A. .~.-.~.—-a—‘

 

 

"1:515

1‘

ST

lIIIIIIIIIIIIIZIII

 

IIIIIIIIIII IIIIIIIIIIIIIIIIIIIIIIIII

165 9315

 

LIBRARY

Mlchigan State I
University

 

 

 

This is to certify that the

dissertation entitled

GALECTIN-l IN THE CELL NUCLEUS:

EVIDENCE FOR A ROLE IN PRE-mRNA SPLICING

presented by

Anandita Vyakarnam

has been accepted towards fulfillment
of the requirements for

Ph . D . degree in Bioghemi 5 try

Mz'ufi

0 Major professor }

Date ”Illa/36

MSU Lt an Affirmative Action/Equal Opportunity Institution 0-12771

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID F INES return on or before date duo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU I. An Afflnnatlvo Action/Equal Opportmuy Intuition
mum

 

GALECTIN-l IN THE CELL NUCLEUS:
EVIDENCE FOR A ROLE IN PRE-mRNA SPLICING
By

Anandita Vyakarnam

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Biochemistry

1 996

ABSTRACT

GALECTIN-l IN THE CELL NUCLEUS:
EVIDENCE FOR A ROLE IN PRE-mRNA SPLICING

By

Anandita Vyakarnam

Galectins are a family of B-galactoside specific lectins that share conserved amino
acid residues in their carbohydrate recognition domains. Two members of this family,
galectin-1 and galectin-3 are widely distributed proteins and are present mainly inside
cells. Previous studies have shown that galectin-3 is localized to the cell nucleus and is a

component of ribonucleoprotein complexes.

Using a cell free assay, galectin-3 has been shown to be a required factor in pre-
mRNA splicing. HeLa cell nuclear extracts that are splicing competent become splicing
deficient when depleted of galectin-3 by adsorption on a lactose affinity column. The
splicing activity can be restored by addition of recombinant galectin-3. However,
depletion of galectin-3 from nuclear extracts by using a monoclonal antibody specific for
galectin-3 did not inhibit splicing activity. Furthermore, the splicing activity of this
galectin depleted extract was still sensitive to saccharide inhibition. This provided a hint
that there may be other galectins present in HeLa nuclear extract that are also involved in

pre-mRNA splicing.

The goal of this dissertation was to define the number and identity of other
galectins in HeLa nuclear extracts that may be involved in pre-mRNA splicing. We show
that galectin-1 is present in the nuclear extract and participates in pre-mRN A splicing in a
cell free system. Depletion of galectin-1 or galectin-3 alone by specific antibody
adsorption failed to remove all splicing activity and the residual activity was still sensitive
to inhibition by sugars. Nuclear extracts depleted of both galectin-1 and galectin-3, either
by lactose affinity adsorption or by double antibody adsorption, are no longer splicing
competent. Either galectin-l or galectin-3 alone is sufficient to reconstitute splicing

activity of the depleted extracts.

Galectin-l has been shown to be present intracellularly. However, the issue of its
nuclear localization is controversial. We have analyzed the intracellular distribution of
galectin-1 using laser confocal microscopy. We find that galectin-1 is present both in the
nucleus and cytoplasm of HeLa cells. In addition, the type of permeabilization reagent
used has a profound effect on the observation of galectin-1 in the nucleus. We have also
compared the distribution of galectin-1 and galectin-3 in the nucleus with the distribution
of the splicing factor SC35, which is localized in a speckled staining pattern in the
nucleus. HeLa cells that were permeabilized and then fixed, were stained simultaneously
with antibodies specific for galectin-l or galectin-3 and a monoclonal antibody against
SC35. Our results indicate that galectin staining in the nucleus colocalizes with the

speckled staining pattern of SC35 although galectins also show a more diffused staining.

Dedicated to my parents

iv

ACKNOWLEDGEMENTS

I would like to thank Dr. John L. Wang for his guidance and support through these
years. He has provided me many opportunities that helped me in becoming a better
researcher and achieve the goals of my education. I would also like to thank the members
of my guidance committee: Drs. Laurie Kaguni, William Deal, Justin McCormick and

Rich Schwartz for their valuable advice and input from time to time during my research.

I would like to express my gratitude to Dr. Ronald Patterson who has been very
helpful to me in setting up splicing assays. His many suggestions and long standing
interest in my thesis project were much appreciated. I am grateful to Dr. Sue Dagher who
was always available for advice and help in troubleshooting. I would also like to thank
Dr. Melvin Schindler and Sharon Grabski for their valuable suggestions and help
especially with confocal microscopy. I would like to thank Karen Lakkides for helping
me with immunofluorescence studies and Lynn Johnson for helping me in the initial

stages of splicing project.

My stay at Michigan State would not have been the same without the friends I
have made here. I would like to thank the members of the Wang and Schindler
laboratories, past and present, for all their help and support. I would like to express my
appreciation to Patty Voss, Mark Kadrofske, Yeou-Guang Tsay and Eric Amoys for their
wonderful friendship and many stimulating discussions. They have made my time in

graduate school a very memorable experience. I would also like to thank a number of

vi

friends outside of the lab, particularly Linda, Carol, Beta and Marty for their support and

constant enthusiasm.

I especially want to thank my parents for their unwavering support and
encouragement throughout the years. My mother’s love and sacrifice has always been an
inspiration for me. I would like to thank my sister for her love, friendship and
understanding over the years. Finally, I want to thank my husband, Murty for all his love,
friendship, encouragement and patience over the years. His support and understanding

made it easier to meet the challenges of graduate school.

TABLE OF CONTENTS

LIST OF TABLES ........................................................................................................... ix
LIST OF FIGURES .......................................................................................................... x
CHAPTER I. Literature Review ..................................................................................... 1
LITERATURE REVIEW ON GALECTINS ...................................................................... 2
Galectins- A Family of Animal Lectins ............................................................... .2
Structural Classification of Galectins ..................................................................... 3
Carbohydrate Recognition Domain of Galectins .................................................... 8
Physical Properties and Subcellular Localization of Galectin—3 ............................. 9
Subcellular Localization of Galectin-l .................................................................. 12
LITERATURE REVIEW ON PRE-mRNA SPLICING ................................................... 14
Pre-mRNA Splicing and SnRNPs ......................................................................... 14
Spliceosome Assembly Pathway ........................................................................... 17
Non-snRNP Splicing Factors ................................................................................ 19
Nuclear Structure and RNA Splicing .................................................................... 26
REFERENCES ................................................................................................................. 28

CHAPTER H. Galectin-l in the Cell Nucleus. 1. Evidence for a Role in Pre-mRNA

Splicing ............................................................................................................................. 35
ABSTRACT ...................................................................................................................... 36
INTRODUCTION ............................................................................................................. 37
MATERIALS AND METHODS ...................................................................................... 39
Preparation of HeLa Nuclear Extracts .................................................................. 39
Antibodies and Affinity Columns ........................................................................ 39
Depletion and Reconstitution of Nuclear Extract ................................................. 40
In Vitro Splicing Assay ......................................................................................... 41
Immunoblot Analysis ............................................................................................ 43
RESULTS ......................................................................................................................... 44
Saccharide Affinity Depletion versus Immunodepletion ...................................... 44

vii

viii

Number and Identity of Galectins in NE ............................................................... 49
Double Antibody Depletion of Galectins-1 and -3 ................................................ 54
Reconstitution of Splicing in the Double Antibody-depleted Extracts ................. 59
Comparison of the Concentration Dependence of Galectin-l and Galectin-3 in
Reconstituting Splicing Activity ......................................................................... 62
DISCUSSION ................................................................................................................... 68
REFERENCES ................................................................................................................. 75

CHAPTER III. Galectin-l in the Cell Nucleus. II. A Comparitive Localization Study

with Other Splicing Components ................................................................................. 78
ABSTRACT ...................................................................................................................... 79
INTRODUCTION ............................................................................................................. 81
MATERIALS AND METHODS ...................................................................................... 83
Immunofluorescence Microscopy ......................................................................... 83
Antibodies ............................................................................................................. 84
Preparation of HeLa Cell Extracts and Immunoblot Analysis .............................. 85
RESULTS ......................................................................................................................... 87
Nuclear and Cytoplasmic Localization of Galectin- 1: Evidence from Laser
Scanning Confocal Microscopy ............................................................................. 87
Detection of Nuclear Galectin-l Dependent on Perrneabilization Detergent ....... 96

Double Immunofluorescence Analyses: Galectins versus Splicing

Factor SC35 ............................................................................... 102
DISCUSSION ................................................................................................................. 108
REFERENCES ............................................................................................................... 112
CHAPTER IV. Concluding Statement ....................................................................... 1 15

CONCLUDING STATEMENT ................................................................. 116

ix

LIST OF TABLES
CHAPTER I.
Table l. Galectin members in mammalian species. .................................................. 5
Table 2. Galectins classified according to polypepetide architecture ....................... 7
Table 3. Major mammalian SnRNPs. ..................................................................... 16
Table 4. SnRNP core proteins .................................................................................. 16

Table 5. Superfamily of pre-mRNA splicing factors containing an SR domain ..... 21

LIST OF FIGURES
CHAPTER I.
Figure 1. Schematic diagram illustrating the polypeptide architecture of galectins. ..4
Figure 2. Schematic diagram of Spliceosome-catalyzed splicing. ............................ 15
Figure 3. The pre-mRNA splicing pathway in mammalian cells. ............................ 18
Figure 4. Schematic diagram summarizing the context and organization of specific
domains in the SR family of proteins. ....................................................... 23
CHAPTER [1.
Figure 1A. Comparison of the splicing activity of NE, NE depleted by Lac affinity
adsorption and depleted NE reconsituted with recombinant galectin-3. ...45
Figure 1B. Analysis of the polypeptide components of the bound fraction when NE is
subjected to affmity adsorption on Lac-agarose columns .......................... 47
Figure 2A. Comparison of the splicing activity of NE after adsorption on protein G-
Sepharose beads conjugated with various antibodies. .............................. 50
Figure 2B. Comparison of the levels of galectin-3 in NE versus the unbound (UB) and
bound (B) fractions of the various immunodepletions ............................... 52
Figure 3A. Comparison of the levels of galectins-1 and -3 in NE versus the unbound
(UB) and bound (B) fractions of immunodepletion using antibodies
directed against galectin-1 or galectin-3 .................................................... 55
Figure 38. Comparison of the splicing activity of NE versus the unbound (UB)
fractions of immunodepletion using antibodies directed against
galectin-1 or galectin-3 .............................................................................. 57
Figure 4. Comparison of the splicing activity of NE, NE immunodepleted by
antibodies against galectin-l and -3, and depleted NE reconstituted with
recombinant galectin-l or recombinant galectin-3. .................................. 60
Figure 5A. Comparison of the splicing activity of NE, NE depleted by Lac affinity
adsorption, and depleted NE reconstituted with recombinant galectin-1,
recombinant galectin-3, or the COOH-terminal domain of galectin-3. ....63
Figure SB. Quantification of the level of mature RNA product formation derived from

the data shown in panel A .......................................................................... 65

xi

CHAPTER 111.
Figure 1A. Western blot analysis of the specificity of the antibody reagents. ............ 88
Figure 18. Immunofluorescence staining of galectins-l and -3 and other nuclear and
cytoplasmic markers in human HeLa cells. .............................................. 90
Figure 2. Subcellular localization of galectin-1 as revealed by laser scanning
confocal fluorescence microscopy ............................................................. 94
Figure 3. Comparison of the immunofluorescence staining pattern of galectin-1 in
HeLa cells permeabilized with Triton X-100, digitonin, or saponin. ....... 98
Figure 4. The effect of Triton X-100 on the immunofluorescence staining pattern of
galectin-1 in saponin-permeabilized HeLa cells ...................................... 100
Figure 5. Double immunofluorescence analysis of galectin-3 and SC35. .............. 103

Figure 6. Double immunofluorescence analysis of galectin-1 and SC35. .............. 106

CHAPTER I

LITERATURE REVIEW

LITERATURE REVIEW ON GALECTINS

Galectins- A Family of Animal Lectins

Lectins are defined as non-enzyme, non-immunoglobulin proteins that have at
least one carbohydrate binding domain (1). Lectins have been identified in various
species ranging from viruses and bacteria to plants and vertebrates. Since
glycoconjugates have recently been found in the cytoplasm and the nucleus (2), there has
been a great interest in the intracellular localization of lectins. This dissertation deals
with the intracellular localization of galectin-1 and -3 and a role for these lectins in pre-

mRNA splicing.

Galectins are a family of animal lectins characterized by: (a) affinity for B-
galactoside-containing glycoconjugates and (b) conserved amino acid residues in their
carbohydrate recognition domain (CRD) (3). They are soluble proteins that lack a signal
peptide and are found mainly intracellularly. However, there is evidence that some
members are also extemalized to the cell surface and in the extracellular matrix.
Galectins do not require metal ions for their carbohydrate binding activity, which
distinguishes them from the C-type lectins that require calcium ions for sugar binding (4).
Members of the galectin family are present in a vast variety of tissues and various species
ranging from lower invertebrates like sponges and nematodes to higher vertebrates
including humans. They have been called by different names by various research groups

(soluble lectins, S-type lectins, galaptins, B-galactoside binding lectins etc.) but recently a

consensus on nomenclature has been reached by investigators in the field to use the term

‘galectin’ (3).

Structural Classification of Galectins

Based on their molecular architecture, members of the galectin family are sub-
divided into 3 categories (Figure l) (5).
(i) Proto Type galectins: These have a polypeptide (Mr ~14,000) that contains a consensus
CRD.
(ii) Chimera Type galectins: These are multi-domain proteins consisting of a N-terminal
domain which is proline and glycine rich and a C-terminal domain which is homologous to
the Proto Type galectin CRD.
(iii) Tandem Repeat Type galectins: These are composed of two tandemly repeated lectin

domains each homologous to the Proto Type galectin CRD.

At present, eight mammalian galectins have been reported (Table l). Galectin-1,
-2, -5 and -7 are Proto Type galectins. Of these, galectin-l and -2 exist as noncovalent
dimers under physiological conditions whereas galectin-5 and -7 are monomeric. Galectin-
] is isolated as a homodimer with subunit polypeptide molecular weight of ~14 kDa and is
not subjected to any post-translational modifications (6). It is abundant in smooth and
skeletal muscle, but is also found in many other cell types. Galectin-2, originally found in
human hepatoma, is isolated as a homodimer and has 43% amino acid identity to human
galectin-l (7). Galectin-5 was isolated from rat lung as a monomer of ~1 8 kDa and has now

been localized to rat erythrocytes (8). Galectin-7 is a 15 kDa monomeric protein found in

 

 

 

 

, o o o
Proto Time. I H-NPR~V-N~WG—E-R~F-G~R
Chimera Type: I (PGAYPGXXX).. I H—NPR~V-N~WG—E-R~F-G~R

 

 

 

O O O O O O O O O O O O
Tandem Repeat “Me: I H-NPR~V-N~WG-E-R~F-G~Rl H-NPR~V-N~WG-E-R~F-G~R

 

Figure 1. Schematic diagram illustrating the polypeptide architecture of galectins.
The Proto Type is composed of a single lectin domain that contains the CRD. The
Chimera Type galectin has two parts, a C-terminal half containing the galectin CRD
and an N-terminal half rich in proline and glycine residues. The Tandem Repeat Type
has two homologous CRD domains. Conserved amino acid residues in galectin CRD
are indicated. Dark circles denote residues that directly interact with the carbohydrate
by hydrogen bonding. 1n the N-terrninal domain of Chimera Type, nine amino acid
sequence motif that is tandemly repeated 8-12 (11) times is shown. The single letter
amino acid code is used. X, any amino acid.

Table l. Galectin members in mammalian species.

 

 

 

Galectin-l Proto 14,500 Homodimer
Galectin-2 Proto 14,600 Homodimer
Galectin-3 Chimera 35,000 Monomer
Galectin-4 Tandem Repeat 36,000 Monomer
Galectin-5 Proto 16,000 Monomer
Galectin-6 Tandem Repeat 34,000 Monomer
Galectin-7 Proto 15,000 Monomer
Galectin-8 Tandem Repeat 35,000 Monomer

 

 

 

 

 

 

human keratinocytes (9). Galectin-3, previously known as CBP35, was first isolated from
mouse 3T3 fibroblasts (10). It consists of a single polypeptide of ~35 kDa and is at present,
the only representative of the Chimera Type of galectins. The carboxy terminal half of
galectin-3 contains the CRD and the amino terminal half is proline and glycine rich because
the sequence PGAYPGXXX is repeated 8-12 times, depending on the species. This domain
shows sequence similarity to heterogeneous ribonucleoproteins (hnRNP) (11). Galectins-4,
-6 and -8 are Tandem Repeat Type galectins. Galectin-4 was found as an abundant rat
intestinal 36 kDa lectin (12). Galectin-6 has been mentioned (3) but its characterization is
yet to be published. Galectin-8 was originally cloned from rat liver with a molecular weight
of 34 kDa. (13). Unlike galectin-4, which is abundant in the intestine, galectin-8 is
expressed in liver, kidney, cardiac muscle, lung and brain. Overall, galectin-8 has 34%

identity to galectin-4.

