Báo cáo khoa học: Involvement of NF-jB subunit p65 and retinoic acid receptors, RARa and RXRa, in transcriptional regulation of the human GnRH II gene pot

Although the hGII-Sil region had a similar gene-repressive effect in both cell lines, slightly different DNA–protein binding patterns were observed in EMSAs using TE671 and JEG-3 nuclear

receptors, RARa and RXRa, in transcriptional regulation

of the human GnRH II gene

Ruby L C Hoo1,*, Kathy Y Y Chan2, Francis K Y Leung1, Leo T O Lee1, Peter C K Leung3 and Billy K C Chow3

1 School of Biological Sciences, University of Hong Kong, Pokfulam Road, Hong Kong, China

2 Department of Paediatrics, Chinese University of Hong Kong, Prince of Wales Hospital, Shatin, Hong Kong, China

3 Department of Obstetrics and Gynecology, University of British Columbia, Vancouver, Canada

In humans, the genes for gonadotropin-releasing

hor-mones (GnRH I and GnRH II) have the same

modu-lar structure, harboring three introns and four exons

The exons encode a precursor polypeptide consisting

of a signaling peptide, the GnRH decapeptide, and the

GnRH-associated peptide (GAP) with unknown

func-tion [1] The promoter region of the human

(h)GnRH II gene is located at the 5¢ flanking region,

the untranslated exon 1, intron 1, and exon 2 The

locations of exon 1, intron 1 and exon 2 are )793 ⁄ )750 (relative to the +1 translation start codon ATG),)749 ⁄ )8 and )7 to +154, respectively

Despite similar gene structures, GnRH I and GnRH II genes are regulated by different regulatory elements Multiple regulatory sites have been identified

in the promoter of the hGnRH II gene In 2001, Chen

et al [2] identified a putative cAMP-response element (CRE) site at nucleotide sequence )860 to )853 The

Keywords

gonadotropin-releasing hormone II; NF-jB

subunit p65; retinoic acid receptors;

silencer; transcriptional regulation

Correspondence

B K C Chow, School of Biological

Sciences, University of Hong Kong,

Pokfulam Road, Hong Kong, China

Tel: +852 2299 0850

Fax: +852 2857 4672

E-mail: bkcc@hkusua.hku.hk

*Present address

Department of Medicine, Li Ka Shing

Faculty of Medicine, University of Hong

Kong, Queen Mary Hospital, Hong Kong

(Received 1 November 2006, revised 19

March 2007, accepted 22 March 2007)

doi:10.1111/j.1742-4658.2007.05804.x

Gonadotropin-releasing hormone (GnRH) I and II are hypothalamic deca-peptides with pivotal roles in the development of reproductive competence and regulation of reproductive events In this study, transcriptional regula-tion of the human GnRH II gene was investigated By scanning mutaregula-tion analysis coupled with transient promoter assays, the motif at )641 ⁄ )636 (CATGCC, designated GII-Sil) was identified as a repressor element Mutation of this motif led to full restoration of promoter activity in TE671 medulloblastoma and JEG-3 placenta choriocarcinoma cells Supershift and chromatin immunoprecipitation assays showed in vitro and in vivo binding of NF-jB subunit p65 and the retinoic acid receptors, RARa and RXRa, to the promoter sequences Over-expression of these protein factors indicated that p65 is a potent repressor, and the RARa⁄ RXRa heterodimer

is involved in the differential regulation of the GnRH II gene in neuronal and placental cells This was confirmed by quantitative real-time PCR Treatment of cells with the RARa⁄ RXRa ligands, all-trans retinoic acid and 9-cis-retinoic acid, reduced and increased GnRH II gene expression in TE671 and JEG-3 cells, respectively Taken together, these data demon-strate the differential roles of NF-jB p65 and RARa⁄ RXRa, interacting with the same sequence in the promoter of the human GnRH II gene to influence gene expression in a cell-specific manner

Abbreviations

ATRA, all-trans retinoic acid; ChIP, chromatin immunoprecipitation; CRE, cAMP-response element; EMSA, electrophoretic mobility-shift assay; GAP, GnRH-associated peptide; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GnRH, gonadotropin-releasing hormone; HDAC, histone deacetylase; L-CoR, ligand-dependent corepressor; N-CoR, nuclear receptor corepressor; RA, retinoic acid; RAR, retinoic acid receptor.

same research group also demonstrated that GnRH II,

but not GnRH I, is potently up-regulated by a cAMP

analog in human neuronal medulloblastoma cells

TE671 From deletion and mutation analysis, it was

concluded that the CRE site is responsible for both the

basal activity and cAMP induction of the hGnRH II

promoter Similarly to the case of cAMP stimulation,

it has been reported that estrogen regulates the

expres-sion of GnRH I and GnRH II differentially Estrogen

treatment down-regulates the promoter activity of

GnRH I but up-regulates GnRH II promoter activity

Indeed, analysis of the promoter sequence has revealed

a partial putative estrogen-responsive element site

and an SP1 site at positions )1252 ⁄ )1256 and

)1726 ⁄ )1717, respectively [3] In addition to cAMP

and estrogen, other hormonal regulation of GnRH II

expression has been investigated In human

granulolu-teal cells, treatment with follicle-stimulating hormone

or human choriogonadotropin was reported to increase

GnRH II mRNA level but decrease GnRH I mRNA

level [4] It is of interest that GnRH II was reported to

be self-regulated in the same study Significant

decrea-ses in GnRH II and GnRH receptor mRNA levels

were observed in cells treated with GnRH II or its

agonist

Our research group has previously identified a

min-imal promoter and two enhancer elements (E-boxes)

and Ets-like element in the untranslated first

exon functioning co-operatively to achieve full

promo-ter activity [5] A silencing element in the first intron,

which has a significant repressive effect on the

GnRH IIgene, has also been reported [6] The present

study aimed to define the cis-acting element and

investigate the protein factors involved in regulation of

the hGnRH II gene in TE61 and JEG-3 cells These

cell lines, which endogenously co-express GnRH I and

GnRH II, are valuable models for examining

tran-scriptional regulation of the GnRH II gene [7,8]

