Isolation, cloning and characterization of the FSH beta gene promoter of the chinook salmon (oncorhynchus tshawytscha)

1. INTRODUCTION

1.1 The gonadotropins
The gonadotropins, follicle-stimulating hormone (FSH) and luteinizing hormone (LH), are
produced in the gonadotropes in the vertebrate pituitary gland. These gonadotropins are
heterodimers comprising a common α subunit and a unique, hormone specific β subunit (Pierce
and Parsons, 1981). After their synthesis and release, FSH and LH bind to specific receptors in the
testes and ovary to stimulate their growth and development, and to regulate their function. In
mammals, FSH is primarily involved in spermatogenesis in the male and folliculogenesis in the
female. In the female, it is also involved in the conversion of androgens to estrogens. LH controls
the synthesis of androgens as well as gamete release (spermiation in males and ovulation in
females). In some species, it may also be responsible for regulating the synthesis and subsequent
release of progesterone by the corpus luteum (Norris, 1997).

In the teleost fish, it was initially thought that only a primordial gonadotropin (GtH)
existed. However in 1979, Ng and Idler reported the presence of two types of gonadotropin in
plaice and flounder pituitaries. Subsequently, two different forms of GtH, GtH-I and GtH-II, were
isolated from salmon and shown to have different steroidogenic activities (Suzuki et al., 1988).
The studies of the amino acid (Itoh et al., 1988) and the cDNA sequences (Sekine et al., 1989)
revealed close similarity of GtH-I and GtH-II to the mammalian FSH and LH, respectively, and
thus established the duality of gonadotropins in the teleost fish. With additional information on the
GtH structures from numerous other fish species, the mammalian nomenclature has been adopted
(Quérat, 1994; 1995).

1


The roles of the pituitary gonadotropins in controlling gonadal function and subsequent
fertility in the teleosts are less clear than in mammals. Nevertheless, studies have shown that in the
salmonid species, the two gonadotropins have distinct roles in reproduction. It has been suggested
that FSH regulates the early stages of gonadal development, vitellogenesis and spermatogenesis,

while LH is responsible for stimulating the events leading to final oocyte maturation and ovulation
in females, and spermiation in males (Suzuki et al., 1988; Swanson et al., 1991; Gomez et al.,
1999). This is in agreement with the observation that FSH levels in the pituitary and plasma
increase with gonadal development and are also correlated to the levels of 11-ketotestosterone and
estradiol in the male and female coho salmon (Swanson et al., 1991). FSH levels decrease while
the LH levels begin to increase and reach a maximum during spawning (Swanson et al., 1991;
Slater et al., 1994; Figure 1). This marked elevation of LH is most probably responsible for the
single and synchronized spawning behavior of the salmonids. This is further supported by electron
microscopy studies which demonstrated the differential biosynthesis of FSH and LH during the
reproductive cycle of the rainbow trout (Naito et al., 1991).

2


Plasma levels of FSH and LH

LH

FSH

Mar

Apr

May

Jun

July

Aug

vitellogenesis

Sep

Oct

Nov

ovulation

Figure 1: Plasma levels of FSH and LH in coho salmon during reproductive
maturation (modified from Dickhoff and Swanson, 1990)

The expression and pulsatile secretion of these hormones are temporally regulated
throughout the reproductive cycle and are subject to the complex control of hypothalamic releasing
factors, paracrine factors from within the pituitary, and through the feedback actions of gonadal
hormones. The secretion of the hypothalamic factors in turn is also regulated by feedback actions
of the gonadal hormones (summarized in Figure 2).

3


Hypothalamus

+ve
GnRH

inhibin, activin, follistatin

- ve
+ve
-ve

Pituitary
-ve
or

+ve
Gonadal peptides
(inhibin, activin, follistatin)
- ve
+ve
-ve

+ve

LH
FSH

Gonads
Steriods
(T, E, P)

Figure 2: The regulation of gonadotropin gene expression in the hypothalamicpituitary-gonadal axis
Gonadotropin releasing hormone (GnRH), synthesized in and released from the hypothalamus, binds to
GnRH receptors on the surface of the pituitary gonadotrope. This leads to the synthesis and secretion of
LH and FSH which stimulate the production of steroid hormones in their gonads. Testosterone (T),
estrogen (E) and progesterone (P) negatively or positively regulate the synthesis of the gonadotropins
directly at the pituitary, or indirectly by modulating GnRH secretion from the hypothalamus. The

gonadal peptides, inhibin, activin and follistatin, also regulate of gonadotropin gene expression by
exerting positive or negative feedback (modified from Brown and McNeilly, 1999).

4


Unlike the mammalian FSH and LH, which are largely co-expressed in the same cells of
the mature pituitary, teleost FSH and LH are produced in distinct cells, in different locations within
the proximal pars distalis (PPD). FSHβ is produced in cells in close association with the
somatotrophs, surrounding the nerve ramifications, while the LHβ subunit is produced in the
peripheral regions of the PPD. The common α subunit, however, is produced in both types of cells
(Nozaki et al., 1990 and Naito et al., 1991).