Galectins were initially investigated mainly as vertebrate lectins from various
species like electric eel, chick, bovine, rat, mouse and human. However, recently
galectins have also been found in invertebrates (Table 2). A Tandem Repeat Type 32
kDa galectin was identified in the nematode Caenorhabditis elegans as the first
invertebrate galectin (14, 15). The internal sequence homology between the two
tandemly repeated domains is 32% amino acid identity between the first and second
domains. Both the first and second parts have 25-30% amino acid identity to vertebrate
lectins. A 16 kDa Proto Type galectin has also been identified in Caenorhabditis elegans
(16). A Tandem Repeat Type galectin has also been found in another member of the

nematode family, Onchocerca volvulus and this galectin has 71% amino acid identity to

Table 2. Galectins classified according to polypepetide architecture.

 

 

  

 

 

Proto Mammals Human, mouse, rat cow, etc. Galectins-1, -2, -5, -7
Birds Chicken, quail 14, 16 kDa galectins
Amphibia Frog (Xenopus, Rana) 14, 16 kDa galectins
Fish Electric eel, conger eel 14, 16 kDa galectins
Nematodes C. elegans 16 kDa galectin
0. voIvqus 16 kDa galectin
Sponges Geodia cydonium 15.1, 15.3 kDa galectins
Chimera Mammals Human, mouse, rat Galectin-3

Tandem-repeat

 

Mammals
Nematodes

 

Human, rat, mouse, pig
C. elegans

 

Galectins-4, -6, -8
32 kDajalectin

 

 

C. elegans 32 kDa galectin (17). Furthermore, two galectin cDNAs were cloned in the
marine sponge Geodia cydonium which code for Proto Type galectins (18). Therefore,

the galectin family is now known to be widely distributed in the animal kingdom.

Carbohydrate Recognition Domain of Galectins

The sequence of each CRD has been shown to be mainly encoded by 3 exons (19,
20). Most of the residues that are conserved among galectins are found in the sequence
encoded by the middle one of these 3 exons. In the crystallographic structure of galectin-
] and -2 (21-23), this sequence includes four contiguous B-strands and intervening loops
and contains all residues that interact directly with the carbohydrate ligand. However,
deletion of sequences encoded by the other two exons that encode the CRD also impairs

activity.

The key structure recognized by galectin CRD is lactose and other related B-
galactosides (24-28). Mechanisms underlying the specific interaction have been
elucidated by binding experiments, site-directed mutagenesis and x-ray crystallography.
X-ray crystal structures of galectin-1 (22, 23) and galectin-2 (21) confrrm that the major
interaction is with the galactose residue in lactose. Some of the highly conserved residues
of the CRD were found to form hydrogen bonds with hydroxyl groups of lactose; 4-OH of
Gal with His45, Asn47 and Arg49, 6-OH of Gal with Asn58 and Glu68 (residue numbers
are those of human galectin-2). However, interaction with glucose residue in lactose is
also significant as reflected in 100-fold higher affinity for lactose compared with

galactose (24). The 3-OH of glucose hydrogen bonds with conserved residues Arg49,

Glu68 and Arg70. Trp67 seems to have been conserved mainly for its structural
importance or stabilization of the binding site by close contact with C-6 of galactose.
Substitution of tyrosine by site-directed mutagenesis slightly reduced lactose binding
affinity but the mutant was still adsorbed strongly on lactose affinity column.
Substitutions of a series of the conserved hydrophilic residues (His45, As47, Arg49,
Asn58, Glu68 and Arg70) led to an almost complete loss of the binding ability (29, 5).
However, both N-terrninal and C-terminal regions of the 14 kDa polypeptide also seem to
be important for carbohydrate binding probably by maintaining structure of CRD, because
mutant proteins of the bovine 14-kDa lectin lacking N-terminal or C-terminal 10 residues

were found to be inactive (30).

Physical Properties and Subcellular Localization of Galectin-3

Galectin-3 was initially isolated from extracts of mouse 3T3 fibroblasts on the
basis of its galactose/lactose specific carbohydrate binding activity(10). Its homologs
have since been found in a variety of tissues and species . Sequence and hydropathy
analysis of the amino acid sequence for galectin-3 indicates that it has two distinct
structural domains (for review, see reference (31)). The C-terminal half, containing the
CRD, exhibits both hydrophilic and hydrophobic regions, as is characteristic of many
globular proteins . In contrast, the N-terminal half consists, in large part, of a repeated 9-
residue motif, PGAYPGXXX; the sequence is repeated eight times in human and
hamster, ten times in mouse and rat and twelve times in canine homolog . This proline-
and glycine-rich sequence is sensitive to collagenases whereas the CRD is very resistant

to proteolysis (32). Collagenase digestion has been used to generate the C-domain and

10

this domain has been shown to be sufficient for carbohydrate binding activity .
Differential scanning calorimetry of galectin-3 and its individual domains yielded
transition temperatures of 39 °C and 56 0C for N- and C-terrninal domains respectively
(33). Lactose binding by the C-domain shifted the transition temperature to 65 0C,
whereas sucrose failed to yield the same effect. These results suggest that the individual
domains of galectin-3 polypeptide are folded independently and that ligand binding by
galectin-3 is accompanied by a conformational change that significantly stabilizes the

polypeptide against thermal denaturation.

Mouse 3T3 fibroblasts, fixed with paraformaldehyde and permeabilized with
Triton X-100, yielded prominent nuclear and variable cytoplasmic staining with anti-
galectin-3 antibodies (34). The nuclear staining appeared to be diffuse, covering the
entire nucleus with the exception of about five black holes that correspond to nucleoli.
At the ultrastructural level, electron microscopic analysis using anti-galectin-3 antibodies
yielded immunogold labeling of interchromatic spaces and at the borders of condensed
chromatin (perichromatic fibrils) (35). This localization is similar to the reported
distribution of the non-snRNP splicing factors which will be discussed in detail later in
this review. The nuclear staining of galectin-3 in immunofluorescence and ultrastructural
experiments was sensitive to treatment of the permeabilized cells with ribonuclease A,
but not to parallel treatment with deoxyribonuclease I (34, 35). The nuclear and
cytoplasmic localization of galectin-3 has also been documented by quantitative
immunoblotting analysis of subcellular fractions, relative to enzyme markers

characteristic of the various fi‘actions (36).

11

The pI of murine galectin-3 is 8.7, as determined both by calculation from the
deduced amino acid sequence and experimentally by isoelectric focusing of recombinant
protein obtained by expression in E. coli (37). When extracts of mouse 3T3 cells were
subjected to two-dimensional gel electrophoresis and immunoblotting, two spots were
observed corresponding to pI values of 8.7 and 8.2. The pI 8.2 species represents a post-
translational modification of the pI 8.7 form by the addition of single phosphate group. It
has been shown that canine homolog of galectin-3 is phosphorylated mainly at Ser6 with
Ser12 carrying a small percent of total radiolabel (38). In addition, Casein Kinase I is
shown to phosphorylate recombinant human galectin-3 in vitro at Ser6. The
phosphorylated (pI 8.2) form of galectin-3 is found both in the cytosol and nucleus,
whereas the unmodified (pI 8.7) species is found exclusively in the nucleus. In quiescent
populations of cell cultures, galectin-3 expression level is low and it is present primarily
in the phosphorylated form, is located predominantly in the cytoplasm (3 7). Serum
stimulated cells have an increased level of galectin-3. The phosphorylated form is present
both in the cytosol and the nucleus but the amount of unmodified form increases

dramatically and is present only in the nucleus.

Studies on human colonic epithelial tissue specimens have shown that galectin-3
expression and its nuclear versus cytoplasmic localization vary along the crypt to surface
axis (39). Galectin-3 is concentrated in the nuclei of differentiated epithelial cells.
During the progression from normal mucosa to adenoma to carcinoma, the expression
and nuclear localization of galectin-3 changed significantly. In the carcinoma cells,

galectin-3 is absent from the nuclei and is localized only to the cytoplasm. These

12

observations suggest that the exclusion of galectin-3 from the nucleus may be related to

the neoplastic progression of colon cancer.

Although the majority of galectin—3 is intracellular, there is also evidence that
galectin-3 can be extemalized. From analyzing the cDNAs corresponding to galectin-3
from a number of species and sources, there does not appear to be a typical signal
sequence for sequestration into the endoplasmic reticulum and the endomembrane
pathway for secretion. Thus it is most probably extemalized by a non-classical
mechanism. Extemalization has been documented in macrophages (40, 41) and in
polarized kidney and intestinal epithelial cells, where the secretion is specifically toward
the apical surface (42, 43). Galectin—3 has also been found at the cell surface of dorsal
root ganglion neurons (44) and in murine fibrosarcoma cells (45). Thus, galectin-3 joins
a growing list of polypeptides that exhibit both intracellular (and, in particular, nuclear)
and extracellular localization, despite the lack of an obvious signal sequence. The best
examples of this list include (3) proteins with nuclear functions, such as the simian virus
40 (SV40) large T antigen (46); (b) members of the fibroblast growth factor (FGF) family
(47); (c) interleukins IL-or and 18 (48); and (d) the annexins (49). More recently, there
has been much interest on reports that the BReast CAncer 1 (BRCAl) gene product can

be found in the nucleus (50), cytoplasm (51), as well as outside of cells (52).

Subcellular Localization of Galectin-1
The subcellular localization of galectin-1 shares features observed for galectin-3:

predominantly intracellular but with some extemalization. Since this issue, particularly

13

with respect to the nuclear versus cytoplasmic localization, is one focus of my
dissertation, it will be discussed in Chapter 3, in the context of the data obtained in my

studies.

LITERATURE REVIEW ON PRE-mRNA SPLICING

Pre-mRNA Splicing and SnRNPs

Pre-mRNA splicing takes place via a two-step mechanism involving sequential
transesterification reactions (Figure 2) ( for reviews, see references (53, 54)). The first
step involves cleavage of pre-mRNA at the 5’ exon-intron boundary generating a free
exon 1 with a 3’-OH terminus and an intron lariat-exon 2. The lariat structure results
from the formation of a 2’-5’ phosphodiester bond linking 5’ terminus of the intron to
the ribose 2’-OH group of an adenosine residue in the intron, usually located 20-40
nucleotides upstream of the 3’ splice site. The second catalytic step is cleavage of 3’

splice site, ligation of the two exons and excision of the intron lariat.

The splicing reaction takes place within a multi-component complex called the
spliceosome. The Spliceosome is comprised of RNA-protein subunits called small
nuclear ribonucleoprotein particles (snRNPs) (Tables 3, 4). Each of the snRNP subunits
contain one or more small nuclear RNAs (snRNAs) and a set of snRNP proteins. In
mammalian species, there are six major snRNAs named U1 through U6 ranging in size
from 107-216 nucleotides. The snRNAs Ul-US are transcribed by RNA polymerase II
and snRNA U6 by RNA polymerase III. With the exception of U6, newly transcribed
monomethyl G-capped UsnRN A is transported to the cytoplasm where the cap is
modified to a trimethyl G structure and the snRNA assembles with proteins to form the
snRNP (for a review see references (53, 54). The major snRNPs are located in the

nucleoplasm except for U3snRNP which is localized in the nucleolus. Each snRNP

14

Spliceosome

 

P.

 

I exon] IPI cxon2 I

no 3,

Figure 2. Schematic diagram of spliceosome-catalyzed splicing.

16

Table 3. Major mammalian SnRNPs.

 

 

 

WP Nuclear “l " ’ strum ” . IProteinCom . “don:
_,,Localiza‘tion3 RNA. size 7 " " RNA ’ “Protein MrT
# -, pl.) .. I . 1: (“06100611“) polymerase: . , .13..» ,. + . > ,7; .-
Ul Nucleoplasm U1 165 11 Core proteins
. U1 70K 70,000
A 34,000
C 22,000
U2 Nucleoplasm U2 189 II Core proteins
A' 31,000
B" 28,000
U3 Nucleolus U3 216 11 U3 34K 34,000
U4 Nucleoplasm U4 139 II Core proteins
U5 Nucleoplasm U5 117 11 Core proteins
U5 100K 100,000
U6 Nucleoplasm U6 107 III

 

 

 

 

 

 

 

 

Table 4. SnRNP core proteins.

 

 

 

 

 

 

 

 

17

particle contains a common set of “core” proteins designated B, B’, D, D’, E, F and G
that range in size from 11-29 kDa. They interact with a conserved sequence present in
U1, U2, U4 and U5 snRNAs (55). Binding of the core proteins to the snRNA is directed
by a sequence motif of AUnG (n = 2-6). Insertion of this sequence into heterologous
RNA has been shown to induce the assembly of the snRNP core proteins suggesting that
this sequence is necessary and sufficient for core protein binding. Patients with
autoimmune diseases like systemic lupus erythmatosus (SLE) possess antibodies against
nuclear antigens called Srn antigens (56). Anti-Sm antibodies recognize antigenic

determinants on B and D polypeptides of the snRNP core proteins.

In addition to the core proteins, each of the spliceosomal snRNPs also contains
specific proteins. For example, the A, C and 70 kDa proteins are specific for UlsnRNP,
while A’ and B” are specific for U2 snRNP. It is not clear what roles these different
proteins might play in the splicing mechanism. It is possible that some of them may be

required for snRN P assembly or transport.

Spliceosome Assembly Pathway

Spliceosome assembly is a multistep process, the key steps of which are shown in
figure 3. Prior to the formation of first splicing complex, newly transcribed pre-mRNA is
bound by hnRNP proteins to form the hnRNP complex or the “H complex”. Formation
of the “H complex” does not require functional 5’ or 3’ splice sites, therefore, it is not
considered as a functional intermediate in the Spliceosome formation (5 7). First step is

the formation of a “commitment complex” which is a stable complex containing

18

5' splice siti branch Willi 1' splice site
5' .3- mm 0 I3.- 3' pre-mRNA

hnRNP proteins
0°C

5' -——.—-3' Complex H
\I

5! m® n m 3. Commitment

 

 

Complex
\1 {gap (non-snRNP splicing factors)
srr + srs
Complex A
Spliceosome
13' mRN A

 

 

 

Complex I T T

Figure 3. The pre-mRNA splicing pathway in mammalian cells.

19

UlsnRNP. Although the pre-mRNA contained within this complex has not yet been
spliced, at this stage it can no longer be displaced by the addition of an excess of
competitor pre-mRNA to the in vitro splicing reaction. The binding of UlsnRNP
involves the formation of base pairs between the intron sequence at the 5’ splice site and
a complementary sequence at the 5’ terminus of UlsnRNA. UlsnRNP is also involved in
interactions with the sequences at the 3’ end of the intron. Such interactions may play a
role in bringing together the separate ends of the intron within the spliceosome. Next a
pre-spliceosome complex called the “A complex” is formed by the ATP dependent
binding of U2snRNP to the pre-mRNA branch site. Binding of U2snRNP requires
UlsnRNP (58) and three protein factors SFl, SF3 and U2 auxiliary factor (UZAF) (59-
61). The spliceosome or “B complex” is formed following the addition of a
preassembled U4/U6.U5 tri-snRNP particle. U4 and U6 snRNAs bind to each other
through an extended complementary region to form the U4/U 6 snRNP complex (62) and
then the U4/U6 snRNP particle associates with U5 snRNP (57). This tri-snRNP is then
bound to “A complex” to form the “B complex” or spliceosome. At this point the two
transesterification reactions (figure 2) take place within the spliceosome complex. The
spliceosome complex dissociates and the newly joined exons are released from the
snRNP-intron “1 complex”. The intron RNA is degraded and snRNPs are recycled for

further splicing.

Non— snRNP Splicing Factors

In addition to the snRNP subunits, spliceosomes are also comprised of ~100 non-

snRN P protein splicing factors which are assembled onto pre-mRNA in an orderly

20

fashion to form spliceosomes. These factors function at multiple steps during
spliceosome assembly. They are involved both in recruiting spliceosomal snRNPs to the
pre-mRNA and in mediating interactions between spliceosomal components to specify
splice sites and define exons and introns. Several of these protein factors have been
identified including ASF/SF2 (63,64), U2AF (U2snRNP Auxiliary Factor) (61) 65-kDa
subunit (U2AF65), U2AF 35-kDa subunit (U2AF35), SC35 (65), UlsnRNP 70-kDa
protein (66) etc. (Table 5). A sequence comparison of these factors reveal a number of
similarities including the presence of RNA recognition motifs (RRM) and SR motifs

making them members of the SR protein family (for a review see reference (67).

SR proteins were originally described by Roth and coworkers based on five
criteria (68). i) They contain a shared phosphoepitope recognized by the monoclonal
antibody mAb104. ii) They copurify in a two-step salt precipitation procedure (soluble in
65% ammonium sulfate and precipitated in 20 mM MgClz). iii) They share a similar
protein structure containing at least one RNA recognition motif (RM) and an RS
domain. iv) Their sizes on SDS-PAGE are conserved from Drosophila to man. v) They
can complement splicing deficient cytoplasmic 8100 extracts, indicating a redundant
function in splicing. A second class of proteins called SR protein related polypeptides
(SRrp) also contain SR domain but lack RRMs. Together, SR proteins and SR protein

related polypeptides constitute the superfamily of RS domain proteins.