Results

Fine mapping of the cis-acting element

at)650 ⁄ )620

To characterize the hGnRH II intronic silencer and to

identify the location of the cis-acting element(s) within

this region, a series of mutant constructs, scanning

mutants Mut1 to Mut10 as shown in Fig 1A, were

generated from the wild-type pGL2-()2103 ⁄ )620)

con-struct The 30 base pairs at )650 ⁄ )620 was a potent

silencing element in both cell lines, significantly

repressing promoter activity to 17.6 ± 0.7% and

31.2 ± 1.6% in TE671 and JEG-3 cells, respectively

In TE671 cells, Mut3, Mut4, Mut5 and Mut6 signifi-cantly [P < 0.001 versus pGL2-()2103 ⁄ )620) (wild-type)] restored promoter activity to 44.0%, 87.1%, 85.1% and 42.0% (compared with full promoter activ-ity), respectively (Fig 1B) Mut4 and Mut5 restored almost full promoter activity (87.1% and 85.1%) Sim-ilar results were observed in JEG-3 cells (Fig 1C): Mut3, Mut4, and Mut5 significantly [P < 0.001 versus pGL2-()2103 ⁄ )620) (wild-type)] restored promoter activity to 44.9%, 81.9% and 84.8%, respectively, with Mut4 and Mut5 restoring almost full promoter activity (81.9 ± 4.9% and 84.8 ± 9.8%, respectively) In con-trast, Mut6 did not show significant restoration of pro-moter activity in JEG-3 cells It is interesting to note that Mut1 led to further significant [P < 0.001 versus pGL2-()2103 ⁄ )620) (wild-type)] repression in both cell lines

hGII-Sil is a novel silencing element of the hGnRH II gene

Mutational analysis of the putative silencing element residing at )650 ⁄ )620 demonstrated the functional significance of the Mut4 and Mut5 region (CATGC-CAG, hGII-Sil) Electrophoretic mobility-shift assays (EMSAs) using the radiolabeled hGII-Sil oligonucleo-tide as DNA probe were then performed to identify whether there is any specific DNA–protein binding Complex in this region Although the hGII-Sil region had a similar gene-repressive effect in both cell lines, slightly different DNA–protein binding patterns were observed in EMSAs using TE671 and JEG-3 nuclear extracts (Fig 2A) Three obvious DNA–protein com-plexes were observed in the EMSA with TE671 nuclear extract (Fig 2A) Formation of Complex A and Com-plex B were dose-dependently inhibited by the unlabe-led DNA probe, and Complex A was completely diminished in 200-fold excess unlabeled competitor This implies that the binding of protein factors in Complex A and Complex B with the putative silencer

is specific In JEG-3 cell lines, three retarded DNA– protein complexes were also observed (Fig 2A) Of these, only Complex C showed a specific interaction because it was the only Complex that was dose-depend-ently inhibited by self competition Furthermore, when

a nonspecific unlabeled oligonucleotide (L8 oligonuc-leotide) was applied as the unlabeled competitor (Fig 2B), formation of Complex A and Complex B was not inhibited The presence of mutant oligonucleo-tides with mutations at the Mut4 and Mut5 region (Mut4+5 oligonucleotide) as the unlabeled competitor

in the binding reaction fails to inhibit the formation of both Complex A and Complex B (Fig 2B)

NF-jB subunit p65 and retinoic acid receptors,

RARa and RXR, interact with hGII-Sil in TE671

cells

According to the results of supershift assays, NF-jB

p65 subunit antibody and RAR antibody abolished

the formation of Complex A, indicating that the p65

subunit and members of the RAR family are involved

in the DNA–protein Complex in TE671 cells

Intrigu-ingly, along with the abolition of Complex A

forma-tion by RAR-specific antibody, there was a

concomitant increase in the intensity of Complex B

(Fig 3A) Subsequent supershift assays using

antibod-ies against different isoforms of RAR (RARa, RARb,

RARc) and RXR were performed to identify which

members of the RAR family were present in the

DNA–protein Complex (Fig 3B) Only RARa-specific antibody and RXR-specific antibody successfully abol-ished the formation of Complex A, indicating the involvement of RARa and RXR in the DNA–protein complex Similarly to the supershift assay described in Fig 3A, abolition of Complex A formation by RARa-specific antibody and RXR-RARa-specific antibody was accompanied by enhancement of Complex B

To show in vivo binding of p65, RAR and RXR

to the hGII-Sil region, chromatin immunoprecipita-tion (ChIP) assays were performed (Fig 4) We observed no PCR signals from the negative controls (No immunoprecipitation, lane 3; anti-rabbit IgG, lane 7; and PCR negative, lane 8) These controls indicate that there was neither nonspecific precipita-tion nor PCR contaminaprecipita-tion Positive PCR signals

