Structural and functional analysis of the endometrial gene encoding insulin-like growth factor-binding protein-2


Material Information

Structural and functional analysis of the endometrial gene encoding insulin-like growth factor-binding protein-2
Physical Description:
xiii, 177 leaves : ill. ; 29 cm.
Song, Sihong, 1959-
Publication Date:


Subjects / Keywords:
Animal Science thesis, Ph. D   ( lcsh )
Dissertations, Academic -- Animal Science -- UF   ( lcsh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1996.
Includes bibliographical references (leaves 154-176).
Statement of Responsibility:
by Sihong Song.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 026491815
oclc - 36741705
System ID:

This item is only available as the following downloads:

Full Text






To my parents with deepest appreciation.


First of all, I would like to thank Dr. Frank A. Simmen for his support, advice,

encouragement and patience. As an outstanding advisor he led me into molecular and cell

biology step by step. It was wonderful to enjoy the pursuit of science under his guidance. I

also would like to especially thank Dr. Rosalia C.M. Simmen for her continuous support and

many helpful suggestions throughout my Ph.D studies. I would like to thank Drs. William

W. Thatcher, Thomas P. Yang and Michael J. Fields for all their helpful discussions and

suggestions. In addition, I want to thank Dr. William C. Buhi for sharing tissues and Dr.

David R. Clemmons for collaborative work on IGFBP-2 cDNA.

I also thank all my colleagues in Drs. Frank A. Simmen and Rosalia C.M. Simmen's

groups, especially Drs. Chul-Young Lee for his help in initiating the research project; Drs.

Lokenga Badinga, Michael L.Green, and Omaththage P. Perera for helpful discussions;

Wemer Collante, Frank J. Michel, and Tricia E. Chung for their technical assistance; Karen

L. Reed for transferring the primary cell culture technique; and Drs. Inho Choi and Wenjun

Liu, Inseok Kwak and Yang Wang for sharing experiences, ideas and friendship throughout

my graduate studies. My special thanks also go to Mary Ellen Hissem, Joyce Hayen and all

other members of the department for the warm and friendly environment they try to create.

I would like to express my appreciation to Dr. Changzheng Wang and all the friends

I made in America, who helped me to stand up in this country, which is also important for

scientific accomplishments. I also want to express my appreciation to all the teachers who

contributed to my early education.

I would like to give my deepest appreciation to my mother and father for their love

and all that they do for me, especially in my early age, when they took my and my brother's

educations as their first concern. My special appreciation also goes to my brother for his love,

help and encouragement. All my accomplishments represent the contributions from my

lovely, warm family. I also want to express my appreciation to my parents-in-laws for their

love and encouragement.

My lovely daughter, Annie Song, and son Alexander Song have given me wonderful

times which helped release me from tiredness, regain energy and hope; therefore they should

be acknowledged.

Lastly, I would like give very special thanks and deepest appreciation to my wife,

Yufei Tang, for her love, encouragement, support, self-sacrifice and patience, especially for

those years, months, days and nights when she was taking care of the family and waiting for

me. This work could not have been done without her support in so many aspects.


ACKNOWLEDGMENTS ................. ............. ........... .... iii

LIST OF TABLES .......................... ............ ............ vii

LIST OF FIGURES ................................... .............. viii

ABSTRACT .................... ............................ xi


1. INTRODUCTION ................. ........................... 1

2 LITERATURE REVIEW ...........................................4

Insulin-like-Growth Factors and Their Binding Protein Family ....... 4
IGFBP-2 Gene and Physiology ............................... 14
Uterine IGF System ................... ...... ............. 25


Introduction ................. ........................... 31
Materials and Methods ...................................... 33
Results .............. .................... .............37
Discussion .............................. ............... 50
Summary .............................................. 56

ACTIVATING EGION OF THE IGFBP-2 GENE ...................... 57

Introduction ............................. ...... ......... 57
Materials and Methods ................ ................... 59
Results .............. ................................. 66
Discussion .............. .................... .......... 84
Summary .................................. .............. 89

IGFBP-2 GENE TRANSCRIPTION ................................. 91

Introduction ............................................. 91
Materials and Methods ...................................... 93
Results ............ ................................. 98
Discussion ............................ ............... 111
Summary .............................................. 120


Introduction ........................................... 122
M materials and M ethods ...................................... 123
Results .............. .................... ............ 124
Discussion .............................................131
Summary ..............................................133

GROWTH FACTOR-BINDING PROTEIN-2 (rpIGFBP-2) .............. 135

Introduction ...........................................135
Materials and Methods ................ ................... 137
Results ................................... ............139
Discussion ............................. ...... ......... 147
Summary ................................... ..........149

8. SUMMARY AND CONCLUSIONS ................................ 150

REFERENCES ............... ............................ 154

BIOGRAPHICAL SKETCH ............... .......................... 177


Table page

3-1. The sequences ofexon-intron boundaries and the 3' end of the IGFBP-2 gene. 45

3-2. Identity of porcine IGFBP-2 amino acid sequence with IGFBP-2 proteins of
other species ................................................. 47


Figure page

3-1. Northern analysis of IGFBP-2 mRNA in pregnant pig uterus ............... 38

3-2. Structural organization of the porcine IGFBP-2 chromosomal locus ......... 40

3-3. DNA sequence and deduced amino acid sequence of porcine IGFBP-2
chromosomal gene and protein ...................................... 41

3-4. Deduced amino acid sequences of pig (P), human (H), rat (R) and mouse (M)
IGFBP-2 proteins ..............................................46

3-5. Identification of 5'-transcriptional start sites ............................ 49

3-6. Summary of the results of the localization of the 5'-ends of IGFBP-2
mRNAs ............. ...................................... 51

3-7. Conserved sequence motifs in the region upstream of the translational
initiation codon of the IGFBP-2 gene ................................. 52

4-1. DNA constructs containing the upstream region of the IGFBP-2 gene fused
to the luciferase reporter gene ..................................... 67

4-2. Reporter gene activities of the Hind-LUCe, Sac-LUCe, Bgl-LUCe and
Sma-LUCe plasmids transfected in different cell lines and uterine cell types ... 68

4-3. Transcriptional activation by the 110 bp upstream region .................. 70

4-4. Map of the GRA probes derived from the 110 bp region .................. 72

4-5. Uterine endometrial nuclear proteins bind the 110 bp fragment, and the
individual A, B and C subregion probes ............................... 73

4-6. Localization of nuclear protein-binding sites within A, B and C subregions
of the 110bp fragment ................ .......................... 75

4-7. Gel retardation assay involving nine overlapping probes .................. 76

4-8. Localization of specific and non-specific protein-binding elements
using GRA ................ ................ ................. 77

4-9. Proteins of the same molecular weight bind a common sequence within
the A2 and B3 oligonucleotides ..................................... 78

4-10. Characterization of the A2 element DNA binding protein ................. 81

4-11. Size determination of the A2 DNA binding protein ...................... 83

5-1. Effects of estrogen and progesterone on IGFBP-2 steady-state mRNA
abundance in vitro and in vivo ..................................... 100

5-2. Effects of estrogen and progesterone on the promoter activities of Sac-LUCe
(-874/+73) and Bgl-LUCe (-764/+73) DNA constructs in primary cultures of
transfected endometrial glandular epithelial cells from Day 18 pregnant pigs 101

5-3. Effects of estrogen and progesterone on reporter gene activity of the
Sac-LUCe (-874/+73) DNA construct in a time-course experiment ......... 102

5-4. Effects of estrogen and progesterone on promoter activity of SaB(ll10)-
pLUC (-874/+764) DNA construct in transfected Day 18 GE cells ......... 103

5-5. Peptide hormonal effects on IGFBP-2 steady-state mRNA abundance in
endometrial glandular epithelial cells from Day 18 pregnant pigs (D8GE) ... 105

5-6. Dose-dependent effects of IGF-II, cAMP and PMA on IGFBP-2
mRNA abundance ........................... ................ 106

5-7. Effects of IGF-II on promoter activities of Hind-LUCe (-1397/+73),
Sac-LUCe (-874/+73), Bgl-LUCe (-764/+73) and Sma-LUCe
(-305/+73) DNA constructs transfected into D18GE cells ................ 108

5-8. Effects of IGF-II on reporter gene activities of SaB(1 10)-pLUC and
SmB(459)-pLUC DNA constructs transfected into D18GE cells ........... 109

5-9. IGF-II receptor density on endometrial cell membranes .................. 110

5-10. Differential expression of IGFBP-2 gene and protein in endometrial cells .... 112

5-11. Speculative model for trophic regulation of the IGFBP-2 gene ............. 119

6-1. All exons of the IGFBP-2 chromosomal gene are G/C rich and exon 1 and
immediate 5' flanking region constitutes a CpG island ................... 125

6-2. Genomic Southern Blot Assay of methylated IGFBP-2 DNA .............. 127

6-3. Schematic summary of the results of Southen blot analysis ............... 129

6-4. In vitro activity of a methylated promoter of the IGFBP-2 gene ............ 130

7-1. Schematic diagram of IGFBP-2 cDNA and genomic DNA ............... 141

7-2. PCR amplification of the cDNA fragment encoding mature IGFBP-2 ....... 142

7-3. Scheme used for the generation of an expression construct encoding for
recombinant porcine IGFBP-2 ............... .................. 143

7-4. Enrichment of recombinant IGFBP-2 ................................ 145

7-5. Recombinant IGFBP-2 binds IGF-I .................................. 146

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



Sihong Song

December 1996

Chairperson: Frank A. Simmen
Major Department: Animal Science

The Insulin-like Growth Factors (IGFs) play important growth regulatory roles in the

uterus. The functions of the IGFs are modulated by their receptors (type-I and type-II) and

at least six distinct IGF binding proteins (IGFBPs), which are expressed in a tissue- and

developmental-specific manner. In porcine uterus of pregnancy, insulin-like growth factor-

binding protein-2 (IGFBP-2) is expressed in endometrial cells and accumulates at the feto-

maternal interface. However, the functions) of this protein and the molecular mechanisms

of IGFBP-2 gene regulation in the uterus remain unknown. To address these questions, the

chromosomal organization of the porcine IGFBP-2 gene was first characterized. This gene

spans -29 kb and is comprised of four exons and three introns. Exons 1, 2, 3 and 4 are ~527,

227, 141 and 528 bp, respectively. Introns 1, 2 and 3 are -24 kb, 746 bp and 2.6 kb,

respectively. The mRNA transcript is -1.6 kb in size, and encodes a 316 amino acid

precursor. The TATA-less GC rich promoter was characterized and shown to have two

clusters of transcriptional start sites (-109 and -78 relative to the transcriptional initiation

codon). Subsequently, 1.4 kb of 5' flanking region was analyzed by deletion mutagenesis and

transient transfection in ECC-I and JEG-3 cell lines, and epithelial and stromal cells isolated

from pregnant pig endometrium. Results identified a 110 bp region from nucleotides-874 to

-765 that has transcriptional stimulatory activity, whereas regions from nucleotides -1397 to

-875 and -764 to -306 appeared to have cell-type dependent inhibitory activities. Within the

110 bp region, two consensus sequences TCAGGG and CCCTGA were identified by gel

retardation assay to bind the same nuclear protein designated A2 DNA binding protein with

molecular weight of 33 kDa as estimated by Southwestern blot analysis. Estrogen stimulated

endometrial IGFBP-2 gene expression in vivo in the ovariectomized pig model and in vitro

in primary cultures of endometrial cells. Ligand blot analysis showed that the binding

capacity of IGF-II receptors on uterine cell membranes was correlated with IGFBP-2 gene

expression in these same cells. IGF-II, Phorbol 12-Myristate 13-Acetate (PMA) and cyclic

AMP increased in IGFBP-2 mRNA abundance in endometrial cells. In contrast, TGF-P 1 and

IGF-I did not alter IGFBP-2 mRNA levels. In vitro DNA methylation at the promoter region

did not alter the rate of transcriptional initiation of this gene in a HeLa cell-free transcription

system. Moreover, no differential DNA methylation between a high IGFBP-2 expressing

tissue (endometrium) and low IGFBP-2 expressing tissues (myometrium and placenta) was

found for several CpG doublets within the 5' end of the gene. Lastly, recombinant porcine

IGFBP-2 protein (rpIGFBP-2) was expressed and purified from E.coli cells. This rpIGFBP-2

exhibited biological activity as monitored by IGF-binding. In conclusion, distal promoter

region of the porcine IGFBP-2 gene is important for this gene regulation. IGF-II may be

responsible for pregnancy stage-specific IGFBP-2 gene expression, whereas estrogen may

be responsible for estrous cycle-related IGFBP-2 gene expression. Availability of functional

rpIGFBP-2 may provide a means to elucidate the endometrial functions of IGFBP-2.


The Insulin-like Growth Factor (IGF) system probably is one of the more complicated

functional control systems required for orchestration of a number of biological processes,

particularly with respect to growth and development. Insulin-like Growth Factor-I (IGF-I)

and Insulin-like Growth Factor-l (IGF-II), as major participants of this system, are

mutifunctional polypeptides with structural similarity to proinsulin. Although some of the

functions of IGF remain unclear, studies have demonstrated that IGFs can affect cell

proliferation, differentiation, cell cycle progression, function and death. These two peptides

are synthesized in a variety of fetal and adult mammalian tissues, and function not only in

an endocrine but also autocrine and paracrine fashion. These actions of IGF are mediated

through two different cell membrane receptors with different affinities for ligand. IGF-I

receptor (also called the type 1 receptor) is a heterotetramer with structural and functional

similarity to the insulin receptor, while IGF-II receptor (also called the type 2 receptor) has

a single transmembrane domain, and is identical to the cation-independent mannose 6-

phosphate receptor. The most complicated aspect of the IGF system is the involvement of

the IGF binding proteins (IGFBPs), a family of related proteins with at least six members

(IGFBPs-1 to -6). In the extracellular fluid, IGFs usually are bound to one or more of these

binding proteins with high affinity. This binding modifies the action of IGFs in several


different ways. IGFBPs can act as carriers in the circulation and thereby prolong IGF

biological half-life. Because they bind to IGFs with higher affinity than do IGF receptors,

IGFBPs can localize IGFs to target cells. Interestingly, IGFBPs appear to have IGF-

independent functions as well. The exact biological roles of each IGFBPs are still not

determined. Functions of the IGFBPs appear to be altered by post-transcriptional

modifications, such as phosphorylation, glycosylation and proteolysis. Therefore, there is no

single model to describe all the features of this system. In different tissues and physiological

conditions, the IGF system can exhibit different functions by virtue of the unique

combination of participating proteins.

The uterus plays an important role in supporting embryo growth and development.

The uterine endometrium is an epithelial-mesenchymal tissue lining the lumen, which

undergoes morphological and biochemical changes during the estrous cycle and throughout

pregnancy. In pregnant animals, the uterine endometrium, in response to signals from the

concepts (embryo and its associated membranes), provides a suitable environment for

implantation and subsequent development. This feto-materal interaction involves both

endocrine controls, such as utero-ovarian interactions, and local autocrine and paracrine

controls involving proteins, such as growth factors and cytokines. The IGF system is

involved in controlling uterine functions. For example, during periimplantation, the porcine

endometrium synthesizes peak amounts ofIGF-I. This IGF-I stimulates embryonic aromatase

gene expression, which in turn enhances estrogen production from the embryo. Estrogen,

as a signal for maternal recognition of pregnancy, acts on the endometrium to block the

PGF,, luteolytic signal production or secretion. Therefore, the corpus luteum is maintained


as is production of progesterone, which is required for pregnancy. IGFs, IGF receptors and

IGFBPs are differentially expressed in uterine and fetal tissues, and undoubtedly have

important roles in uterine cell proliferation and differentiation.

IGFBP-2 exhibits a particularly interesting expression pattern in porcine uterus. This

gene is expressed in the endometrium, but not in the myometrium and placenta of pregnant

pigs. In endometrium, this gene is expressed at low levels during early pregnancy and is

highly expressed during mid- and late-pregnancy. This uterine-synthesized IGFBP

accumulates on the surface of endometrial epithelial cells. These observations raise two

important questions: 1). What is the molecular mechanisms) for this tissue- and

developmental stage-specific pattern of gene expression? 2). What is the function of this

protein at the feto-materal interface? The objectives of this study were to gain a better

understanding of the molecular basis of uterine IGFBP-2 gene expression, and to provide

information concerning the functional aspect of endometrial IGFBP-2. The specific

hypotheses examined were as follows: 1). The distal regulatory region of the endometrial

IGFBP-2 gene is important for its transcriptional expression. 2). Use of this gene as a model

may lead to identification of novel transcription factors) and cis-element(s) underlying

endometrial-associated gene expression.


Insulin-like-Growth Factors and Their Binding Protein Family

Insulin-like Growth Factors

Background. The Insulin-like Growth Factors (IGFs) were discovered on the basis

of three separate biological activities. Sulfation factor activity (SFA) was defined by Salmon

and Daughaday (1957), who observed that the growth-promoting action of Growth Hormone

was mediated by a substance in serum that stimulated the uptake of sulfate by costal cartilage

in explant cultures. The term nonsuppressible insulin-like activity (NSILA) was coined by

Froesch et al. (1966) based on the observation that the insulin-like action of serum on muscle

and adipose tissue could not be abolished by inclusion of anti-insulin antibody. BRL-3A rat

hepatoma cell conditioned medium had mitogenic activity, and this was termed

multiplication-stimulating activity (MSA) by Dulak and Temin (1973). In 1972, the term

somatomedin was used to denote uncharacterized factors that mediated the action of Growth

Hormone in the stimulation of somatic growth and that displayed insulin-like activity

(Daughaday et al., 1972).