The length and composition of the SR domain and its location within the primary

sequence of the protein is variable (figure 4). For example, the SR motif is located at the

21

Table 5. Superfamily of pre-mRNA splicing factors containing an SR domain.

 

 

 

Igproteins ISR protein related polypeptides
IHumans Other Species Humans Other Species

SRp20 x16 (mouse) = SRp20 U170k Tra (Dros .)

SRp30a/SF2/ASF pr264 (chicken) = SC35 U2AF65 Tra-2 (Dros .)

SRp30b/SC35 HRS (rat) = Srp40 U2AF35 SWAP (Dros .)

SRp30c B52 (Dros .) = SRpSS HCCl xUl70k (Xenopus) = U170k
SRp40 RBP-l (Dros .) HRHl dU170k (Dros .) = U170k
SRpSS SR-l (plant) CLK-l mU2AF65 (mouse) = U2AF65
SRp75 Npl3 (S. cerevisiae) CLK-2 dU2AF50 (Dros .) = U2AF 65
968 CLK-3 dU2af38 (Dros .) = U2AF35

 

 

 

 

 

 

Dros., Drosophila
=, homologous

Table modified from reference (67)

 

22

N-terminus of U2AF65, while in ASF/SF2 and SC35 the SR motif occurs at the C-
terminus. The SR motif itself can either consist of an uninterrupted stretch of SR
dipeptides or of a more dispersed SR-rich region. RRM motifs have been found in
diverse number of proteins that bind RNA. RRM includes two conserved sequences,
RNP-l and RNP-2. The RRM domain has been shown to be necessary for high affinity

and sequence specific binding of protein to RNA.

In addition to the SR protein family members described above, an RS domain is
present in other essential splicing factors and regulators. In fact, the prototype for RS
domain proteins are the Drosophila splicing regulators SWAP (69), Tra (70), Tra-2 (71).
These three proteins are not essential splicing factors, but are specific splicing regulators.
Tra and Tra-2 are required for female-specific splicing of the doublesex (sz) transcript
(for review see reference (72)), and SWAP undergoes autoregulation by suppressing the
splicing of its own transcript. These and other SR domain proteins are collectively

referred to as SR related polypeptides, or ‘SRrps’.

Among the SRrps is the U1 snRNP-associated 70-kDa protein (U 1 70k) that
contains an RS domain (67). UZSnRNP auxiliary factor (U2AF) is a protein splicing
factor required for stable binding of U25nRNP to pre-mRNA during spliceosome
assembly (62). It binds specifically to polypyrimidine tract/3’ splice site region of pre-
mRNAs. U2AF is comprised of two associated polypeptides of 35 kDa (U2AF 35) and 65
kDa (U2AF 65) which are present in equimolar amounts. U2AF 65 is an essential splicing

factor that has an N-terminal RS domain followed by three RNA-binding domains at the

23

 

 

 

 

 

-I RRM —E ®— ASF/SFZ
-I RRM ‘ PG @ SC35

 

 

 

 

 

 

 

 

 

 

U2AF 65

 

 

 

@—
®J U2AF 35

 

 

IQI

Figure 4. Schematic diagram summarizing the context and organization of
specific domains in the SR family of proteins.

The RRM box indicates the RNA recognition domain. The relative location
of the serine/arginine rich domain (SR) is shown for each protein. PG and G
indicate proline/ glycine and glycine rich regions, respectively.

24

C-terminus (73). However, similar to Tra and SWAP, the small subunit of U2AF (known
as U2AF 35 in human and U2AF38 in Drosophila) contain an SR domain but lacks an
RNA-binding motif, and appears unnecessary for constitutive splicing in nuclear extracts
from which the U2AF is removed (74). However, recent studies show that U2AF35 is
involved in specific protein-protein interactions with SR proteins (75), and both subunits
of U2AF are specifically associated with the mammalian commitment complex (76).
Finally, the most recent addition to the superfamily of SR domain proteins is a kinase
called CLK/STY which can interact with SR proteins in yeast two-hybrid system and
phosphorylate these proteins. It is a dual specificity kinase that can autophosphorylate at
both serine/threonine and tyrosine residues (77, 78). Human cells express two other
related kinases CLK-2 and CLK-3. These kinases contain an N-terrninal domain
embedded with multiple non-adjacent RS or SR dipeptides and a C-terminal catalytic
core present in all serine/threonine kinases. It remains to be tested whether all of these

kinases can interact and phosphorylate SR proteins.

In cellular extracts lacking SR proteins, such as 8100 (79, 80) no specific splicing
complex could be detected, indicating that SR proteins act early in spliceosome assembly.
These observations qualify SR proteins as essential splicing factors. Individual SR
proteins can each complement this splicing deficient extract suggesting that SR proteins
have redundant functions in splicing (81). However, despite their redundant function in
complementing 8100, individual SR proteins have distinct specificity and efficiency in
splicing different pre-mRNA substrates. For example, HIV tat pre-mRNA splicing in

HeLa nuclear extracts depends on SF2/ASF but not SC35 (82). In another case, SC35 but

25

not SF2/ASF can replace the 968 antigen to restore splicing in 9G8 depleted nuclear
extracts (83). In addition, different SR proteins display distinct activity in 5’ splice site
selection (84). For example, Srp40 and Srp55 promote the use of a distal 5’ splice site for
SV40 large T splicing, whereas SC35 stimulates the use of a proximal splice site for the
SV40 small t splicing. A specific SR protein effects 5’ splice site selection of a given
pre-mRNA by either selectively recruiting U1 snRNP to a specific site or by promoting
U1 snRNP occupation of all potential binding sites, where the most proximal site is

selected by an additional mechanism (85).

U1 snRNP depleted nuclear extracts are inactive for pre-mRNA splicing.
Interestingly, addition of excess SR proteins to these extracts rescues both spliceosome
assembly, in particular the association of U2 snRNP with the branch site, and splicing
activity, demonstrating that SR proteins can bypass the requirement for U1 snRNP (86).
U1 snRNP independent specification of the 5’ splice site in the reconstituted reaction

occurs after the U2 snRNP-branch site interaction and is mediated by U6 snRNA.

Commitment of a pro-mRNA to splicing requires the initial binding of a specific
SR protein to the RNA molecule. However, in contrast to the U2AF 65, which binds
specifically to the polypyrimidine tract at the 3’ splice site, the RNA-binding specificity
for most SR proteins remains unknown. In fact, purified SC35 alone was shown to bind
nonspecifically to RNA, but in splicing complexes SC35 was found to be associated with

a complex bound to the 3’ splice site (87), indicating that other spliceosomal components

26

may cooperate with SC35 to acquire an RNA-binding specificity during spliceosomal

assembly.

Nuclear Structure and RNA Splicing

The subnuclear locations of a number of proteins and RNAs important for RNA
splicing have been studied by light and electron microscopy. The most striking feature is
the concentration of snRNPs U1, U2, U4-U6 and U5 in a specific distribution within the
nucleus that has been termed the speckled pattern (88). The speckled pattern is composed
of 20-50 intensely stained and irregularly shaped regions in the nucleus that are connected
in places and set against a diffuse nuclear labeling . The diffuse nuclear staining of
snRNPs may represent an excess soluble population, snRNPs in transit to or from nascent
transcripts, or it may represent snRNPs in transit to speckles from their assembly sites in
the cytoplasm. The use of antibodies to trimethylguanosine cap of some of the snRNAs
and of antisense probes to the snRNA (89, 90) revealed that the RNA components of the
snRNPs are also organized in the nucleus as a speckled pattern. Several non-snRNP
splicing factors such as SC35, SF2/ASF have also been localized to speckled nuclear

regions (for a review, see reference (88)).

Electron microscopy with antibodies to snRNPs and SC35 showed that the
speckled pattern seen by immunofluorescence corresponds to structures termed
interchromatin granules and perichromatin fibrils (91, 92). Interchromatin granules
corresponds to the larger intensely stained and irregularly shaped speckles seen by

immunofluorescence; they have little to no [3 H] uridine labeling in their interior and may

27

represent the sites of splicing factor storage and/or assembly. Perichromatin fibrils are
found on the surface of and between interchromatin granule clusters; they are rapidly
labeled with [3H] uridine and are thought to represent nascent pre-rnRNA transcripts. It
was proposed that snRNPs from interchromatin granule clusters move to the sites of
active transcription (perichromatin fibrils) to splice nascent pre-mRNA transcripts (93,
94). Three dimensional reconstruction techniques have shown that the snRNPs are not
present in isolated islands, instead portions of speckled pattern are connected to form a
latticework that extends between the nucleolar surface and the nuclear lamina-envelope.
The organization of splicing factors in a latticework is probably dynamic and reflects the
physiological state of the cell. Therefore, at any given time the shape, connections

between, and organization of, speckles may vary (95).

To analyze the spatial and temporal organization of pre-mRNA splicing in
mammalian cells, several individual RNA species transcribed from endogenous templates
have been localized and compared to the localization of splicing factors. The induced
expression of c-fos transcripts was found to be closely associated with splicing factors
(96). In addition it has been shown that the association of nascent RNA transcripts with
splicing factors is intron dependent druing transient or stable expression suggesting a

close link between transcription and splicing (95).

(1)
(2)

(3)

(4)
(5)
(6)
(7)

(8)

(9)

(10)

(11)
(12)

REFERENCES

Barondes, SH. (1988) Trends Biochem. Sci. 13, 480-482.

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

Barondes, S.H., Castronovo, V., Cooper, D.N.W., Cummings, R.D., Drickamer,
K., Feizi, T., Gitt, M.A., Hirabayashi, 1., Hughes, R.C., Kasai, K., Leffler, H., Liu,
F.-T., Lotan, R., Mercurio, A.M., Monsigny, M., Pillai, S., Poirer, F ., Raz, A.,
Rigby, P., Rini, J. M., and Wang, J.L. (1994) Cell 76, 597-598.

Drickamer, K. and Taylor, ME. (1993) Ann. Rev. Cell Biol. 9, 237-264.
Hirabayashi, J. and Kasai, K. (1993) Glycobiology 3, 297-304.

Barondes, SH. (1984) Science 223, 1259-1264.

Gitt, M.A., Massa, S.M., Leffler, H., and Barondes, SH. (1992) J. Biol. Chem.
267, 10601-10606.

Gitt, M.A., Wiser, M.F., Leffler, H., Herrman, J ., Xia, Y.R., Massa, S.M., Cooper,
D.N.W., Lusias, A.J., and Barondes, SH. (1995) J. Biol. Chem. 270, 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. 270, 5823-5829.

Roff, CF. and Wang, J.L. (1983) J. Biol. Chem. 258, 10657-10663.

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

Oda, Y., Herrman, J., and Gitt, A.M., Turck, C.W., Burlingame, A.M., Barondes,

SH, and Leffler, H. (1993) J. Biol. Chem. 268, 5929-5939.

28

(13)

(14)

(15)

(16)

(17)

(18)

(19)

(20)

(21)

(22)

(23)

(24)

(25)

(26)

29

Hadari, Y.I., Paz, K., Dekcl, R., Mestrovie, T., Accili, D., and lick, Y. (1995) J.
Biol. Chem. 270, 3447-3453.

Hirabayashi, J., Satoh, M., Ohyama, Y., and Kasai, K. (1992) J. Biochem. 111,
553- 555.

Hirabayashi, J ., Satoh, M., and Kasai, K. (1992) J. Biol. Chem. 267, 15485-
15490.

Hirabayashi, J ., Ubukata, T. and Kasai, K. (1996) J. Biol. Chem. 271, 2497-2505.
Klion, A. D. and Donelson, J .E. (1994) Mol. Biochem. Parasitol. 65, 305-315.
Pfeifer, K., Haasermann, M., Gamulin, V., Bretting, H., Fahrenholz, F., and
Muller, W.E.G. (1993) Glycobiology 3, 179-184.

Gitt, MA. and Barondes, SH. (1991) Biochemistry 30, 82-89.

Gritzmacher, C.A., Mehl, V.S., and Liu, F.T. (1992) Biochemistry 31, 9533-
9538.

Lobsanov, Y.D., Gitt, M.A., Leffler, H., Barondes, SH, and Rini, J .M. (1993) J.
Biol. Chem. 268, 27034-27038.

Liao, D.I., Kapadia, G., Ahmed, H., Vasta, GR, and Herzberg, O. (1994) Proc.
Natl. Acad. Sci. USA 91, 1428-1432.

Bourne, Y., Bolgiano, B., Liao, D.-I., Strecker, G., Cantan, P., Herzberg, 0., Feizi,
T., and Cambillau, C. (1994) Struct. Biol. 1, 863-870.

Leffler, H. and Barondes, SH. (1986) J. Biol. Chem. 261, 10119-10126.

Abbott, W.M., Hounsell, ER, and Fiezi, T. (1988) Biochem. J. 252, 283-287.
Lee, R.T. Ichikawa, Y., Allen, H.J., and Lee, Y.C. (1990) J. Biol. Chem. 265,

7864-7871.

(27)

(28)

(29)
(30)

(31)

(32)

(33)

(34)

(35)

(36)

(37)

(38)

(39)

(40)

(41)

(42)

30

Sato, S. and Hughes, RC. (1992) J. Biol. Chem. 267, 6983-6990.

Knibbs, R.N., Agrwal, N., Wang, J.L., and Goldstein, 1.]. (1993) J. Biol. Chem.
268, 14940-14947.

Hirabayashi, J. and Kasai, K. (1994) Glycoconjugate J. 11, 43 7-442.

Abbot, W.M., and Feizi, T. (1991) J. Biol. Chem. 266, 5552-5557.

Anderson, R.L. and Wang, J .L. (1992) Trends Glycosci. Glycotechnol.
4, 43-52.

Raz, A., Pazerini, G., and Carmi, P. (1989) Cancer Res. 49, 3489-3493.

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

Wang, S.Y., Voss, P.G., Patterson, R.J., and Wang, J .L. (1995) Antibody,
Immunoconjugate, and Radiopharrnaceuticals 8, 311-324.

Hubert, M., Wang, S.Y., Wang, J.L., Seve, AP, and Hubert, J. (1995)
Exp. Cell Res. 220, 397-406.

Moutsatsos, I.K., Davis, J.M., and Wang, J .L. ( 1986) J. Cell Biol. 102,
477-483.

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

Huflejt, M.E., Turck, C.W., Lindstedt, R., Barondes, SH, and Leffler, H.
(1993) J. Biol. Chem. 268, 26712-26718.

Lotz, M.M., Andrews, C.W., Kor'zelius, C.A., Lee, E.C., Steele, G.D., Clarke,
A., and Mercurio, A.M. (1993) Proc. Natl. Acad. Sci. USA 90, 3466-3470.

Cherayil, B.J., Weiner, SJ, and Pillai, S. (1989) J. Exp. Med. 170, 1959-1972.
Ho, MK. and Springer, TA. (1982) J. Immunol. 128, 1221-1228.
Lindstedt, R., Apodaca, G., Barondes, S.H., Mostove, K., and Leffler, H. (1993) J.

Biol. Chem. 268, 11750-1 1757.

31

(43 ) Sato, S., Burdett, J ., and Hughes, RC. (1993) Exp. Cell Res. 207, 8-18.

(44)

Hynes, M.A., Gitt, M.A., Barondes, S.H., Jessell, T.M., and Buck, LB.

(1990) J. Neurosci. 10, 1004-1013.

(45) Raz, A. and Lotan, R. (1987) Cancer Mestastasis Rev. 6, 433-452.

(46)

(47)

(48)

(49)

(50)

(51)

(52)

(53)

(54)

(55)

(56)

(57)

Butel, J.S. and Jarvis, D.L. (1986) Biochim. Biophys. Acta 865, 171-195.

Mason, 1.]. (1994) Cell 78, 547-552.

Siders, W.M., Klimovitz, J.C., and Mizel, SB. (1993) J. Biol. Chem. 268, 22170-
22174.

Kojima, K., Ogawa, H., Seno, N., Yamamoto, K., Irimura, T., Osawa, T., and
Matsumoto, I. (1992) J. Biol. Chem. 267, 20536-20539.

Scully, R., Ganesan, S., Brown, M., DeCaprio, J.A., Cannistra, S.A., Feunteun, J .,
Schnitt, S., and Livingston, D.M. (1996) Science 272, 123-125.

Chen, Y., Chen, C.-F., Riley, D.J., Allred, D.C., Chen, P.L., VonHoff, D.,

Osborne, C.K., and Lee, W.H. (1995) Science 270, 789-791.

Jensen, R.A., Thompson, M.E., Jetton, T.L., Szabo, C.I., van der Meer, R., Helon,

B., Tronick, S.R., Page, D.L., King, M.C., and Holt, IT (1996) Nature Genetics
12, 303-308.

Lamond, A.I. (1993) BioEssays 15, 595-603.

Sharp, P.A. (1994) Cell 77, 805-815.

Luhrmann, R., Kastner, B. and Bach, M. (1990) Biochim. Biophys. Acta 1087,
265-292.