A

pGL2-Basic

pGL2–(-2103/-650)

pGL2–(-2103/-620)

pGL2–(-2103/-620) Mut1

pGL2–(-2103/-620) Mut2

pGL2–(-2103/-620) Mut3

pGL2–(-2103/-620) Mut4

pGL2–(-2103/-620) Mut5

pGL2–(-2103/-620) Mut6

pGL2–(-2103/-620) Mut7

pGL2–(-2103/-620) Mut8

pGL2–(-2103/-620) Mut9

pGL2–(-2103/-620) Mut10

pGL2-Basic pGL2–(-2103/-650) pGL2–(-2103/-620) pGL2–(-2103/-620) Mut1 pGL2–(-2103/-620) Mut2 pGL2–(-2103/-620) Mut3 pGL2–(-2103/-620) Mut4 pGL2–(-2103/-620) Mut5 pGL2–(-2103/-620) Mut6 pGL2–(-2103/-620) Mut7 pGL2–(-2103/-620) Mut8 pGL2–(-2103/-620) Mut9 pGL2–(-2103/-620) Mut10

Relative promoter activity (% change)

Relative promoter activity (% change)

TE671

JEG-3

Fig 1 Fine mapping of the putative silencing element in the first intron of the hGnRH II gene The series of mutational constructs (A) were cotransfected (1 lg each) with 0.5 pSV-b-gal vector into TE671 cells (B) and JEG-3 cells (C) using Lipofection Reagent GeneJuice At 48 h post-transfection, cell lysate was prepared and used for luciferase and b-galactosidase assays Luciferase values are normalized by b-galac-tosidase expression and are shown as percentage changes in relative promoter activities compared with that of pGL2-( )2103 ⁄ )650), the hGnRH II promoter region with the putative 30-bp silencing element deleted, which is designated as having 100% promoter activity Values are mean ± S.E.M from at least three independent experiments each in triplicate *Significant difference (P < 0.001) versus control pGL2-( )2103 ⁄ )620).

were detected using p65, RAR and RXR antibodies

(lane 4–6) and in the positive control (lane 2) In

summary, data from the ChIP assays indicate in vivo

interaction of p65, RAR and RXR with the hGII-Sil

promoter

Over-expression of NF-jB p65 subunit down-regulated hGnRH II gene expression Functional assays were carried out to identify the effect of these trans-acting elements on expression of

Competitor (fold)

Nuclear extract (µg)

Complex A

Complex B

Non-specific binding

Free probe

Complex C Non-specific binding Non-specific binding Free probe

15µg

0 µg

Competitor

(fold)

Nuclear extract (µg)

Complex A

Complex B

Non-specific binding

Free probe

TE671

50× 100× 200×

0 µg

15µg

0 µg

Fig 2 Specific interaction of nuclear factors from TE671 and JEG-3 cells with the Mut4 and Mut5 (hGII-Sil) region in the putative silencer (A) EMSAs to characterize the pro-tein factor(s) binding to the Mut4 and Mut5 (hGII-Sil) region in the putative intronic silen-cing element in TE671 and JEG-3 cells Syn-thetic oligonucleotides of hGII-Sil were annealed to form dsDNA before radiolabe-ling with c32P The radiolabeled 24-bp DNA probe (0.2 omol, 200 000 cpm) was incuba-ted with 15 lg nuclear extracts from TE671

or JEG-3 cells during the binding reaction Increasing concentrations (0–200-fold excess) of unlabeled hGII-Sil oligonucleo-tides were applied as unlabeled competitors

to allow self-competition (B) L8 nonspecific oligonucleotide and mutant oligonucleotide (Mut4+5 oligonucleotide) were used as competitors.

A

Nuclear extract TE671 Nuclear extract (µg) 0µg 15µg Antibody -ve +ve p65 c-Jun RAR

Complex A

Complex B

Non-specific binding

Free probe

B

Probe Nuclear extract Nuclear extract (µg)

hGII-Sil TE671 0µg 15µg Antibody -ve +ve RARα RARβ RARγ RXR

Complex A

Complex B

Non-specific binding

Free probe

Fig 3 Protein factors NF-jB subunits p65, RARa and RXR family interact with the putative silencer in TE671 cells Supershift assay to iden-tify protein factors that bind to the Mut4 and Mut5 (hGII-Sil) region in the putative intronic silencing element in TE671 cells Synthetic oligo-nucleotides of hGII-Sil were annealed to form dsDNA before radiolabeling with c 32 P In each reaction, 15 lg TE671 nuclear extract was incubated with specific antibodies against different transcription factors to allow specific protein–antibody interactions The radiolabeled 24-bp DNA probe (0.2 omol, 200 000 cpm) was then incubated with the nuclear extracts for the binding reaction –ve, No antibody incuba-tion; + ve, 0.2 lg BSA applied as positive control.

the hGnRH II gene in vivo These were conducted

by transient transfection coupled to luciferase assay

using TE671 and JEG-3 cells Co-transfection of

the silencer-containing promoter constructs

pGL2-()2103 ⁄ )620) with p65 expression vector

(pCMV4-p65) led to a dramatic decrease in promoter activity in

both cell lines in a dose-dependent manner (Fig 5)

Even 0.1 lg of the p65 expression vector produced a

significant (P < 0.001) decrease in promoter activity

in both cell lines, indicating the strong potency of repression induced by p65

Effect of unliganded RARa and RXRa and retinoic acid (RA) treatments on GnRH II promoter activity