1.2 Genomic organization of the gonadotropins
The gonadotropins are encoded by separate genes localized in different chromosomes
(Naylor et al., 1983; Julier et al., 1984). The genomic sequences encoding the LHβ and FSHβ
genes of several mammalian and a few teleost species have been elucidated, and analysis has
revealed that like the α subunit genes, the genomic organization of β subunit genes is generally
well conserved. Unlike the α subunit gene however, which is composed of four exons and three
introns (Fiddes and Goodman, 1979; Godine et al., 1982; Nilson et al., 1983; Gordon et al., 1988),
all the β subunit genes comprise three exons and two introns. Furthermore, the β genes are smaller
due to the considerably smaller introns, with a total length of about 1.5 kb for LHβ and 4 kb for
FSHβ (Figure 3). The positions of the β subunit introns are also conserved. The first intron
interrupts the region of either the signal peptide or the 5’ untranslated sequences whereas the
second intron is located three amino acids downstream from the fifth cysteine residue in the coding
region (Jameson et al., 1984; Talmadge et al., 1984; Virgin et al., 1985; Gharib et al., 1989; Ezashi
et al., 1990; Xiong et al., 1994a; Rosenfeld et al., 1997; Sohn et al., 1998) or the sixth cysteine
residue in the goldfish (Yoshiura et al., 1997). However, compared to the LHβ subunit gene, the
genomic structure of the FSHβ subunit gene is less conserved, particularly in regard to the size of
5

introns and 3’ untranslated region (3’ UTR; Figure 3). Compared to the 0.3 – 0.35 kb size of intron
1 of the LHβ genes, the size of the corresponding intron in the FSHβ genes is more variable,
ranging from 0.28 kb to 1.35 kb (Figure 3). Furthermore, the mammalian FSHβ is different from
other β subunit genes in that it possesses an extremely long 3’UTR that is over 1 kb, as seen in the
rat (1 kb), cow (1.2 kb) and human (1.5 kb). Within this 3’UTR, there are five highly conserved
segments (Gharib et al., 1989) whose significance is presently not known but it is possible that
they may be involved in determining RNA stability, as shown for other genes (Shaw and Kamen,
1986).

Limited information is available for the teleost FSHβ genomic sequence as only two FSHβ
genes (goldfish and tilapia) have been isolated and sequenced to date (Rosenfeld et al., 1997; Sohn
et al., 1998). The genomic organization in both fish conforms to the three exons and two introns,
as reported in the mammalian FSHβ genes, although the exon 3 of the goldfish and tilapia FSHβ
genes contains a much shorter 3’UTR than those of mammals (Figure 3). The location of the first
and second introns in both fish shows a well-conserved pattern similar to that of mammalian FSHβ
genes. For the tilapia, the gene is found to be coded by a single copy gene. In contrast, two distinct
genes (GTHIβ-1 GTHIβ-2) encoding the FSHβ in the goldfish were isolated and sequence analysis
revealed high sequence identity between the coding regions. Initial studies revealed that both
goldfish FSHβ genes are expressed, although one of these is expressed in higher proportion in
sexually immature goldfish, while both are expressed in equal amounts in the mature fish. This
suggests differential regulation of the goldfish GTHIβ genes at various reproductive stages.

6


Comparison of the amino acid sequence of the tilapia FSHβ and LHβ with β subunits from
other fish species showed that the FSHβ subunit protein sequence is less conserved than that of the
LHβ gene. The tilapia LHβ amino acid sequence shares the highest homology with the grouper at

96%, while the highest homology of the tilapia FSHβ amino acid sequence is with that of the
bonito and is only 60% (Rosenfeld et al., 1997).

7


LHβ subunit gene organization
E1
Porcine
Human
Bovine

E2
In 2

0.30

0.30

0.35

0.28

0.30

Rat
Equine
Tilapia

E3

In 1

0.20

0.30

0.20

0.35

0.28

0.35

1.25

FSHβ subunit gene organization
1
Goldfish
2
Mouse
Rat
Ovine
Bovine
Porcine
Human
Tilapia

0.32

0.79

0.28

0.70
0.62

1.25

0.64

1.26

0.64

1.60

0.62

1.56

0.92
1.10
1.35

1.57
1.40
0.9

Figure 3: Comparison of the LHβ and FSHβ subunit genes
The open reading frames are represented by yellow boxes and the untranslated region of exons (E) are
represented by open boxes. Blue lines between the boxes indicate introns (In). Numbers below the introns
show the approximate length in kb. References for LH β subunit genes: porcine (Ezashi et al., 1990);
human (Talmadge et al., 1984); bovine (Virgin et al., 1985); rat (Jameson et al., 1984); equine (Sherman,
1992); tilalia (Rosenfeld et al., 1997). References for FSH β subunit genes: goldfish (Sohn et al., 1998);
mouse (Kumar et al., 1995); rat (Gharib et al., 1989); ovine (Guzman et al., 1991); bovine (Kim et al.,
1988); porcine (Hirai et al., 1990); human (Jameson et al., 1988); tilapia (Rosenfeld et al.., 1997).
(Modified from Bousfield et al., 1994).

