-mRN A splicing remain unclear.

To address these questions, I have developed two levels of study to demonstrate that
galectin-3 is associated with components of the splicing machinery, specifically, the Sm
epitopes found on polypeptides of the snRNPs. At the single cell level, the
iMunofluorescence patterns of galectin-3 were shown to be similar to that of the snRNP
Sm polypeptides during the course of cell cycle and under a variety of conditions. Double
labeling experiments also demonstrated that galectin-3 colocalized with Sm antigens in the
cell nucleus. At the biochemical level of protein-protein interaction, I have shown that a
monoclonal antibody anti-Mac 2, which is specific for galectin-3, could
coimmunoprecipitate a fraction of Sm B polypeptides from nuclear extracts of HeLa cells
and mouse 3T3 fibroblasts. Nuclear extracts depleted of galectin-3 by prior adsorption on a
lactose affinity resin failed to yield Sm B in the anti-Mac 2 precipitate. These results

demonstrate that the coprecipitation of Sm B with galectin-3 by anti-Mac 2 was dependent

149

 

150

demonstrate that the coprecipitation of Sm B with galectin-3 by anti-Mac 2 was dependent
on the presence of the lectin. The coprecipitation was not affected by the addition of lactose;
nor was it perturbed by prior treatment of the nuclear extract with ribonuclease. Taken
together, these results suggest that at least a fraction of the Sm B polypeptides in the
splicing extract can interact with galectin-3 through protein-protein interactions.
Identification of the RNP component with which galectin-3 is associated should be the next

level of study to address the function of galectin-3 in the splicing machinery.

I have also analyzed the composition and identities of polypeptides bound to the
affinity resin lactose-agarose under conditions used to deplete the nuclear extract of splicing
activity. In addition to galectin-3, the bound fraction of the lactose-agarose also contains
galectin-1, whose presence in the nucleus of HeLa cells was confirmed by confocal
fluorescence microscopy. These findings provide an explanation for previous observations
that while lactose-agarose can deplete nuclear extracts of splicing activity, anti-Mac 2
precipitation failed to yield the same effect. Together with the observation that galectin-1,
as well as galectin-3, can alone reconstitute splicing activity in a lactose-agarose depleted
extract, these results suggest that the activities of galectin-1 and galectin-3 are redundant in

the cell nucleus.

 

MICHIGAN 511an UNIV. LIBRARIES
llllllllll lllllll 1111111111 1111111 11111111111111
31293014172252