Lerner, MR. and Steitz, J.A. (1979) Proc. Natl. Acad. Sci. USA 76, 5495-5499.

Konarska, M.M. and Sharp, P.A. (1987). Cell 49, 763-774.

(58)

(59)
(60)

(61)

(62)
(63)
(64)
(65)
(66)
(67)

(68)

(69)

(70)

(71)
(72)
(73)

- (74)

32

Barbino, M.L., Blencowe, B.J., Ryder, U., Sproat, BS, and Lamond, A1. (1990)
Cell 63, 293-302.

Kramer, A. and Utans, U. (1991) EMBO J. 10, 1503-1509.

Ruskin, B., Zamore, PD. and Green, MR. (1988) Cell 52, 207-219.

Zamore, PD. and Green, MR. (1989) Proc. Natl. Acad. Sci. USA 86, 9243-

9247.

Hashimoto, C. and Steitz, J .A. (1984) Nuc. Acids. Res. 12, 3283-3293.

Ge, H. and Manley, J.L. (1990) Cell 62, 25-34.

Krainer, A., Conway, GO and Kozak, D. (1990) Cell 62, 35-42.

Fu, X.D.,and Maniatis, T. (1990) Nature 343, 437-441.

Krainer, A., Mayeda, A., Kozak, D., and Binns, G. (1991) Cell 66, 383-394.
Fu, X.D. (1995) RNA 1, 663-680.

Zahler, A.M., Lane, W.S., Stolk, J .A., and Roth, M.B. ( 1992) Genes Dev. 6, 837-

847.

Chou, T.B., Zachar, Z. and Bingham, RM. (1987) EMBO J. 6, 4095-4104.
Boggs, R.T., Gregor, P., Idriss, S., Belote, J .M., and McKeown, M. (1987) Cell

50, 739-747.

Amrein, H., Gorman, M. and Nothiger, R. (1988) Cell 55, 1025-1035.

Baker, BS. (1989) Nature 314, 521-524.

Zamore, P.D., Patton, J .G., and Green, MR. (1992) Nature 355, 609-614.
Zhang, M., Zamore, P.D., Carmo-Fonseca, M., Lamond, AL, and Green, MR.

(1992) Proc. Natl. Acad. Sci. USA 89, 8769-8773.

(75)

(76)

(77)

(78)

(79)

(80)

(81)

(82)

(83)

(84)

(85)

(86)

(87)

(88)

(89)

33

Wu, J.Y. and Maniatis, T. (1993) Cell 75, 1061-1070.

Stakins, D. and Reed, R. (1994) Mol. Cell Biol. 14, 7670-7682.

Ben-David, Y., Letwin, K., Tannock, L., Bernstein, A., Pawson, T. (1991) EMBO
J. 10, 1273-1278.

Howell, B.W., Afar, D.E., Lew, J ., Douville, E.M., Icely, P.L., Gray, D.A., and
Bell, J. (1991) Mol. Cell Biol. 11, 568-572.

Krainer, A., Conway, G.C. and Kozak, D. (1990) Genes Dev. 4, 1158-1171.
Zahler, A.M., Neugebauer, K., Stolk, J.A., Roth, MB. (1993) Mol. Cell Biol. 13,

4023-4028.

Mayeda, A., Zahler, A.M., Krainer, A, Green, MR. (1992) Proc. Natl. Acad. Sci.
USA 89, 1031-1300.

Fu, X.D. (1993) Nature, 365, 82-85.

Cavaloc, Y., Popielarz, M., Fuchs, J .P., Gattoni, R., and Stevenin, J. (1994)
EMBO J. 13, 2639-2649.

Zahler, A.M., Neugebauer, K., Stolk, J .A., Roth, MB. (1993) Science 260, 219-
222.

Eperon, I.C., Ireland, D.C., Smith, R.A., Mayeda, A., and Krainer, AR. (1993)
EMBO J. 12, 360703617.

Tarn, W.Y. and Steitz, J.A. (1994) Genes Dev. 8, 2704-2717.

Fu, X.D.,and Maniatis, T. (1992) Proc. Natl. Acad. Sci. USA 89, 1725-1729.
Spector, D.L. (1993) Annu. Rev. Cell Biol 9, 265-315.

Carmo-Fonseca, M., Pepperkok, R., Carvalho, M.T., and Lamond, AL (1992) J.

Cell Biol. 117, 1-14.

(90)
(91)
(92)
(93)

(94)

(95)

(96)

34

Huang, S. and Spector, D. (1992) Proc. Natl. Acad. Sci. USA 89, 305-308.
Spector, D., Fu, X.D., and Maniatis, T. (1991) EMBO J. 10, 3467-3481.

Fakan, S., Leser, G. and Martin, T.E. ( 1984) J. Cell Biol. 98, 358-363.
Jimenz-Garcia, LP. and Spector, D.L. (1993) Cell 73, 47-57.

Pombo, A., Ferreira, E., Bridge, E., and Carmo-Fonseca, M. (1994) EMBO J. 13,
5075-5085.

Huang, S. and Spector, D.L. (1996) J. Cell Biol. 133, 719-732.

Huang, S. and Spector, D.L. (1991) Genes Dev. 5, 2288-2302.

CHAPTER II

Galectin-l in the Cell Nucleus. 1. Evidence for a Role in Pre-mRNA Splicing
Anandita Vyakarnam', Sue F. Dagher”, John L. Wang], and Ronald J. Patterson4

Departments of 1Biochemistry and 4Microbiology and 2Genetics Program,
Michigan State University, East Lansing, MI 48824

3 Present address: Department of Biochemistry, Dartmouth Medical School,
Hanover, NH 03755

Acknowledgments: We thank Drs. Sam Barondes and Hakon Leffler for their generous
gifts of recombinant rat galectin-1 and rabbit antiserum directed against recombinant rat
galectin-1. This work was supported by grants MCB 91-22363 fiom the National Science
Foundation, GM-38740 from the National Institutes of Health, No. 99 fi'om the Mizutani
Foundation for Glycoscience, and an Interdisciplinary Research Award from the Cancer
Center at Michigan State University.

35

ABSTRACT

Galectins are a family of B-galactoside-binding proteins that contain characteristic
amino acid sequences in the carbohydrate recognition domain of the polypeptide. The
polypeptide of galectin-1 contains a single domain, the carbohydrate recognition domain.
The polypeptide of galectin-3 contains two domains, a carbohydrate recognition domain
fused onto a proline-, and glycine-rich domain. In previous studies, we had shown that
galectin-3 is a required factor in the splicing of nuclear pre-mRNA, assayed in a cell-free
system. We now document that: (a)nuclear extracts derived from HeLa cells contain both
galectins-1 and -3; (b)depletion of both galectins from the nuclear extract, either by
lactose affinity adsorption or by double antibody adsorption, resulted in the concomitant
loss of splicing activity; (c)depletion of either galectin-1 or galectin-3, by specific
antibody adsorption, failed to remove all of the splicing activity and the residual activity
was still saccharide inhibitable; (d)either galectin-1 or galectin-3 alone was sufficient to
reconstitute, at least partially, the splicing activity of nuclear extracts depleted of both
galectins; and (e)although the carbohydrate recognition domain of galectin-3 (or galectin-
1) is sufficient to restore splicing activity to galectin-depleted nuclear extract, the
concentration required for reconstitution was far greater than that of the full-length

galectin-3 polypeptide.

36

INTRODUCTION

Galectins are a family of widely distributed proteins that: (a)bind to B-
galactosides; and (b)contain characteristic amino acid sequences in the carbohydrate
recognition domain (CRD) of the polypeptide. At present, eight mammalian galectins
have been reported and classified into three subgroups, according to the content and
organization of the domains in their respective polypeptides (for reviews, see references 1
and 2). The Prototype subgroup consists of polypeptides (~ 14 kD) with a single domain,
the CRD. Galectins-1, -2, -5, and -7 are members of this subgroup. Another subgroup is
the Tandem Repeat type, which has three members: galectins-4, -6, and -8. These
galectins have two domains, each a CRD, connected by a linker region. Finally, the
Chimera subgroup is, at present, represented by a single member, galectin-3. Its

polypeptide contains two domains, a CRD fused onto a Pro-, Gly-rich domain.

In previous studies, we had reported the localization of galectin-3 to the cell
nucleus, in the form of a ribonucleoprotein (RNP) complex (3, 4). We had also identified
it as one of the many proteins required for the splicing of pre-mRNA, assayed in a cell-
free system. This conclusion was based on several key findings (5): (a)nuclear extracts
(NE) derived from HeLa cells, capable of carrying out splicing, contain galectin-3; (b)NE
depleted of galectin-3 by affinity adsorption on lactose (Iac)-agarose become deficient in
splicing; (c)the activity of the Lac-agarose depleted extract could be reconstituted by the

addition of purified recombinant galectin-3; and (d)saccharides that bind galectin-3 with

37

38

high affinity inhibit the cell-free splicing reaction. These results strongly suggested that

the lectin is a required factor in cell-free splicing of pre-mRNA.

When we attempted to deplete the splicing activity using a galectin-3 specific
monoclonal antibody, the anti-Mac-Z (anti-M2) antibody (6, 7), we were surprised to find
that there was some, but not complete, loss of splicing activity in the galectin-3-depleted
extract. We thus performed a series of experiments to resolve this apparent dilemma. In
the present communication, we report that, in addition to galectin-3, galectin-1 is also a
component of NE and it also plays a role in the splicing of pre-mRNA. The
accompanying chapter documents the nuclear localization of galectin-1, using laser

confocal scanning microscopy.

MATERIALS AND METHODS

Preparation of HeLa Nuclear Extracts
HeLa S3 cells were obtained from American Type Culture Collection (CCL 2.2)

and grown in suspension culture in Minimum Essential Medium containing 10% defined
/ supplemented bovine calf serum (HyClone), 100 U/ml penicillin and 100 U/ml
streptomycin at 37°C in humidified atmosphere of 5% C02. Nuclear extract (NE) was
prepared 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 phenyl methylsulfonyl fluoride (PMSF), 0.5 mM
dithiothreitol (DTT)] as described (8). NB was frozen as aliquots in dry ice/ethanol bath
and stored at -80°C. Typically NEs with protein concentrations of 10-13 mg/ml were
prepared. Protein concentrations were determined by the method of Bradford using

bovine serum albumin as the protein standard (9).

Antibodies and Affinity Columns

Anti-M2 is a rat monoclonal antibody directed against the Mac-2 antigen (6)
which has been shown to be galectin-3 (7). The antibody was purified from serum-free
cell culture supernatant derived from hybridoma line M3/38.1.2.8.HL.2 obtained from
ATCC (TIB 166). Anti-transferrin receptor (TR) antibody is an isotype matched rat
monoclonal antibody (Ingax) that was used as a control for anti-M2. The hybridoma
(R17 217.1.3) producing the anti-(TR) antibody was obtained from ATCC (TIB219). The

polyclonal rabbit antiserum against recombinant rat galactin-l (anti-G1) was a gift from

39

40

Dr. Hakon Leffier (UCSF, CA). Human autoimmune serum reactive with the Sm antigens

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

For saccharide affinity column, or-lactose insolubilized on 6% beaded agarose
(Lac-agarose) was purchased from Sigma. Cellobiose-agarose was used as a control for
NE depletion. For immunodepletions, antibodies were immobilized on pre-swollen
protein G-Sepharose beads (Sigma). Beads were washed with 20 mM Hepes (pH 7.9), 0.5
M NaCl. 150 pl of washed beads were mixed with 150 pl serum or 100 pg monoclonal
antibody. The mixture was adjusted to 20 mM Hepes (pH 7.9), 0.5 M NaCl and incubated
for 1 h at room temperature with continual rocking. The beads were washed with 1 ml of
0.2 M sodium borate, pH 9.0 and resuspended in the same buffer. Dimethylpimelimidate
(Pierce) was added to final concentration of 20 mM to covalently crosslink antibodies to
protein G-Sepharose. After 1 hour incubation at room temperature, the beads were
washed and incubated in lml of 0.2 M ethanolamine (pH 8.0) for one hour. The coupled
antibody-Sepharose mixture was washed five times with 1 ml of 20 mM Hepes (pH 7.9),

0.5 M NaCl and used for immunodepletions.

Depletion and Reconstitution of Nuclear Extract

NEs were depleted of galectins by adsorption on either a saccharide affinity
column or an antibody affmity column. For saccharide affinity depictions, 150 pl of Lac-
agarose beads were washed with 20 mM Hepes (pH 7.9), 0.5 M NaCl and 30 pl NE were
added and the mixture incubated on ice for 20 min in disposable spin columns

(Millipore). For imunodepletions, 30 p1 of NE were incubated with either 60 pl of anti-

41

Mac-2-Sepharose or 90 pl of anti-galectin-l-Sepharose beads for single antibody
depletions and with a mixture of both for double antibody depletions. After incubation on
ice for 20 min with appropriate affinity matrices, the unbound fraction was removed. The
beads were washed with 12 pl of buffer 1 [60% Dignam buffer D adjusted to 0.42 M
NaCl, buffer D: 20 mM Hepes (pH 7.9), 20% v/v glycerol, 0.1M KCl, 0.2 mM EDTA,
0.5 mM PMSF and 0.5 mM DTT] and this wash was added to the unbound fraction. The
beads were washed three times with 1 ml of buffer 1 and the bound material eluted by

boiling in 100 pl of Laemmli sample buffer (10).

Aliquots of nondepleted NE and unbound fiactions of saccharide depletions or
immunodepletions were dialyzed in a microdialyzer against 60% buffer D for 40 min at
4°C. The final protein concentration of the depleted and dialyzed extracts was 6-8 mg/ml.
The dialyzed fractions were then assayed for splicing activity. In reconstitution
experiments, recombinant proteins were added to the unbound fractions prior to dialysis.
In experiments directly comparing the efficiency of various recombinant proteins in
reconstituting splicing activity, aliquots taken from the same unbound fraction were
reconstituted with different amounts of proteins. In experiments testing the effect of
exogenously added carbohydrates, the extracts were incubated with thiodigalactoside for

5 min at room temperature after dialysis but prior to the addition of pre-mRNA substrate.

In Vitro Splicing Assay

MINX pre-mRNA, used as a substrate in the splicing assay, was transcribed from

the plasmid provided by Susan Berget (Baylor College of Medicine, Houston) using SP6

42

polymerase (Gibco). The MINX pre-mRNA was labeled with 32P-GTP and the

monomethyl cap was added during transcription.

Splicing reaction mixtures (10 pl) contained dialyzed NE (4 pl) or unbound
fraction (8 pl), ”9-me pre-mRNA, 2.5 mM Mgcrz, 1.5 mM ATP, 20 mM creatine
phosphate, 0.5 mM DTT and 20 units of RNasin (Promega). Splicing reactions were
carried out at 30°C for 45 min. Proteinase K/SDS was added to a final concentration of 4
mg/ml proteinase K and 0.1% SDS and the sample was incubated at 37°C for 15 min.
Each splicing sample was diluted to 100 pl with 125 mM Tris (pH 8.0), 1 mM EDTA, 0.3
M sodium acetate. RNA was extracted with 200 pl of phenol/chloroform (50:50 v/v)
followed by 200 pl of chloroform. RNAs were precipitated with 400 pl of ethanol at
-80°C. The extracted RNAs were subjected to electrophoresis through 13%
polyacrylamide (bisacrylamide/acrylamide, 1.9:50 wat), 8.3 M urea gels followed by

autoradiography.

The intensities of the bands on the gels were quantitated by direct B-particle
counting using an AMBIS Systems system. The percent product formation was calculated
by dividing the radioactivity present in the final product (ligated exon l-exon 2) by the
total radioactivity present in the pre-mRN A substrate (exon l-intron-exon 2), the splicing

intermediates (exon 2-lariat, exon 1) and the product at the end of the incubation.

43

Immunoblot Analysis

For immunoblot analysis, protein samples were resolved on 12.5% SDS-PAGE as
described by Laemmli (10) and electrophoretically transferred to Irnmobilon-P membrane
(Millipore) in a transfer buffer containing 25 mM Tris, 193 mM glycine and 20%
methanol. The membrane was blocked for several hours with 10% non-fat dry milk in T-
TBS (10 mM Tris pH 7.5, 0.5 M NaCl, 0.05% Tween-20). After brief washes with T-
TBS twice, the membrane was incubated with primary antibody diluted in 1% non-fat dry
milk/T-TBS for two hours at room temperature followed by five washes in T-TBS for 15
min each. The membranes were incubated with secondary antibodies conjugated to
horseradish peroxidase for 30 min and washed in T-TBS extensively. Proteins were
visualized using Enhanced Chemiluminescence detection system (Amersham). To probe
a membrane sequentially with two different antibodies, the manufacturer’s instructions

were followed.