In 1978, Rinderknecht and Humbel purified two distinct somatomedin peptides from

human serum. Base on the sequence and structure homologies with human proinsulin, IGF-I

and IGF-II were used to designate these peptides. In 1987, the term IGF-I was officially


recommended to refer to the somatomedins A and C, which were found to be identical to

IGF-I (Klapper et al., 1983; Enberg et al., 1984) and IGF-II was recommended to replace the

designation MSA (Daughaday et al. 1987).

Peptides. IGF-I and IGF-II are single-chain polypeptides with three intrachain

disulfide bonds consisting of 70 (7.6 kDa) and 67 (7.5 kDa) amino acids, respectively

(Rinderknecht and Humbel, 1978a, b).These two peptides can be subdivided into B, C, A and

D domains beginning from the N-terminus. The structure of the A and B domains are

homologous to the A and B chains of insulin. Similar to proinsulin, the C domain separates

the A and B domains. Unlike for proinsulin, however, a D domain is found at the C-terminus

of IGFs.

Gene structures. The genes encoding IGF-I and IGF-II are single-copy (Tricoli et al.,

1984; Brissenden et al., 1984). IGF-I genes have been characterized in human (Rotwein et

al., 1986), rat (Shimatsu and Rotwein, 1987), Sheep (Dickson et al., 1991), and chicken

(Kajimoto and Rotwein, 1991). The IGF-I gene consists of six exons (reviewed by Sara and

Hall, 1990; LeRoith and Roberts, 1993). Exons 1 and 2 encode alternative 5'-untranslated

regions and ATG translation initiation codons for the precursor proteins. Exons 3 and 4

encode the B, C, A, and D domains of the peptide. Exons 5 and 6 encode the extended E

peptide, which is post-translationally processed, and contain 3-untranslated sequences. The

chicken IGF-I gene does not have an exon 2 encoding a alternative 5'-untranslated region nor

an exon 5 encoding the E peptide. Multiple transcription start sites in exons 1 and 2,

alternative splicing in exons 1 and 5, and multiple polyadenylation sites in exon 6 produce


a collection ofIGF-I mRNAs containing divergent 5'- and 3'-untranslated regions as well as

the region encoding the E peptide.

The structure of the IGF-II gene has been elucidated for the human (Bell et al.,

1985; de Pagter-Holthuizen et al., 1988), rat (Frunzio et al., 1986; Chiariotti et al., 1988) and

mouse (Rotwein and Hall., 1990). Similar to the IGF-I gene, the structure of the IGF-II gene

consisting of 9 exons is quite complicated (reviewed by LeRoith and Roberts, 1993). Exons

1 through 6 encode divergent 5'-untranslated regions. There are 4 different promoters (P1-4)

in front of exons 1, 4, 5, and 6, respectively. Exons 7 and 8 encode the IGF-II peptide

domains B, C, A, and D. Exon 9 encodes part of the E peptide and also contains the 3'-

untranslated region. The combination of differential promoter usage, alternative splicing and

multiple polyadenylation sites results in 6 different IGF-II mRNA species with the sizes of

6.0,5.3, 5.0, 4.8,2.2 or 1.8 kb. All of these transcripts, except for the exon 9-derived 1.8 kb-

species, encode the same IGF-II precursor. Promoter 1 is active in hepatic cells after birth.

Promoters 2, 3 and 4 are predominantly utilized in embryonic cells (Sara and Hall, 1990).

Interestingly, the IGF-II gene is closely linked to the insulin gene, which is located only 1.4

kb from exon 1 of the IGF-II gene (Bell et al., 1985).

Actions of the IGFs. The biological functions of IGF-I and -II have been reviewed

intensively (Sara and Hall, 1990; Lowe 1991; Giudice, 1992; Cohick and Clemmons, 1993;

and Jones and Clemmons, 1995). The in vitro biological actions of IGF-I exhibited the

following aspects: 1. Effects on cell cycle progression. IGFs stimulate DNA synthesis and

cell replication causing cells to traverse successive phases of the cell cycle, in particular from

Go to G,. 2. Effects on cell proliferation. A wide variety of cell types demonstrate a


mitogenic response to exogenous IGF-I. 3. Inhibition of cell death. 4. Stimulation of cell

differentiation and function. For example, IGF-I and -II stimulate hormone synthesis and

secretion by ovarian granulosa and theca cells (Giudice, 1992).

The in vivo studies of the effects of IGF administration have clearly demonstrated

the insulin-like action of IGFs (Tomas et al., 1992, reviewed by Jones and Clemmons,

1995). Mice with null mutations of the IGF-I and IGF-II genes had birth weights that were

60% of normal mice while relative body proportions were normal (DeChiara et al., 1990; Liu

et al., 1993a). Mice with a null mutation for both IGF-I and IGF-II had 30% of normal body

weight, and all such mice died within minutes after birth due to respiratory failure (Baker et

al., 1993). Overexpression of human IGF-I in transgenic mice increased body weight by 30%

over control mice (Mathews et al., 1988). Overall, in transgenic mice models, IGFs exhibited

prenatal growth-promoting activities.

Insulin-Like Growth Factor Receptors

There are two known receptors that specifically interact with the IGFs. The IGF-I

receptor (also called the type 1 IGF receptor) shares a high degree of sequence similarity with

the insulin receptor. The IGF-II receptor (also called the type 2 IGF receptor) is identical to

the cation-independent mannose 6-phosphate receptor.

IGF-I receptor. Intensive studies of the IGF-I receptor were reviewed recently by

LeRoith et al. (1995). The IGF-I receptor contains two a-subunits and two P-subunits that

form a heterotetrameric structure similar to that of the insulin receptor (Massague and Czech,

1982). The entire a subunit protrudes extracellularly and contains a cysteine-rich domain,

which is required for ligand binding. The p subunit contains the transmembrane spanning

domain and a cytoplasmic region with the highly conserved tyrosine kinase catalytic domain

as well as several tyrosine residues which can be autophosphorylated. The a and P subunits

are linked by a single disulfide bond to form an ap-half-receptor. Subsequently, two of these

ap-half-receptors become linked by a disulfide bond between the a subunits to form a a232

heterotetrameric structure. The a and P subunits are generated from a single precursor

encoded by a chromosomal gene with 21 exons (Abbott et al., 1992). The peptide cleavage

site for generation of the subunits is the basic tetrapeptide Arg-Lys-Arg-Arg, which is

encoded by exon 11.

Because of the similarity between IGF-I and insulin receptors, the ap-half-receptor

of these receptors can form hybrid insulin/IGF-I receptors. This hybrid receptor has been

isolated from solubilized placental membranes (Soos et al., 1989; 1993) and cultured cells

(Moxham et al., 1989). The affinities of receptor hybrids were lower for insulin than for IGF-

I (Frattali and Pessin, 1993; Soos et al., 1993).

Signal transduction from the IGF-I receptor appears to be similar to that from the

insulin receptor. IGF-I binding to the a subunits induces IGF-I receptor autophosphorylation

of three tyrosine residues in the kinase domain of the P-subunit, and this results in activation

of the intrinsic tyrosine kinase activity of the IGF-I receptor (Gronborg et al., 1993; Kato et

al., 1993; 1994). The activated IGF-I receptor is able to phosphorylate other tyrosine-

containing substrates. Interestingly, IGF-I receptor shares a common motif "PL-X4-

NPXYXSXSD" (so-called insulin/IL-4R motif) with insulin and interleukin-4(IL-4) receptors

within the intracellular domain (Wang et al., 1993a, b; White and Kahn, 1994; Keegan et

al., 1994). After ligand binding, this motif becomes (auto-)phosphorylated and consequently


binds IRS-1 and other proteins to transduce the signal to the cell interior. IRS-1, the major

substrate of the activated insulin receptor, is a predominant substrate of the IGF-I receptor

as well (LeRoith et al., 1988; Myers et al., 1993a, b). When the gene encoding IRS-I was

inactivated, the IRS(-/-) mice were both insulin and IGF-I resistant, as manifested by marked

intrauterine growth retardation (-50% of normal body weight) and a blunted response to the

hypoglycemic action of injected IGF-I and insulin (Kahn et al., 1995). The residual biological

effects in those mice were associated with the appearance of an alternative substrate of the

receptor, IRS-2 which also appears to be 4PS, a substrate of the IL-4 receptor. IRS-1, which

has multiple tyrosines in YMXM motifs, is considered to be a "docking" protein, that can

bind SH2 domain-containing proteins. IGF-I receptor can potentially influence multiple

intracellular signaling pathways through phosphorylation of IRS-1 (Sun et al., 1993), and its

subsequent binding to proteins such as PI-3 kinase, MAP kinase, Grb-2, Nck (also an adaptor

protein) and Syp (a phosphotyrosine phosphatase). These substrates and multiple signaling

pathways, which also involve many other components, may be important for the diversity of

biological functions described for IGF ligands and receptors in different tissues. In different

cell types and different developmental stages, the presence of these substrates and other

components for each pathway may differ, thereby resulting in different cell responses


IGF-II receptor. Un like the IGF-I and insulin receptors, the IGF-II/Man-6-P receptor

is a monomeric glycoprotein. Human IGF-II receptor has a predicted molecular mass of 270

kDa including a long extracellular domain comprised of 15 Cys-rich repeats (~150 amino

acids for each repeat), a small transmembrane region (23 amino acids) and a short


intracellular domain (164 amino acids) (Morgan et al., 1987). This receptor has distinct

binding sites for IGF-II and Man-6-P or Man-6-P-containing glycoproteins, and can bind

IGF-II and Man-6-P simultaneously (Braulke et al., 1988). Cell surface IGF-II receptors have

a high affinity for IGF-II, a low affinity for IGF-I and no affinity for insulin (Nissley et al.,

1991). There is a 240 kDa truncated form of IGF-II receptor extracellularr domain) that is

found in the blood stream (MacDonald et al., 1989; Bobek et al., 1991). The function of this

truncated soluble receptor form is, however, unknown. The gene encoding IGF-II receptor

(Igf2r) is subject to maternal imprinting, whereas the gene encoding IGF-II is paternally

imprinted (Barlow et al., 1991; DeChiara et al., 1991; Filson et al., 1993). When the IGF-II

receptor gene was inactivated, the mice (Igf2r-/-) developed edema, and 100% of the fetuses

died in utero at day 15 of gestation. However, when both the IGF-II receptor and IGF-II

genes were knocked out, -60% of the double mutant mice survived to birth (Filson et al.,

1993). These authors speculated that excess IGF-II may have toxic effects in utero. In the

case of mutation of the IGF-II receptor, the toxic effect of IGF-II may predominate, whereas

in the case of the double mutation this toxic effect is no longer evident. However, this model

remains speculative.

In contrast to the IGF-I receptor, the signal transduction pathway(s) of the IGF-II

receptor is less clear. Unlike the IGF-I receptor, the short intracellular domain of the IGF-II

receptor lacks tyrosine kinase catalytic activity (Czech, 1989; Ullrich and Schlessinger,

1990). Studies done by Nishimoto and colleagues have contributed to some understanding

of the signaling by this receptor. In 1987, these workers found that IGF-II stimulated Ca"

influx in BALB/c 3T3 fibroblasts (Nishimoto et al., 1987). Subsequently, they reported that


the IGF-II receptor directly interacts with a G protein family, G,2 (Nishimoto et al., 1989;

Murayama et al., 1990; Okamoto et al., 1990b). Subsequent analysis (Okamoto et al., 1990a)

revealed a Gi couplone, which later was found to be a consensus region in the intracellular

domain of most (if not all) G-protein-coupled receptors (Nishmoto et al., 1993; Nishimoto

1993).This couplone with a length of between 10 and 26 residues requires 2 basic residues

on the N-terminal end and the sequence B-B-X-B or B-B-X-X-B at the C-terminus (where

B is a basic residue and X, a nonbasic residue). This couplone has G-protein activating

function. Moreover, it was found that these G-protein signals are interchangeable, by creating

a chimeric IGF-II receptor (pIII-2/IGF-II receptor) containing a pII-2 couplone sequence

(Takahashi et al., 1993).

IGF Binding proteins (IGFBPs)

IGFs in the circulation are always associated with their cognate binding proteins. This

phenomenon was discovered some 30 years ago, with the observation that the M, of NSILA

became reduced after treatment with 5 M acetic acid (Burgi et al., 1966). IGF binding

proteins that bound [125I]IGF with high affinity were first reported by Zapf et al. (1975).

However, in only the most recent decade were the genes of at least six distinct IGFBPs

cloned and intensive studies done on these proteins functions, gene expression pattern and

genetic regulation (reviewed by Drop et al., 1992; Rechler and Brown, 1992; Jones and

Clemmons, 1995).

Characteristics of the IGFBPs. All six IGFBPs are secretary proteins that generally

do not accumulate in intracellular site(s). Therefore it can be difficult to measure the

intracellular levels of these proteins. The different IGFBPs are clearly distinct molecules that


do share regions of strong similarity (Drop et al., 1992). Specifically, the hydrophobic

cysteine-rich N-terminal and C-terminal regions are conserved. The alignment of these 18

cysteines is strongly conserved across the six IGFBPs with the exception of IGFBP-6 (lacks

2 cysteines in the human protein) and IGFBP-4 (contains 2 extra cysteines in the human and

rat proteins). The middle one-third region, where no cysteines are present, is the most

divergent among IGFBPs. IGFBPs-3, -4, -5 and -6 are glycosylated proteins. IGFBP-3 and

-4 are N-glycosylated, whereas IGFBPs-5 and -6 are O-glycosylated. IGFBPs-1, and -2

contain an RGD sequence near the carboxyl terminus, which is involved in binding to the

extracellular matrix protein receptors or integrins (Jones et al., 1993; Delhanty and Han,


Functions of IGFBPs. It has been demonstrated that IGFBPs have multiple functions

(recently reviewed by Jones and Clemmons, 1995) including the transport of IGFs in the

circulation, mediation of IGF transport from the vascular compartment, localization of IGFs

to specific cell types, modulation of IGF binding to receptors, and subsequent biological

actions. IGFBP-1 is a 25 kDa protein found in high concentrations in amniotic fluid, and is

also secreted by hepatocytes. IGFBP-2 has a significantly higher affinity for IGF-II than for

IGF-I, and is the major fetal IGF binding protein. IGFBP-3 is the main carrier of IGF-I and

-II in postnatal and adult serum. At least 95% of the total content of IGF-I and IGF-II in

serum is bound to IGFBP-3. In plasma, IGFBP-3 and IGFs can form a 150 kDa complex.

IGFBP-4 is a 24 kDa protein that was identified in serum and seminal plasma. IGFBP-5 and

IGFBP-6 have recently been cloned, and these proteins exhibit a higher affinity for IGF-II

than for IGF-I (Shimasaki, et al, 1991a, b).


The functions of IGFBPs appear to be altered by posttranslational modifications, such

as glycosylation, phosphorylation and proteolysis. In human pregnancy serum, IGFBP-3

proteolytic activity is responsible for the disappearance of intact IGFBP-3 from serum as

determined by western ligand blotting, with no change in IGFBP-3 immunoreactivity

(Guidice et al., 1990). In human seminal fluid, prostate specific antigen (PSA) can function

as a potent IGFBP-3 protease (Cohen et al., 1994). The proteolytic cleavage of IGFBP-5 to

lower molecular weight forms reduced its affinity for IGF-I (Camacho-Hubner et al., 1992).

Phosphorylation of IGFBP-1 seriess 119 and 169) cause a six-fold increase in its affinity

for IGF-I (Jones et al., 1993). It has been proposed that, when IGFBP-1 is phosphorylated,

it has a greater affinity for IGF-I than that of IGF-I receptor, and this favors the binding of

IGF-I by the binding protein; when IGFBP-1 is dephosphorylated, it has a lower affinity for

IGF-I, and IGF-I binding to its receptor is favored.

It is also the emerging consensus that IGFBPs may have IGF-independent functions.

IGF binding proteins most likely exert this function through RGD interactions with

extracellular matrix and integrins on the cell surface (Jones et al., 1993; Delhanty and Han


Gene structures. Similar to the protein structures, the genes encoding the six IGFBPs

share several common features. First, the localization of the IGFBP gene family in the

genome is closely associated with the HOX gene family ( Allander et al., 1993; Allander et

al., 1994). Both these families localize to four regions in the human: 1). 2q31-2q34 contains

IGFBPs-2 and -5, and the HOXD gene cluster (Cannizzaro et al., 1987; Ehrenborg et al.,

1991). 2). 7q15-7q12 contains IGFBPs-1 and -3, and the HOXA gene cluster (Ehrenborg et


al., 1992). 3). 17q12-17q22 contains IGFBP-4 and the HOXB cluster (Bajalica et al., 1992).

4). chromosome 12 contains IGFBP-6 and the HOXC gene cluster (Shimasaki et al., 1991b;

Scott, 1992). Secondly, the genes encoding the IGFBPs have similar structures: the coding

regions are divided into four exons; the corresponding exons are similar in size and

sequence; all genes have a relatively large first intron; and all the promoters, with the

exception of that for the IGFBP-2 gene, contain TATA boxes (Rechler and Brown 1992;

Allander et al., 1993; Zhu et al., 1993; Gao et al., 1993).

In summary, the IGFBPs, in all likelihood, have important roles in the IGF axis by

virtue of their endocrine, autocrine and paracrine actions, tissue and developmental stage-

specific expression, posttranslational modifications and potential ligand-independent


IGFBP-2 Gene and Physiology

IGFBP-2 Gene

Structure. The IGFBP-2 gene and/or complementary DNA has been cloned and

sequenced from the human (Binkert et al., 1989; Ehrenborg et al., 1991; Binkert et al., 1992),

rat (Brown et al., 1989; Margot et al., 1989; Brown and Rechler, 1990; Kutoh et al., 1993),

mouse (Landwehr et al., 1993), cow (Bourner et al., 1992), sheep (Delhanty and Han, 1992)

and chicken (Schoen et al., 1995). The IGFBP-2 chromosomal gene structures, in those

species that were studied (human, rat, mouse and chicken), are highly conserved. This gene

contains 4 exons and 3 introns. Because of the large first intron (23-32 kb), these genes span

28 to 38 kb in length. The size of the third exon is identical among animals, whereas the


other exons and introns are very similar in size, except for exon 4 of the chicken which

contains about 800 bp of additional noncoding sequence.