To investigate the in vivo effect of RARa and RXR

on the transcriptional regulation of the hGnRH II gene, the expression vector of human RARa (pCMX-hRARa) and⁄ or human RXRa (pCMX-hRXRa) were cotransfected with the silencer-containing promoter construct pGL2-()2103 ⁄ )620) into TE671 and JEG-3 cells Neither the transfection of RARa or RXRa nor the cotransfection of both receptors had a significant effect on the GnRH II promoter activity in TE671 cells (Fig 6) In contrast, transfection of RARa and cotransfection of RARa and RXRa significantly (P < 0.001) alleviated the gene repression of the silen-cer-containing promoter constructs in JEG-3 cells (Fig 6) Surprisingly, RXRa alone might not be responsible for the repression, as its over-expression, without RARa, did not have any significant effect on promoter activity

To further elucidate the regulation of GnRH II gene

by RARs, TE671 and JEG-3 cells were treated with all-trans retinoic acid (ATRA; a ligand of RARs) and⁄ or 9-cisRA (a ligand of RXRs) before the meas-urement of GnRH II promoter activity In TE671

Marker Input No IP p65 RAR RXR IgG -ve

274bp Antibody

Fig 4 ChIP assay of GnRH II promoter on TE671 cells It shows

binding of p65, RAR and RXR to GnRH II promoter in the context

of chromatin Chromatin from TE671 cells was formaldehyde

cross-linked and immunoprecipitated with p65, RAR and RXR antibodies

(lanes 4–6) After reversal of the cross-linking, the purified DNA

fragments were subjected to PCR using primers to amplify a

274-bp segment spanning the hGII-Sil region of the GnRH II promoter.

Immunoprecipitation without antibody (No IP, lane 3) and using a

nonspecific antibody against rabbit IgG (IgG, lane 7) was carried out

as negative controls Lane 8, a negative control for PCR (without

any DNA template) Input DNA from fragmented chromatin before

immunoprecipitation was used as a positive control (lane 2).

Lane 1, DNA size standards (100-bp DNA ladder; Invitrogen).

120

100

80

60

40

20

pGL2-(-2103/-620)

pCMV4- p65

0

+ 0µg

+ 0.1µg

+ 0.25µg

+ 0.5µg

+ 1µg

*

*

*

*

TE671 JEG-3

e (% change)

Fig 5 In vivo dose-dependent effect of over-expressing NF-jB

p65 subunit on the promoter activity of hGnRH II gene in TE671

and JEG-3 cells pCMV4-p65 expression vector was transfected to

each sample at different doses, and their effects on

pGL2-( )2103 ⁄ )620) were evaluated Values are shown as percentage

changes in relative promoter activities compared with that of the

positive control [pGL2-( )2103 ⁄ )620) without over-expression of

p65] The promoter activity of the positive control is regarded as

100% Values are mean ± S.E.M from at least three independent

experiments each performed in triplicate *P < 0.001, significant

difference from control pGL2-( )2103 ⁄ )620) p65, pCMV4-p65

expression vector.

200 180 160 140 120 100 80 60 40 20 0

hRARa+ hRXRa hRXRa hRARa –

*

*

TE671 JEG-3

Retinoic Acid Receptors

e (% change)

Fig 6 Effects of unliganded RARa and RXRa on hGnRH II promo-ter activity in TE671 and JEG-3 cells Supershift assay to identify which members of the RAR family bind to the Mut4 and Mut5 (hGII-Sil) region in the putative intronic silencing element in TE671 cells Synthetic oligonucleotides of hGII-Sil were annealed to form dsDNA before radiolabeling with c 32 P In each reaction, 15 lg TE671 nuclear extract was incubated with specific antibodies against RARs to allow specific protein–antibody interactions The radiolabeled 24-bp DNA probe (0.2 pmol, 200 000 cpm) was then incubated with the nuclear extracts for the binding reaction –ve, No antibody incubation + ve, 0.2 lg BSA applied as positive control.

cells, the application of both ATRA and 9-cisRA

sig-nificantly down-regulated the promoter function to

61.8 ± 5.0% and 57.5 ± 3.3%, respectively (Fig 6)

(P < 0.001) Interestingly, cotreatment with ATRA

and 9-cisRA had a similar repressive effect

(54.1 ± 0.8%; P < 0.001) In contrast with the

repres-sive effect in TE671 cells, neither cotreatment nor

treatment with ATRA or 9-cisRA alone had an

obvi-ous effect on the promoter activity of hGnRH II gene

in JEG-3 cells (Fig 7)

Differential effects of ligand-activated RARa and

RXRa on the promoter activity of the hGnRH II

gene in TE671 and JEG-3 cells

Over-expression of RARs together with the application

of RAs led to significant (P < 0.001) down-regulation

of the promoter activity in TE671 cells (Fig 8) A

synergistic repressive effect was observed in cells when

compared with cells only treated with retinoic acid

(RAs) (Figs 7 and 8) (P < 0.05 or P < 0.001) It was

also demonstrated that RARs alone have no effect on

the promoter activity of hGnRH II in TE671 cells

(Fig 6) The repressive effect of RAs and the

synergis-tic effect observed in Fig 7 therefore imply that

lig-and-bound RAR and RXR are responsible for the

repression of GnRH II gene in TE671 cells On the

other hand, over-expression of RARs together with

the application of RAs in JEG-3 cells led to significant

(P < 0.001) alleviation of the promoter activity from

the repressed state (Fig 8) When compared with the samples that were only transfected with RARa and⁄ or RXRa (without RA treatment), simultaneous ligand activation of RARa and RXRa (cotransfection of RARa and RXRa together with treatment of both ATRA and 9-cisRA) provided further up-regulation of the promoter activity (P < 0.001) Although a syner-gistic effect was observed in simultaneous ligand-acti-vated RARa and RXRa, RXRa alone, in either its unliganded or ligand-bound state, had no effect on the promoter activity Finally, the endogenous transcript levels of the GnRH II gene in TE671 and JEG-3 cells were further evaluated by quantitative RT-PCR Con-sistent with the results obtained from luciferase assays, ligand-bound RARa and RXRa led to a significant (P < 0.05) decrease in endogenous GnRH II gene expression in TE671 cells In contrast, ligand-bound RARa and RXRa led to a significant (P < 0.001) increase in GnRH II gene expression in JEG-3 cells (Fig 9)