8


1.3 Transcriptional regulation of the gonadotropin subunit genes
The regulation of gonadotropin genes comprises basal gene expression which targets these
genes specifically to the gonadotropes, and hormonal stimulation in which GnRH and gonadal
steroids up-regulate expresson at puberty. To date, much less is known about the regulation of
these genes in teleost fish than in mammals and information regarding the regulation of the FSHβ
subunit gene expression lags far behind what is known about the LHβ subunit.

1.3.1 Regulation of gonadotropins by GnRH
The decapeptide GnRH plays a critical role in reproductive development and function in
vertebrates by stimulating the biosynthesis and secretion of the pituitary gonadotropins.
In the teleost, in vivo and transfection studies demonstrated the stimulatory effect of GnRH on LH
and FSH gene expression, as seen in the goldfish (Khakoo et al., 1994; Klausen et al., 2001),
tilapia (Melamed et al., 1996; Gur et al., 2002), striped bass (Hassin et al., 1995), sea bream
(Kumakura et al., 2003) and salmonid species (Weil and Marcuzzi, 1990; Kitahashi, et al., 1998;
Dickey and Swanson, 1998; Melamed et al., 2002).

The degree of stimulatory effects of GnRH on the gonadotropin genes is dependent largely

on the reproductive stage of the fish and the species, and to a lesser extent, the gender. In the study
of gonadotropin response to GnRH during sexual ontogeny of the common carp, gonadotropin
levels in juvenile fish were unresponsive to GnRH administration. In contrast, FSH- and LHβ
mRNA of the maturing females increased up to three fold over controls, while there was only a
slight increase of the LHβ mRNA in the male counterparts. In the post-vitellogenic females, LHβ
but not the FSHβ mRNA levels increased dramatically whereas in the male fish of the same age,

9


increased FSHβ and LHβ mRNA levels were not observed (Kandel-Kfir et al., 2002). In immature
striped bass, GnRH in combination with testosterone, stimulated the β subunit mRNA levels to
various extents but had no effect in the maturing fish (Hassin et al., 2000). A study on female
rainbow trout revealed that GnRH did not significantly stimulate FSHβ secretion at any stage of
gametogenesis, even when the FSHβ levels increased after ovulation, whereas LHβ secretion
increased following mid-vitellogenesis (Breton et al., 1998). Likewise, in the sham- and
testosterone-implanted sea bass, GnRH had no effect on FSHβ mRNA levels, but increased LHβ
mRNA levels (Mateos et al., 2002). In contrast, GnRH treatment induced the expression of all
three gonadotropin subunit genes in the immature sea bream (Kumakura et al., 2003). In the
goldfish, however, GnRH treatment increased both gonadotropin subunit mRNAs in the sexually
matured fish but decreased the level of FSHβ mRNA in the sexually regressed fish (Khakoo et al.,
1994; Sohn et al., 2001). Taken together, these studies indicate that the gonadotropin genes are
regulated through various GnRH-induced signaling pathways and are also influenced by the
reproductive stage of the fish.

1.3.1.1 GnRH signaling pathway
The stimulation of gonadotropin subunit genes by GnRH is activated through intracellular
signaling transduction pathways following the binding of GnRH to G protein-coupled, seventransmembrane receptors (GnRHR) on the pituitary gonadotrope (Marshall and Kelch, 1986;
Gharib et al., 1990). In mammals it has been shown that the receptor activates L-type calcium
channels, allowing extracellular calcium into the cell (Naor, 1990). Phospholipase C (PLC) is also

activated, leading to cleavage of phosphatidylinositol-diphosphate (PIP2) located in the cell
membrane, into inositol 1,4,5-triphosphate (IP3), which mediates calcium release from

10


intracellular stores, and diacylglycerol (DAG). Increased concentrations of intracellular calcium
together with DAG production lead to activation of protein kinase C (PKC), which in turn leads to
activation of other protein kinases, such as the ubiquitous mitogen-activated protein kinases
(MAPK) known to be involved in various cellular functions including cell growth, differentiation,
transformation, cell cycle and apoptosis. (Roberson et al, 1995; Haisenleder et al, 1998). The
activated MAPKs translocate to the nucleus and stimulate sequence-specific transcription factors
via phosphorylation. Once the transcription factors are activated, they bind to promoter DNA
sequences and trigger the expression of the gonadotropin subunit genes (Naor et al., 2000 and
Shacham et al., 2001).