The promoter of the IGFBP-2 gene is GC rich and lacks TATA and CAAT boxes,

whereas the promoters of the other IGFBP genes contain TATA boxes. The promoter region

of IGFBP-2 gene is highly conserved across species and contains at least 3 G/C boxes. In the

rat, it has been shown that three clustered G/C boxes bind SP-1 transcription factor and are

required for efficient transcriptional initiation of this gene (Boisclair et al.,1993). In the

human, the transcriptional start site was mapped to -113 2 bp relative to the translational

start codon ATG (Binkert et al., 1992). In the rat, the transcriptional start site was first placed

at -88 nt using primer extension with a primer from +4 to -22 nt. However, it was mapped

to -151 nt using primer extension with primers from -65 to -104 nt and -76 to -99 nt. The

latter position was confirmed by RNase protection assay. Therefore, the transcriptional start

site was reported to correspond to -151 nt (Brown and Rechler, 1990). In 1993, however,

Schwander and colleagues reported a transcriptional start site for the rat IGFBP-2 gene at

-90 nt as identified using primer extension and Si nuclease protection assays (Kutoh et al.,

1993). Using reverse ligation-PCR (RLPCR), Boisclair and Brown (1995) provided

corroborative evidence that transcription of the rat IGFBP-2 gene initiates at a cluster

extending from -86 to -90 nt (relative to the ATG, +1), instead of the -151 nt. In the mouse,

this start site was mapped to -85 nt (Landwehr et al., 1993), which corresponds to -84 nt of

the rat IGFBP-2 gene. Nevertheless, this highly G/C rich promoter and exon may exhibit a

secondary structure at the mRNA cap site, which may make it difficult to denature and

reverse-transcribe IGFBP-2 cDNA at the 5' end. In addition, promoters lacking TATA-motifs


can exhibit multiple transcriptional initiation sites (for reviews see McKeon et al., 1990; Lu

et al., 1994).

Mammalian IGFBP-2 mRNAs are usually in the size range of~1.5 -1.6 kb. However

in human fetal liver, HepG2 and embryonic liver cell lines, a 4.4 kb mRNA was also found

(Zapf et al., 1990; Badinga, L. unpublished data). It is not clear why these human cells have

a different size of IGFBP-2 transcript. In the chicken, IGFBP-2 mRNA is 2.3 kb in length

because of a larger 3-untranslated region (Schoen et al., 1995).

Localization. In the human, this gene has been localized to chromosome 2q33-34

(Ehrenborg et al., 1991), where it is somewhat close to a cluster ofhomeobox genes at 2q31-

2q37 which are active in the regulation of development (Acampora et al., 1989), a gene

encoding cAMP response element binding protein 1 (CREB1) at 2q32.3-2q34 (Taylor et al.,

1989), and a gene encoding inhibin a at 2q33-2qter (Barton et al., 1989). Moreover, human

IGFBP-2 and IGFBP-5 genes are closely linked in a tail to tail orientation and are about 20-

40 kb apart (Allander et al., 1994). Similar to the human, the mouse IGFBP-2 and IGFBP-5

gene are linked in tail-to-tail arrangement separated by only 5 kb, and co-localized to

chromosome 1 (Kou et al., 1994). Interestingly, IGFBP-1 and IGFBP-3 genes were also

found in a tail-to-tail arrangement in the human (Ehrenborg et al., 1992) and mouse (Kou

K., 1994). These gene pairs may prove useful as models to study long-range chromatin

structure involved in genetic regulation.

IGFBP-2 Gene Expression and Regulation

General. Several studies showed that IGFBP-2 gene expression is controlled, at least

in part, at the transcriptional level (Babajko et al., 1993; Ooi et al., 1993), as transcriptional


rate and steady state RNA levels were closely correlated and increases in mRNA abundance

were found without correspondent changes in mRNA stability (Mouhieddine et al., 1996).

Unlike the IGF-II and IGF-II receptor genes, the IGFBP-2 gene is normally expressed from

both alleles and is not imprinted (Wood et al., 1994).

Developmental changes. IGFBP-2 gene expression is developmentally regulated in

those species where examined. The level of IGFBP-2 mRNA is higher in liver, kidney,

intestine and lung of fetal rats at late gestation than in these same tissues of adult rats

(Brown et al., 1989). IGFBP-2 mRNA is expressed at high levels in the livers of term-

gestation and early neonatal rats, and was greatly reduced in Day 21 postnatal and adult rats

(Ooi et al., 1993; Babajko et al., 1993). IGFBP-2 in chicken embryo serum increased

progressively between embryonic stages E10 to E22 (Yang et al., 1993). In the pig, hepatic

IGF-II and IGFBP-2 mRNA levels were higher at late gestation and birth than at 21 and 49

days of age (Kampman et al., 1993). Levels of IGFBP-2 in the fetal pig circulation increased

from fetal Days 45 to 110 (McCusker et al., 1988). IGFBP-2 was 2- to 3-fold more abundant

in fetal serum than in postnatal serum (Lee et al., 1991), while IGFBP-2 levels in pregnant

gilt serum were unaltered during pregnancy (Lee et al., 1993). In the ovine fetus, expression

of the IGFBP-2 gene was ubiquitous in tissues before 80 days of gestation, whereas this

became restricted to the liver, kidney and choroid plexus after Day 80 of gestation (Delhanty

and Han, 1993a).

In the human ovary, the follicular fluid of dominant follicles contains lower levels

of IGFBP-2 than subordinate or atretic follicles (Schuller et al., 1993; San Roman and

Magoffin, 1992). Similarly, levels of IGFBP-2 in dominant follicles were lower than in


subordinate follicles of cattle (Thatcher et al., 1996). In porcine ovarian follicles, IGFBP-2

mRNA level was positively correlated with day of cycle and follicle diameter (Samaras et

al., 1993)

IGFBP-2 mRNA abundance in the livers of lactating ewes is markedly increased

compared to that for pregnant ewes, whereas in mammary glands IGFBP-2 mRNA

expression was lower in lactating than pregnant ewes (Klempt et al., 1993).

Tissue specific expression. IGFBP-2 gene expression appears to be somewhat tissue

restricted. It has been shown that the IGFBP-2 gene is expressed in certain tissues including

the liver (Margot et al., 1989), ovary (Samaras et al., 1993), testis (Lin et al., 1993a),

mammary gland (Simmen et al., 1992), pituitary (Bach and Bondy, 1992; Michels et al.,

1993), bone (Schmid et al., 1992), lymphocytes (Neely et al., 1991; Nyman and Pekonen,

1993) and endothelial cells ( Moser et al., 1992). IGFBP-2 mRNA is also detectable in

adipose tissue of mid-pregnant gilts (Simmen et al., 1992). In rat intestinal epithelial cells,

IGFBP-2 is the major IGFBP produced (Park et al., 1992), whereas in the intestine of mid-

pregnant gilts, IGFBP-2 is not highly expressed (Simmen et al., 1992). In post-implantation

rat embryos, IGFBP-2 is expressed in cell populations that are rapidly dividing or in regions

that direct the growth and differentiation of neighboring cells and tissues (Wood et al., 1992;

Streck et al., 1992; Green et al., 1994). Studies of the porcine ovary showed that IGFBP-2

mRNA was most abundant in granulosa cells, lower in theca cells, and lowest in luteal cells.

However, another study found that IGFBP-2 was not detectable in the medium of rat

granulosa cells (Liu et al., 1993b). In mid-pregnant gilts, the IGFBP-2 gene is highly


expressed in uterus, liver and mammary gland, whereas there is little or no expression in

heart, muscle, pancreas and skin (Simmen et al., 1992).

In cancer cells. IGFBP-2 is expressed by some ovarian carcinoma cell lines, such as

PEO4, EFO-21, MFO-35, MFO-36, while not expressed in other ovarian carcinoma cell

lines, such as EFO-27 (Krywicki et al., 1993; Hofmann et al., 1994). IGFBP-2 mRNA

abundance was 2- to 30-fold higher in malignant ovarian tumors than in benign ovarian

neoplasms (Kanety et al., 1996). The human breast cancer cell lines MCF-7 and BT-20

synthesize and secrete this protein into the medium, whereas the MDA-MB-231 cell line

does the not express this particular IGFBP (Kim et al., 1991).

Metabolic effects. Fasting can induce IGFBP-2 mRNA abundance in the livers of

adult rats (Orlowski et al., 1990; Ooi et al, 1993). This induction can be reversed by

refeeding (Ooi et al., 1993). Starvation increased the fetal plasma levels of ovine IGFBP-2

(Gallaher et al., 1992). Liver IGFBP-2 mRNA levels were elevated in intrauterine growth-

retarded (IUGR) piglets at day 90 of gestation and at birth (Kampman et al., 1993). However,

another study indicated that maternal fasting did not alter the abundance of IGFBP-2 mRNA

in fetal rat liver (Straus et al., 1991). In dairy cows, the negative nutrient balance during early

lactation increased the serum concentration of IGFBP-2 (Vicini et al., 1991). A study

(Pucilowska et al., 1993) on measurement of IGFs and binding proteins before and after 21

days of refeeding of 22 undernourished Bangladesh children (2-4 years of age), showed that

the IGFBP-2 serum concentration before refeeding was twice that in controls and became

normalized after refeeding of a high protein diet, but remained high in these individuals fed


the normal protein diet, suggesting that circulating IGFBP-2 levels are sensitive to dietary

protein intake.

IGFBP-2 mRNA in livers of diabetic rats was increased over that for normal animals

(Ooi et al., 1992). This increase was, however, reversed by treatment with insulin.

Hormonal effects. Using cultured human endometrial stromal cells, Giudice and

colleagues have shown that the combination of estrogen plus progesterone increased IGFBP-

2 secretion by 10 to 15-fold and that progesterone alone stimulated IGFBP-2 mRNA steady-

state levels by 12 to 15-fold (Giudice et al., 1991b). This stimulation by progesterone was

blocked by the receptor antagonist, RU 486. EGF plus progesterone also had stimulatory

effects on IGFBP-2 gene expression in these same cells (Giudice et al., 1992). In the anterior

pituitary of ovariectomized rats, estrogen increased IGFBP-2 mRNA abundance (Michels et

al., 1993). In ovarian cancer cells, however, estrogen minimally depressed IGFBP-2 mRNA

abundance (Krywicki et al., 1993). Similarly estrogen did not influence IGFBP-2 mRNA

abundance in the human breast cancer cell line MCF-7 (Kim et al., 1991). In this same study,

it was shown that a-difluoromethylomithine (DFMO) an inhibitor of polyamines (PA),

inhibited IGFBP-2 gene expression in the BT-20 and MDA-MB-231 cell lines by an as yet

unknown mechanism.

In alveolar epithelial cells, TGF-pl was a potent stimulator of IGFBP-2 gene

expression (Cazals et al., 1994). In rat hepatocytes, retinoic acid stimulated IGFBP-2 gene

expression, and insulin blocked both the basal and retinoic acid-induced IGFBP-2 expression

(Schmid et al., 1992). A most recent study has shown that retinoic acid (RA) inhibits

proliferation of bovine mammary epithelial cells and increases levels of IGFBP-2 in


conditioned medium and in plasma membrane preparations (Woodward et al., 1996).

However, such effects of retinoic acid were not observed in osteoblasts, suggesting that this

regulation is tissue-specific (Schmid et al., 1992).

In rat leydig cells, the expression of both IGFBP-2 and its mRNA were decreased

by hCG in a dose-dependent manner (Wang et al., 1994). Treatment with 10 ng/ml hCG

reduced by 32% the transcription rate of the IGFBP-2 gene, whereas the half-life of the

mRNA remained unchanged. Forskolin decreased IGFBP-2 mRNA abundance and protein

synthesis in MDBK bovine kidney epithelial cells (Cohick et al., 1991). Phytohemagglutinin

(PHA), in contrast, increased IGFBP-2 mRNA production in human lymphocytes (Nyman

and Pekonen, 1993)

Treatment with bovine somatotropin decreased the serum concentration of IGFBP-2

in cows at early lactation, late lactation and the dry period (Vicini et al., 1991). Clemmons

and colleagues found that the serum concentration of IGFBP-2 is usually elevated in patients

with Growth Hormone deficiency (GHD), whereas IGF-I and IGFBP-3 concentrations are

lower than normal (Clemmons et al., 1991; Smith et al., 1993). However, the ratio of IGFBP-

2/IGF-I failed to constitute a reliable test for diagnosis or exclusion of GHD in all short

children. Moreover, it was reported that Growth Hormone deficiency in 'little' mice does not

affect the serum level of IGFBP-2 (Donahue and Beamer, 1992). Using primary cultures of

hepatocytes, Schwander and colleagues demonstrated that insulin was a negative regulator

of hepatic IGFBP-2 mRNA expression, whereas Growth Hormone had no effect (Boni-

Schnetzler et al., 1990).


A developmental switch from high expression of IGFBP-2 in fetal rats to low

expression in adult rats led investigators to examine the effects of thyroid hormone on the

expression of this gene. It was reported that hypothyroid pups continue to manifest high

levels of serum IGFBP-2 and IGFBP-2 mRNA in liver up to 19 Days of age, and that

treatment with thyroid hormone decreased this high expression, thereby indicating that

thyroid hormone may inhibit IGFBP-2 gene activity (Nanto-Salonen et al., 1991).

Recent studies showed that glucocorticoid (GC) rapidly decreased DNA synthesis and

proliferation of lung alveolar epithelial cells, which was associated with accumulation of

IGFBP-2 in the culture medium, and increases of IGFBP-2 mRNA in these cells

(Mouhieddine et al., 1996). Transfection using a 1.4 kb promoter region of the rat IGFBP-2

gene demonstrated that this region can respond to GC treatment with increased luciferase

reporter gene activity.

Summary of IGFBP-2 gene expression and regulation. 1). Although IGFBP-2 is

expressed in a number of tissues, it tends to be more expressed in epithelia and the

supporting stroma cells. 2). Since this gene often is up-regulated during certain physiological

conditions such as fasting, starvation, negative nutrient balance, cancer, and pregnancy,

IGFBP-2 may have some protective (homeostatic) function. 3). Estrogen, progesterone, TGF-

P, phorbol ester, GC and retinoic acid appear to be stimulatory, whereas GH, insulin, thyroid

hormone and hCG, appear to be inhibitory for this gene's expression. 4). Control of

expression of this gene appears to be complicated and divergent between tissues and across


Physiology of IGFBP-2

Protein. IGFBP-2 has been purified from conditional medium of the buffalo rat liver

cell line (BRL 3a) (Mottola et al., 1986) and the Madin-Darby Bovine Kidney epithelial cell

line (MDBK) (Szabo et al., 1988). Therefore, this protein was called the BRL3A cell line-

derived IGF BP, MDBK cell line-derived IGF BP or IBP-2 (Drop et al., 1992). The new

nomenclature has been used since 1991, as mandated by participants in the second

International IGF/Somatomedin Symposium in San Francisco. The mature IGFBP-2 in

human (Binkert et al., 1989), rat (Brown et al., 1989; Margot et al., 1989), mouse (Landwehr

et al., 1993), cow (Bourner et al., 1992), sheep (Delhanty and Han, 1992) and chicken

(Schoen et al., 1995) has 289,270,289,284,284 and 275 amino acids and 31 kDa, 29.5 kDa,

34 kDa, 34 kDa, 31 kDa and 33.5 kDa molecular weights, respectively. This protein from

all species studied contains 18 conserved cysteine residues and an RGD sequence. Mutations

of these cysteines influence the affinity of IGF binding (Coulter et al., 1995). Infusion of pure

IGFBP-2 in rats demonstrated that the half-life for this protein in the bloodstream was 144

32 min (Young et al., 1992).

Function. IGFBP-2 purified from MDBK cells had higher affinity for IGF-II than

IGF-I (Bourner et al., 1992). IGF-I binds to IGFBPs-3 and -1 with higher affinity than to

IGFBP-2 (McCusker et al., 1991). IGFBP-2 protein enhanced the DNA synthesis response

of porcine aortic smooth muscle cells to the IGF-I present in platelet-poor plasma (ppp).

However, in serum-free medium, it blocked the stimulation by IGF-I. Similarly, an early

study showed that MDBK IGFBP-2 inhibited the ability of IGF-II to stimulate DNA

synthesis, and protein accumulation in chick embryo fibroblasts, while a lesser effect on IGF-


I was observed (Ross et al., 1989). Bovine IGFBP-2 inhibited binding of IGF-I to the cell

surfaces of human fetal fibroblasts (GM10 cells) and porcine smooth muscle cells

(McCusker et al., 1991). Human recombinant IGFBP-2 with a single amino acid mutation

([Cys2S.]rhlGFBP-2) has been expressed and purified from the conditioned medium of a

clonal Chinese hamster ovary cell line (Feyen et al., 1991). This protein inhibited IGF-I-

stimulated cell proliferation in a dose-dependent manner, whereas it had no effect on insulin

stimulated cell proliferation. In addition, IGFBP-2 inhibited both basal and IGF-I stimulated

bone collagen synthesis. It has been reported that soluble IGFBP-2 inhibited the binding of

IGF-I and IGF-II to SCLC and NSCLC cells and inhibited IGF-stimulated DNA synthesis

in NSCLC cells (Reeve et al., 1993).

Accumulating evidence indicates that IGFBP-2 may have IGF independent functions.