Discussion

The hGnRH II gene was first identified by White and his colleagues in 1998 [9] Although GnRH II and its

140

**

**

120

100

80

60

40

20

ATRA + 9-cisRA

TE671 JEG-3

9-cisRA ATRA

– Retinoic Acids Treatment

0

e promoter activity (% change)

Fig 7 Differential effects of RA treatment on the promoter activity

of the hGnRH II gene in TE671 and JEG-3 cells In vivo effect of

over-expressing RARa and RXRa on the promoter activity of the

hGnRH II gene in TE671 and JEG-3 cells Values are shown as

per-centage changes in relative promoter activities compared with that

of the positive control [pGL2-( )2103 ⁄ )620) without over-expression

of transcription factors] The promoter activity of the positive

con-trol is regarded as 100% Values are mean ± S.E.M from at least

three independent experiments each performed in triplicate

*Signi-ficant difference (P < 0.001) versus control pGL2-( )2103 ⁄ )620).

hRARa, pCMX-hRARa expression vector; hRXRa, pCMX-hRXRa

expression vector.

250 TE671 JEG-3 200

150

100

50

pGL2-(-2103/-620) Retinoic Acids Receptor

–

hRARa*

hRXRa

*

♦

♠

♦

♦

♦

hRARa

ATRA

hRXRa hRARa*

hRXRa Retinoic Acids

Treatment

0

Fig 8 Differential effects of ligand-activated RARa and RXRa on the promoter activity of the hGnRH II gene in TE671 and JEG-3 cells In vivo effect of RARa and RXRa with their ligands, ATRA and 9-cisRA, on the promoter activity of the hGnRH II gene in TE671 cells and JEG-3 cells Values are shown as percentage chan-ges in relative promoter activities compared with that of the posit-ive control [pGL2-( )2103 ⁄ )620) without treatment] The promoter activity of the positive control is regarded as 100% Values are mean ± S.E.M from at least three independent experiments each performed in triplicate *, r,“represent significantly different val-ues (* and“, P < 0.001; * and r, P < 0.01;“and r, P < 0.05 or above) hRARa, hRARa expression vector; hRXRa, pCMX-hRXRa expression vector.

first isoform share 70% homology, they are encoded

by different gene loci and possess distinct tissue

expres-sion patterns and biological functions It is widely

expressed in various parts of the brain and the

periph-eral tissues Its expression and potent antitumor

activ-ity in various normal and cancerous cells have received

much attention [10,11] In contrast with the

well-stud-ied GnRH I, the gene regulation, expression patterns

and biological role of GnRH II are still largely

unclear Several studies have focused on the gene

activation mechanisms of hGnRH II It has been

demonstrated that expression of GnRH II can be

up-regulated by cAMP [2], gonadotropins [4] and estrogen

[3] Two AP-4-interacting E-boxes, and an Ets-like

ele-ment have been identified in the untranslated first exon

and found to be responsible for the minimal promoter

activity of GnRH II [5] In addition to these activating

elements, our research group has located a putative

silencer-like element in the first intron at )650 ⁄ )620

(relative to the +1 translation start site) [6] In the

present study, a novel cis-acting element was first

iden-tified by deletion analysis and designated hGII-Sil

(GATGCC, position at )641 ⁄ )636) It was found to

be a major responsible element that mediates the

repressive effect, which, when mutated, led to a almost complete restoration of promoter activity in two GnRH II-expressing cell lines: medulloblastoma TE671 and placental cell JEG-3

The hGII-Sil site does not show significant homo-logy with any known consensus repressor binding site The highest similarity was suggested on comparison with a novel repressive element SNOG (AATGG GGG) of human growth-associated protein 43 gene (hGAP43) with 50% homology [12] The nucleotides, ATG, in the hGAP43 SNOG element, which have been reported to be crucial for the repressive effect, coincide with the core sequence hGII-Sil identified in our study Although the protein factors of hGAP43 SNOG have not yet been identified, it is possible that the two repressive elements in hGAP43 and GnRH II gene have the same or a very similar mechanism

EMSAs and supershift assays performed in this study indicated specific protein factors that bind to the hGII-Sil region in a cell-specific manner It was dem-onstrated that NF-jB p65, RARa and RXR are responsible for forming Complex A, or, at least, are members of the Complex in TE61 cells This was con-firmed by ChIP assays which provided evidence for

in vivo interaction of these protein factors (p65, RAR and RXR) with the hGII-Sil region It is noteworthy that, when Complex A was abolished by RAR anti-body and RXR antianti-body, another specific Com-plex (ComCom-plex B) increased in intensity This may imply competition binding between RAR and⁄ or RXR and Complex B on the hGII-Sil silencing element Sim-ilar results have been reported in other in vitro mam-malian promoter studies of the Fas gene, in which multiple protein factors and cofactors were involved [13]