In the mammalian gonadotrope, GnRH is able to differentially regulate LH and FSH
biosynthesis through preferential sensitivity of the gonadotropin subunit gene promoters to either
calcium influx or PKC/MAPK cascade (Roberson et al., 1993; Schoderbek et al., 1993; Weck et
al., 1998; Saunders et al., 1998). In the report by Saunders et al (1998), PKC-dependent pathways
seemed to play a larger role in the GnRH-mediated transcriptional control of the LHβ- and FSHβ
genes while the α-subunit appeared to be more sensitive to calcium influx. However, in a recent
study involving exposure of tilapia pituitary cells to GnRH, the PKC-extracellular signal-regulated
kinase (ERK) cascade elevated α- and LHβ mRNAs, whereas FSHβ transcript induction was ERKindependent and was regulated by the cAMP-PKA pathway (Gur et al., 2002). The frequency and
amplitude of the GnRH pulses delivered to the anterior pituitary, which vary during different
phases of the estrous cycle, is also known to direct differential regulation of the gonadotropin gene
expression (Levine and Ramirez, 1982; Crowley et al., 1985). In general, higher GnRH amplitude
stimulates LH release while FSH production is increased at lower GnRH doses (Papavasiliou et

11

al., 1986). The effect of the variation in GnRH pulse pattern is also associated with differential LH
and FSH release, with higher frequency resulting in greater LH secretion while lower frequency
stimulates FSH release (Dalkin et al., 1989).

1.3.1.2 GnRH-responsive elements on the common α subunit promoter
Gonadotropin α-subunit promoter elements required to mediate GnRH responsiveness have
been mapped in human (Kay and Jameson, 1992), cow (Hamernik et al., 1992), and mouse
(Schoderbeck et al., 1993), but not in fish. In the mouse, two DNA elements, the pituitary
glycoprotein hormone basal element (PGBE) and GnRH response element (GnRH-RE;
TGCCTGTT) were found to be important for GnRH stimulation (Schoderbeck et al., 1993). It was
postulated that the GnRH-RE identified at the proximal promoter (-406 to -399 bp) might respond
to GnRH by binding an Ets-like transcription factor that is activated through the MAPK pathway
(Roberson et al., 1995), while the PGBE, at a more distal region, could be targeted by a LIMhomeodomain protein (Roberson et al., 1994). In the human α-subunit promoter, two or more
GnRH responsive regions have been mapped to between -420 to -244 bp, and these are responsible
for transcriptional activation by calcium influx (Holdstock et al., 1996). There are two cAMP
response elements (CRE) within this region whose mutation greatly reduces the basal activity, and
are thus likely to be involved in the GnRH responsiveness of this promoter (Andersen et al., 1990).
cAMP response element binding protein (CREB) is known to bind the CRE and a recent report
showed that GnRH stimulates the phosphorylation of CREB in α-T3 cells, which was necessary
for GnRH stimulation of the human α-subunit gene (Duan et al., 1999).

12


1.3.1.3 GnRH-responsive elements on the LHβ subunit promoter
Sequence analysis of the teleost LHβ gene promoters reveals a number of cis-acting
elements, some of which have also been shown to be essential in the basal and GnRH-stimulated
transcription in the mammalian homologs. Since the first teleost genomic sequence for the LHβ

gene was isolated from the Chinook salmon (csLHβ; Xiong and Hew, 1991), most functional
studies on the putative elements have so far been restricted to this fish. The csLHβ gene promoter
contains a strong proximal tripartite element targeted by steriodogenic factor-1 (SF-1), pituitary
homeobox-1 (Ptx-1) and estrogen receptor (ER) whose synergistic interaction, in estradioldependent manner, drives the expression of the gene (Le Dréan et al., 1996, 1997; Melamed et al.,
2002). Moreover, this proximally bound ER appears to interact with another ER bound at a distal
ERE (dERE), and both EREs were suggested to be involved in mediating the GnRH effect (Liu et
al., 1995). It was further suggested that this interaction is facilitated by the bending of the target
DNA, induced by Ptx-1 dimers binding four Ptx-1 binding sites located between these two EREs
(Melamed et al., 2002).

GnRH responsive regions have been identified in the LHβ subunit promoter of several
mammalian species including sheep (Brown et al., 1993), cow (Keri et al., 1994) and rat (Fallest et
al., 1995). Unlike in teleosts however, mammalian LHβ gene promoters contain two early growth
response factor-1 (Egr-1) REs which cooperate synergistically with SF-1 and Ptx-1 to mediate
GnRH stimulation (Tremblay and Drouin, 1999; Halvorson et al., 1999). In addition, the rat LHβ
promoter contains two binding sites for Sp-1 which were demonstrated to play an important role in
conferring GnRH responsiveness (Kaiser et al., 1998). Further investigation showed that the 3’ Sp1 site is also critical for basal expression and that the Sp-1 binding sites act synergistically with the

13


downstream SF-1 and Egr-1 binding sites to incite maximal response to pulsatile GnRH (Weck et
al., 2000).