In vitro studies suggested that IGFBP-2 binds to cell membrane proteoglycans through its

glycosaminoglycan binding domain (PKKLRP) (Russo et al., 1996). It has been

demonstrated that IGFBP-2 stimulates Growth Hormone receptor binding of Growth

Hormone and mRNA abundance of Growth Hormone receptor (Slootweg et al., 1995).

Inhibition of endogenous IGFBP-2, by transfection of the cDNA antisense expression

construct, stimulates proliferation of intestinal epthelial cell lines (Corkins et al., 1995).

Inactivation of one or both alleles of the IGFBP-2 gene in mice did not affect pre-

natal and post-natal growth, whereas in these animals the serum concentrations of IGFBP-3

and IGFBP-1 were increased (Wood et al., 1994). These data support the concept that

IGFBPs may have overlapping functions and be compensatory to some extent.


Collectively, IGFBP-2 may have several potential functions including inhibition

and/or modulation of IGF actions in soluble form and in association with cell membranes or

IGF-independent functions. However, since only few functional studies of IGFBP-2 have

been conducted, the function of IGFBP-2 remains relatively undefined.

Uterine IGF System

The uterus is an important reproductive organ with many physiological functions in

pregnancy and fetal growth and development. These functions which include uterine gene

expression, protein synthesis and secretion, maternal-fetal nutrient exchange and utero-

ovarian interactions are controlled not only by systemic hormones but also by locally

produced factors. Uterine IGFs and their binding proteins are implicated in the control of

uterine cell proliferation and differentiation in cyclic and pregnant animals. This family of

growth factors together with other hormones, growth factors and cytokines constitutes a

complicated and efficient uterine growth control network.

Differential Expression and Distribution in Uterine Tissues

A number of studies have demonstrated that IGFs, IGF receptors and IGFBPs are

expressed in uterine tissues and cells (reviewed by Simmen et al., 1995). Expression of this

system exhibits tissue and developmental specificity.

IGFs and IGF receptors. In the pig (Simmen et al., 1992; Song et al., 1996), IGF-I and

IGF-II genes are expressed in the endometrium and myometrium. Expression of the IGF-I

gene is higher in tissues of early than later pregnancy, whereas expression of IGF-II is lower

in the tissues of early than later pregnancy. Peak levels of IGF-I mRNA occur on Day 12 of

the cycle and pregnancy in pigs. IGF-I receptor numbers are low and do not change with


stage of pregnancy (Simmen et al., 1992). IGF-II receptor expression in pig uterus has not

been examined. In the human (Guidice et al., 1993), IGF-I mRNA is primarily expressed in

proliferative and early secretary endometrium. Abundant IGF-II mRNAs are expressed in

mid-late endometrium and early pregnancy decidua. IGF-I and IGF-I receptor mRNAs are

abundantly expressed in secretary endometrium and early pregnant decidua. In the rat (Zhang

et al., 1994), both IGF-I and IGF-II mRNAs are expressed in the uterus during the

periimplantation period.

IGFBPs. Similar to the IGFs and their receptors, IGFBPs are also differentially

expressed in the uterus. In the rat (Girvigian et al., 1994), IGFBP-2 and IGFBP-4 mRNAs

were localized in luminal epithelium of endometrium. IGFBP-2 mRNA is more highly

expressed during pro- and early estrus than at other stages of the cycle, while IGFBP-4

mRNA is present only at diestrus. IGFBP-3 mRNA is found in stromal cells and is highly

expressed on Day 12 of pregnancy. IGFBP-5 and IGFBP-6 mRNAs are maximally expressed

in myometrium. Maximal levels of IGFBP-5 mRNA are at estrus, whereas IGFBP-6 mRNA

abundance peaks at estrus. IGFBP-I expression in the rat uterus was detected in one study

(Ghahary et al., 1993), but was undetectable in another study (Girvigian et al., 1994). In the

pig, IGFBP-5, and -6 mRNA are expressed in both the endometrium and myometrium (Song

et al., 1996). IGFBP-2 is expressed at low levels in endometrium of early pregnancy and at

high levels in endometrium of late pregnancy of pigs, with little or no expression in

myometrium (Simmen et al., 1992). IGFBP-1 is the major IGFBP produced by the

endometrium and decidua of the human and baboon, which have a haemochorial placentation

(Giudice et al., 1991a, b; Tarantino et al., 1992; Tang et al., 1994). IGFBP-2 and -3 are less


abundantly expressed, although still at significant levels in these tissues. Interestingly,

IGFBP-1 gene expression is not detectable in the uterus of the pig, a species which exhibits

epitheliochorial placentation (F. A. Simmen, unpublished observations).

Potential Uterine Functions of IGFs

In maternal fetal interactions. One example of how IGFs might be involved in

maternal:fetal interactions is the paracrine role of IGF-I to stimulate estrogen production

from periimplantation conceptuses in the pig model. During the periimplantation period, the

porcine endometrial IGF-I gene exhibits a peak of expression at Day 12 (Simmen et al.,

1992). At this same time, IGF-I receptors are expressed in endometrium (Hofig et al., 1991)

and concepts (Green et al., 1995). This unique expression pattern is temporally coincident

with a transient period of secretion of estrogens from porcine conceptuses. An in vitro study

showed that treatment with IGF-I stimulated the abundance of mRNA for P450 aromatase

in Day 12 filamentous conceptuses (Green et al., 1995). These data support the model

proposed by Simmen et al. (1995). In this model, IGF-I from endometrium stimulates

concepts aromatase gene expression by binding to IGF-I receptors on the concepts.

Aromatase serves as a key enzyme in the pathway of estrogen synthesis. Estrogens during

this specific physiological condition are known to serve as the biological signals for

maternal recognition of pregnancy in pigs (Bazer et al., 1991). Estrogens also may be

involved in inducing or reducing uterine expression of genes, such as c-jun and c-fos

(Cicatiello et al., 1993), which in turn probably regulate uterine functions.

In contrast to enhancing embryo development, IGF-I may also be responsible, in part,

for early embryo losses (Katagiri et al., 1996). Superovulation increases early embryo losses


and IGF-I levels in uterine luminal fluid (ULF) of the rat. Both normal ULF infused with

IGF-I and ULF from superovulated rats impaired embryo development in vitro. Anti-IGF-I

antibody infusions after superovulation reversed the detrimental effect of superovulation.

Evidence supporting an IGF contribution to embryo losses is from an early study in

the pig model (Simmen et al., 1992). During the periimplantation period, expression of IGF-

II mRNAs are lower in the endometria from the Meishan breed with low concepts mortality,

than from the Large White breed exhibiting high concepts mortality. However, in this case,

IGF-I mRNA abundance was high in both breeds. IGF effects may parallel the morphogen

retinoic acid. The latter molecule is believed to be an embryotoxic factor when transported

in larger than normal quantities into the uterine lumen (reviewed by Roberts et al., 1993).

In implantation. A major function of the endometrium is to participate in implantation

of the concepts and support subsequent pregnancy. In the epithelialchorial placental species,

such as the pig, the role of luminal and glandular epithelial cells is predominant, whereas in

the hemochorial species, in which trophoblast invasion is extensive, differentiation of the

stromal compartment with formation of the decidua is characteristic. Temporal expression

of IGFs and their binding proteins in the endometrium provides a possibility that they may

be important for implantation in these different species. A recent study (Markoff et al., 1995)

showed that mouse IGFBP-4 mRNA becomes strongly expressed at each implantation site

and extends throughout the decidua, 24 hours after implantation, whereas no expression in

the uterine tissue between implantation sites was observed. This anatomically and temporally

characteristic expression pattern suggests a physiological role for IGFBP-4 in the

implantation process. Similarly in the baboon, IGFBP-1 and -2 are the predominant IGFBPs


in the endometrium. During early pregnancy, IGFBP-1 is only present in stromal cells that

are in intimate contact with the trophoblastic tissue (Tarantino et al., 1992). IGFBP-1

contains an RGD sequence, which is commonly seen on extracellular matrix protein

receptors and has been shown to be the recognition site for integrin cell surface receptors

(Jones et al., 1993). Since extracellular matrix is important for contact and migration of

trophoblast and endometrium (Kliman and Feinberg, 1990), IGFBP-1 may serve as a special

extracellular matrix binding protein during the implantation process.

Porcine Uterine IGFBP-2 and Remainine Problems

Uterine IGFBP-2 gene exhibits several interesting expression features (Simmen et

al., 1992, Song et al., 1996). First, the expression of this gene is tissue specific. It is restricted

in endometrium, while no expression is detected in the myometrium and placenta. In

endometrium, glandular and luminal epithelial cells as well as stroma express IGFBP-2.

Second, the expression of this gene is developmentally and temporally regulated. In cyclic

pigs, endometrial IGFBP-2 expression is high during estrus, and low during diestrus and

exhibits a cyclic wave. In endometrium of pregnant pigs, IGFBP-2 expression is low at early

pregnancy, increases from periimplantation, and reaches maximal levels at mid-pregnancy.

Third, this gene appear to be hormonally regulated. In cyclic pig endometrium, IGFBP-2

expression follows changes in estrogen concentration in the circulation, and is negatively

correlated to the systemic changes in progesterone concentration.

IGFBP-2 appears to have a specialized function in the feto-matemal interface.

Immunohistochemical localization showed that IGFBP-2 accumulates on the surface of

endometrial epithelial cells of pregnancy (Song et al., 1996). This observation implicates


IGFBP-2 in feto-matemal interactions and perhaps as a specific modulator of extracellular

matrix protein-cellular interactions through its RGD and/or other sequence motifs.

Based on its novel gene expression pattern and protein localization, a number of

interesting questions remain to be investigated: 1). What are the molecular mechanisms,

including cis- and trans-element interactions, DNA methylation, chromatin structure,

underlying the temporally-regulated and endometrial-specific transcription of the IGFBP-2

gene? 2). What hormones regulate IGFBP-2 gene expression at the level of transcription? 3)

Are there endometrial specific transcription factors? 4) What are the functions of uterine

IGFBP-2, especially at the feto-matemal interface? Answers to these questions may be

helpful for furthering our understanding of uterine gene expression and protein synthesis

during pregnancy.



The IGF binding proteins (IGFBPs) are a class of six growth-regulatory molecules

that exhibit sequence relatedness and the capacity to bind the mitogens IGF-I and IGF-II

(reviewed by Rechler, 1993; Jones and Clemmons, 1995). These proteins can associate with

cell membranes and extracellular matrix, where they influence IGF:IGF-receptor interactions

and possibly exert IGF-independent functions (Jones and Clemmons, 1995). Each IGFBP

is encoded for by a separate gene containing a unique set of transcriptional regulatory

sequences. Nevertheless, a particular cell-type may synthesize more than one IGFBP

(Giudice et al., 1991a, b; Girvigian et al., 1994; Ko et al., 1994b). The biological role(s) of

each IGFBP for a given tissue or cell-type remains relatively undefined by virtue of the

apparent redundancy in the expression and postulated autocrine/paracrine actions of this

protein family.

Previous work documented the temporally regulated expression during pregnancy of

mRNAs encoding components of the uterine IGF system (Simmen and Simmen, 1990;

Simmen et al., 1990, 1992; Geisert et al., 1991; Tarantino et al., 1992; Green et al., 1995).

The nature of the major IGFBP(s) produced by uterine tissues may vary considerably for


mammalian species during pregnancy. In the human, baboon and rat, species which exhibit

haemochorial placentation, IGFBP-1 is the major IGFBP synthesized in the endometrium

and decidua (Giudice et al., 1991a, b; Tarantino et al., 1992; Tang et al., 1994). IGFBP-2

and -3 are less abundantly expressed in uteri of these species. The pig is characterized by

noninvasive, epitheliochorial placentation and has been used by our laboratories to examine

the involvement of the uterine IGF system in embryo and concepts development (reviewed

in Simmen et al., 1993; 1995). Pig uterine endometrium maximally express mRNAs

encoding IGF-I and IGF-II at periimplantation (Day 8-12 post-mating) and post-implantation

stages of pregnancy (114-115 days, total length), respectively (Simmen and Simmen, 1990;

Simmen et al., 1992). The IGFBP-2 gene exhibits abundant mRNA expression in the porcine

endometrium and this is maximal at mid/late-pregnancy (Simmen et al., 1992). In sexually

mature, nonpregnant pigs, uterine endometrial IGFBP-2 mRNA abundance varies in parallel

with circulating estrogen concentrations and is negatively correlated with serum progesterone

concentration (Simmen et al., 1992). It also was shown that IGFBP-2 mRNAs are expressed

in uterine luminal epithelial, glandular epithelial and stromal cells, with little or no mRNA

in myometrium and placenta. However, immunohistochemical localization studies

demonstrated that IGFBP-2 protein is produced and/or accumulated in the luminal and

glandular epithelium but not in the endometrial stroma (Song et al., 1996).

In order to provide the requisite foundation for elucidating the molecular mechanisms

underlying this interestinting uterine gene expression phenotype, the pig IGFBP-2

chromosomal gene was cloned and characterized in this chapter.


Materials and Methods


Gilts were monitored twice daily for onset of estrous activity. Gilts exhibiting two

consecutive estrous cycles of normal duration (18-22 days) were mated at estrus with boars

and again 12 and 24 h later. The day of onset of estrus was defined as Day 0 of pregnancy.

Animals were sacrificed at the University abattoir on the indicated days of pregnancy.

Reproductive tracts were removed, immersed in ice and trimmed from the mesometrium.

Endometria, myometria and placentae were obtained by dissection as previously described

(Simmen et al., 1990; Ko et al., 1994a). Animal use protocols were approved by the

University of Florida Institutional Animal Care and Use Committee.

Northern Hybridization

Procedures used for extraction of total cellular RNA, isolation of poly(A)*-RNA,

formaldehyde-agarose gel electrophoresis, blotting of RNA to nylon membranes, preparation

of gel-purified, 32P-labeled IGFBP cDNA fragment and blot-hybridization were described

previously (Lee et al., 1993a, b; Ko et al., 1994b). Cloned cDNA insert used as probe

encoded rat IGFBP-2 (Brown et al., 1989). Sizes of IGFBP-2 mRNAs were calculated based

on the migration positions of the 18S and 28S ribosomal RNAs (2 and 5 kb, respectively).

Isolation and Mapping of Pig IGFBP-2 Cosmid Clones

A library of porcine genomic DNA fragments (derived from one male pig) cloned in

the cosmid vector pWE15 was obtained from Clontech Laboratories, Inc. (Palo Alto, CA).

Cosmids were screened by colony-hybridization with a rat IGFBP-2 cDNA fragment (SacI

X Small, 530 base pairs; Brown et al., 1989) labeled with 2P-dCTP by nick-translation. Five


positive clones were obtained after several rounds of re-screening and three of these

underwent detailed restriction analysis. The relative locations of restriction-endonuclease

cleavage sites in cosmid DNA inserts were determined using the Cosmid Mapping System

(Gibco-BRL, Inc., Gaithersburg, MD).

DNA Subcloning and Sequencing

Exons were preliminarily assigned to restriction fragments of cosmid clones by

Southern blot-hybridization with end-labeled, exon-specific oligodeoxynucleotide probes

designed by computer analysis of conserved mammalian IGFBP-2 cDNA sequences.

Restriction fragments that were positive by Southern blot were isolated from cosmid clones

by agarose gel electrophoresis and subcloned in pGEM-3Z (Promega Corp., Madison, WI).

DNA sequences of plasmid subclones were determined using the Sanger dideoxy procedure,

Sequenase (USB, Amersham Life Science, Cleveland, OH) and a primer-walking strategy.

Oligodeoxynucleotide primers were synthesized by the DNA Synthesis Core of the

Interdisciplinary Center for Biotechnology Research at the University of Florida. All DNA

sequences were confirmed on both strands and in some cases, sequencing reactions were

modified (e.g., by altering the temperature of denaturation, extension and/or termination

steps or by inclusion of deoxyinosine triphosphate) in order to sequence through problematic

G/C-rich areas. DNA sequences were compiled and analyzed using the Sequence Analysis

Software Package of the Genetics Computer Group, Inc. (GCG Package, Version 7, 1991).

The genomic sequences encompassing pig IGFBP-2 exons 1, 2-3 and 4 have GenBank

Accession Numbers U21117, U21118 and U21119, respectively.

Analysis of Intron 3

A fragment encompassing all of intron 3 was amplified from endometrial DNA (of

one pig) by high fidelity PCR (Expand High Fidelity PCR System, Boehringer Mannheim,

Indianapolis, IN). The upstream primer was 5'- TGACAAGCATGGCCTGTACAACCTC-3'

(within exon 3) and the downstream primer was 5'-ACGCTGCCCATTCAGAGACATC

TTG-3' (within exon 4). The PCR fragment was cloned using the TA cloning Kit

(Invitrogen, San Diego, CA) and analyzed by restriction endonuclease-digestion and partial

DNA sequencing.

Primer Extension

The procedure used for the primer extension assay was described previously (Simmen

et al., 1989). Briefly, the 32P-end labeled primer 5'-GGCAGCATGTTGGCG-3' was

coprecipitated with poly(A)* RNA, total cellular RNA or yeast RNA (as negative control),

and the precipitate dissolved in 30 pl of hybridization solution ( 50% formamide, 40 mM

Tris-HC1, pH 7.5, 400 mM NaC1, and 1 mM EDTA). This solution was incubated at 65 C

for 30' and then at 42 C for 16 hours. Nucleic acids were precipitated and dissolved in

extension buffer (50 mM Tris-HC1, pH 8.0,50 mM KC1, 5 mM MgCl2, 5 mM DTT, 4 mM

dNTPs and 1U RNase inhibitor). The extension reaction was initiated by addition of 40 U

AMV reverse transcriptase (Invitrogen, San Diego, CA), and the samples were incubated at

42 C for 90 min. Extended DNA products were separated on a 6% denaturing,

polyacrylamide gel. Sequencing reactions using the same primer (without 32P-label) and

plasmid H6 (see Results) as template were electrophoresed on the gel as size markers.