NF-jB subunit p65 was demonstrated in this study

to act as a potent repressor of hGnRH II promoter

in both neuronal and placental cells NF-jB com-plexes comprise homodimers or heterodimers of the family including p65 (RelA), c-Rel, RelB, p50 (p105⁄ NF-jB1), and p52 (p100 ⁄ NF-jB2) Different heterodimers bind to their specific promoters to regu-late transcription of a wide range of genes to control immune responses, cell apoptosis⁄ survival and tissue repair [14–16] A classic model of NF-jB activation involves the p50⁄ p65 heterodimer, which interacts with the jB site and the CRE of the promoters Being the active partner of the heterodimer in the nucleus, p65 is able to establish interactions with various transcription factors such as CBP⁄ p300 and histone deacetylases (HDACs) [17] It has also been reported that NF-jB

is involved in gene repression through differential

JEG-3

5

4

3

2

1

0

dimer

‡

‡

‡

GnRH II mRNA / GAPDH mRNA (r

Fig 9 Effects of ligand-activated RARa and RXRa on hGnRH II

gene expression The effect of ligand-bound RARa and RXRa on

the endogenous hGnRH II transcript level in TE671 and JEG-3 cells,

using quantitative real-time PCR analysis Cells were treated with

ATRA and 9-cisRA 24 h after the transfection of expression vectors

of RARa and RXRa The transcript level of untreated cells is defined

as 1.0 Total RNAs were harvested 24 h after drug treatment

First-strand cDNAs were prepared from total RNAs as described and

used for quantitative PCRs The hGnRH II transcript level of cells

treated with ligand-bound RARa⁄ RXRa was compared with that of

untreated cells The GnRH II mRNA ⁄ GAPDH mRNA ratio was

calculated by the 2 –DDCt method, using the GAPDH mRNA

concen-tration measured by quantitative PCR as the internal control Data

are the mean ± SEM from three experiments, each performed in

duplicate.

phosphorylation of p65 [15] or the association of

CBP⁄ p300 to form a repressor Complex [18–20] Other

studies have suggested that the binding of p65 to the

cofactors renders them unavailable for gene activation

[20–22] In the context of the GnRH II promoter,

there is a cluster of enhancing elements, including a

functional CRE, within 200 bp upstream of the

hGII-Sil site The enhancers have been suggested to be

responsible for the basal and stimulatory transcription

level of the gene [5,7] Accordingly, it is possible that

p65 is an active member of the repressor Complex at

the hGII-Sil site, which then interacts with the

activa-tors involved to down-regulate promoter activity

Furthermore, the nuclear receptor RAR⁄ RXR

het-erodimer was found to be involved in the regulation of

GnRH II gene in both cells, yet, differential response

occurs in the two cell types in the presence of the

receptor’s ligands Ligand-activated RARa and RXR

were found to contribute to the repressed expression of

GnRH II in the neuronal TE671 cells; ligand-activated

RARa, on the other hand, up-regulated gene

expres-sion in JEG-3 cells To our knowledge, this study is

the first to demonstrate the presence of differential

transcriptional regulation of the GnRH II gene in

dif-ferent GnRH II-expressing human cell types RAs play

important roles in development, differentiation, and

homeostasis in a tissue-specific manner [23,24] The

actions of RAs are highly diversified because their

sig-nals can be transduced through different RARs In

addition, the nuclear receptors are able to cross-talk

with cell surface receptor signaling pathways, and the

RARs and RXRs can interact with multiple

coactiva-tors and⁄ or corepressors These combinatorial effects

result in the pleiotropic effects of RAs For

RA-induced genes, unliganded RAR⁄ RXR heterodimers

bind to corepressors such as the silencing mediator of

retinoid and thyroid hormone receptor (SMRT)

and⁄ or nuclear receptor corepressor (N-CoR) SMRT

and N-CoR in turn function as bridging factors that

recruit other coregulator proteins to form a larger

corepressor Complex [25,26]

Conversely, addition of hormone agonist leads to

the release of corepressor Complex by the receptor,

which then recruits a series of coactivator proteins,

such as steroid receptor coactivator 1 (SRC-1),

gluco-corticoid receptor interacting protein 1 (GRIP1),

acti-vator of thyroid and retinoic acid receptor (ACTR)

and p300⁄ cAMP response element binding

protein-binding protein (CBP⁄ p300) [26–28] This may explain

the up-regulation (alleviation of the repressing effect of

GII-Sil) of the hGnRH II promoter by the

ligand-acti-vated RAR⁄ RXR heterodimer in JEG-3 cells It is

noteworthy that RXR, even in its ligand-activated

state, did not induce up-regulation of the gene without over-expression of RAR In contrast, ligand-activated RAR itself was able to up-regulate the gene to a signi-ficant level without the aid of RXR This phenomenon

of RXR acting as the silent or ‘nonpermissive’ partner

in an RXR⁄ nuclear receptor heterodimer has often been observed These dimers do not respond to RXR ligands but are only sensitive to RAR ligand activation [23,29–31] In the case of RAR⁄ RXR, it was observed that RXR can acquire the ability to respond to its own ligand only if RAR is activated by ATRA before-hand In this situation, simultaneous addition of lig-ands for both RAR and RXR leads to synergistic activation of the heterodimer [31,32], which agrees with the observation in the present study Moreover, quantitative RT-PCR analysis demonstrated this up-regulation by ligand-activated RAR⁄ RXR at the tran-scriptional level of the hGnRH II gene Therefore, the RAR⁄ RXR heterodimer interacts with the hGII-Sil silencer and is probably responsible for its gene-repres-sive effect in JEG-3 placental cells