RESULTS

Saccharide Affinig Depletion versus Immunodepletion

We have optimized conditions for the splicing of the MINX pre-mRNA substrate
by NE derived from HeLa cells. In the experiments to be reported below, our typical
assay shows that approximately 20-3 0% of the input substrate is converted to product
(Fig. 1A, lane 1). As reported in our previous study (5), splicing activity was depleted
from NE when the latter was subjected to adsorption, under conditions of high ionic
strength (0.42 M NaCl), on Lac affinity beads. The unbound fraction of the Lac column
yielded less than 5% product formation (F ig. 1A, lane 3). Parallel control experiments, in
which NE was subjected to adsorption on beads containing cellobiose, which does not
bind to galectins, failed to yield the same result (5). Splicing activity in the extract
depleted on the Lac matrix was restored by the addition of recombinant galectin-3 (Fig.
1A, lane 4). In both the original NE, as well as in the reconstituted fraction, the splicing
activity was sensitive to inhibition by thiodigalactoside, a saccharide that binds to
galectin-3 with high affinity (Fig. 1A, lanes 2 and 5). These results formed the basis for

our suggestion that galectin-3 is a required factor in the splicing of pre-mRN A.

When we attempted to deplete the splicing activity using anti-M2, a monoclonal
antibody specific for galectin-3 (6, 7), there was some, but not complete, loss of splicing
activity in the galectin-3-depleted extract. An isotype-matched monoclonal, directed

against the transferrin receptor (TR), was used as a control for the anti-M2 antibody. The

44

45

Figure 1A: Comparison of the splicing activity of NE, NE depleted by Lac affinity
adsorption, and depleted NE reconstituted with recombinant galectin-3.
Lane 1: control splicing reaction of nondepleted NE in the presence of ATP.
Lane 2: components of lane 1 plus 50 mM thiodigalactoside (TDG).
Lane 3: the unbound (UB) fraction when NE is subjected to Lac affmity
adsorption.

Lane 4: components of lane 3 plus 12 pM recombinant galectin-3 (rG3).
Lane 5: components of lane 4 plus 150 mM TDG.

Products of the splicing reactions were analyzed by electrophoresis through a
13% polyacrylamide-urea gel and autoradiography. The positions of
migration of the pre-mRNA substrate, the splicing intermediates (exon 1 and
lariat-exon 2) and mature RNA product are indicated on the right.

46

M «9+
B
U

 

12345

47

Figure 13: Analysis of the polypeptide components of the bound fraction when NE is
subjected to affinity adsorption on Lac-agarose columns. The material bound
to the Lac-agarose beads was solubilized by SDS-PAGE sample buffer and
was electrophoresed on 12.5% polyacrylamide gels. The resolved
polypeptides were detected by silver staining or by immunoblotting with
specific antibodies.

Lane 1: silver stain. Lane 2: immunoblot with rat monoclonal antibody
against galectin-3 (aM2). Lane 3: immunblot with rabbit antibody against rat
galectin-l (aGl). The binding of the primary antibodies in lanes 2 and 3
were revealed with horseradish peroxidase-conjugated goat anti-rat and goat
anti-rabbit immunoglobulin, respectively, using the Enhanced

Chemiluminescence system. The positions of migration of human galectins-
1 and -3 are shown on the right.

48

Immunoblot
Silver

Stain (1M2 ocG1

<— Galectin-3

<— Galectin-1

 

49

unbound fraction of the anti-M2 column retained much of the splicing activity, compared
to the corresponding unbound fraction of the anti-TR column (Fig. 2A, lane 4 (15%
product) versus lane 1 (21% product». As a positive control for immunodepletion, the
snRNPs were removed from NE by an autoimmune serum reactive against the Sm
antigens on the core polypeptides B and D of the snRNPs (designated as anti-Sm) (11).
As expected, the Sm-depleted extract (unbound fraction of anti-Sm column) was
completely deficient in splicing activity (Fig. 2A, lane 7). Western blot analyses
indicated that anti-M2 removed >95% of galectin-3 (galectin-3 band in the bound fraction
and undetectable in the unbound fraction), while neither anti-TR nor anti-Sm removed the
galectin-3 polypeptide (Fig. ZB). These results indicate that removal of galectin-3 from

NE did not result in the simultaneous depletion of splicing activity.

Number and Identity of Galectins in NE

A hint at the resolution of this apparent dilemma was derived from the finding
that the splicing activity of the galectin-3-depleted extract, derived from the unbound
fraction of the anti-M2 adsorption, was still sensitive to TDG inhibition (Fig. 2A, lane 6).
Sucrose, a control disaccharide that does not bind to any of the galectins, failed to yield
the same effect (Fig. 2A, lane 5). These results suggest that other Lac-binding proteins,
besides galectin-3, may be present in the splicing extract and they are depleted, along

with galectin-3, by Lac adsorption but not by anti-M2.

This prompted us to analyze the protein components in the bound fraction of the

original Lac affinity adsorption. Using silver staining, we detected two bands whose

50

Figure 2A: Comparison of the splicing activity of NE after adsorption on protein G-
Sepharose beads conjugated with various antibodies. Lanes 1-3: the unbound
fraction of rat monoclonal anti-transferrin receptor (aTR) adsorption. Lanes
4-6:the unbound fraction of rat monoclonal anti-galectin-3 (aM2) adsorption.
Lanes 7-9: the unbound fraction after adsorption by human autoimmune
serum reactive against the Sm epitopes of the snRNP polypeptides (aSm).
Splicing reactions with the immunodepleted extracts were assayed in the
absence of saccharide (lanes 1, 4, 7), in the presence of 50 mM sucrose (lanes
2, 5, and 8),and in the presence of 50 mM thiodigalactoside (TDG) (lanes 3,
6, and 9). The positions of migration of the pre-mRNA substrate, splicing
intermediates, and mature RNA product are highlighted on the right.

51

U3 aTR UB (1M2 UB orSm

 

OPT
$983+
00.?
$225+

00.:
$0.53...

 

123456789

52

Figure 2B: Comparison of the levels of galectin-3 in NE versus the unbound (UB)
and B (bound) fiactions of the various immunodepletions. Samples were
subjected to SDS-PAGE on 12% polyacrylamide gels and immunoblotting
with rat monoclonal anti-galectin-B.

53

orTR orM2 orSm

 

UBBUBBUBBNE

 

54

mobilities corresponded to polypeptides with Mr ~30,000 and ~15,000 (Fig. 1B). The
identity of the ~30 kD polypeptide was ascertained to be galectin-3 by immunoblotting
with the anti-M2 antibody. On the basis of the apparent molecular weight of the second
silver-stained band, we suspected that it might correspond to galectin-1. This notion was
confirmed by immunoblotting with a rabbit antiserum raised against recombinant rat
galectin-1 (anti-GI) (Fig. 1B). Because this antibody does cross-react weakly with
galectin-2, we carried out direct comparisons between galectins-1 and -2 in terms of:
(a)the intensities of their immunoblotted bands when probed with the anti-G1 antibody;
and (b)the intensities of the silver-stained hands when known amounts of the respective
recombinant proteins were electrophoresed in SDS gels. The level of immunoreactivity
in the Western blot, relative to the intensity of the silver-stained gel (Fig. 1B), indicated
that the ~15 kD polypeptide was galectin-1. These results suggest that galectin-1 was the
other Lac-binding protein in NE and could be responsible for the residual saccharide-
inhibitable splicing activity after depletion of galectin-3. Indeed, we were able to detect,
by Western blotting analysis, galectin-1 in NE, as well as in the unbound fraction of the

anti-M2 column (Fig. 3A, lanes 1 and 2).

Double Antibody Depletion of Galectins-l and -3

The above results implicate that when NE is subjected to Lac affinity adsorption,
both galectins-l and -3 are bound on the Lac matrix, with concomitant depletion of
splicing activity, whereas an antibody directed against any single galectin would fail to
completely deplete the activity. Thus, a comparison was made of subjecting NE to

incubation with beads derivatized with: (a)anti-MZ; (b)anti-Gl; (e)anti-M2 plus anti-G1;

55

Figure 3A: Comparison of the levels of galectins-l and -3 in NE versus the unbound
(UB) and bound (B) fractions of immunodepletion using antibodies directed
against galectin-1 or galectin-3. Lane 1: NE. Lanes 2 and 3: rat monoclonal
antibody against galectin-3 (aM2). Lanes 4 and 5: rabbit antibodies against
rat galectin-1 (aGl). Lanes 6 and 7: orM2 and aGl used in combination.
Lanes 8 and 9: control antibodies, rat antibody against the transferrin receptor
(aTR) and rabbit preimmune serum (PI), used in combination. Samples were
subjected to SDS-PAGE on 12.5% polyacrylamide gels and immunoblotting
with orM2 or (1G1. Separate panels are shown for the a6] and (1M2
immunoblots, corresponding to the regions of migration for galectin-l and
galectin-3, respectively.

56

_n_ + mbo

$05 + N25

v05

N26

NEUBBUBBUBBUBB

2 3.4...5 "e rs 9T

 

1

57

Figure 3B: Comparison of the splicing activity of NE versus the unbound (UB)
fractions of immunodepletion using antibodies directed against galectin-l
or galectin-3. Lane 1: NE. Lane 2: unbound fraction of the aM2 depletion.
Lane 3: unbound fraction of the aGl depletion. Lane 4: unbound fraction
of immunodepletion using both antibodies, aM2 and 0:61. Lane 5:
unbound fraction of immunodepletion using controls antibodies, orTR
and PI. The positions of migration of the pre-mRNA substrate, splicing
intermediates, and mature RNA product are highlighted on the right.

58

‘—
(D
5
+
N
2
ZS
m
3

U3 0LTR+PI

 

59

and (d) anti-TR plus rabbit preimmune serum. Western blotting analyses yielded the
expected presence of galectins-1 and -3 in the various unbound and bound fractions (Fig.
3A). For example, galectin-3 was quantitatively adsorbed in the bound fractions of
matrices containing anti-M2. Similarly, no galectin-1 could be detected in the unbound
fractions of matrices containing anti-galectin-l. Finally, both galectins-l and -3 were
found exclusively in the unbound fractions of the beads containing the control antibodies,

anti-TR plus preimmune rabbit serum.

The unbound fraction of the double antibody column, anti-M2 plus anti-GI,
showed complete loss of splicing activity (Fig. 3B, lane 4). This should be compared to
the level of product formation in the unbound fraction of the control adsorption (Fig. 3B,
lane 5), as well as that of the original NE (Fig. 3B, lane 1). Together, all of these results
suggest that both galectin-l and -3 contribute independently to the splicing activity of NE
and that their complete removal, either by Lac affinity chromatography or by double

antibody adsorption, was necessary to deplete the splicing activity.

Reconstitution of Splicing in the Double Antibody-depleted Extracts

The unbound fraction of the anti-M2 plus anti-GI double antibody column,
devoid of splicing activity, could be reconstituted by the addition of either recombinant
galectin-1, or recombinant galectin-3 (Fig. 4). Recombinant galectins-1 and -3 were
derived from E. coli expression systems for the rat and mouse cDNAs, respectively (12,
13). Addition of either recombinant protein to the double antibody-depleted extract

restored the splicing activity (Fig. 4, lanes 3 and 5). Moreover, the splicing activity of the

60

Figure 4: Comparison of the splicing activity of NE, NE immunodepleted by antibodies
against galectins-l and -3, and depleted NE reconstituted with recombinant
galectin-1 or recombinant galectin-3.

Lane 1: control splicing reaction of nondepleted NE.

Lane 2: the unbound (UB) fraction when NE is subjected to immunodepletion
with rat monoclonal anti-galectin-3 (orM2) and rabbit antibodies against rat
galectin-1 ((rGl).

Lane 3: components of lane 2 plus 12 pM recombinant galectin-3 (rG3).

Lane 4: components of lane 3 plus 150 mM thiodigalactosdie (TDG).

Lane 5: components of lane 2 plus 26 pM recombinant galectin-1 (rGl).

Lane 6: components of lane 5 plus 150 mM TDG.

The positions of migration of the pre-mRN A substrate, the splicing inter-
mediates, and the mature RNA product are highlighted on the right.

61

U3 ocM2+0LG1

on? For
5....
we... no:
8...

m2

 

123456

62

double antibody-depleted extract, reconstituted with either galectin, was again sensitive to
TDG inhibiton (Fig. 4, lanes 4 and 6). The effects of galectin-1 or galectin-3 addition
were concentration dependent. Compared on an equal molar concentration basis,
recombinant galectin-3 was more potent than the corresponding galectin-1 in
reconstitution of splicing activity (see below). Finally, neither recombinant galectin-1
nor recombinant galectin-3 had an effect on the splicing assay when added to a
nondepleted NE. All of these results indicate that galectin-1 and galectin-3 may serve

redundant functions in the splicing activities of the NE.

Comparison of the Concentration Dependence of Galectin-1 and Galectin-3 in

Reconstituting Splicing Activig

In this series of experiments, we compared the concentrations of galectin-1 and

galectin-3 required to reconstitute a splicing deficient NE, depleted of the galectins by
Lac affinity adsorption. We also compared the intact polypeptide of galectin-3 (Mr
~30,000) with its COOH-terminal CRD (Mr ~15,000) (13) in terms of splicing
reconstitution. Using galectin-3 concentrations ranging from 2 pM to 18 pM, we found
that the minimal threshold concentration required for reconstituting splicing activity was
~3 pM. The optimal concentrations for reconstitution were 6 - 12 pM (Fig. 5A, lanes 3
and 4 and Fig. 5B). Product formation was inhibited at 18 pM, consistent with our

previous observations that high concentrations of galectin-3 inhibit the cell-free splicing

assay (5).

63

Figure 5A: Comparison of the splicing activity of NE, NE depleted by Lac affinity
adsorption, and depleted NE reconstituted with recombinant galectin-l, recom-
binant galectin-3, or the COOH-terminal domain of galectin-3.

Lane 1: control splicing reaction of nondepleted NE.

Lane 2: the unbound (UB) fraction when NE is subjected to Lac affmity
adsorption.

Lanes 3-5: components of lane 2 plus 6, 12, and 18 pM recombinant galectin-3
(rG3), respectively.

Lane 6-8: components of lane 2 plus 13, 26, and 52 pM COOH-terminal
domain derived fiom recombinant galectin-3 (C-domain), respectively.

Lane 9-11: components of lane 2 plus 13, 26, and 52 pM recombinant galectin-
] (rGl), respectively.

The positions of migration of pre-mRN A substrate, splicing intermediates, and
mature RNA product are highlighted on the right.

64

U8 LAC

+COOH-
"63 domain ”G1

 

 

% I 61218132652132652

 

’5‘- .Ireumrfl gm III

1234567891011

65

Figure 5B: Quantitation of the level of mature RNA product formation derived from
the data shown in panel A. The intensities of the bands on the gels were deter-
mined by direct B-particle counting using an AMBIS system. The percent
product formed was calculated by dividing the radioactivity present in the
RNA product by the total radioactivity present in the pre-mRNA substrate, the
splicing intermediates, and the mature RNA product. The numbers at the
bottom indicate the concentrations of rG3, C-domain, and rGl used to
reconstitute the UB fraction of the Lac affinity adsorption.

66

For.
an 8 2

59:06
-1000...

mm mm Mr

 

mg...
2. N_. o

 

...N

1v

 

r2.

iOnPO-ld‘yo

67

On the basis of our observations that a higher concentration of recombinant
galectin-1 than recombinant galectin-3 was required to reconstitute the splicing deficient
NE derived from double antibody depletion experiments (Fig. 4), the concentration range
tested for galectin-1 was 13 pM to 52 pM. In contrast to the results obtained with
galectin-3, 13 pM was insufficient to reconstitute splicing activity (Fig. 5A, lane 9 and
Fig. SB). Splicing activity was observed with 26 pM galectin-1 (Fig. 5A, lane 10 and
Fig. 5B), comparable or even better than that observed with the optimal concentration of
galectin-3. Finally, 52 pM showed inhibition of splicing, as was observed with galectin-3

at the high concentration end.

Although the COOH-terminal CRD of galectin-3 can reconstitute the splicing
activity of a deficient NE, it required a much higher concentration of protein than was
required of the intact galectin-3 polypeptide. Thus, we observed reconstitution of
splicing only at the highest concentration tested (52 pM) (Fig. 5A, lane 8 and Fig. 5B).
These results suggest that the galectin-3 polypeptide, with a Gly-, Pro-rich domain fused
onto the CRD (corresponding to galectin-1 or the COOH-terminal domain of galectin-3),

was most efficient in the reconstitution of splicing activity.

DISCUSSION

The key findings documented in the present study include: (a) NE capable of
carrying out pre-mRNA splicing, assayed in an in vitro system, contain both galectins-1
and -3; (b) depletion of both galectins from NE, either by Lac affinity adsorption or by
double antibody adsorption, resulted in the concomitant loss of splicing activity; (c)
depletion of either galectin-l or galectin-3, by specific antibody adsorption, failed to
remove all of the splicing activity and the residual activity was still sensitive to
saccharide-specific inhibition; (d) either galectin-1 or galectin-3 alone is sufficient to
reconstitute, at least partially, the splicing activity of NE depleted of both galectins; and
(e) although the CRD is sufficient to restore the splicing activity to galectin-depleted NB,
the concentration required for reconstitution was far greater than that of the full-length

galectin-3 polypeptide.