S -Nuclease Protection Assay

Sma-LUCe plasmid DNA (see Chapter 4, Materials and Methods), which contains

an IGFBP-2 genomic fragment spanning from -305 to +73 relative to the ATG and 3 bp of

HindI linker sequence, was linearized at the SmaI site and used as template for synthesis

of labeled antisense probe by asymmetric PCR. The oligonucleotide primer used in PCR was

5'-GGCAGCATGTTGGCG-3'. The 32P-labeled, single-stranded probe was purified by

denaturing polyacrylamide gel electrophoresis. Hybridization of probe to RNA and

subsequent treatment with SI nuclease was performed using reagents provided in the S1-

Assay kit (Ambion Inc., Austin, Texas). The products of the S -nuclease protection assay

were separated on 6% denaturing, polyacrylamide gels.

RNase Protection Assay

The plasmid H6 was cleaved with NotI at position +73 relative to the ATG initiation

codon. The adhesive ends were made blunt-ended by use of the Klenow fragment of DNA

polymerase I. The DNA was ligated to a synthetic HindIIl linker (Cat. # 2172, Promega

Corp., Madison, WI) and then was cleaved with HindIII. A fragment, spanning from -1397

to +73 relative to the ATG, and containing 3 bp of HindIll linker sequence at the 3'-end was

isolated from an agarose gel and ligated to pGEM-3Z plasmid (Promega Corp., Madison,

WI) previously linearized at the HindIII site. The resultant plasmid in which the SP6

promoter is adjacent to the 5'-end of the insert was designated BP2-19. The plasmid in

which the SP6 promoter is adjacent to the 3'-end of the insert was designated BP2-18.

Plasmid BP2-18 was then cleaved with SmaI. The fragment containing the SP6 promoter and

the insert, spanning from +73 to -305, was purified by agarose gel electrophoresis and DEAE


paper, and used as template for synthesis of antisense RNA probe for RNase protection

assay. The 32P-labeled probe was synthesized using the MAXIscript SP6/T7 in vitro

transcription Kit (Ambion Inc., Austin, Texas). RNase protection assay was performed

using the HybSpeedM RPA Kit (Ambion Inc., Austin, Texas). Products of the RNase

protection assay were resolved on 6% denaturing, polyacrylamide gels.

Hela Cell In Vitro Transcription System

Plasmid BP2-19 was cleaved with Smal + HindlI. The fragment spanning from the

Small to HindIl sites was purified by agarose gel electrophoresis and DEAE paper, and used

as template for in vitro transcription using the Hela Cell Extract Transcription System

(Promega Corp., Madison, WI). The "P-labeled transcripts were separated on 6%

denaturing, polyacrylamide gels.


IGFBP-2 mRNAs in Pregnant Pig Endometrium

Previously, it was reported that IGFBP-2 mRNA abundance (i.e., steady-state level)

was highly induced in the endometrium of pigs after the implantation period as analyzed by

the RNA dot blot-hybridization technique (Simmen et al., 1992). In order to confirm this

gene expression pattern, Northern blot-hybridization for IGFBP-2 mRNA was performed.

The pig endometrial IGFBP-2 transcript (-1.6 kb in length) was confirmed to be of higher

relative abundance after implantation (i.e., Days 30-90) than prior to or during the

periimplantation period (Days 8-12) (Figure 3-1).

0 LU

w w www
C Co 0 o


- l IGFBP-2 (1.6 Kb)

Figure 3-1. Northern analysis of IGFBP-2 mRNA in pregnant pig uterus. Thirty u'g total
cellular RNA from pig endometrium at Days 0, 8, 10, 12, 30, 60 and 90 of pregnancy was
subjected to Northern analysis with "P-labeled rat IGFBP-2 cDNA fragment. Gestation in
the pig is 114-115 days.


Characterization of the Pig IGFBP-2 Chromosomal Locus

The above and previous studies identified several unique characteristics of the

uterine-expressed IGFBP-2 gene. Of particular interest were: 1) the marked induction in

IGFBP-2 mRNA accumulation with later stages of pregnancy, and 2) the endometrial (as

opposed to myometrial) specificity of expression of this gene. Therefore, as a first step to

identifying the underlying molecular mechanism responsible for this uterine IGFBP gene

phenotype, the porcine IGFBP-2 locus was cloned and characterized. Three cosmids

(designated 6-1, 4-9 and 9-7) that collectively spanned this gene were isolated and

characterized by restriction endonuclease mapping and Southern blot-hybridization using rat

cDNA and synthetic oligodeoxyribonucleotide probes (Figure 3-2). Appropriate plasmid

subclones were constructed from these cosmids and the four exons were then identified first,

by blot-hybridization with consensus exon-specific oligonucleotide probes and subsequently,

by DNA sequencing (Figure 3-3). The intron-exon junctions and putative polyadenylation

signal were found to be highly conserved across the IGFBP-2 genes of species where

characterized (Table 3-1). In addition, the complete cDNA (-1.4 Kb in length) defined from

sequencing analysis of the four genomic exons was in excellent agreement (upon addition

of 150-200 nucleotides of poly A) with estimates of length of the corresponding endometrial

mRNA (above) and encoded an open reading frame for porcine IGFBP-2 that exhibited high

identity with IGFBP-2 proteins from other species (Figure 3-4 and Table 3-2). Comparisons

of IGFBP-2 exonic sequences across mammals indicated differential degrees of sequence

conservation between the four exons (order of conservation: exons 3 > 4 > 2 > 1).

El E2E3 E4
[- 1 f1l

Sac I .


Barn HI i i i i i I i

I s 11 I II

Hind III

Not I




H1-12 S29

10 Kb

Figure 3-2. Structural organization of the porcine IGFBP-2 chromosomal locus. Cosmids
6-1 and 4-9 overlap exon (E) 1, E2 and E3. Cosmid 9-7 contains E4 and a large segment of
DNA flanking the 3' side of E4. Plasmids H6 and Hl-12 (Hind III fragments, -6 kb, derived
from Cosmid 6-1,) were subjected to dideoxy sequencing to obtain the DNA sequences of
El, E2 and intron 2. Plasmid B11 (Bam Hi fragment, -4 Kb, derived from cosmid 4-9) was
analyzed to obtain sequence of E3. Plasmid S29 (Sac 1 fragment, -1.5Kb, derived from
cosmid 9-7) was analyzed to obtain the sequence of E4. Plasmid IN3 contains the entire
intron 3 fragment which was generated from pig genomic DNA by PCR. Lengths of El, E2,
E3 and E4 are -527 bp, 227 bp, 141 bp and 528 bp, respectively. Introns 1, 2 and 3 are -24
Kb, 746 bp and 2.6 Kb, respectively.

' '

Figure 3-3. DNA sequence and deduced amino acid sequence of porcine IGFBP-2
chromosomal gene and protein, respectively. Shown are the sequences of -1.4 Kb of
upstream flanking DNA, exon 1, exon 2, all of intron 2, exon 3 and exon 4. The sites of
transcriptional initiation and the polyadenylation signal are capitalized and underlined,
respectively. Intron sequences and sequences upstream of the translational initiation codon
are depicted in lower case letters.


A. Exon 1

gcttcaatttcaattctgatcccacccctgaaacccccatcacctttcta 50
cacctgacaaaagctgcttcagtaaagtcttgaagttggaaagagctgga 250
aatgattcagtaaatcaaacagaagaataaagggaggccattcttgttca 500
aagggagcaggtttctcagtgtccagaaactgcatttcaggtacttacat 750
cggccacatgggaagcgcgcaaacgaagtgcttccgaattgaaccgaaaa 1000
gcgcgggagtgtcgggggaagggggtggtctccaaaagggggaggggaga 1250
gccacctgccgct tcgcgcctcgcc gctcaccgccgccaacATGC
I MetL

3 euProArgLeuGlyGlyThrAlaLeuSerLeuLeuProLeueeuLeuLeu

19 LeuLeuGlyThrGlyGlyArgGlyAlaArgAlaGluValLeuPheArgCy

36 sProProCysThrProGl uSerLeuAl aAl aCysArgProProProAl aA

53 1 aProProSerAlaGlyAlaGlyProAlaGlyAspSerArgAlaProCys

69 G6uLeuValArgGluProGlyCysGlyCysCysSerValCysAlaArgLe

86 uGluGlyGluArgCysGlyValTyrThrProArgCysAlaG]nGlyLeuA

103 rgCysTyrProHisProGlySerGluLeuProLeuGlnAlaLeuValLeu

119 GlyGluGlyThrCysGluLysArgArgAspAlaGluTyrGlyAlaSerPr

136 oGluGlnValAla


Figure 3-3 -- continued

B Exon 2-Exon 3

acctgcagctcttccttgcttgctctttgcagACAATGGCGACGATGCTG 50
140 AspAsnGlyAspAspAlaG

147 luGlyGlyLeuValG1uAsnHisValAspGlyAsnValAsnLeuLeuGly

163 GlyThrGlyGlyAlaGlyArgLysProLeuLysSerGlyMetLysGluLe

180 uAl aValPheArgGluLysValThrGl uG1nHisArgG1nMetGlyLysG

197 lyGlyLysH1sHi sLeuGlyLeuGluGl uProLysLysLeuArgProPro

213 ProAlaArg

tttgaaatgggctgagattatgggttcttgaagtccttccttctaaaac 500
gggaagagaaactctagaggcagcctcttgaatttaggatttgcaagtcg 750
gtttcacgatccagttctcactctaagcatcctcttggctcgcctgtgcc 1000
216 ThrProCysGlnG nGluLeuAspGlnValLeuGluArgl eSer

231 ThrMetArgLeuProAspGl uArgGlyProLeuGl uHi sLeuTyrSerLe

248 uHislleProAsnCysAspLysHisGlyLeuTyrAsnLeuLysGln

gagctgagccggagacctagtcttcatgagccaggggtggggatgtccgc 1250


Fimure 3-3 -- continued

C Exon 4

cttagccctgcagaaccgcaagcacacccagacctgccgaagggcatctg 50
263 CysLysM

266 etSerLeuAsnGlyG1nArgGlyGluCysTrpCysValAsnProAsnThr

282 GlyLysLeulleGinGlyAlaProThrIleArgGlyAspProGluCysHi

299 sLeuPheTyrAsnGluGlnG nGlyAl aArgGlyAlaHisThrGInArgM

316 etGln

gatgcccagctggagatgcccgcaccctccctcctggaatccctgctggg 500
cttcagctttccccggaccctgggggaaatgggagaggcaagaatgggtt 750

Table 3-1. The sequences of exon-intron boundaries and the 3' end of the IGFBP-2 gene.

5' splicing site 3' splicing site
Consensus AG/GTAAGT Pyrmidine-rich NCAG/G


3' end of the gene


IGF Binaing





Figure 3-4. Deduced amino acid sequences of pig (P), human (H), rat (R) and mouse (M)
IGFBP-2 proteins. Putative IGF binding site, protein kinase C phosphorylation sites (PKC
PS), tyrosine phosphorylation sites (Tyr. PS) and Arg-Gly-Asp (RGD) integrin-binding
sequence were identified by computer-assisted sequence analysis.

Table 3-2. Identity of porcine IGFBP-2 amino acid sequence with IGFBP-2 proteins of
other species.

Exons Human" Bovineb Ovine' Rat' Mouse' Overall

El 84.9 84.2 84.2 72.7 71.9 79.6

E2 90.8 94.7 89.5 81.6 81.6 87.6

E3 100 100 100 97.9 95.7 98.7

E4 96.3 96.3 98.2 87.0 87.0 93.0

Overall 90.8 91.1 90.2 81.0 80.0
a Binkert et al., 1989
b Bourner et al., 1992
c Delhanty and Han, 1992
d Brown et al., 1989
" Landwehr et al., 1993


It was recently reported that the human IGFBP-5 gene is closely linked (distance of

20 to 40 kb) with the IGFBP-2 gene in a tail to tail orientation in chromosomal region 2q33-

34 (Allander et al., 1994). Southern blot analysis using a human IGFBP-5 cDNA probe did

not indicate hybridization to the pig cosmid 9-7, the insert of which contains exon 4 at one

end and spans about 30 kb to the 3' side (data not shown). Therefore in the pig genome, the

IGFBP-2 gene is separated from the IGFBP-5 gene by more than 30 kb.

Identification of 5'-Transcription Start Site(s)

In order to identify the transcriptional start site(s) of the pig IGFBP-2 gene, primer

extension was initially employed. A 15 bp synthetic oligodeoxynucleotide primer was end-

labeled and used in hybridization to poly(A) -RNA, total cellular RNA and yeast RNA (as

negative control). Extension products from poly(A) and total cellular RNA but not from

yeast RNA were observed with 5'-ends at positions corresponding to -96 to -87 (relative to

the ATG initiation codon), suggesting multiple clustered initiation sites (Figure 3-5A).

However, these signals were relatively weak and unresolved. Therefore, Sl-nuclease

protection was used to corroborate or refute the results from primer extension. Again,

multiple bands were observed that indicated heterogeneous 5' termini at positions -96 to -87

(Figure 3-5B). Interestingly, when the Hela cell in vitro transcription system and an RNase

protection assay (RPA) were used to further confirm the locations of these transcription start

sites, strong signals were found (especially in the RPA) not only at positions -96 to -87 but

also slightly upstream at positions -109 to -105 (Figure 3-5C and D). Overall, the results

from these four independent experimental approaches indicated that the TATA box-less and

GC-rich promoter region of the pig IGFBP-2 gene exhibits two clusters of transcriptional

1 2 3 4 1 2 3 1 2 3

S20 182-
> 18 178-


100- ii100-

Figure 3-5. Identification of 5'-transcriptional start sites. Panel A. Results of primer
extension assay. Primers were annealed to 2 pg of poly (A)W RNA (lane 1) or, 50 pg of total
cellular RNA (lane 2) from endometrium of a Day 60 pregnant pig, 50 pg of yeast tRNA as
control (lane 3) or no RNA (lane 4), respectively. Sequencing reactions using the same
primer were co-electrophoresed and the sequence corresponding to the extended products is
shown on the left. Panel B. Results of SI nuclease protection assay. The "P-labeled probe
was hybridized to 10 pg of total cellular RNA from endometrium of Day 30 (lane 1), Day
60 (lane 2) and Day 90 (lane 3) pregnant pigs. DNA size markers (100 bp ladder, Gibco
BRL, Life Technologies, Inc., Grand island, NY) were run on the same gel and their
migration positions are indicated on the left. Panel C. Results of Hela cell in vitro
transcription assay. Templates used for the in vitro transcriptions were 100 ng (lane 1) or 200
ng (lane 2) of Smna + HindIII fragment, or 100 ng of CMV immediate early promoter DNA
(lane 3, as positive control). RNA size markers transcribed by T7 RNA polymerase using
Century Marker Templates (Ambion, Inc., Austin, Texas) were run on the same gel and their
migration positions are indicated on left. Panel D. Results of RNase protection assay. Five
ug of total cellular RNA from the endometrium of a Day 60 pregnant pig was hybridized to
the "P-labeled probe. The positions of the RNA markers are indicated on the left.


start sites (at -109 to -105 and -96 to -87) of which the cluster at -96 to -87 appears to be the

predominantly utilized one (Figure 3-6).

Insights from Analysis of the Upstream Region of the IGFBP-2 Gene

Comparisons of the sequences of the porcine, human, rat and mouse IGFBP-2 gene

5' flanking regions (to the ATG translation initiation codon) were performed using the

currently available sequences. The highest overall homology was observed for the DNA

region spanning the G/C boxes to the ATG codon (Fig 3-7). Smaller blocks of conserved

sequence were apparent in the more upstream region, several of which overlapped with

transcription factor binding site motifs: e.g., a DBP site (Faisst and Meyer, 1992), a

nonconsensus estrogen-response element (ERE) and two ERE half sites (Dana et al., 1994)

(Figure 3-7). Two non-conserved progesterone-response element (PRE) half sites (Lieberman

et al., 1993) were also identified in the 5' flanking region of the pig IGFBP-2 gene.


This study has extended previous investigations of the pig uterine IGF/IGFBP system

during pregnancy. A close temporal association of IGFBP-2 mRNA and maximal circulating

estrogens and an inverse relationship of IGFBP-2 mRNA with circulating progesterone in

sexually mature pigs was noted previously (Simmen et al., 1992). However, estrogen or the

combination of estrogen and progesterone were ineffective in causing an induction (or

repression) in uterine IGFBP-2 mRNA level in prepubertal pigs (Simmen et al., 1990).

Results of the present study indicated maximal IGFBP-2 mRNA abundance in pig

endometrium during the mid-pregnancy. This pregnancy-stage specific gene expression

seems not simply correlated with neither estrogen nor progesterone levels in maternal serum.

z + X
Linker sequece

Products of primer extension and S1 assay
I Primer

Products of RPA
182 178 169 160

Products of in vitro transcription assay

Figure 3-6. Summary of the results of the localization of the 5'-ends of IGFBP-2 mRNAs. Clusters of transcriptional start sites are boxed.