However, contrary to the results observed in placen-tal cells and the current paradigm of the role of RA-induced activation, the application of RAs and introduction of ligand-activated RAR⁄ RXR hetero-dimer further down-regulated the hGnRH II gene in TE61 cells In fact, there are examples of ligand-bound nuclear repressors exerting transrepression, rather than activation, over their regulating genes For instance, thyroid hormone receptor b, which is closely related to RARs, is found to markedly repress the thyroid-stimu-lating hormone b promoter after being bound by its cognate ligand thyroid hormone, via HDAC recruit-ment [33,34] Indeed, there is increasing evidence that the ligand–nuclear receptor–corepressor relationship is often not a simple switching on–off model The func-tions of the ‘corepressors’ may depend on cell type, combinations of neighboring regulatory factors, and the phase of the cell cycle [35] Most corepressors have been found to be promiscuously but not specifically expressed [24,36] It has also been reported that coacti-vators can act as corepressors of liganded RAR and thyroid hormone receptor in the context of epidermal keratin genes, and vice versa [24] It has long been known that the interaction between coregulators and nuclear receptors (liganded versus unliganded) is deter-mined by the cis-acting elements Hence, different tran-scriptional responses can be elicited in various promoter contexts even when the same ligands and receptors are involved [24,37–39] Another possible hypothesis on the ligand-dependent transrepression mechanism of hGnRH II observed here may involve the newly discussed theory of ligand-dependent

corepressor (L-CoR) L-CoR is a distinct class of

core-pressor which causes gene repression through

ligand-bound nuclear receptors [35,40,41] Fernandes et al

[35] found a wide expression pattern of L-CoR in

var-ious human adult and fetal tissues, including kidney,

placenta, cerebellum and corpus callosum of the brain

at the transcription level L-CoR interacts with both

HDAC2 and another corepressor C-terminal binding

protein, mediating strong gene repression through both

HDAC-dependent and HDAC-independent

mecha-nisms Such a ligand-induced repression mechanism is

important as a means of attenuating and

counterbalan-cing hormone-induced transactivations, acting

transi-ently as part of a cycle of cofactors at the target

promoters, and allowing hormone-induced target gene

repression [35,41]

The present study discusses the transcriptional

regu-lation of the hGnRH II It is also the first report of

differential regulation of the gene in two GnRH

II-expressing cell types It may give useful cues about the

expression pattern of the largely unknown GnRH II

We also provide evidence of the newly discussed

mech-anism of L-CoR Little is still known about the

molecular basis of ligand-induced transrepression

[35,41,42] As knowledge on this topic accumulates, a

more detailed elucidation of GnRH II transcriptional

regulation is expected

Methods and Materials

Cell lines

TE671 (human medulloblastoma cell line) and JEG-3

(human placental cell line) were maintained in Dulbecco’s

modified Eagle’s medium (Gibco-BRL, Invitrogen, Grand

Island, NY, USA) and Medium 199 (Gibco-BRL,

Invitro-gen), respectively, supplemented with 10% fetal bovine

serum (Gibco-BRL, Invitrogen) All cells were incubated at

37C with 5% CO2 in medium supplemented with

100 UÆmL)1 penicillin G and 100 lgÆmL)1 streptomycin (Life Technologies, Carlsbad, CA, USA)

Promoter-luciferase constructs

The full-length hGnRH II promoter construct

pGL2-2103⁄ +1-Luc was generated by PCR amplification from human genomic DNA using sequence-specific primers followed by subsequent cloning into the promoter-less pGL2-Basic vector (Promega, Madison, WI, USA) [5] The deletion mutants p-2103⁄)650 Luc, which contains the core promoter region and enhancing elements but lacks a silen-cing element, and p-2103⁄)620 Luc, which also includes the silencing element, were generated by PCR amplification using sequence-specific primers with p-2103⁄ +1 Luc as the template All mutant clones of the 30-bp silencer (scanning mutants Mut 1–10) were generated by PCR amplifications using mutagenic reverse primers and GLprimer1 forward primer with wild-type pGL2-()2103 ⁄ )620) construct as the template (Table 1) The purified PCR products were sub-cloned into a pGL2-Basic vector (Promega) at KpnI and HindIII restriction sites All mutant plasmids were verified

by big dye terminator DNA sequencing analysis Plasmid DNAs used for transfection experiments were prepared using the Nucleobond AX preparation kit (Macherey-Nagel, Duren, Germany) Enzymes and oligoprimers were purchased from Life Technologies and the Genome Research Centre, University of Hong Kong, respectively

Transfection and drug treatments

Two days before transfection, cells were seeded on to a 35-mm well (six-well plate; Costar, San Diego, CA, USA) The seeding densities used for TE671 and JEG-3 were 1.5· 105

cells⁄ well and 2.5 · 105 cells⁄ well, respectively The transfection mixture containing 1 lg promoter–lucif-erase constructs, 0.5 lg pSV-b-gal or pCMV4-b-gal, an appropriate amount of GeneJuice Transfection Reagent (Novagen, Darmstadt, Germany) and 500 lL serum-free medium was prepared For assays of the effect of NF-jB

Table 1 Primer sequences for construction of scanning mutants The mutated nucleotides are underlined.