1.3.1.4 GnRH-responsive elements on the FSHβ subunit promoter
Inspection of the 1.7 kb 5’-flanking region of the tilapia FSHβ gene has revealed several
putative regulatory sequences which have been shown to be essential for inducible and tissuespecific transcriptional regulation of other gonadotropin genes (Rosenfeld et al., 2001). These
regulatory sequences include the gonadotrope-specific element (GSE), the binding site for
steroidogenic factor-1 (SF-1); half sites of estrogen response elements (½ ERE) and a cAMP
response element (CRE), the recognition site for the Fos and Jun transcription factors or activating

protein-1 (AP-1). Transient transfection studies of the tilapia FSHβ promoter fused to a luciferase
reporter vector showed that the promoter responds to GnRH treatment in a dose-dependent manner
although the region that mediates this stimulation was not fully defined. 5’ deletion analysis of the
gene construct showed that this 1.7 kb DNA sequence contains both positive and negative
regulatory elements (Rosenfeld et al., 2001; Yaron et al., 2001). Examination of the 5’-flanking
region of the two goldfish FSHβ genes also revealed the presence of several putative cis-acting
elements including a GSE, GnRH REs and several half androgen responsive elements (½ ARE)
and EREs located contiguously between –187 and –124 bp upstream from a TATAA sequence,
which may be involved in the transcriptional control of the goldfish FSHβ gene (Sohn et al.,
1998). However, functional studies of these putative GnRH REs in the goldfish FSHβ subunit gene
promoter have not been performed.

14


The mammalian FSHβ gene has been more extensively analyzed. In recent studies, AP-1
recognition sites (TGAC/GTCA) located at -120 and –83 bp of the ovine FSHβ (oFSHβ) promoter
were shown to mediate the GnRH response (Strahl et al., 1997, 1998; Huang et al., 2001b). Using
HeLa and COS-7 cells, both AP-1 sites were required by GnRH to induce a threefold increase in
oFSHβLuc transcription, and mutation of these AP-1 sites completely abolished the GnRH effect
(Strahl et al., 1997, 1998). It was also shown that these AP-1 sites could support GnRH-mediated
induction of oFSHβLuc in pituitary cultures derived from transgenic mice harboring the
oFSHβLuc DNA constructs. (Huang et al., 2001b).

A Ptx-1 binding site on the rat FSHβ (rFSHβ) promoter that is similar to the consensus Ptx1 recognition sequence (TAAT/GCC; Zakaria et al., 2002), was also found to be important for
both basal and GnRH-stimulated expression of the gene. Conversely, the GnRH-stimulated
promoter activity was significantly reduced when the Ptx-1 site was mutated suggesting that Ptx-1
could mediate GnRH induction of the FSHβ gene (Zakaria et al., 2002). Ptx-1, a member of the
Bicoid protein family of TFs (Lamonerie et al., 1996), is expressed in almost all cell-types in the
pituitary including the gonadotropes. Ptx-1 has been found to activate the promoters of most

pituitary-specific genes (Szeto et al., 1996; Tremblay et al., 1998).

1.3.2 Regulation by gonadal steroids
The action of gonadal steroids, testosterone and estrogen, on gonadotropin production in
fish can be negative or positive, depending largely on the stage in the reproductive cycle and
possibly also the species. Studies in several species showed that treatment of immature fish (e.g.
salmon, European eel, goldfish and catfish) with gonadal steroids resulted in positive feedback

15


effects on the LHβ mRNA levels (Trinh et al., 1986; Dickey and Swanson, 1998; Quérat et al.,
1991; Huggard et al., 1996, 2002; Rebers et al., 1997). Similar results were observed in cultured
pituitary cells of immature rainbow trout, European eel, tilapia and goldfish (Xiong et al., 1994b;
Huang et al., 1997; Melamed et al., 1997; Huggard et al., 1996, 2002). Contrary to the results
observed in immature fish, treatment of pituitary cells from the mature or spawning fish with
steroids did not affect LHβ transcript levels (Xiong et al., 1994b; Huggard et al., 1996; Sohn et al.,
1998). Furthermore, circulating LH levels were reduced in the mature salmon injected with
estrogen (Billard et al., 1977). There seems to be a shift in the sensitivity of the pituitary gland to
steroid treatment between early and late recrudescent Atlantic Croaker and goldfish (Khan et al.,
1999; Huggard et al., 2002) although this is not true for the Atlantic salmon (Borg et al., 1998). In
the early recrudescence phase of the gonadal cycle of Atlantic Croaker, estrogen and testosterone
exerted positive effects on LH secretion. In the early recrudescent goldfish, however, high doses of
estrogen were needed to achieve maximal stimulation of both α- and LHβ subunit mRNAs, but a
lower concentration of estrogen was required for similar response in late recrudescent fish
(Huggard et al., 2002).