One important consideration in the interpretation of these results is that galectin-1
is indeed in the nucleus of a cell. Although it is generally accepted that galectin-1 is
predominantly an intracellular protein (1), observations and interpretations pertaining to
its nuclear localization are somewhat more difficult to establish (see discussion in chapter
111). There are studies that specifically show the localization of galectin-1 in the nuclei, as
well as the cytoplasm. There are other studies, however, at the immunofluorescence and
ultrastructural levels, which have led to explicit statements that antibodies against

galectin-1 failed to label the cell nucleus. To obviate these difficulties in interpretation,

68

69

we carried out a study on the intracellular localization of galectin-1 on the very same
HeLa cells from which the NE for splicing were derived. Using laser scanning confocal
microscopy, we have documented that galectin-1 is indeed in the nucleus of the HeLa
cells and that this observation of nuclear localization for galectin-1 is dependent on the
detergents used for the permeabilization procedures to allow accessibility of the probing

antibody reagent (Chapter H1).

The nuclear localization of galectin-1 and the results documented in the present
study, demonstrating that depletion of both galectins-1 and -3 are required to remove
splicing activity from NE and that either galectin-l or galectin-3 can restore the splicing
activity to a galectin-depleted extract, indicate the two galectins are redundant, at least in
terms of this function. The notion of functional redundancy between galectins-1 and -3 is
consistent with the results of experiments using transgenic mice in which a null mutation
in the gene encoding galectin-l has been introduced by homologous recombination in
embryonic stem cells (14). Homozygous animals carrying the mutant allele lack galectin-
1 but there was no apparent damage in terms of development. The mice were viable and
fertile. Thus, the function(s) assigned to galectin-1 seems to have been taken over by a

redundant relative, possibly galectin-3.

In this connection, it is important to consider the patterns of tissue specific
expression of the two galectins. This has been studied carefully during mouse
embryogenesis by Poirier and her colleagues (15-17). Both galectins-l and -3 are first

detected on day 4 of mouse development and their expression appear to be limited to the

7O

trophectoderrn of the hatched blastocyst. Thus, galectins-l and -3 overlap in terms of
their expression during early embryogenesis. Following gastrulation, however, their
patterns of expression no longer parallel each other. Galectin-1 is found in muscle cell
precursors of somites while galectin-3 is restricted to the notochord. During the later
parts of mouse embryogenesis, galectin-1 expression can be detected in many tissues of
the kidney, gut, lung, liver and muscle but not in the chondrocytes of cartilage. In
contrast, galectin-3 could be found in the cartilage of the vertebrae, with the hypertrophic
chondrocytes exhibiting higher levels of expression than differentiated chondrocytes.
Galectin-3 is also found in the suprabasal layer of the epidermis while no transcripts for
galectin-1 could be detected. Finally, while galectin-l is found in the motomeurons and
in the sensory neurons of the dorsal root ganglia, galectin-3 could not be observed in the

central nervous system.

On the basis of detailed comparisons of the carbohydrate-binding specificities and
certain “conserved” versus “variable” amino acid residues within the CRD, Ahmed and
Vasta (l 8) have proposed that the family of galectins actually exhibit two types of CRDs.
Galectin-l contains Type I CRD, with strict conservation of amino acids corresponding to
His-52, Asp-54, and Arg-73 and with strict requirements for equitorial -OH groups at C-3
or C-4 of the reducing end of a disaccharide (e.g. low affinity for galactoseBl,3-N-
acetylgalactosamine). On the other hand, galectin-3' contains Type II CRD, in which
residues corresponding to His-52, Asp-54, and Arg-73 can be substituted with a different
amino acid and with considerable affinity for galactoseB1,3-N-acetylgalactosamine. Our

present results suggest both Type I and Type II CRDs can function in the splicing assay.

71

This conclusion is based on the facts that: (a) both galectins-l and -3 must be
simultaneously depleted from NE in order to remove splicing activity; (b)cither lectin
alone can reconstitute the splicing activity of a galectin-depleted NE; and (c)the COOH-

terminal CRD of galectin-3 can also restore splicing activity of a galectin-depleted NE.

Although polypeptides corresponding to the CRD are sufficient to reconstitute
splicing activity, the minimum concentrations required are 4-8 times higher than that of
the intact galectin-3 polypeptide, which contains a Gly-, Pro-rich domain fused onto the
CRD. Several possibilities need to be considered to account for this apparent difference
in efficiency. First, it is known that endogenous cellular galectin-3 exists as two
isoelectric variants: a pl 8.7 species corresponding to the unmodified polypeptide and a pI
8.2 species representing a singly phosphorylated derivative (19). Chemical studies with
the canine homolog have identified a major site of phosphorylation to be Ser-6 (~90%)
and a minor site at Set-12 (~10%) (20). In the recombinant murine galectin-3 used in the
present studies, residue 12 is Ala so a reasonable possibility is that the polypeptide can
undergo phosphorylation at Ser-6 due to kinase(s) present in the galectin-depleted NE,

even during the steps of reconstitution, including dialysis and assay. A polypeptide
devoid of the Ser-6 phosphorylation site in NHz-terminal domain (galectin-l or the

COOH-terrninal CRD of galectin-3) would thus lack this phosphate group.

Although immublotting of subnuclear fractions and immunofluorescence of
permeabilized cells have suggested that galectin-3 is associated with RNP complexes (21,

22) and although there is a recent report that galectin-3 interacts directly with RNA and

72

single-stranded DNA (23), the specific spliceosomal component that interacts with
galectin-3 has not yet been identified. Seve and co-workers have shown that galectin-3 is
associated with a glucose-binding protein (24, 25). Athough this interaction can be
disrupted with Lac but not by glucose, it is thought the association is via protein-protein
interactions rather than protein-carbohydrate recognition. It should be noted, however,
that in our previous experiments, the control affinity matrix, composed of cellobiose
(glucoseBl,4glucose)-agarose, did not remove galectin-3 from NE and the unbound
fraction of the cellobiose-adsorbed NE was splicing competent (5). In any case, the role
of phosphorylation, in terms of enhancing galectin-3 ’s interaction with protein and/or
RNA components of spliceosome, is not known. Therefore, we can only speculate that if
phosphorylation indeed plays a role, then galectin-1 and COOH-terrninal CRD of

galectin-3 must overcome their lack of a phosphorylation site on the basis of mass action.

Second, similar arguments apply to the other structural feature of the Pro-, Gly-
rich NHz-terminal domain of galectin-3. The Pro-Gly-Ala-Tyr-Pro-Gly-Xxx-Xxx repeats
in this domain (1) may play a role in the interaction between galectin-3 and components

of the splicing machinery. Again, galectin-1 and the COOH-tenninal CRD of galectin-3,

both of which lack this structural feature, must compensate by simple mass action.

The third consideration is that, besides the interaction between the galectin and
other spliceosome components, these structural features (phosphorylation and/or Pro-,
Gly-rich repeats) may play a role in self-association of the galectin polypeptide. Self-

association of the rat and mouse homologs of galectin-3 has been inferred from the

73

concentration-dependent hemagglutination activity (26, 27) and positive cooperativity in
the binding of the lectins to multivalent glycoproteins (26, 28). Both the cooperative

binding experiments (26, 28) and cross-linking studies carried out with hamster galectin-3

(29) implicate the NHz-tenninal domain being responsible for the oligomerization of the

lectin. Thus, polypeptides lacking the structural features of the NHz-terminal domain

(galectin-1 or the COOH-terrninal CRD of galectin-3) would require other mechanisms,
including self-association via sequences contained in the CRD, to achieve the same

degree of multivalency.

Several studies have indicated that galectin-1 from various species formed
homodimers (3 0-33) and recent crystallographic structures of bovine galectin-1 are
consistent with this conclusion (34, 35). In a detailed analysis, Cho and Cummings (36)
showed that both recombinant galectin-l produced in an E. coli expression system, as
well as galectin-1 endogenous to the cytosol of Chinese hamster ovary cells, existed in a
reversible and active monomer-dimer equilibrium. The equilibrium dissociation constant
and equilibration time were estimated to be about 7 pM and 10 hours, respectively.
Although the concentration dependence was studied in less detail, it has also been shown
that the COOH-terminal CRD of galectin-3 can undergo monomer-dimer association
(3 7). Inasmuch as our present data showed that 26 pM recombinant rat galectin-l was
required to reconstitute the splicing activity of a galectin-depleted NE, it appears that
protein was fimctional in the dimeric state. We should hasten to note, however, that the

concentration of galectin-1 in NE is estimated to be in the nanomolar range and much of

74

the requirement for high concentration of the reconstituting protein may have to do with
the inefficient steps of dialysis and reassembly of the spliceosome, as discussed

previously (14).

(1)

(2)
(3)
(4)
(5)

(6)

(7)

(8)

(9)

REFERENCES

Barondes, S.H., Cooper, D.N.W., Gitt. M.A. and Leffler, H. (1994) J .Biol. Chem.
269, 20807-20810.

Kasai, K. and Hirabayashi, J. (1996) J. Biochem. 119, 1-8.

Moutsatsos, I.K., Davis, J.M., and Wang, J.L. (1986) J.Cell Biol. 102, 477-487.
Laing, J .G. and Wang, J .L. (1988) Biochemistry 27, 5329-5334.

Dagher, S.F., Wang, J.L. and Patterson, R.J. (1995) Proc. Natl. Acad. Sci. USA 92,
1213-1217.

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.
Dignam, J .D., Lebovitz, RM. and Roeder, R.G. (1983) Nucleic Acids Res. 11,
1475-1489.

Bradford, M.M. (1976) Anal. Biochem. 72, 248-254.

(10) Laemmli, UK. (1970) Nature. 227, 680-685.

(1 l) Lerner, MR. and Steitz, J .A. (1979) Proc. Natl. Acad. Sci. USA 76, 5495-5499.

(12) Cooper, D.N.W., Massa, SM. and Barondes, SH. (1991) J. Cell Biol. 115, 1437-

1448.

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

14939.

(14) Poirier, F. and Robertson, E]. (1993) Development 119, 1229-1239.

75

76

(15) Poirier, F., Timmons, P.M., Chen, C.T.J., Guenet, J .L., and Rigby, P.W.J. (1992)
Development 115, 143-155.

(16) Fowlis, D., Colnot, C., Ripoche, M.A., and Poirier, F. (1995) Dev. Dyn. 203, 241-
251.

(17) Colnot, C., Ripoche, M.A., Scaerou, F., Fowlis, D., and Poirier, F. (1996) Trans.
Biochem. Soc. 24, 141-146.

(18) Ahmed, H. and Vasta, GR. (1994) Glycobiology 4, 545-549.

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

(20) Huflejt, M.E., Turck, C.W., Lindstedt, R., Barondes, SH, and Leffler, H. (1993) J.
Biol. Chem. 268, 26712-26718.

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

(22) Wang, S.Y., Voss, P.G., Patterson, R.J., and Wang, J.L. (1995) Antibody,
Immunoconjugates, and Radiopharrnaceuticals 8, 311-324.

(23) Wang, L., Inohara, H., Pienta, K.J., and Raz, A. (1995) Biochem. Biophys. Res.
Commun. 217, 292-303.

(24) Seve, A.P., Felin, M., Doyenette-Moyne, M.A., Sahraoui, T., Aubery, M., and
Hubert, J. (1993) Glycobiology 3, 23-30.

(25) Seve, A.P., Hadj-Sahraoui, Y., Felin, M., Doyennette-Moyne, M.A., Aubery, M.,
and Hubert, J. (1994) Exp. Cell Res. 213, 191-197.

(26) Hsu, D.K., Zuberi, R.I., and Liu, RT. (1992) J. Biol. Chem. 267, 14167-14174.

(27) Ochieng, J., Platt, D., Tait, L., Hogan, V., Raz, T., Carmi, P., and Raz, A. (1993)

Biochemistry 32, 4455-4460.

77

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

(29) Mehul, B., Bawumia, S., Martin, SR, and Hughes, RC. (1994) J. Biol. Chem. 269,
18250-18258.

(30) Briles, E.B., Li, B, and Kornfeld, S. (1977) J. Biol. Chem. 252, 1107-1116.

(31) Beyer, E.C., Zweig, SE, and Barondes, SH. (1980) J. Biol. Chem. 255, 4236-4239.

(32) Teichberg, V1. and Levi, G. (1981) J. Biol. Chem. 256, 5735-5740.

(33) Roff, CF. and Wang, J.L. (1983) J. Biol. Chem. 258, 10657-10663.

(34) Liao, D.I, Kapadia, G., Ahmed, H., Vasta, GR, and Herzberg, O. (1994) Proc. Natl.
Acad. Sci. USA 91, 1428-1432.

(35) Bourrne, Y., Bolgiano, B., Liao, D.I., Strecker, G., Cantau, P., Herzberg, 0., Feizi,
T., and Cambillau, C. (1994) Struct. Biol. 1, 863-870.

(36) Cho, M. and Cummings, RD. (1995) J. Biol. Chem. 270, 5198-5206.

(37) Wang, J.L., Anandita (Vyakamam), Wang, S.Y., and Agrwal, N. ( 1993) in Lectins:
Biology, Biochemistry, Clinical Biochemistry, Vol. 9 (P. Chakrabarti, J. Basu, and

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

CHAPTER III

Galectin-1 in the Cell Nucleus. II. A Comparative Localization Study with Other
Splicing Components

Anandita Vyakarnam', Karen M. Lakkides', Ronald J. Pattersonz, and John L. Wang1

Departments of 1Biochemistry and 2Microbiology, Michigan State University
East Lansing, MI 48824

Acknowledgments: We thank Drs. Sam Barondes and Hakon Leffler for their generous
gifts of recombinant rat galectin-1 and rabbit antiserum directed against recombinant rat
galectin-1. This work was supported by grants MCB 91-22363 from the National Science
Foundation, GM-3 8740 from the National Institutes of Health, No. 99 from the Mizutani
Foundation for Glycoscience, and an Interdisciplinary Research Award from the Cancer
Center at Michigan State University.

78

ABSTRACT

Using both conventional and laser confocal fluorescence microscopy, the
intracellular distribution of galectin-1 in HeLa cells was analyzed and compared with the
localization of previously documented markers of the nucleus and cytoplasm. The Sm
epitopes of the small nuclear ribonucleoprotein complexes (snRNPs) and the non-snRNP
splicing factor SC35 yielded only nuclear staining. On the other hand, the enzyme lactate
dehydrogenase was cytoplasmic. In contrast to these patterns in which nuclear versus
cytoplasmic localizations appeared to be mutually exclusive, galectin-1, as well as
galectin-3, yielded both nuclear and cytoplasmic staining simultaneously. Whereas
galectin-3 exhibited prominent labeling of the nucleus and weak, diffuse staining of the
cytoplasm, the nuclear versus cytoplasmic distribution of galectin-1 was just the opposite,
with predominant staining in the cytosol. Confocal microscopy showed galectin-1
fluorescence throughout most of the sections from the top of the cell to the bottom.
Through the middle sections, as the plane of focus cuts through the nucleus, there was
definite fluorescence staining in the nuclear compartment. Double immunofluorescencc
analysis showed that, within the nucleoplasm, galectin-3 can be found coincident with the
speckled structures observed with SC35. Similar results were also obtained with
galectin-1, although in this case, there were areas of galectin-1 devoid of SC35 and vice

versa. These results establish the presence of galectin-1 in the nuclei of HeLa cells, a

79

80

conclusion consistent with the identification of the protein in nuclear extracts of the same

cells and with its documentation as a factor in pre-mRN A splicing.

INTRODUCTION

In the accompanying manuscript (1), we documented that: (a) nuclear extracts
(NE) derived from HeLa cells contain both galectins-1 and -3; (b) depletion of both
galectins from NE, either by lactose (Lac) affinity adsorption or by double antibody
adsorption, resulted in the concomitant loss of splicing activity; (c) depletion of either
galectin-1 or galectin-3, by specific antibody adsorption, failed to remove all of the
splicing activity and the residual activity was still saccharide inhibitable; and ((1) either
galectin-1 or galectin-3 alone is sufficient to reconstitute, at least partially, the splicing
activity of NE depleted of both galectins. All of the results suggest that the activities of

galectin-1 and galectin-3 in the nucleus may be redundant.

A number of lines of evidence have been accumulated to indicate that galectin-3 is
found in both the cytoplasm and nucleus of cells. Immunofluorescence staining, using
monoclonal, as well as polyclonal, antibodies specifically directed against galectin-3, was
carried out on formaldehyde fixed, Triton X-100 permeabilized 3T3 fibroblasts. There
was prominent labeling of the nucleus and variable staining of the cytoplasm (2, 3). This
dual localization of galectin-3, in the nucleus and cytoplasm, has been confirmed by
immunoelectron microscopy (4). Moreover, the nuclear staining of galectin-3 in
immunofluorescence and ultrastructural experiments was sensitive to treatment of the

permeabilized cells with ribonuclease A (RNase), but not to parallel treatment with

81

82

deoxyribonuclease I (DNase) (3, 4). Finally, the nuclear and cytoplasmic localization of
galectin-3 has been documented by quantitative immunoblotting analysis of subcellular

fractions, relative to enzyme markers characteristic of the various fractions (2).