on promoter activity, 0.25 lg pCMV4-p50 and⁄ or

pCMV4-p65 was cotransfected per well For assays

inves-tigating the effect of RAR and RXR on promoter

activ-ity, 0.5 lg pCMX-hRARa and⁄ or pCMX-hRXRa was

cotransfected per well Appropriate amounts of the empty

vector, pcDNA3.1, were cotransfected so that the same

amounts of DNA were transfected in each sample The

500 lL transfection mixture was added to 1.5 mL 10%

fetal bovine serum supplemented medium per well After

37C incubation for 48 h, cell lysates were prepared by

first washing the cells twice with ice-cold NaCl⁄ Pifollowed

by the addition of 200 lL reporter lysis buffer according

to the manufacturer’s protocol (Promega) For assays

investigating the effect of RAs on promoter activity, 1 lm

ATRA and⁄ or 9-cisRA were added to the seeded cell

cul-tures in 35-mm wells, 24 h after the transient transfection

of promoter–luciferase constructs and⁄ or human RA

expression vectors (0.5 lg pCMX-hRARa and⁄ or

pCMX-hRXRa) After 24 h of drug treatment, cell lysates were

harvested as described previously

Luciferase assay

A 100-lL sample of luciferase substrate solution (Promega)

was automatically injected into 20 lL cell lysate, and

luciferase activity was measured as light emission using a

luminometer (Lumat LB9507; EG & G Berthold, Bad

Wildbad, Germany) b-Galactosidase activity was

deter-mined by incubating the cell lysate (50 lL) in 100 mm

sodium phosphate buffer, pH 7.3, containing 1 mm MgCl2,

50 mm 2-mercaptoethanol and 0.7 mgÆmL)1 o-itrophenyl

galactoside for 15 min at 37C A420was measured using a

spectrophotometer (U-2800; Hitachi High-Technologies

Corporation, Tokyo Japan) For each transfection assay,

luciferase activity was determined and normalized on the

basis of b-galactosidase activity Each plasmid was tested

at least nine times in three separate experiments

Electrophoretic mobility-shift and supershift

assays

Nuclear proteins were extracted from TE671 cells and

JEG-3 cells as described previously [4JEG-3] The double-stranded

probe corresponding to hGII-Sil was end labeled using the

Ready-To-Go T4 polynucleotide kinase labeling kit

(Amer-sham Pharmacia Biotech, Arlington Heights, IL, USA) with

[c-32P]ATP (5000 CiÆnmol)1; Amersham Pharmacia

Bio-tech) Unlabeled nucleotides were removed by passing the

sample through a microspin column G-25 (Amersham

Pharmacia Biotech) at 3000 g Binding reactions were

per-formed by incubating the 10 mg nuclear extracts with the

binding buffer (10 mm Tris⁄ HCl, pH 7.5, 0.1 mm EDTA,

1 mm magnesium acetate, 0.1 mm dithiothreitol, 5%

gly-cerol, 60 mm KCl), 1 lg poly(dI-dC), and 0.5 pmol

(200 000 cpm) labeled probe for 15 min at room

tempera-ture For competition assays, 50-fold, 100-fold and 200-fold molar excess of the unlabeled wild-type oligonucleotide, hGII-Sil (5¢-CCTCCACCCCTGAACCATGCCAGA-3¢), and nonspecific L8 oligonucleotide were used For the supershift assay, specific antibodies (rabbit polyclonal IgG; Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA) against known transcription factors were incubated with the nuclear extract in the presence of 1· binding buffer at room temperature for 45 min before binding to the labeled probe Free and bound probes were separated by electro-phoresis for 2 h at 200 V in a 5% nondenaturing poly-acrylamide gel in 0.5· Tris ⁄ borate ⁄ EDTA buffer (45 mm Tris-borate, 0.1 mm EDTA) After electrophoresis, the gel was dried and autoradiographed (Biomax MR film; East-man Kodak Co., Rochester, NY, USA) for 16 h at )70 C with intensifiers (Amersham Pharmacia Biotech)

ChIP assay

ChIP assays were performed essentially as described by Lee

et al [44] TE671 cells were cross-linked with 1% formal-dehyde Cells were harvested by centrifugation and resus-pended in lysis buffer (1% SDS, 10 mm EDTA, 50 mm Tris⁄ HCl, pH 8.1, 1 mm phenylmethanesulfonyl fluoride,

1 lgÆmL)1 aprotinin and 1.5 lgÆmL)1 pepstatin A) After sonication in Sonifier 450 (Branson, Danbury, CT, USA),

4 lg antibody and 20 lL Protein G⁄ agarose (Santa Cruz Biotechnology) were added to precipitate the DNA–protein complex Precipitated DNA–protein Complex was washed in the ChIP buffer (0.1% SDS, 1% Triton X-100, 0.1% sodium deoxycholate, 140 mm NaCl, 1 mm phenylmethanesulfonyl fluoride, 1 lgÆmL)1aprotinin and 1.5 lgÆmL)1pepstatin A) and eluted in the elution buffer (1% SDS and 0.1 m

NaH-CO3) The mixture was incubated at 65C for 4 h to reverse the formaldehyde cross-linking Protein was removed by pro-teinase K digestion (200 lgÆmL)1) and phenol⁄ CHCl3

extraction The extracted DNA was used for PCR using forward (GnRH II-F, 5¢-GGGTGGAGCTGCCTGGTC TATA-3¢) and reverse (GnRH II-R, 5¢-CAGGGGCAACA AGCACAAGA-3¢) primers

Quantitative RT-PCR

Transfected cells were treated with the drug 1 day after transfection for 24 h as described above, and total RNA was isolated using the TriPure Isolated Reagent (Roche Molecular Biochemicals, Basel, Switzerland) Total RNA (5 lg) was reverse-transcribed with an oligo-dT primer and Superscript III reverse transcriptase (Invitrogen) One-fifth

of the first-strand cDNA was used for real-time quantita-tive PCR The transcript levels of GnRH II were measured with the SYBR Green Master Mix (Applied Biosystems, Foster City, CA, USA) with specific primers; for GnRH, forward primer 5¢-GCCCACCTTGGACCCTCAGAG-3¢ and reverse primer 5¢-CGGAGAACCTCACACTTTAT