A. PRE ...-...-..... tottct
P -1272 gtaottttTgt -1261
P -801 ggqttctgqq -789
B. ERE --------------. ootca
P -1035 :.:rttgTiadatcdajarcc4qdqaatgdg -1003
R -1050 ll d iiIll|llll l c 9 -1017
H -1073 Il lll llllllllllll -4 1041
ERE ............taacc
P -833 gctaidlictaacCtctEagatcctLCad)jd. -800
R -854 t, all III I I l l '| I Ii t 2l 2 -822
S -865 t la lll 1111 I lllci I r 1. -833
ERE -.--.--------.. -- --- -atca(n)taacc
P -486 9Ctg ddadctactt cIIQtcaCt igcgaci
H -528 I 'I l l ll l lI ldadlg 4lCrll ttadC ttrtCgl tc
R -509 tic 't I I I Ir lI i a C Ctl a gittc
H -529 ct c i't It ll Ill ldla lll c ttitqigittc
P icctCigi :cacIat g c dajcqcccadadqagajjtg -417
H cCta4iCrgJ ( 4 I -429
R ccag a. d-:.I I d t -422
H c cgqc.Ucttliqt d aI a I t -442
C. DBP .acaaaaca
P -39 aca.dga caaaa, caaccca ccLttgiccag -367
R -1 6 II ill lllil 13., -367
H -4 11111111 i I I a -38
Idin],. 388
D. GC boxes
Box 4
P -240 aau 9TPat.Jggcgadtqacq= laigcqCd gidi Ldga g:c gc
H -256 6 I | | | I IIIc I ll l l| ac c.C tl la
R -243 a IIIIlIl itl I Cl ccg I al I
8 -264 I I llll l ll d lidia t 4l I'
Box 3
P 7NlaL; qt l gggaaaaot iatctcaddadq
H III W Ig I llialilIl l I I[
R ol I LIg i.a, q, I, 1J1
K sIII CaggJga(HCqfloc HC44C4CtdH504H IllallllIlli
Box 2 box 1
P agosamaQajag rq. ga00uc4aaadjdagCd 94CCC t JC Qg
H IIIIII I ll III 1111111a I lilll I
NM ilii l l IiI Ii i I
P r adgddgc~cqL dCiqog.L9 iL LCqC9GC da4 3JiCC4gtCCiCCdCqCCCgCICCgCcoE

P ............ gccgtcgcgctcaccccgccaacatg +3
H ~cccCCHcCi c|cH IC |Il|l |C HallI | 'g I +3
Slidg itC g I acrdj l +3

Figure 3-7. Conserved sequence motifs in the region upstream of the translational initiation
codon of the IGFBP-2 gene. Corresponding sequences of the porcine (P), human (H)
(Binkert et al., 1992), rat (R) (Brown and Rechler, 1990) and mouse (M) (Landwehr et al.,
1993) IGFBP-2 genes were aligned by use of the Group and Pileup programs of the GCG
Package. Two PRE half sites in the porcine sequence are shown in A. Localized regions of
homology which overlap ERE half sites and imperfect palindrome (B), DBP site and G/C
boxes (D) are shaded. The transcription start sites are boxed.


In human endometrium, IGFBP-2 and IGFBP-3 mRNA abundance and protein

synthesis and secretion are greater in the secretary than proliferative phase of the estrous

cycle, suggesting a positive association with increased serum progesterone (Giudice et al.,

1991a, b). Moreover, cultured endometrial stromal cells from women exhibit enhanced

IGFBP-2 and IGFBP-3 secretion and correspondingly increased mRNA levels at -6 days

after administration of estrogen and/or progesterone (Giudice et al., 1991 a). The effects of

steroid hormones on uterine production of IGFBPs may be immediate (i.e., steroid receptor

transactivation of IGFBP promoter), indirect, perhaps via the induction or repression of

intermediate proteins such as endometrial-specific transcription factors or a combination of

both mechanisms.

Results from this and earlier studies (Simmen et al., 1990, 1992; Geisert et al., 1991)

identified marked inductions in endometrial IGFBP-2 mRNA steady-state levels beginning

just after initiation of implantation. It is interesting that pig endometrial IGFBP-2 transcripts

and epithelial IGFBP-2 are maximally expressed after day 30, the time at which the porcine

placenta becomes completely formed (King, 1993). The pig exhibits diffuse epitheliochorial

placentation with no invasiveness of the trophoblast. Therefore, IGFBP-2 may be an

important mediator of endometrial/placental interactions that underlie degree of trophoblast

invasiveness, placental development and/or nutrient transfer across the placenta.

IGFBP-2 has been localized on membranes of cells that synthesize this protein

(Reeve et al., 1993; Chapter 5) where it competes for binding of IGFs to receptors. A

negative correlation of endometrial IGF-I and IGFBP-2 mRNA abundance was recently

observed (Song et al. 1996) and may indicate that IGF-I is an inhibitor of endometrial


IGFBP-2 gene expression. One interesting discrepancy relates to the absence of

immunostaining for IGFBP-2 in endometrial stroma at mid-pregnancy (Song et al., 1996).

These cells have been shown to express the IGFBP-2 gene (Simmen et al., 1992). This may

indicate either a block in translation of the stromal IGFBP-2 mRNA or the rapid transfer of

IGFBP-2 from stromal cells to endometrial epithelial surfaces during its biosynthesis and


The coding and deduced amino acid sequences, predicted exon-intron junctions and

polyadenylation site for pig IGFBP-2 reported here are in complete agreement with the

corresponding sequences of an unpublished, full-length porcine IGFBP-2 cDNA cloned from

vascular smooth muscle cells in the laboratory ofD.R. Clemmons (personal common., D.R.

Clemmons). In addition, Coleman et al. (1991) reported a partial amino-terminal amino acid

sequence for porcine serum IGFBP-2. Their sequence initiated 3 amino acid residues

downstream of the glutamic acid residue considered to be the first residue of mature IGFBP-

2 (Brown et al., 1989; Bourner et al., 1992; Drop et al., 1992) and was in agreement with that

elucidated here by sequencing of the first exon of the pig IGFBP-2 gene. The porcine

IGFBP-2 protein exhibits the conserved pattern of 18 cysteine residues clustered at both

termini of the protein as well as the Arg-Gly-Asp (RGD) motif in its carboxyl-terminus.

Putative protein kinase C and tyrosine kinase phosphorylation sites were found in this

protein by computer analysis. However, the functions of these sites are yet to be tested. The

overall sequence conservation of IGFBP-2 molecules of different mammalian species is very

high (80-91%).


The promoter of the IGFBP-2 gene is TATA-less and GC-rich. Computer assisted

sequence analysis indicated that the corresponding mRNAs can form hair-pin secondary

structures (data not shown). In this study, primer-extension analysis yielded a very weak

signal due to the low efficiency of reverse transcription possibly caused by mRNA secondary

structures. Although the DNA sequences of the IGFBP-2 gene promoter region are highly

conserved in the pig (this study), the human (Binkert et al., 1992), the rat (Brown and

Rechler, 1990; Kutoh et al. 1993; Boisclair and Brown,1995) and the mouse (Landwehr et

al., 1993) (Figure 3-6), the transcriptional start sites are not the same. TATA and CAAT-less

promoters generally have multiple transcriptional start sites (for reviews see McKeon et al.,

1990; Lu et al., 1994). In the present study, two clusters of transcriptional start sites were

identified. The major cluster at -96 to -87 is located at a similar position as the transcription

start site(s) of the human, rat and mouse IGFBP-2 genes. However, the minor cluster at-109

to -105 may represent initiation sites unique to the pig IGFBP-2 gene. Although this cluster

of start sites was not clearly demonstrated by primer extension and Sl-nuclease protection

assays, in vitro transcription and RNase protection assays clearly demonstrated its


The IGFBP-2 exon-1 region and 5' flank is characterized by a number of G/C boxes

(GGGCGG or the reverse complement). These motifs can constitute cis-acting elements for

a number of different DNA-binding proteins such as Spl and related factors and were

previously shown to be essential for basal transcriptional activity of the rat IGFBP-2 gene

(Boisclair et al., 1993; Kutoh et al., 1993). A number of G/C box-binding proteins, including

SPI, SP3 and BTEB are present in porcine uterine endometrium of pregnancy (Wang, Y. and


Simmen, R.C.M., unpublished results). Spl binding to cis elements may also serve as part

of a mechanism to protect CpG islands from de novo methylation (Brandeis et al., 1994;

Macleod et al., 1994). For tissue-specific genes, a strong correlation exists between

expression and undermethylation (Razin and Cedar, 1991). The location of this CpG island

(bracketing the transcriptional initiation site, translational start site, leader peptide and

remainder of exon 1), its strong evolutionary conservation, and potential for differential

methylation point to its probable function in transcriptional and perhaps, post-transcriptional

regulation of expression of this gene in utero.


In this study, endometrial-specific and temporally regulated expression of the pig

uterine IGFBP-2 gene was confirmed. In particular, maximal expression of the IGFBP-2

gene was found around mid-pregnancy. As a prerequisite to understanding the molecular

mechanisms of this gene's regulation, the genomic structure of the IGFBP-2 gene, which is

comprised of four exons spanning -29 kb and which encodes a 316 amino acid precursor

protein, was elucidated. The TATA and CAAT box-less promoter region of this gene

exhibits two clusters of transcriptional start sites located -109 and -87 relative to the

translational initiation codon. Lastly, putative transcription factor binding sites in the

upstream 1.4 kb region flanking the IGFBP-2 gene were identified. Several of these motifs

may be associated with molecular mechanisms by which circulating steroid hormones affect

endometrial IGFBP-2 mRNA and protein steady-state levels during the estrous cycle and




IGF-I and -II undoubtedly play important regulatory roles in feto-matemal

interactions (Simmen et al., 1995) as in many other physiological processes. IGF binding

proteins (IGFBPs), a family of at least six members, are modulators of IGF actions in the

circulation and at the cellular surface (Jones and Clemmons, 1995). Functions and

expression of each these IGF binding proteins exhibit tissue-specificity, developmental-

specificity and hormonal regulation (Rechler and Brown, 1992; Clemmons, 1993). Inthe pig

uterus, the IGFBP-2 gene has a relatively unique pattern of mRNA expression during

pregnancy. The expression of this gene becomes induced at periimplantation to reach

maximal levels at mid-pregnancy. Unlike the other uterine-expressed IGFBPs such as

IGFBP-3, -4 and -5 and other uterine genes such as uteroferrin and IGF-I (Simmen et al.,

1992; Song et al., 1996), the IGFBP-2 gene is highly expressed in uterine endometrium, with

little orno mRNA expressed in the corresponding myometrium during pregnancy. Therefore,

this gene may constitute a useful model system to study endometrial-specific and pregnancy-

dependent modes of transcriptional and post-transcriptonal regulation as well as identify

novel uterine endometrial transcription factors.


The genomic organization of the IGFBP-2 locus has been characterized in human

(Ehrenborg et al., 1991; Binkert et al., 1992), rat (Brown and Rechler, 1990; Kutoh et al.,

1993), murine (Landwehr et al., 1993) and chicken (Schoen et al., 1995). Unlike the other

IGFBPs, the IGFBP-2 gene has a TATA box-less promoter, which is highly GC-rich and

conserved across species (see Figure 3-7.). There are several GC boxes in the conserved

promoter region. In rat, it has been shown that four GC boxes are required for binding of

transcription factor SP1 to elicit basal transcription. Moreover, the further upstream or distal

region also contains some conserved sequences and putative transcription factor binding sites

such as half EREs, half PREs, DBP and API sites. In pregnant pig endometrium, SPI

expression is constitutively low and is not correlated to the pattern of expression of IGFBP-2

RNA during pregnancy (Y. Wang and R.C.M. Simmen, unpublished data). Although this

GC rich gene promoter may interact with SP1 or other GC box binding proteins in uterine

endometrium, the distal upstream region may also be involved in the regulation of this gene.

However, the potential effects of this distal region on transcriptional activity of the IGFBP-2

gene have not been previously studied in any species.

In the previous chapter, the structure of the porcine IGFBP-2 gene was characterized.

This gene was shown to contain four exons and three introns spanning -29 kb and encoding

for a 316 amino acid IGFBP-2 precursor protein. Two clusters of transcriptional start sites

were identified at about 100 bp upstream of the ATG initiation codon in exon 1 and

approximately 1.4 kb of the 5'-flanking region was sequenced. The present study has

extended these structural findings by focussing on the identification of functional upstream


regions) and corresponding cis- and trans-elements within this region that may confer, in

part, the high level endometrial expression of the IGFBP-2 gene during pregnancy.

Materials and Methods

Reagents and Enzymes

All reagents used were of molecular biology grade. Oligonucleotides were

synthesized by the DNA Synthesis Core Laboratory of the Interdisciplinary Center for

Biotechnology Research (ICBR) at the University of Florida. Restriction enzymes and

poly(dI-dC) were purchased from Boehringer-Mannheim Biochemicals (Indianapolis, IN);

Radionucleotides were purchased from ICN Pharmaceuticals, Inc. (Irvine, CA); RPMI 1640

medium, Hanks' Balanced Salt Solution and Pancreatin were purchased from Gibco BRL,

Life Technologies, Inc. (Grand island, NY); DME/F-12 (1:1 mixture), Calf thymus DNA,

fetal bovine serum (FBS) and ABAM (Antibiotic-Antimycotic) were purchased from Sigma

Chemical Co. (St. Louis, MO).

Plasmid Construction

The plasmid H6 (see Chapter 3) was cleaved with Notl (at nucleotide position +73

relative to the translational initiation codon). The adhesive ends were made blunt-ended by

treatment with the Klenow fragment of DNA polymerase I and ligated to a synthetic Hindl

linker (Promega, Corp., Madison, WI) and the DNA cleaved with HindII. A fragment,

spanning from -1389 to +73 relative to the ATG, and also containing 3 bp of linker sequence

on the 3'-end, was isolated from an agarose gel and cloned in the pGL2-enhancer vector

(Promega, Corp., Madison, WI) at the HindIll site to construct the plasmid Hind-LUCe.

Hind-LUCe was cleaved with SacI, BglII or Smal. The large fragments containing the


upstream regions from -874, -764, or -305 to +73, respectively and the linked vectors were

isolated from agarose gels and religated to create Sac-LUCe, Bgl-LUCe, and Sma-LUCe,

respectively. Sac-LUCe was cleaved with Sad and BglII, a fragment of 110 bp (spanning

from -874 to -765) was isolated from an agarose gel using DEAE paper, and inserted in the

sense orientation to the pGL2-promoter vector (Promega, Corp., Madison, WI) at the unique

SacI and BglI sites to create SaB (110)-pLUC. Bgl-LUCe was cleaved with BglI and Smal,

a fragment of 459 bp spanning from -764 to -306 was isolated from an agarose gel using

DEAE paper, and this was inserted in an antisense orientation in pGL2-promoter vector to

construct SmB (459)-pLUC. PCR products from -874 to -765, generated with a 5'-primer


GCAG-3', were treated with SI nuclease and inserted into the pGL2-promoter vector at the

unique Smal site. Ten clones were subjected to sequence analysis and two of these were

found to have inserts. One contained an insert of 42 bp fragment from -829 to -788 in the

antisense orientation and was named 42(+)Sma-LUCe. The other contained an insert of 25

bp from -812 to-788 in the sense orientation, and was named 25(-)Sma-LUCe. All DNA

constructs were sequenced to confirm orientation and identity of chimeric inserts. Plasmid

DNAs used for transfection were prepared using a DNA purification kit (Qaigen Inc.

Chatsworth, CA) or by the cesium chloride method (Sambrook et al., 1989).

Propagation of Primary Cell Cultures and Cell Lines

Gilts were monitored twice daily for onset of estrous activity. Gilts exhibiting two

consecutive estrous cycles of normal duration (18-22 days) were mated at estrus with boars

and again 12 and 24 h later. The day of onset of estrus was defined as Day 0 of pregnancy.


Animals were sacrificed at the University abattoir on the indicated days of pregnancy.

Reproductive tracts were removed, immersed in ice and trimmed from the mesometrium.

Endometria, myometria and placentae were obtained by dissection as previously described

(Simmen et al., 1990b; Ko et al., 1994a). Animal use protocols were approved by the

University of Florida Institutional Animal Care and Use Committee.

Primary cultures of uterine cells were obtained as described previously (Zhang et al.,

1994; Reed et al. 1996). Briefly, the uterus was removed and flushed with Phosphate-

Buffered Saline (PBS) containing 1% ABAM (Antibiotic Antimycotic: 500 U/ml penicillin,

500 pg/ml streptomycin, 25 lg/ml amphotericin B). Endometrium was obtained by

dissection and washed three times with Hank's Balanced Salt Solution (HBSS) containing

1% ABAM and 50 lg/ml gentamicin. The tissue was incubated in digestion solution I (4.8

mg/ml dispase and 12.5 mg/ml pancreatin in HBSS) for 2.5 hours at room temperature. The

tissue was then washed six times in HBSS containing 1% ABAM to obtain luminal epithelial

cells. The residual tissue was incubated in digestion solution II (0.04% trypsin, 0.06%

collagenase, 0.01% DNase I in HBSS) for 50 minutes at 37 C to dissociate into a single cell

suspension. The cell suspension was separated into glandular epithelial and stromal cells by

passage through a 38 im stainless steel sieve and separated cells were seeded in 35 mm well,

6-well plates in RPMI 1640 medium containing 10% FBS and 1% ABAM. Medium was

changed every 2 to 3 days until cultures were ready to be transfected.

The human endometrial cell line ECC-1 was propagated in RPMI 1640 medium

containing 10 % FBS, 1 % ABAM, 25 pg/ml transferring, 6 mM glutamine, 10 ng/ml cholera

toxin, 0.11 % Na bicarbonate, 20 mM Hepes, and 200 mg/L glucose. The human


trophoblastic choriocarcinoma cell line JEG-3 was propagated in DME/F-12 medium

containing 1 % ABAM and 10 % FBS.