On the other hand, however, the notion that galectin-l can be found in the nucleus
is controversial. For example, in the original publication reporting the purification and
characterization of the chicken homolog, there was an explicit statement that antibodies
to the chicken heart protein corresponding to galectin-1 failed to label the cell nucleus of
chicken embryo fibroblasts under immunofluorescence (5). More recently,
immunofluorescence analysis of hamster galectin-1 showed that it was exclusively
cytoplasmic (6). Indeed, the main difficulty in the analysis of galectin-1 is that, unlike
galectin-3, the predominant portion of the former protein is found in the cytoplasm. As a
result, a strong fluorescent antibody probe staining the cytoplasm 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 confocal microscope to analyze the intracellular distribution of galectin-1. In
the course of these investigations, we also found that the type of detergent used to
permeabilize the cell had a profound effect on the observation of galectin-1 in the
cytoplasm versus the nucleus. The results of these studies are documented in the present

communication.

MATERIALS AND METHODS

Immunofluorescence Microscopy

HeLa cells were grown as monolayers in Minimum Essential Media containing
10% defined/supplemented bovine calf serum (HyClone), 100 U/ml penicillin and 100
U/ml streptomycin at 37°C in humidified atmosphere of 5% C02. For
immunofluorescence microscopy, cells were seeded onto glass coverslips which were
placed in 6-well (8 cmz/well) cluster dishes. Cells grown to 50-70% confluency were
washed twice with phosphate buffered saline (PBS, 140 mM NaCl, 2.68 mM KCl, 10
mM Na2HP04, 1.47 mM KH2P04, pH 7.4). They were fixed for 20 min at room
temperature with 4% paraformaldehyde in PBS. Cells were washed twice in PBS and
permeabilized with 0.2% Triton X-100 in PBS for 5 min. After permeabilization, cells
were again washed twice with PBS and incubated with 0.2% gelatin in PBS for at least 1
hour at room temperature. After washing in T-TBS (10 mM Tris pH 7.5, 500 mM NaCl,
0.05% Tween-20), cells were incubated with primary antibody at an appropriate dilution
in 0.2% gelatin/T-TBS for 1 hour. Cells were washed three times (15 min each) with T-
TBS and incubated for 1 hour with the secondary antibody conjugated with fluorescein
isothiocyanate (FITC) at appropriate dilution in 0.2% gelatin/T-TBS. Cells were washed
three times in T-TBS for 15 min each and mounted in Perma F luor (Immunon) on glass
slides. In some experiments, cells were permeabilized with 0.05% saponin/PBS for 1

hour at 37°C (6) or 20 pg/ml (0.002%) digitonin/PBS for 10 min at room temperature (7).

83

84

In experiments requiring sequential permeabilization with two reagents, fixed
cells were first permeabilized with 0.05% saponin for 1 hour at 37°C. Cells were washed
three times with PBS and then incubated with 0.2% Triton X-lOO/PBS for 5 min at room

temperature.

For double immunofluorescence analysis, cells were first permeabilized with
0.05% Triton X-100 in permeabilization buffer (300 mM sucrose, 100mM NaCl, 3 mM
MgC12 and 10 mM PIPES pH 7.2) for 5 min at 4°C (8). Cells were rinsed twice with cold
permeabilization buffer and then twice with cold PBS. Cells were fixed with 4%
paraformaldehyde for 20 min at room temperature, washed extensively with T-TBS and
then treated for immunofluorescence staining as described above. Samples were analyzed

using a Meridian Instruments (Okemos, MI) Insight confocal laser scanning microscope.

Antibodies

Anti-galectin-l is a polyclonal rabbit antiserum raised against recombinant rat
galectin-1 (anti-GI) (9) and was a gift from Dr. Hakon Lefiler (UCSF, CA). Human
autoimmune serum reactive with the Sm antigens of snRNPs (ENA anti-Sm) was
purchased from The Binding Site (San Diego). Anti-SC35 (Sigma) is a mouse
monoclonal antibody against the non-snRNP splicing factor SC35 (10). Anti-Mac-Z
(anti-M2) is a rat monoclonal antibody directed against galectin-3 (11, 12). The antibody
was purified from cell culture supernatant derived from hybridoma line M3/38.1.2.8.HL.2
obtained from ATCC (TIB 166). Anti-lactate dehydrogenase (anti-LDH) is a polyclonal

rabbit antiserum against pig muscle LDH and was a gift from Dr. John Wilson (Michigan

85

State University). Secondary antibodies conjugated with fluorescein isothiocyanate
(FITC) were used at following dilutions: FITC-goat anti-human immunoglobulin (E.Y.
Labs) at 1:500, FITC-goat anti-rabbit immunoglobulin preadsorbed with human serum
proteins (Sigma) at 1:500, FITC-goat anti-mouse immunoglobulin at 1:500 and FITC-
goat anti-rat immunoglobulin (Sigma) at 1:250. Cy3 conjugated sheep-anti-mouse

immunoglobulin (Sigma) was used at 1:250 dilution.

Preparation of HeLa Cell Extracts and Immunoblot Analysis

For preparation of cell extract, HeLa cells were pelleted at 5000g, resuspended in
RSB (RSB: 10 mM Tris, pH 8.0, 3 mM Mng ,10 mM NaCl) and incubated on ice for
10 min. Cells were again pelleted at 2000 rpm for 5 min and resuspended in lysis buffer
(25 mM Tris pH 7.4, 50 mM NaCl, 1% NP-40, 0.2% SDS and 200 U/ml aprotinin).
Protein concentration of extracts was determined using DC Protein Assay Reagent (Bio-

Rad) using BSA as the protein standard (13).

For immunoblot analysis, 50 pg total protein was loaded in each lane. Protein
samples were resolved on 12.5% SDS-PAGE as described by Laemmli (14) and
electrophoretically transferred to Irnmobilon-P membrane (Millipore) in a transfer buffer
containing 25 mM Tris, 193 mM glycine and 20% methanol. The membrane was blocked
for several hours with 10% non-fat dry milk in T-TBS ( 10 mM Tris pH 7.5, 0.5 M NaCl,
0.05% Tween-20). After brief washes with T-TBS twice, the membrane was incubated
with primary antibody diluted in 1% non-fat dry milk/T-TBS for two hours at room

temperature followed by five washes in T-TBS for 15 min each. The membrane was then

86

incubated with secondary antibodies conjugated to horseradish peroxidase for 30 min and
washed in T-TBS extensively. Proteins were visualized using Enhanced

Chemiluminescence detection system (Amersham).

RESULTS

Nuclear and Cfloplasmic Localization of Galectin-1: Evidence from Laser Scanning

Confocal Microscopy

Because the interpretation of the immunofluorscence experiments to be
documented below depends on the specificity of the antibody reagents used, the number
of components in HeLa cells that react with each antibody reagent was determined by
immunoblotting. Anti-G1, a polyclonal rabbit antiserum raised against recombinant rat
galectin-1, immunoblotted one predominant polypeptide (Mr ~14,000) whose mobility
corresponded to that of authentic galectin-1 (Fig. 1A, lane 1). Much weaker reactivity was
observed with a few polypeptides at ~50 kD, which were also observed with rabbit
preimmune serum (Fig. 1A, lane 2). Anti-M2, a rat monoclonal antibody directed against
galectin-3, yielded a single band corresponding to the human homolog in HeLa cells (Fig.
1A, lane 3). Antibody reagents directed against components that serve as markers for
the cytoplasmic and nuclear compartments were also tested. Rabbit anti-pig muscle LDH
immunoblotted a single polypeptide (Mr ~37,000) (Fig. 1A, lane 4), consistent with the
molecular weight of the enzyme (15). Human autoimmune serum anti-Sm
immunoblotted the Sm-B doublets (Mr ~29,000) (Fig. 1A, lane 5), known polypeptide
components of the snRNP complexes (16). Finally, mouse monoclonal anti-human SC35
yielded a single band (Mr ~3 5,000) (Fig. 1A, lane 6), corresponding to this member of the

SR family of splicing factors (10).

87

88

Figure 1A: Western blot analysis of the the specificity of the antibody reagents.
Extracts of HeLa cells (50 pg protein) were subjected to SDS-PAGE and
immunoblotting with: (l)rabbit antiserum against recombinant rat galectin-1
(orGl) at 1:1000 dilution; (2)rabbit preimmune serum (PI) at 1:1000 dilution;
(3)rat monoclonal antibody against galectin-3 (aM2) at 12.5 pg/ml; (4)rabbit
antiserum against pig muscle LDH (aLDH) at 1:1000 dilution; (5)human auto-
immune serum reactive against the Sm epitopes of the core polypeptides of
snRNPs (aSm) at 1:10,000 dilution; and (6)mouse monoclonal antibody
against human SC35 (aSC35) at 1:500 dilution. The binding of the primary
antibody was revealed with horseradish peroxidase-conjugated goat anti-rabbit
immunoglobulin, goat anti-rat immunoglobulin, goat anti-human immuno-
globulin, or goat anti-mouse immunoglobulin, using the Enhanced
Chemiluminescence system. The positions of migration of molecular weight

markers are indicated on the right.

89

—83

.—51

—35
., _29

 

_21

90

Figure 1B: Immunofluorescence staining of galectins-l and -3 and other nuclear and
cytoplasmic markers in human HeLa cells. Cells were fixed with paraform-
aldehyde (4%) and permeabilized with Triton X-100 (0.2%) and then stained
with antibodies: orGl at 1:50 dilution; Plat 1:50 dilution; orM2 at 12.5 pg/ml;
aLDH at 1:50 dilution; aSm at 1:750 dilution; and aSC35 at 1:250 dilution.
The binding of the primary antibody was detected by the appropriate
fluorescein-conjugated secondary antibody.

 

 

orG1 orM2
Pl aLDH
aSm aSC35

 

92

HeLa cells cultured on coverslips were fixed with paraformaldehyde,
permeabilized with Triton X-100 and stained with various antibodies. Both antibodies
directed against bona fide splicing components, anti-Sm and anti-SC35, yielded nuclear
staining. The anti-Sm staining exhibited diffuse fluorescence throughout the nucleus,
with the exception of ~5 circles devoid of fluorescence (Fig. 1B). These “black holes”
correspond to nucleoli (17, 18). Anti-SC35 yielded the characteristic punctate or
speckled pattern reported previously (10, 17, 18). Anti-LDH, on the other hand, yielded
diffuse distribution of fluorescence in the cytoplasm. The staining obtained with anti-Sm
(or anti-SC35) and anti-LDH provided reference patterns expected for a nuclear and
cytoplasmic protein, respectively. Finally, anti-M2 yielded labeling of the nucleus, as
well as staining of the cytoplasm, in accord with our previous reports (2, 3). As with anti-
Sm, the staining of galectin-3 in the nucleus was diffuse throughout the nucleoplasm,

with the exception of ~5 “black holes.”

The staining patterns obtained with anti-G1 resembled neither those obtained with
anti-Sm or anti-M2 nor those seen with anti-LDH. There was extensive staining
throughout the cytoplasm, with slightly weaker staining of the nucleus (Fig. 18). In
contrast, the rabbit preimmune serum yielded little or no staining. The galectin-l
staining data appear to be similar to those presented in the original localization of the
chicken homolog to galectin-1 in chicken embryo fibroblasts, although the authors
explicitly stated that there was no labeling of the cell nucleus (5). Because of the

complications arising from cytoplasm overlying and underlying the nucleus, the

93

fluorescent cells were visualized by laser confocal microscopy, collecting images through

consecutive focal planes.

Anti-Sm was used as a positive control for nuclear staining. 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 (Fig. 2). As was observed
under the conventional fluorescence microscope (Fig. 1B), the Sm antigens showed
diffuse distribution throughout the nucleus, except for the ‘black holes.” Below the
nucleus, the various planes of focus no longer showed fluorescence staining. Similar
results were also obtained with anti-SC35 (Fig. 2). Anti-M2 also yielded, through the
middle sections, bright nuclear staining with the exception of the “black holes” (Fig. 2).
Thus, the nuclear distribution of galectin-3 and the Sm epitopes of snRNPs are very
similar, a conclusion documented previously (3). Staining for LDH, serving as a negative
control, showed that the enzyme was cytoplasmic. Through the middle sections, the anti-
LDH yielded a distinct pattern, with inside of the nucleus devoid of fluorescence (Fig. 2).
Finally, the staining for galectin-1 showed fluorescence throughout most of the sections
(Fig. 2). The middle sections showed fluorescence in both the nucleus and the
cytoplasm. All of these results suggest that, although the predominant portion of
galectin-1 is found in the cytoplasm, there was definitely nuclear fluorescence due to the

same protein.

94

Figure 2: Subcellular localization of galectin-1 as revealed by laser scanning confocal
fluorescence microscopy. Human HeLa cells were fixed with paraform-
aldehyde (4%) and permeabilized with Triton X-100 (0.2%) and then
stained with rabbit anti- galectin-1 (0:61). The staining of galectin-3 (aM2),
snRNP Sm polypeptides (aSm), non-snRNP splicing factor SC35 (aSC35),
and LDH (orLDH) provided reference patterns for nuclear and cytoplasmic
localizations. The antibodies were used at the same dilution as in Figure 18.
For each staining, images were collected fiom 9 consecutive focal planes,
with the increment of 0.5 pm for each step in the z-direction. Three
sections representing the bottom, middle, and top are displayed

95

och

222.2

EON—“om

9690

Ian—6

 

N25

 

v05

96

Detection of Nuclear Galectin-1 Dependent on Permeabilization Detergent

The conventional immunofluorescence results we obtained on galectin-1 are, in
fact, similar if not identical to the very early study on the immunofluorescence
localization of the homolog in chicken embryo fibroblasts (5). The authors of that study
apparently discounted the nuclear localization in their statement that “Fluorescence was
unifome distributed throughout the cytoplasm...(but not in the nucleus)” The apparent
discrepancy between our results and the most recent localization of galectin-1 in Chinese
hamster ovary cells (6) was, however, harder to dismiss. A hint to the resolution of this
dilemma was derived from the recent development of assays for studying the import of
proteins bearing nuclear localization signals. These assays used digitonin to permeabilize
the plasma membrane so that the fluorescently labeled import substrates can gain access
to the nulcear translocation machinery at the nuclear pore complex (7, 19). However, the
nuclear membrane was not permeabilized so that the import substrates can be transported
only in a signal dependent fashion. 0n the basis of the fact that saponin, the detergent
used by Cho and Cummings (6) to permeabilize cells, belonged to the same family of
cholesterol-like drugs as digitonin, the possibility was raised that antibodies against
nuclear components can gain access in Triton X-100 permeabilized cells of this study
while the same reagents were excluded from the nuclear antigens in digitonin/saponin-

permeabilized cells.

97

Thus, parallel cultures of HeLa cells were fixed with paraformaldehyde and then
permeabilized with: (a) Triton X-100 (0.2%); (b) digitonin (0.002%; 20 pg/ml); and (c)
saponin (0.05%). When subjected to staining with anti-G1, there was a clear difference in
the localization observed. Triton-permeabilized cells showed nuclear and cytoplasmic
staining (Fig. 3), as was observed in Figures 18 and 2. 0n the other hand, both digitonin-

and saponin-permeabilized cells showed exclusively cytoplasmic staining (Fig. 3).

To confirm the notion that, in saponin-permeabilized cells, galectin-1 was indeed
in the nucleus and simply not accessible to the antibody staining reagents, HeLa cells
were fixed with paraformaldehyde and then permeabilized with saponin. These saponin-
perrneabilized cells were then incubated in the presence or absence of Triton X-100. In
those cells that were permeabilized only with saponin (no Triton X-100), anti-GI showed
exclusively cytoplasmic staining (Fig. 4). In contrast, saponin permeabilization followed
by Triton X-100 yielded both nuclear and cytoplasmic staining. More striking results
were obtained with the Sm polypeptides, our reference for nuclear antigens. In the
absence of Triton X-100 treatment, the saponin-permeabilized cells showed no anti-Sm
staining; these same cells revealed the Sm antigens upon Triton X-100 addition (Fig. 4).
Finally, Triton X-100 had no effect on the staining for a cytosolic marker enzyme, anti-
LDH. All of these results strongly implicate that galectin-1 is indeed in the nucleus but
that it could be detected by immunofluorescence only if the permeabilization procedure

allows access of the antibody staining reagents.

Figure 3:

98

Comparison of the immunofluorescence staining pattern of galectin-1 in
HeLa cells permeabilized with Triton X-100, digitonin, or saponin. The cells
were fixed with paraformaldehyde (4%) and the permeabilized with Triton
X-100 (0.2%), digitonin (0.002%; 20 pg/ml), or saponin (0.05%) and then
stained with rabbit antiserum against rat galectin-l (1:20 dilution). The
binding of the primary antibody was revealed with fluorescein-conjugated
goat anti-rabbit immunoglobulin.