Fewer studies have been carried out on the effects of steroids on FSHβ gene expression.
However, for the tilapia FSHβ subunit gene, response to testosterone depends on the reproductive
stage of the fish. In cultured pituitary cells from immature or maturing tilapia, FSHβ mRNA levels

increased after exposure to low doses of testosterone (Melamed et al., 1997). This effect was not
observed in regressed fish while those at the end of the spawning season showed reduced mRNA
levels after treatment with testosterone or estrogen (Melamed et al., 1998). In salmonids, treatment
with testosterone had no significant effects on the FSHβ levels in immature rainbow trout and coho

16


salmon (Xiong et al., 1994b; Dickey and Swanson, 1998), and likewise, steroid treatment did not
affect immature coho salmon FSHβ mRNA levels, although the level decreased in early
recrudescent fish (Swanson et al., 1991).

Steroids may regulate gonadotropin gene expression by acting on the hypothalamus or the
pituitary. At the hypothalamic level, steroids negatively regulate the transcription of mammalian
gonadotropin genes by altering the GnRH pulse pattern. A direct effect of steroids has been shown
on the rat FSHβ gene promoter, which contains progesterone-like responsive elements (PREs)
sequences. Gel shift assays showed that these PRE-like sequences bind the progesterone receptor
with high affinity and specificity. Moreover, transient transfection studies on the rat FSHβ
promoter showed that progesterone positively regulates the rat FSHβ gene (O’Conner et al., 1997,
1999). Dihydrotestosterone-bound androgen receptor also acts at the pituitary to negatively
regulate basal and GnRH-stimulated α-subunit gene expression (Clay et al., 1993). This
transcriptional repression occurs through direct interaction of the ligand bound receptor with Sp1
and possibly Egr-1, so disrupting the interactions of these transcription factors with their binding
sites on the promoter (Curtin et al., 2001). Jorgensen et al (2001a ; 2001b) reported that the
suppression of the bovine LHβ promoter activity by the ligand bound androgen receptor is
mediated through interactions with the SF-1 transcription factor; while for the human α-subunit, it
is mediated via interactions with the c-jun and activation transcription factor 2 (ATF2). In other
studies, estrogen potentiates GnRH analog-induced LHβ secretion in goldfish and black porgy fish
(Trudeau et al., 1991, 1993; Yen et al., 2002). The mechanism was suggested to be through acting
on the GnRH receptor number (Hamernik et al., 1995), post-translational processes (Gardner et al.,

2000), or on the GnRH-induced signaling in the gonadotrope (Colin and Jameson, 1998).

17


1.3.2.1 Molecular mechanisms of steroid actions
The transcriptional regulation of gene expression by steroid hormones is based on a twostep model that involves high affinity binding of the hormone to specific steroid hormone
receptors, usually within the cytoplasm of the target cells, followed by activation of the hormonereceptor complexes which translocate into the nucleus. The activated receptors bind to the
hormone responsive elements of target promoters as monomers (Johnston et al., 1997),
homodimers (Vanacker et al., 1999) or heterodimers (Beato and Klug, 2000) to activate
transcription.

While both positive and negative steroid effects on gonadotropin production have been
observed in various fish species, the molecular mechanisms of these steroidal actions are still
relatively unknown and most studies have been restricted to the csLHβ subunit gene promoter. In
this promoter, one full ERE and several ½ EREs have been mapped, of which the proximal ERE
(pERE) is suggested to de-repress the proximal silencer resulting in increased transcription of the
LHβ subunit gene (Xiong et al., 1994b). There is also evidence that the other ½ EREs may be
involved in estrogen responsiveness through the synergistic cooperation with the proximal ERE
(Liu et al., 1995). The ER also interacts synergistically with SF-1 to stimulate the csLHβ subunit
gene promoter in an estradiol-dose dependent manner (Le Dréan et al., 1996). In the study, SF-1
alone could stimulate promoter but the stimulation was dramatically enhanced when combined
with ER. This synergistic SF-1/ER effect was dependent on binding to the ERE and two putative
SF-1 binding sites located in the proximal region of the promoter (Le Dréan et al., 1996).

18


Recently, transient transfection studies in α-T3 cells revealed that ER can also be activated
by the PKA or the PKC/MAPK signaling pathways leading to the transactivation of an ERresponsive promoter in a ligand-independent manner (Demay et al., 2001; Schreihofer et al.,

2001). This ligand-independent ER activity likely depends on its phosphorylation in AF-1,
allowing interaction with specific co-activators (Hall et al., 2001).

1.3.3 Regulation by activin
Activin, a member of the transforming growth factor-β (TGFβ) superfamily, is a homo- or
heterodimer of inhibin B, and is produced in a wide variety of tissues including the pituitary
gonadotropes (Roberts, et al., 1989). It stimulates the biosynthesis and release of FSH (Ling et al.,
1986; Weiss et al., 1992). In primary pituitary cell cultures, activin caused a dose-dependent
increase in FSHβ mRNA levels and FSH secretion that was not observed in LHβ and common α
subunit levels (Carroll, et al., 1989). Moreover, GnRH stimulation of FSHβ mRNA may be
activin-dependent, because the increase in FSHβ mRNA level induced by GnRH was attenuated
when the activin-binding follistatin was added to the perifusion system (Besecke, et al., 1996).
Conversely, female rats treated with a GnRH antagonist secrete FSH when stimulated by activin
(River, et al., 1991). Although activin signaling in a variety of cells has been studied to some
extent, the specific signal pathways through which activin enhances FSHβ gene expression are at
present unclear and have not been studied in the fish.