Transfection of Primary Cell Cultures and Cell Lines

Cells were detached from plates by treatment with trypsin and collected by

centrifugation at 2000 rpm for 3 minutes. The cell pellet was resuspended in transfection

buffer [Hepes Buffered Solution (HBS): 21 mM Hepes, pH 7.05, 137 mM NaCI, 5 mM KC1,

0.7 mM Na2HPO4, 6 mM Glucose)]. Cells were transfected with luciferase reporter plasmids

(20 gg) by electroporation (Cell-Porator, GIBCO BRL, Life Technologies, Inc., setting at

200-250 voltage and 330 giF) at room temperature. The transfection efficiency was

normalized for total protein concentration in cell extracts or for pCAT activity which was

determined by co-transfection of pCAT plasmid (10 pg)(Promega, Corp., Madison, WI)

with the Luciferase reporter plasmids. Forty-eight hours after transfection, cells were washed

twice with PBS, and lysed in 200 tl of cell lysis buffer (Promega, Corp., Madison, WI). One

hundred [l of cell lysate was saved for protein or CAT assay. Another 100 il was

centrifuged at 12000 rpm for 5 min. Thirty ul of supernatant was used to determine luciferase

activity with luciferase substrate (Promega Corp., Madison, WI) for 3 min in a luminometer

(AutoLumat, LB953, EG&G Berthold). Protein concentrations were determined by the

Bradford method using bovine serum albumin as standard. For CAT assays, 100 i.l of lysate

was incubated at 65 C for 10 min and centrifuged at 12000 rpm for 10 min. Eighty p1 of

supernatant was adjusted to a volume of 125 ui of 0.25 mM Tris-HC1, pH 8.0, 0.08 mCi 4C

chloramphenicol and 25 gg of n-butyryl CoA, and incubated at 37 C overnight. After one


extraction with mixed xylenes and one back extraction with 0.25 mM Tris-HC1, 200 gl of

xylenes containing the radiolabeled CoA was counted in a liquid scintillation counter.

Transfection data were analyzed by least squares analysis of variance using the

General Linear Models procedure of the Statistical Analysis System (SAS; Barr et al. 1979).

In the model, the effects of experiment or pig (as random effect), constructs and their

interactions on luciferase activities were considered. Protein concentration and/or pCAT

activity were used as covariates to normalize the luciferase activity. Differences between

constructs were determined using orthogonal contrasts.

Preparation and Fractionation of Nuclear Extracts

Nuclear extracts were prepared as described previously (Gonzalez et al. 1994) with

some minor modifications. Briefly, 2 g of frozen tissue was minced in a Waring Blender

using 20 ml of homogenization buffer (10 mM Hepes, pH 7.6, 25 mM KCI, 1 mM EDTA,

1.5 M sucrose, 10% glycerol) supplemented with 0.15 mM spermine, 0.5 mM spermidine,

4 ug/ml aprotinin, 0.1 mM phenylmethlsulfonyl fluoride (PMSF) and 0.2 mg/ml trypsin

inhibitor. The homogenate was transferred to a Dounce homogenizer on ice and further

disrupted by three strokes. Nuclei were centrifuged twice through 10-ml sucrose cushions

(30,000 X g for 30 min). Nuclear pellets were resuspended in a 6 X volume of nuclear lysis

buffer [10 mM Hepes, pH 7.6, 100 mM KCI, 3 mM MgCI,, 0.1 mM EDTA, 1 mM

dithiothreitol (DTT), 0.1 mM phenylmethlsulfonyl fluoride (PMSF), 10% glycerol], and 4

M ammonium sulfate was added dropwise with stirring to a final concentration of 0.3 M,

followed by gentle mixing for 30 min at 4 C using a rocking platform. Extracts were

centrifuged at 150,000 X g, for 1 hour at 4 C. Ammonium sulphate (0.3 g/ml of extract)


was then added to the supernatant with gentle stirring. The nuclear protein precipitate was

pelleted by centrifugation at 50,000 X g for 15 min and was dialyzed against buffer (25 mM

Hepes pH 7.6,40 mM KCL, 0.1 mM EDTA, 1 mM DTT, 10% glycerol) overnight. Protein

concentrations were determined by the Bradford method using bovine serum albumin as

standard. Nuclear extracts were subdivided into 25 ug aliquots and stored at -80 OC until use.

Fractionation of nuclear extracts followed the procedure described by Gonzalez et al.

(1994). Nuclear extracts were loaded on a pre-equilibrated (100 mM NaCI in dialysis buffer)

DEAE BioGel A column (Bio-Rad, Richmond, CA) at 4 OC and eluted with a linear salt

gradient (0.1 to 1.0 M NaCI) in dialysis buffer. Fractions were collected and the protein

concentration of each fraction was measured. The flow-through fractions were pooled and

designated as Fraction 1.The eluted fractions were separated and pooled as Fractions 2 to 6.

Gel Retardation Assay (GRA)

The procedure used for the gel retardation assay was modified slightly from that

described by Gonzalez et al. (1994). Briefly, 5 to 20 jig of crude or fractionated nuclear

proteins were preincubated in binding buffer (10 mM Tris-HC1, pH 7.4, 60 mM KC1, 1 mM

MgCI2, 0.5 mM EDTA, 0.5 mM DTT, and 10% glycerol) also containing 4 ug of sonicated

calf thymus DNA and 2 ug of poly(dl-dC) with or without unlabeled competitor for 15 min

on ice. DNA probes ["P-labeled (1.0-1.5 X 10' cpm) restriction enzyme-generated DNA

fragments or synthetic oligonucleotides] were added to the binding mixture and incubated

on ice for an additional 30 min. Binding complexes were monitored in low-ionic strength,

4 or 6 % polyacrylamide gels (acrylamide/bis-acrylamide = 29:1) with TAE buffer (10 mM


Tris-HC1, pH 7.4, 50 mM HoAc, 1 mM EDTA) and electrophoresed at room temperature.

Gels were dried and exposed to X-ray film for overnight to 4 days.

All double-stranded oligonucleotides, except for probes A and C, were generated

by annealing complementary oligodeoxynucleotides in annealing buffer (40 mM Tris-HC1,

pH 7.5, 20 mM MgCl,, 50 mM NaCI) first by heating at 65 OC for 20 min, and then

incubating at 37 C for 20 min followed by placement at room temperature for 20 min. Probe

A was generated by annealing of an 18 nt primer (SB5) and a 40 bp oligonucleotide, of

which the 3'-end was the complement of the SB5 primer. The hybrid was made double-

stranded using the Klenow fragment of DNA polymerase. Similarly, Probe C was generated

from the SB3 primer and a 40 nt oligonucleotide, of which the 3'-end was the complement

of SB3.

UV Cross-Linking

Standard procedures for UV cross-linking as previously described by Williams and

Konigsberg (1991) were used. The 32P-end labeled A2 probe was incubated with nuclear

extracts in the same buffer used for GRA by placement on ice for 30 min. The final reaction

volume was 15 ul. The binding reactions were exposed to UV light (254 nm, using a hand-

held short wave UV light box, 115 V, 60 Hz, 0.16 A) at a distance of 10 cm for 5 min. The

cross-linked products were examined by SDS-PAGE. Gels were dried and exposed to X-ray

film autoradiographyy).

Southwestern Blot

A protocol modified from that described by Harrison et al. (1991) was used. A total

of 50 to 200 ug nuclear protein or fractions of nuclear extracts were treated with 3-


mercaptoethanol and subjected to SDS-PAGE using a Bio-Rad mini gel apparatus. The

proteins in the gels were electrotransfered to nitrocellulose filters. The membranes were

incubated in TEN buffer (10 mM Tris-HCI pH, 7.4, 1 mM EDTA, 60 mM NaCI) containing

5 % Carnation non-fat dry milk by gently rocking at room temperature for 20 min. After

blocking, the membrane was washed twice with TEN and then incubated in TEN containing

sonicated calf thymus DNA (5 pgiml) and the presence or absence of unlabeled competitor

DNA for 15 min. Then, radiolabeled probe (A2; 7 X 106 cpm/ml) was added, and

incubation at 4 OC continued for 24 hours. After two 5 min washes in TEN, the membrane

was exposed to X-ray film.


Identification of Functional Upstream Regulatory Regions of the IGFBP-2 Gene

In cancer cell lines. As an initial step for the identification of functional upstream

regulatory regions, the promoter constructs Hind-LUCe (-1397/+73), Sac-LUCe (-874/+73),

Bgl-LUCe (-764/+73) and Sma-LUCe (-305/+73) (Figure 4-1) were separately transfected

into the human endometrial carcinoma cell line ECC-1 (known to express the endogenous

IGFBP-2 gene) and the human trophoblastic cell line JEG-3 (Figure 4-2). Data from these

transient transfections showed that deletion from -874 to -765 decreased luciferase activity

(2 to 2.5 fold), while deletion from -764 to -306 increased luciferase activity in both cell

lines. However, in Ecc-1 but not JEG-3 cells, deletion from -1397 to -875 increased reporter

gene activity. These data therefore indicate that the region from -874 to -765 may have

positive effects on promoter activity, whereas regions from -1397 to -875 (in ECC-1), and

from -764 to -306 (in both cell types), may have transcriptional inhibitory effects.

Hindlll Sad Bglll Smal Notl
-139 lUC e Hind-LUCe

-874 LUCCSV4e Sac-LUCe

-764 iUCi SV4e BgI-LUCe

-305 1UC~ S Sma-LUCe


LUC SaB(110)-pLUC



Figure 4-1. DNA constructs containing the upstream region of the IGFBP-2 gene fused to
the luciferase reporter gene. The top line presents the restriction endonuclease map of the
upstream region of the pig IGFBP-2 gene. The number for each construct designates the 5'
end of the fragment relative to the translational initiation codon (ATG) of IGFBP-2 (+1). The
open boxes represent the luciferase coding sequence (LUC). The reporter genes (pGL2-
enhancer) with SV40 enhancer (SV40e) and without SV40 promoter are labeled as LUCe;
the reporter genes (pGL2-promoter) with SV40 promoter and without enhancer are labeled
pLUC. Hind, Sac, Bgl, and Sma represent cleavage sites for HindIlI, SacI, Bglll and Smal


S Ecc-1 JEG-3

H Sa B Sm H Sa B Sm

D18 GE D30 GE

H Sa B Sm H S. B Sm

D18 ST D30 ST
o 10.0 -


oo H Sa B Sm H S B S.

Figure 4-2. Reporter gene activities of the Hind-LUCe, Sac-LUCe, Bgl-LUCe and Sma-
LUCe plasmids transfected in different cell lines and uterine cell types. Each bar represents
the Least Squares Mean (n=9) of triplicate data in arbitrary light units (normalized for protein
concentration of cell extract) from three experiments (n=3 pigs). Hind-LUCe, Sac-LUCe,
Bgl-LUCe and Sma-LUCe are labeled as H, Sa, B and Sm, respectively. The open bars
represent the data from transfected cancer cell lines. In ECC-1 cells, Sa is significantly higher
than H (P<0.05) and B (P<0.01), and Sm is significantly higher than B (P<0.05). In JEG-3
cells, B is significantly lower than both Sa (P<0.05) and Sm (P<0.05). The black bars
represent the data from transfected uterine endometrial glandular epithelial cells of pigs. In
transfected glandular epithelial cells from the endometrium of Days 18 (D 18 GE) and Day
30 (D30 GE) pregnant pigs, no statistically significant difference was found between
constructs. The stippled bars represent the data from transfected uterine endometrial stromal
cells. In stromal cells from endometria of Days 18 (D18 ST) pregnant pigs, Sm is
significantly higher than B (P<0.05). In stromal cells from endometrium of Day 30 (D30 ST)
pregnant pigs, Sa is significantly higher than B (P<0.05).


In primary cultures of uterine endometrial cells. Continuous porcine endometrial cell

lines are not currently available. Therefore, the above constructs were transfected into

primary cultures of uterine endometrial glandular epithelial (GE) and endometrial stromal

(ST) cells isolated from endometrium of Day 18 and Day 30 pregnant pigs. Results from

DNA transfections of the two uterine cell types representing the two stages of pregnancy are

shown in Figure 4-2. Similar to the results from, deletion of sequence from -874 to -765

decreased luciferase activity, whereas deletion from -1397 to -875 increased luciferase

activity. However, effects of deletion from -764 to -306 were observed differentially in

transfected D18 GE, D30 GE, D18 ST and D30 ST cells. GE cells from Day 18 pregnant pigs

yielded the highest overall IGFBP-2 promoter activity. Therefore, D18 GE cells were

utilized for subsequent experiments.

Although the above data for the pig endometrial cells suggested that the 110 bp

region from -874 to -765 may have an inducing or activating function, the relatively large

variation in the data may have diminished the overall statistical significance of this effect.

Therefore, the constructs Sac-LUCe, Bgl-LUCe and LUCe ( promoterless control) were

cotransfected with pCAT into GE cells from Day 18 pregnant pigs (to correct for variability

of transfection efficiency). The normalized data clearly showed that deletion from -874 to

-765 decreased the reporter gene activity by 75%, again suggesting that this region is

stimulatory for IGFBP-2 promoter activity in these cells (Figure 4-3).

Enhancing Activity of the 110 bp Region Extending from -874 to -765

To examine the activity of the 110 bp region within the context of a heterologous

promoter, the construct SaB(l 10)-pLUC was generated by inserting the DNA fragment from

12 3000

10 2500

S8- 2000

6 1500

4 |1000 -

2 500

0 0

Figure 4-3. Transcriptional activation by the 110 bp upstream region. Bars represent the
Least Squares Mean (n=9) (normalized by assay ofcotransfected pCAT) of triplicate plates
for each of three experiments (n=3 pigs). Left (black bars). Sac-LUCe had 4-fold higher
activity than Bgl-LUCe (p<0.05), whereas there was no difference (p-0.53) between Bgl-
LUCe and LUCe, in transfected endometrial glandular epithelial cells of Day 18 pregnant
pigs. Right (open bars). SaB( 10)-pLUC had 2 to 5-fold higher activity (p<0.05) than
SmB(459)-pLUC and pLUC, respectively, whereas there was no difference (p=0.16) between
SmB(459)-pLUC and pLUC, in transfected endometrial glandular epithelial cells from Day
18 pregnant pigs.


-874 to -765 into the pGL2-promoter vector in the sense orientation (Figure 4-1). This DNA

was cotransfected with the pCAT plasmid (control for transfection efficiency) into primary

cultures of GE cells from Day 18 pregnant pig endometrium. As controls, the SmB(459)-

pLUC construct, which was generated by inserting the DNA fragment from -764 to -306 into

the pGL2-promoter vector in the antisense orientation (Figure 4-1), and the pGL2-promoter

plasmid (pLUC) were also transfected. These results (Figure 4-3) indicated that the 110 bp

fragment increased SV40 promoter activity by approximately 5-fold, whereas the 459 bp

fragment did not significantly increase the SV40 promoter's activity, in agreement with the

previous results in Figure 4-2. These data supported the idea that positively-active cis-acting

elements may exist within the 110 bp region and also showed that the activating function

may not be promoter-specific.

Identification of Potential cis-Elements within the 110 bp Activating Region

As an initial step to examine for the binding of nuclear proteins to the 110 bp region,

the DNA fragment generated by cleavage of Sac-LUCe with Sacl + BglI (Figure 4-4) was

used in gel retardation assay of nuclear extract from endometrium of a Day 18 pregnant pig.

This preliminary experiment confirmed that this fragment specifically bound nuclear proteins

(Figure 4-5). Because of the relatively large size of this fragment, the binding complex

theoretically could contain multiple proteins. Therefore to clarify the nature of this binding,

the 110 bp region was subdivided into the A (40 bp), B (30 bp) and C (40 bp) subregions and

used in GRA (Figure 4-4). Surprisingly, all three subregions bound to endometrial nuclear

protein(s) with different patterns and multiple bands observed (Figure 4-5).

(40 bp) (30 bp) I (40 bp)

Al A3 B1 B3 C1 C3
A2 B2 \ C2

/ \ / -.
I \ / \ 2 "
1Al r- A3\ / B2- rC2



Figure 4-4. Map of the GRA probes derived from the 110 bp region. The top thick line
represents the 110 bp fragment and the relative positions of the A, B and C probes. Probes
Al, A2, A3, C1, C2, and C3 are 18 bp long. A2 has 7 bp on each end overlapping with Al
and A3. C2 has 7 bp on each end overlapping with C1 and C3. The probes B1, B2, and B3
are 15 bp long. B2 has 6 and 7 bp overlapping with Bl and B3, respectively. The A2 and B3
common sequence is boxed. The half progesterone receptor binding element (half site) is
indicated as /2 PRE. The leukemia virus factor c (Lvc) binding element (Speck et al., 1987)
is indicated. The probes BC, Cl' and Cl'M have overlapping sequences with B3 and C1. BC
contains a palindromic sequence with a 3 bp spacer. Cl' contain 4 bp of B3 and 11 bp of C1.
CI'M has only one bp different from C1' at the 5th position from the 5' end. Three other
oligonucleotide probes corresponding to TCAGGG-containing subregions (designated S1,
S2 and S3) outside of the 110 bp region and flanking the IGFBP-2 promoter, are shown on
the bottom. The location of each probe relative to the ATG translational initiation codon is




.~ iI

1 2 3 4


D18E D18E


- ~~i~1i

5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Figure 4-5. Uterine endometrial nuclear proteins bind the 110 bp fragment, and the individual A, B and C subregion probes. Using
the gel retardation assay as described in Material and Methods, each probe was incubated with no nuclear extract (lanes 5, 10, and
15), with 10 tg of nuclear extract (lanes 1, 6, 11 and 16) or with 10 |tg of nuclear extract preincubated with 50 X (lanes 2, 7, 12
and 17), 100 X (lanes 3, 8, 13 and 18) or 150 X (lanes 4, 9, 14 and 19) of a molar excess of unlabeled homologous competitor DNA,


In order to localize the individual protein-binding DNA elements, three overlapping

double-stranded oligonucleotide probes within each of the A, B and C subregions were

designed and synthesized (Figure 4-4). These probes were then used in GRA. Probes A2 and

A3 competed with fragment A for binding to nuclear proteins; similarly probe B1 competed

with fragment B; and probe C1 competed with fragment C (Figure 4-6). When all nine

probes were individually labeled and used in GRA. the B2 and C3 probes were little or not

shifted and therefore were eliminated from further study (Figure 4-7). Surprisingly,

competition assays with the seven remaining probes showed that only the complex formed

from the A2 probe and the largest complex formed from the B3 probe were completely

inhibited by inclusion of excess unlabeled A2 and B3 oligonucleotides, respectively. All

other complexes were not inhibited by excess unlabeled homologous competitor

oligonucleotides and therefore were considered to represent nonspecific binding (Figure 4-

8). The complexes identified with the A2 and B3 probes appear to have the same size as

judged by co-migration in gels, suggesting that these two probes may bind the same protein.