99

Triton X-100 Digitonin Saponin

 

Figure 4:

100

The effect of Triton X-100 on the immunofluorescence staining pattern of
galectin-1 in saponin-permeabilized HeLa cells. The cells were first fixed
with paraformaldehyde (4%) and then permeabilized with saponin (0.05%).
Half of the samples were then treated with Triton X-100 (0.2%). All samples
were then subjected to staining with rabbit anti-galectin-l (aGl), rabbit anti-
LDH (aLDH), or human autoimmune serum reactive with Sm (aSm), using

the same conditions as that in Figure 13

101

Saponin Saponin + Tx-100

    

aGl

 

aLDH

aSm

 

102

Double Immunofluorescence Analyses: Galectins versus Splicing Factor SC35

Although the immunofluorescence patterns obtained with fixed and permeabilized
cells suggest that the nuclear galectins were diffusely distributed within the nucleoplasm
(Fig. 1B and Fig. 2), previous studies have shown that conditions can be found such that
galectin-3 and the Sm antigens of snRNPs (3), as well as non-snRNP splicing factors
such as SC35 (17, 18), exhibited “speckled” staining patterns. The differences between
the diffuse staining pattern and the speckled pattern 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), the staining intensity could be so
strong that it covers the entire nucleus, obscuring distinct subnuclear localization. With
loss of antigen fi'om the nuclei during permeabilization and extraction procedures, the
staining of discrete structures/regions becomes more distinct, giving rise to a speckled

pattern.

HeLa cells were permeabilized with Triton X-100 (without fixation), then fixed
and stained simultaneously with anti-M2 and anti-SC35. The binding of anti-M2 was
revealed with F I'TC-conjugated goat anti-rat immunoglobulin, yielding the “green”
fluorescence pattern in Figure 5. The binding of anti-SC35 was revealed with cy3-
conjugated sheep anti-mouse immunoglobulin, yielding the “red” fluorescence pattern in
the same figure. In the composite of the double immunofluorescence patterns, the

“yellow” represents regions of coincidence of the “green” and “red” stains. Careful

I ‘ $0.3.tlr

Figure 5:

103

Double immunofluorescence analysis of galectin-3 and SC35. HeLa cells
were permeabilized with 0.05% Triton X-100 (without fixation) and then
fixed with paraformaldehyde (4%) and stained simultaneously with: aM2, rat
monoclonal antibody against galectin-3 at 12.5 pg/ml; and aSC35, mouse
monoclonal antibody against SC35 at 1:250 dilution. The binding of (1M2
was revealed by FITC-conjugated goat anti—rat immunoglobulin, yielding

the panel with the “green” fluorescence; the binding of aSC35 was revealed
by cy3-conjugated sheep anti-mouse immunoglobulin, yielding the panel
with the “red” fluorescence. A composite of the the two panels is also shown.

104

mmUm5

mmUw5\N25

 

N25

105

analysis of such doubly stained photomicrographs indicates: (a)the “red” anti-SC35
staining appears speckled; (b)there is coincidence between the “red” anti-SC35 spots and
the “green" anti-M2 spots, leading to a speckled “yellowish” composite; and (c)the anti-
M2 staining for galectin-3 appears to be more diffuse, giving rise to “green” areas in the
composite containing no “red” staining. The latter point is important, for it suggests that
the sheep anti-mouse immunoglobulin reagent does not cross-react with the rat
monoclonal anti-M2 antibody. This conclusion is supported by control experiments,
checking for cross-reactivity between the sheep anti-mouse and goat anti-rat with the rat

and mouse immunoglobulins, respectively.

A similar double immunofluorescence analysis was canied out with anti-G1 and
anti-SC35. The binding of anti-G1 was revealed with FI'TC-conjugated goat anti-rabbit
immunoglobulin (“green” fluorescence); the binding of anti-SC35 was revealed with cy3-
conjugated sheep anti-mouse immunoglobulin (“red” fluorescence) (Fig. 6). The
speckled SC35 pattern contains regions of coincidence with the anti-G1 staining, giving
rise to “yellowish” spots on the composite. However, there are also spots where SC35
staining is not matched by the presence of galectin-1. These appear as “reddish orange”
spots on the composite. Finally, there are “green” areas that appear to be devoid of any

“red” SC35 staining.

Figure 6:

106

Double immunofluorescence analysis of galectin-1 and SC35. The experi-
mental conditions of Figure 5 were used. orGl, rabbit antiserum directed
against recombinant rat galectin-1, was used at 1:5 dilution and was
revealed by FIT C-conjugated goat anti-rabbit immunoglobulin (“green”
panel). aSC35, mouse monoclonal antibody against SC35 was used at
1:250 dilution and was revealed by cy3-conjugated sheep anti-mouse

immunoglobulin (“red” panel). A composite of the two panels is also
shown.

107

mm0m5

 

mmUm52U5

 

~95

DISCUSSION

Comparisons of the immunofluorescence staining of live cells versus cells fixed
with formaldehyde, followed by permeabilization, indicate that the prototype subgroup of
galectins (galectins-1 and -2) are found predominantly in the intracellular compartment
(see, for example, reference (5)). 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, there are studies that specifically show the localization of the prototype
galectins in the nuclei as well as the cytoplasm. Cryostat sections of tissues subjected to
immunofluorescence showed labeling of both nuclei and cytoplasm in the following
cases: anti-CLL-I staining of adult chicken kidney (20) and anti-bovine heart galectin-l
staining of calf pancreas (21). In a series of studies on chick embryonic skin, Akimoto et
al. (22, 23) reported the chick homolog of galectin-1 in the cytosol and nuclei of cells in
the intermediate layer of the epidermis and in some dermal fibroblasts. This nuclear
localization, observed at the light microscope level, was confirmed at the ultrastructural

level. Similar results were also reported for normal human skin (24).

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

108

109

galectin-1. Dorsal root ganglion neurons were subjected to immunofluorescence with
anti-rat lung galectin-l and with a monoclonal antibody directed against a lactoseries
glycoconjugate (25). 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 (26). Similarly, reassessment of previously
published immunocytochemical studies in non-neuronal cells has suggested the presence

of galectin-1 in nuclei as well as the cytoplasm (27).

Finally, immunofluorescence and ultrastructural studies have also led to explicit
statements that anti-galectin-l antibodies failed to label the cell nucleus. Two separate
studies on mouse C2 myoblasts showed that galectin-l was found in a diffuse distribution
in the cytoplasm but no nuclear localization could be detected (28, 29). Moreover, upon
fusion to form myotubes, the intracellular cytoplasmic staining decreases with the
appearance of the lectin at the cell surface. In chicken embryo fibroblasts, anti-chicken
heart galectin-1 staining was observed only in cells fixed and permeabilized, but not in
unfixed cells, suggesting that most of the lectin was intracellular (5). The intracellular
staining was ascribed only to the cytoplasm, although the immunofluorescence
micrographs are very similar, if not identical, to the present documentation that galectin-1
can be found in nucleus and cytoplasm of HeLa cells. Most recently,
immunofluorescence analysis of Chinese hamster ovary cells showed that galectin-1 was

exclusively cytoplasmic (6).

 

 

110

On the basis of our present data, it seems likely that at least one parameter that
might account for differences in observing nuclear plus cytoplasmic localization of
galectin-1, versus exclusive cytoplasmic localization, is the detergent used to
permeabilize the cells prior to antibody staining. The use of Triton X-100 permeabilizes
both the plasma membrane as well as the nuclear compartment, allowing antibodies
reactive against nuclear components to gain access to the antigenic epitopes. On the other
hand, saponin (and digitonin) fails to permeabilize the nuclear membrane to allow
antibody accessibility. This notion is consistent with the body of literature on the
development of a permeabilized cell assay for studying the properties and requirements of
protein import through the nuclear pore complex (7, 19). This assay uses digitonin to
permeabilize the plasma membrane so that the fluorescently labeled import substrates can
bind to the nuclear translocation machinery. At the same time, however, the nuclear
membrane is not permeabilized such that the import substrates can be transported only if

they carry a nuclear localization signal.

The present microscopic localization of galectin-1 in the HeLa cell nucleus is
consistent with the biochemical data indicating that NE contain galectin-1, as well as
galectin-3 (1). Both lines of evidence provide the basis for the , observations that
depletion of both galectins was necessary to achieve a concomitant loss of splicing
activity and that either galectin-1 or galectin-3 alone is sufficient to reconstitute the
splicing activity of NE depleted of both galectins. The identification of galectins-1 and -3
as splicing factors is also supported by the colocalization of each galectin with speckled

patterns observed in immunofluorescence analyses of snRNP antigens and non-snRNP

lll

splicing factors such as SC35 (17, 18). These speckled structures have been found to
correspond, at the ultrastructural level, to interchromatin granule clusters and
perichromatin fibrils (30). Perichromatin fibrils are readily labeled with short pulses of
[3H]uridine (31) and anti-RNA polymerase 1] antibodies (32), suggesting that they
represent nascent transcripts at the sites of mRNA synthesis and early events of pre-
mRNA processing. Indeed, immunogold labeling at the ultrastructural level has found
galectin-3 in perichromatin fibrils (33). It would, therefore, be of obvious interest to

determine if a similar localization can be achieved for galectin-l.

(1)
(2)
(3)

(4)

(5)

(6)

(7)

(8)
(9)

REFERENCES

Chapter 2

Moutsatsos, I. K., Davis, J.M., and Wang, J .L. (1986) J. Cell Biol. 102, 477-483.
Wang, S.Y., Voss, P.G., Patterson, R.J., and Wang, J .L. (1995) Antibody, Immuno-
conjugates, and Radiopharmaceuticals 8, 311-324.

Hubert, M., Wang, S.Y., Wang, J.L. Seve, AR, and Hubert, J. (1995) Exp. Cell
Res. 220, 397-406.

Barak-Briles, E., Gregory, W., Fletcher, P., and Komfeld, S. (1979) J. Cell Biol.
81, 528-537.

Cho, M. and Cummings, RD. (1995) J. Biol. Chem. 270, 5207-5212.

Adam, S.A., Marr, RS, and Gerace, L. ( 1990) J. Cell Biol. 111, 807-816.
Xing,Y. and Lawrence, JR (1991) J. Cell Biol. 112, 1055-1063.

Cooper, D.N.W., Massa, S.M., and Barondes, SH. (1991) J. Cell Biol. 115, 1437-

1448.

(10) Fu, X.D. and Maniatis,T. (1990) Nature 343, 437441.

(11) Ho, MK. and Springer, T.A. (1982) J. Immunol. 128, 1221-1228. w

(12) Cherayil, B.J., Weiner, S.J., and Pillai, S. (1989) J. Exp. Med. 170, 1959-1972.

(13) Lowry, 0.H., Rosenbough, N.J., Farr, A.L., and Randall, R.J. (1951) J. Biol. Chem.

193, 265-275.

(14) Laemmli, UK. (1970) Nature 277, 680-685.

112.

113

(15) Weber, K. and Osborn, M. (1969) J. Biol. Chem. 244, 4406-4412.

(16) Lerner, MR. and Steitz, J .A. (1979) Proc. Natl. Acad. Sci. USA 76, 5495-5499.

(17) Spector, D.L., Fu, X.D., and Maniatis, T. (1991) EMBO J. 10, 3467-3481.

(18) Spector, D.L. (1990) Proc. Natl. Acad. Sci. USA 87, 147-151.

(19) Moore, MS. and Blobel, G. (1992) Cell 69, 939-950.

(20) Beyer, EC. and Barondes, SH. (1980) J. Supramol. Struct. 13, 219-227.

(21) Childs, RA. and Feizi, T. (1980) Cell Biol. Int. Rep. 4, 775.

(22) Akimoto, Y., Kawakami, H., Oda, Y., Obinata, A., Endo, H., Kasai, K. and Hirano,
H. (1992) Exp. Cell Res. 199, 297-304.

(23) Akimoto, Y., Obinata, A., Hirabayashi, J., Sakakura, Y., Endo, H., Kasai, K., and
Hirano, H. (1995) Cell & Tissue Res. 279, 3-12.

(24) Akimoto, Y., Hirabayashi, J ., Kasai, K., and Hirano, H. (1995) Cell & Tissue Res.
280, 1-10.

(25) Regan, L. J ., Dodd, J ., Barondes, SH, and Jessell, TM. (1986) Proc. Natl. Acad.
Sci. USA 83, 2248-2252.

(26) Hynes, M.A., Gitt, M.A., Barondes, S.H., Jessell, T.M., and Buck, LB. (1990)
J. Neurosci. 10, 1004-1013.

(27) Barondes, SH. (1988) Trends Biochem. Sci. 13, 480-482.

(28) Cooper, D.N.W. and Barondes, SH. (1990) J. Cell Biol. 110, 1681-1691.

(29) Harrison, FL. and Wilson, T.J.G. (1992) J. Cell Sci. 101, 635-646.

(30) Fakan, S., Leser, G., and Martin, TE. (1984) J. Cell Biol. 98, 358-363.

(31) Fakan, s. and Puvion, E. (1980) Int. Rev. Cytol. 65, 255-299.

114

(32) Spector, D.L., O’Keefe, R.T., and Jimenez-Garcia, LE (1993) Cold Spring Harbor
Symp. Quant. Biol. 58, 799-805.

(33) Hubert, M., Wang, S.Y., Wang, J.L., Seve, AP, and Hubert, J. (1995) Exp. Cell

Res. 220, 397-406.

CHAPTER IV

Concluding Statement

115

CONCLUDING STATEMENT

The identification of nuclear galectin-1 and galectin-3 as novel splicing factors
raises many exciting and provocative questions. For example, what is the identity of
potential nuclear ligand/partner(s) for galectins-l and -3? Three classes of molecules
could interact with nuclear galectins:

i) RNA (either the pre-mRNA or the snRNA of the snRNPs). Several proteins
involved in splicing bind directly to the pre-mRNA or snRNAs required for splicing.
RNA-binding motifs have been identified in many of these proteins. Galectins lack
identifiable RN A-binding motifs. Moreover, direct interaction of galectins with RNA
species has not been demonstrated.

ii) A nuclear glycoconjugate. The obvious candidate ligand for nuclear galectins
is a glycoconjugate containing galactose. To date, no such nuclear glycoconjugate has
been identified. Recently, glycoproteins with 0-linked N-acetylglucosamine (GlcNAc)
have been shown to localize to nuclei. Although galectins have no affinity for GlcNAc, it
has been shown that GlcNAc residues of nucleoporins are acceptors for galactose
addition by galactosyltransferases in a cell-free system. If similar galactose transfer can
occur in vivo, the resulting N-acetyllactosamine would exhibit high affinity for galectins.

iii) The nuclear partner for the galectins may be a protein that interacts with the

CRD or some other region of the protein. Precedence for the former has been described

116

117

for the plant lectin concanavalin A. Hexapeptides from a phage display library have been
identified that have an affinity for concanavalin A similar to its normal saccharide ligand.
Alternatively, another nuclear splicing component may bind to the galectins through
protein-protein interactions. In this case, lactose inhibition of splicing would be an
indirect effect, inducing a conformational change resulting in the dissociation of the
galectin from its partner. Identification of the nuclear splicing ligand for galectin-l and

-3 is crucial to understand the role of these galectins in splicing.

Another question that remains to be answered: Is the carbohydrate-binding
activity of the nuclear galectins required for splicing activity? The ability to bind [3-
galactoside containing glycoconjugates is the distinguishing feature of galectins.
However, since no such conjugates have been identified in mammalian nuclei, it is
premature to invoke a specific role for carbohydrate recognition in pre-mRN A
processing. Tools are now available to directly test this question. Site-directed
mutagenesis of the CRD for galectin-1 has been used to produce mutant forms unable to
bind lactose. It needs to be determined whether these mutant galectins restore splicing
activity to a galectin-depleted nuclear extract. If carbohydrate binding is not required for
splicing function, one possible explanation for inhibition of splicing by galectin-specific
saccharides is that binding of lactose to the CRD induces a conformational change in the
galectin. This may causes a dissociation of the galectin from its splicing partner, thus
inhibiting splicing. Alternatively, specific saccharides could displace a nuclear molecule

that may bind to the CRD via molecular mimicry.

118

Our studies show that one member of the Proto Type and the single member of
Chimera Type galectins are involved in pre-mRNA splicing. It would be interesting to
determine if other members of the galectin family particularly, Tandem Repeat type

galectins, also exhibit splicing activity.

This study demonstrates nuclear localization of galectin-1. Localization of
galectin-3 in the nucleus is well documented in literature. However, further studies are
required to determine the mechanism of nuclear localization of galectins-1 and -3. Three
general modes of nuclear protein import have been described: 1) presence of a nuclear
localization sequences (NLS) on the polypeptide, ii) diffusion and iii) piggy-back entry.
Both galectins-1 and-3 lack NLS. The polypeptide of galectin-1 can associate to form 30
kDa dimers. Galectin-3 remains as monomeric species. At these sizes, both galectins
could enter the nucleus by diffusion. Once within the nuclear interior, binding to splicing
factors or spliceosomes could result in nuclear retention. Alternatively, these galectins
could associate in the cytoplasm with a nuclear-destined protein containing an NLS and
enter as a complex. The challenge now is to distinguish between these possibilities,
delineate their relative contribution to the nuclear import process and to analyze their

regulation.

It is clear that these and many other aspects of galectin function and localization

will become apparent as the details of the role of galectins in splicing become unraveled.