1.3.3.1 Mechanisms of activin-stimulated gene expression
Activin regulates FSH transcription through interaction with receptors that are members of
the TGFβ family of transmembrane receptor kinases, resulting in both the elevation of FSHβ

19


mRNA levels and FSH secretion (Matthews and Vale, 1991; Attisano et al., 1992). Activin signal
transduction is mediated by a heterotetrameric receptor complex consisting of two ligand-binding
type II receptors (ActRII) and two signal transducing type I receptors (ActRI; Massagué and
Wotton, 2000). Upon ligand binding, the ActRII phosphorylates the ActRI on multiple serine and
threonine residues located in a conserved GS-rich motif located just N-terminal to the kinase
domain. The phosphorylated receptor is then activated to signal to its downstream targets, the

Smad family of proteins which transmit TGFβ signals from the cell surface into the nucleus
(Attisano and Wrana, 2000; Zimmerman and Padgett, 2000).

The Smad family of proteins is classified into three major classes. The first class is the
receptor-regulated Smads (R-Smads) that includes the activin receptor-targeted Smad 2 and 3.
These Smad proteins are phosphorylated by the type I receptor on the last two serines at the
extreme carboxy terminus (Abdollah et al., 1997). The second class of Smad proteins is the
common mediator Smad (Co-Smad) comprising of the Smad4 found in mammals and the highly
related Smad4β/Smad10 in Xenopus. These Co-Smads associate with phosphorylated R-Smads
and the resulting R-Smad/Co-Smad complex then translocates to the nucleus to induce specific
gene responses (Lagna et al., 1996; Howell, et al., 1999; Masuyama, et al., 1999). Activin
signaling is prevented by the third class of Smad proteins known as the inhibitory Smads (ISmads). These proteins antagonize signaling by interacting directly with the receptor to prevent
access to and phosphorylation of R-Smads, or by interfering with R-Smad/Smad4 complex, or by
recruiting an E3 ligase that induces degradation of the receptor complex (Welt et al., 2002).

20


Once in the nucleus, the R-Smad/Co-Smad complex may bind specific gene promoters
through the consensus motif GCTC or AGAC (Massagué and Wotton, 2000). However, this
binding is of low affinity and specificity, and might not be required on all promoters. For example,
Pardali et al. (2000) showed that Smad3 and Smad4 functionally cooperate with Sp-1 to activate
the human p21 promoter in hepatoma HepG2 cells. This activation requires the ubiquitous Sp-1 to
bind to the proximal promoter. Using Smad3 and Smad4 DNA-binding site mutants, the Smad
proteins could still transactivate the p21 promoter as efficiently as wild type Smads indicating that
the Smad3 and Smad4 binding sites are not crucial. The same Smad proteins synergize with cjun/c-fos at the AP-1 binding site of the collagenase I promoter to induce transcriptional activation
in response to TGF-β. Mutational analyses of the c-jun protein and the AP-1 binding site in the
promoter revealed that the interaction of c-jun with DNA is necessary for the transcriptional
activation, whereas similar analysis of Smad3 revealed that its binding to the DNA, is required
although there appears to be little importance for the actual DNA binding sequence (Qing et al.,

2000).

These studies, together with the short consensus Smad recognition sequence, suggest that
Smad-dependent gene expression only involves weak binding of Smad to its recognition sequence,
and requires interactions with other transcription factors (Welt et al., 2002). The activin-mediated
regulation of gene expression therefore likely involves a series of interconnecting signal
transduction pathways, culminating with the concerted actions of the DNA-binding Smad proteins
and other transcription factors and associated co-factors.

21


It was reported that the Smad3 and Smad4 transcription factors are responsible for
mediating the stimulation of the rat FSHβ promoter by activin in LβT2 cells. Functional analysis
identified the target site of the promoter through which the transcription factors act. Moreover, the
pituitary-specific Ptx-2 transcription factor and its corresponding binding site on the promoter
were also shown to be involved in regulating the FSHβ subunit gene in a tissue-specific manner.
This was the first paper to identify Smad transcription factors acting in concert with a pituitaryspecific nuclear transcription factor to mediate the activin stimulation of the FSHβ subunit gene
(Suszko, et al., 2003). In another recent study in primary ovine pituitary and LβT2 cells, activin
activated several signaling pathways with different time course but only the Smad2 pathway
appeared to be directly implicated in the expression and release of FSH in LβT2 cells (Dupont et
al., 2003). Taken together, these two studies suggest that Smad3/Smad4 complex is at least partly
responsible for the activin-induced expression of the FSHβ gene via the Smad2 pathway.

1.3.3.2 Hormonal modulators of activin action on the FSHβ subunit gene
The actions of activin are regulated by the extracellular proteins inhibin and follistatin.
Inhibin interferes with the activin transduction pathway by disrupting the formation of functional
activin-ActRII-ActRI complexes. This inhibitory regulation occurs through at least two inhibinactivated intracellular pathways. The first involves betaglycan, a type III TGFβ receptor that
functions as an inhibin co-receptor that binds to the ActRII with high affinity, thus preventing the
activin from doing so (Lewis et al., 2000). The second is the inhibin-binding protein, InhBP

(p120), that associates strongly with the ActRI (type IB) in a activin-responsive manner. It
antagonizes the activin signal transduction through modulating the activin heteromeric receptor
complex assembly (Chapman and Woodruff, 2001). Follistatin binds to activin with high affinity

22


and prevents it from binding to its own receptor on the cell surface (Nakamura et al., 1990; de
Winter et al., 1996). Huang and co-workers (2001a) showed that both inhibin and follistatin
decreased the activin-stimulated basal expression of the oFSHβ in the LβT2 cell cultures. Similar
results were also obtained with the mouse FSHβ promoter using activin and follistatin in the LβT2
cells (Jacobs et al., 2003). In this respect, it is reasonable to assume a regulatory function of these
extracellular proteins for the fish FSHβ subunit gene transcription, although the detailed
mechanism remains to be elucidated.

1.4 Hypothesis
The promoter of the Chinook salmon FSHß gene contains regulatory sequences that are
responsible for the cell-specific and basal expression of this subunit, and also those that mediate
the stimulatory effects of GnRH and activin during puberty and sexual maturation.

1.5 Aim
The aim of this project is to initiate research on the mechanism of transcriptional regulation
of the FSHβ subunit gene promoter of the Chinook salmon. The Chinook salmon FSHβ subunit
gene and 5’ flanking sequence will be isolated, sequenced and the transcriptional start site mapped
using primer extension. The promoter will then be cloned to a reporter construct to test for
functionality in the pituitary-derived LβT2 as well as a heterogeneous cell line. Subsequent
transfection studies will be carried out to determine responsiveness of the promoter and the
sequence will be analyzed to identify putative regulatory elements that may mediate these effects.

23

2. MATERIALS AND METHODS

2.1 Extraction and isolation of genomic DNA
Female Chinook salmon brain (200 mg) in 4 mL of DNAzol reagent (Life Technologies)
was homogenized using a tissue grinder according to the manufacturer’s instructions. The
homogenate was centrifuged for 10 min at 10, 000 x g at 4 0C. Following centrifugation, the
resulting supernatant was transferred to a fresh 1.5 mL microcentrifuge tube. DNA was
precipitated from the lysate/homogenate by adding 2 mL of 100 % ethanol. The sample was mixed
by inversion and stored at room temperature for 2 min. The DNA precipitate was spooled with a
pipette tip and transferred to a clean 1.5 mL microcentrifuge tube. Following this, the DNA
precipitate was washed twice with 1 mL 75 % ethanol. At each wash, the DNA was suspended in
ethanol by inverting the tube three times. The DNA was allowed to settle to the bottom of the tube
before removing the ethanol by pipetting. After the DNA was air dried for 15 s, it was dissolved in
400 µL of 8 mM NaOH.

The integrity of the genomic DNA (gDNA) was first determined. One µg of gDNA and
control human gDNA were loaded onto a 1.0 % agarose/EtBr gel to check for the size of the
gDNA. The quality of the genomic DNA was also analyzed by Dra I digestion. The
electrophoresis was carried out at 100 V for 1 h. A 1 kb DNA ladder (GeneRulerTM, MBI
Fermentas) was used as a molecular weight marker for size estimation of the DNA fragments. The
EtBr-stained linear DNA fragments were observed under UV illumination and the gel was
photographed using a digital camera.

24


2.2 Genome walking
The gDNA sequence upstream of a partial cDNA sequence of Chinook salmon csFSHβ

gene from a cDNA library (Suzuki and Hew, unpublished) was isolated using the Universal
GenomeWalkerTM Kit (Clonetech Laboratories). The principle of this kit is based on two-rounds of
PCR amplifications from the blunt-end gDNA fragments linked with known adaptor primers (AP1
and AP2) provided in the kit. The first step was to generate libraries of uncloned, adaptor-ligated
gDNA fragments. Aliquots of gDNA were completely digested with different blunt-end restriction
enzymes before ligation to the GenomeWalker adaptors. Each adaptor-ligated library was then
used as a template for primary PCR reaction using the adaptor primer 1 (AP1) and a gene-specific
primer (GSP1), to amplify the fragment of interest. Subsequently, the primary PCR product was
diluted and used as the template for a secondary or “nested” PCR with the nested adaptor primer
(AP2), and a nested gene-specific primer (GSP2). The overall procedure is shown in Figure 4.

25