When the sequences of these two probes were closely examined, a common sequence 5'-

TCAGGG-3' was found (reverse orientation in B3). To examine whether this common

sequence in the A2 and B3 probes binds the same nuclear protein, 32P-labeled A2 was used

as probe and all nine unlabeled oligonucleotide probes were separately used as competitors

in GRA. As predicted, only the A2 and B3 probes exhibited competition with labeled A2

(Figure 4-9), again suggesting that A2 and B3 bind the same nuclear protein herein

designated as "A2 binding protein". Since A2 has 7 bp sequences overlapping with Al and



D18E D18E

- Al A2 A3 B1 B2 B3 C1 C2 C3

1 2 3 4 5

6 7 8 9 10 11 12 13 14 15

Figure 4-6. Localization of nuclear protein-binding sites within A, B and C subregions of the 110 bp fragment. Competitions of
unlabeled Al, A2 and A3 oligonucleotides with A; Unlabeled B1, B2 and B3 oligonucleotides with B; and C1, C2, C3
oligonucleotides with C are shown. A 100 X molar excess of each competitor was preincubated with 10 pg of nuclear extract for
15 minutes prior to addition of the radiolabeled probe.

Probe Al A2 A3 B1 B2 B3 C1 C2 C3


- + -+ -+ -+ -+ -+ -+ -+ +

Figure 4-7. Gel retardation assay involving nine overlapping probes. Each probe was incubated without or with endometrial nuclear
extract from a Day 18 pregnant pig. The gel for Al, A2 and A3 was exposed to X-ray film for 4 days, while the gel for B1, B2, B3,
C 1, C2 and C3 was exposed for 2 days.

Probe Al A2
Extract + + + +
Competitor Al A2

# <

A3 B1 B3 C1 C2
- ++ ++ ++ -++ ++
- A3 81 B3 C1 C2

)I Iww

1 2 3 4 5 6
1 2 3 4 5 6

7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Figure 4-8. Localization of specific and non-specific protein-binding elements using GRA. Each probe was incubated without
nuclear extract, with nuclear extract, or with nuclear extract which was preincubated with a 100 X molar excess of unlabed
competitor oligonucleotide. The nuclear extract used in this experiment was from endometrium of a Day 18 pregnant pig.

Competitor -

D18 Endometrium
-Al A2 A3 B1 B2 B3 C1 C2 C3

6W so


1 2 3 4 5 6 7 8 9 10 11


D12 Ovary

- A2 B3 C1 C1' CGt' BC S1 S2 S3

1 2 3 4 5 6 7 8 9 1011

Figure 4-9. Proteins of the same molecular weight bind a common sequence within the A2
and B3 oligonucleotides. In both the top and bottom panels, a 50-fold molar excess of each
unlabeled probe was used as competitor for binding of the A2 probe and nuclear extract
prepared from the indicated tissue. Competitors used were defined in Figure 4-4.


A3 at both ends and B3 has 7 bp sequences overlapping with B2 at the 5'-end (Figure 4-4),

the lack of competition by Al, A3 and B2 of A2-DNA binding suggests that the sequence

5'-TCAGGG-3, overlapping the middle of A2 and the 3'-end of B3 is, in all likelihood, the

DNA binding element.

Using one base mismatch patterns at every position of the TCAGGG sequence to

search the entire 110 bp region's sequence, two similar sequences were identified. One

sequence, TCtGGG, is in the overlapping region of C1 and C2. The other sequence,

gCAGGG, is at the 5'-end of C1. Since neither Cl nor C2 competed with A2 for binding, the

sequence TCtGGG is judged to be non-binding, thereby suggesting that the A at the third

position may be important for this interaction. However, the relative importance of the T at

the first position of TCAGGG could not be inferred because this sequence is at the very end

of probe Cl. In this regard, it is possible that binding of proteins to this element requires

flanking nucleotides. Moreover, a half PRE resides adjacent to this element (Figure 4-4).

It is therefore possible that progesterone receptor interrupts the binding of this element to

other nuclear proteins. The sequences CCCTGA in B3 and gCAGGG in Cl form a

palindromic sequence 5'-CCCTGnnnCAGGG-3' with 3 bp spacer. To examine whether or

not the gCAGGG element binds to the same protein as does A2, a new probe, C containing

4 bp at the 5' end of gCAGGG and 5 bp at the 3'-end but missing the half PRE (Figure 4-4),

was designed and used as a potential competitor of A2 in GRA. Cl' had little or no

competition with A2 (Figure 4-9). In addition "P-labeled C was not shifted when incubated

with nuclear extract (data not shown). These data therefore suggest that gCAGGG and its

surrounding sequence has little or no interaction with A2 binding protein or other proteins,


such as Leukemia virus factor c (Lvc) (Figure 4-4) (Speck et al., 1987). The CI'M probe,

which had the same sequence except for a 1 bp mutation from G to T (Figure 4-4) exhibited

some competition with A2, but this was not as strong as for B3 and A2 (Figure 4-9). Thus

the T at the first position of TCAGGG may be important for interaction with A2 binding

protein. Moreover, the probe BC, which has both CCCTGA and gCAGGG competed for

binding with A2 and exhibited the same size complex in GRA (Figure 4-10), again

suggesting that the sequence CCCTGA (the reverse sequence of TCAGGG) is the only

sequence interacting with the A2 binding protein.

Using computer-assisted sequence analysis, the sequence TCAGGG was found at 3

other positions in the sequenced 1.4 kb upstream region of the IGFBP-2 gene (Figure 4-4).

Oligonucleotide probes corresponding to each of these regions were used as competitors in

GRA. Only S2 and S3 competed with A2 but with differing affinities, whereas S1 exhibited

little or no competition with A2 (Figure 4-9), suggesting that not only the TCAGGG but also

the flanking nucleotides can affect interaction with the A2 binding protein.

Characterization of the A2 Binding Protein

In order to examine the tissue and possible stage of pregnancy-dependence of the A2

binding protein, nuclear extracts from endometria of Days 12, 18, 21, 60 and 90 pregnant

pigs, placenta of a Day 90 pregnant pig and ovary of a Day 12 pregnant pig were used in

GRA analysis. Data from this experiment showed that the A2 binding protein is present in

all tissues examined (Figure 4-10). However, the relative binding in early pregnant pig

endometrium was greater than in mid- or later-pregnant endometrium. The binding in the

Day 12 ovary appeared to be the stronger. The 3P-labeled A2 probe also was shifted by

Probe A2
Extract "
Competitor -- ----


Extract D12 E
Fraction 1 2 3 4 5


D 60E >
6 1 2 3 4 5 6

win. 4

Figure 4-10. Characterization of the A2 element DNA binding protein. Panel A, nuclear
extracts from Day 12 (D12E), Day 18 (D18E), Day 21 (D21E), Day 60 (D60E) and Day 90
(D90E) endometrium, Day 90 placenta (D90P), and Day 12 ovary (D12V) were used in GRA
with radiolabeled A2 and BC probes. The right-most lane demonstrate the competition by
a 100 X molar excess of unlabeled BC with the 32P-labeled BC probe. Note that BC and A2
have the same size complex. Panel B, nuclear extracts from Day 12 and Day 60 endometria
were fractionated on DEAE Bio-Gel A columns (see Materials and Methods). Fraction 1 is
the flow-through prior to the application of the salt gradient. Fractions 2 to 6 were eluted in
a salt gradient (0.1 M to 1.0 M). The nuclear extract from ovary of a Day 12 pregnant pig
(D12V) was used as positive control. Probe A2 was used in all lanes of this GRA.


nuclear proteins from the (human hepatocarcinoma) HepG2 cell line with two smaller

complexes apparent (data not shown). These same complexes were also observed in GRA

analysis of some endometrial nuclear extract preparations.

When nuclear extracts from Day 12 and Day 60 endometria were fractionated on

DEAE BioGel A columns and analyzed by GRA, two observations were made. First, A2

binding protein was present in Fraction 2 of Day 12 endometrial nuclear extract, whereas

A2 binding protein was in Fraction 3 of the Day 60 endometrial nuclear extract (Figure 4-

10). This delay in elution may indicate that the A2 binding protein in mid-pregnancy

endometrium has more net negative charge than the corresponding protein in Day 12

endometrium. This increase in negative charge is possibly due to protein phosphorylation.

Second, smaller A2 binding complexes were observed in some fractions of both the D12 and

Day 60 endometrial nuclear extracts. However, these smaller complexes were most abundant

in Fraction 3 of the Day 60 endometrial nuclear extract.

In order to characterize the A2 binding protein as to its molecular weight, UV cross-

linking was employed. "P-labeled A2 oligonucleotide and nuclear proteins from ovary of

Day 12 and endometrium of Day 18 of pregnancy were utilized in this experiment. Three

cross linked products with sizes of 35 to 40 kDa were observed (Figure 4-11A). Similarly,

Southwestern blotting (Figure 4-11B) showed that the A2 probe bound to a 33 kDa protein

in nuclear extract from Day 12 ovary, Day 90 placenta, and in Fraction 2 of Day 12

endometrial extract. The specificity of this binding was also confirmed using competitive

Southwestern blot with Fraction 2 of Day 12 endometrial nuclear extract (Figure 4-11C).


2M 2 D12E-F2
> wi > ML ,' ,' 5
O )0 0C 0 0

S- 200 kDa
-200 kDa "
S4 -118 --200 kDa
-118 78
-118 118
78 47 78
31 47
'-.47 i6 -- 47
lii -19
-31 m' 31
8 26
-26 19

Figure 4-11. Size determination of the A2 DNA binding protein. A, UV cross-linking.
Nuclear proteins from D12 ovary (D 2V, 20 pg) and from Day 18 endometrium (D18E, 10
pig) of pregnant pig, were used in the assay as described in Materials and Methods. B.
Southwestern blot. Nuclear proteins from Day 12 ovary (D12V, 100 Ig), Day 90 placenta
(D90P, 50 ig), Fraction 2 from Day 12 endometrium (D12E-F2, 200 pg) and Fraction 3 from
Day 60 endometrium (D60E, 100 jig) of pregnant pigs were used in the assay as described
in Materials and Methods. C. Competitive Southwestern blot. Nuclear proteins in Fraction
2 from Day 12 endometrium (D12E-F2, 175 jig) was used for each lane. The left lane was
preincubated with a 50 x molar excess of cold A2 probe. Protein size markers are indicated
on the right side of each gel.


However, in Fraction 3 of the Day 60 nuclear extract, the A2 probe bound to a 13 kDa

protein (Figure 4-11B).


Although the 5' flanking sequence of the IGFBP-2 gene has been isolated for the

human (635 bp, Binkert et al., 1992), rat (1290 bp, Kutoh et al., 1993), murine (1980 bp,

Landwehr et al., 1993), chicken (661 bp, Schoen et al., 1995) and porcine (1397 bp, Chapter

2), few studies on the functions of this region have been conducted. Schwander and

colleagues, took the -579 to +1 (ATG, +1) fragment of the rat IGFBP-2 gene, and using

deletion and transfection analysis followed by gel retardation assay demonstrated the

presence of three GC boxes that bind transcription factor SPI and display cooperativity in

transcriptional assays (Kutoh et al., 1993). Similarly, Rechler and colleagues also

demonstrated that these three SPI sites are required for efficient transcriptional initiation of

the rat IGFBP-2 gene (Boisclair et al., 1993). Recently, Schwander and colleagues showed

that cell density can affect IGFBP-2 gene expression and similarly affects the promoter

activity of a 1.5 kb flanking fragment of the rat IGFBP-2 gene (Kutoh et al., 1995). All of

these studies mainly used rat BRL-3A liver cells. To date no comparable studies have been

conducted on other species or tissues. In this regard, SP1 gene and protein expression in

pregnant pig endometrium is constitutively low and is not correlated with the temporal

pattern of endometrial expression of IGFBP-2 gene (Wang, Y. and Simmen, R.C.M.,

unpublished observations). Moreover, potential distal sequence contributions to the

transcriptional regulation of this gene, in concert with or independent of SP1, were not

previously studied.


As an extension of previous work on the tissue and developmental expression and

characterization ofIGFBP-2 gene, the present study examined the distal region flanking the

IGFBP-2 promoter for transcription regulation. Transient transfection assays using two

cancer cell lines (ECC-I and JEG-3) and primary cultures of pig endometrial cells (GE and

ST) at two different pregnancy stages, indicated that deletion from -874 to -765 (Sac-LUCe

vs Bgl-LUCe) decreased the reporter gene activity. Fusion of this region to the SV40

promoter, increased the promoter activity by 5-fold, clearly demonstrating that the 110 bp

region from -874 to -765 has transcriptional stimulatory activity. Moreover, this stimulatory

activity has also been observed in HepG2 cells (Badinga, L., Song, S., Simmen, F.A., and

Simmen, R.C.M., unpublished data). Thus, this stimulatory activity appears not to be cell

type- or developmental stage-specific. The position- and orientation- dependence of this

stimulatory activity has yet to be examined. However, results demonstrate that this

stimulatory activity is functional for two different promoters.

Transfection experiments also suggested that regions from -1397 to -875 and -764

to -306 have transcriptional repressing activities, which appear to be cell-type and/or

developmental stage-specific. In Ecc-1, JEG-3 and D18 ST cells, the region from -764 to -

306 (459 bp) exhibited a strong repressing activity (B vs Sm). This repressing activity was

also noted in D30GE cells, although this inhibitory activity was not statistically significant.

However, transfected D18GE and D30ST cells exhibited no repressive effect with the -764

to -306 construct. In DI8GE cell, this region did not decrease SV40 promoter activity again

suggesting no repressing affect in this cell type.The inhibitory activity of the region from -

1397 to -875 (523 bp) was manifested in ECC-I cells and in all pig endometrial cell cultures,


but was not observed in JEG-3 cells. Moreover, in HepG2 cells, this region has

transcriptional stimulatory activity (Badinga, L., Song, S., Simmen, F.A., and Simmen,

R.C.M., unpublished data). These observations taken together suggest that this 523 bp

region exhibits cell type-dependent regulatory activity. However, these two regions remain

to be further investigated and may prove useful for the eventual identification of tissue and

developmental regulatory cis- and trans-elements required for IGFBP-2 gene expression.

Transfection data from all cell types showed that the Sma-LUCe construct has

relatively high and constant reporter gene activity (about 5000 arbitrary luciferase light

units). This observation may indicate that the region from -305 to +73 covers the basal

promoter region, which is controlled by ubiquitous transcription factors, such as SPI (Faisst

and Meyer, 1992; Kutoh et al., 1993; Boisclair et al., 1993; Y. Wang and R.C.M. Simmer,

unpublished observations).

There were two reasons why primary cultures of endometrial cells were used for the

present study. First, there are no continuous pig endometrial cell lines available. Second,

primary cell cultures are perhaps closer to the normal physiological condition, which may

be important for transcriptional activity and regulation of the IGFBP-2 gene. The IGFBP-2

gene is highly expressed in Days 30 and 60 pregnant pig endometrium, and is expressed at

much lower levels in Day 12 pig endometrium (Simmen et al., 1992; Chapter 3). In this

regard, Day 60 GE cells may be the best cell type for identification of IGFBP-2 gene

regulatory regionss. However, Day 60 GE cells did not survive the electroporation process,

and consequently reporter gene activity was very low (data not shown). In the pig,

placentation initiates around Days 14-16 of pregnancy (Dantzer et al., 1985; Keys and King,


1990). At this time, signals from the embryo are thought to trigger the temporal expression

of a battery of uterine genes (Roberts et al., 1993; Simmen et al., 1995). Since IGFBP-2 gene

expression is induced by Day 18 of pregnancy, a similar embryo-maternal signaling

mechanism may be involved. In the present study, glandular epithelial and stromal cells

from both Day 18 and Day 30 of pregnancy were successfully used for transfections.

However, GE cells from Day 18 pregnant pig endometrium exhibited the highest reporter

gene activity after DNA transfection.

Using nine overlapping oligonucleotide probes, two potential cis-elements TCAGGG

and their common binding protein were identified within the 110 bp transcriptional inducing

region. Although mutation on each position for the consensus element has not been done,

two native mutants TcTGGG and gCAGGG in the 110 bp region did not show competition

with A2, suggesting that "G" at the first and "T" at the third position from the 5-end are

important for binding. This TCAGGG sequence was also found at other positions (S1, S2

and S3). However, different degree of competition with A2 were observed, suggesting that

flanking sequences at each end of this element are also important for protein binding.

It remains possible that there are other important cis-elements within this 110 bp

region. Although the competition study with nine oligonucleotide probes showed incomplete

self competition by all probes except for A2 and B3, partial self competitions by A3, B1 and

Cl oligonucleotides were observed. A3, BI and Cl also exhibited some competition with

A, B and C, respectively. These data may indicate that the binding affinities are low under

conditions presently used. Collaborative studies (Badinga, L., Song, S., Simmen, F.A., and

Simmen, R.C.M., unpublished data) have shown that in HepG2 human liver cells, A3 and

Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd