Báo cáo khoa học: Cloning and functional analysis of 5¢-upstream region of the Pokemon gene pptx

Deletion analysis and a DNA decoy assay indicated that the NEG-U and NEG-D elements were involved in negative regulation of the Pokemon promoter, whereas the POS-D element was mainly res

of the Pokemon gene

Yutao Yang, Xiaowei Zhou, Xudong Zhu, Chuanfu Zhang, Zhixin Yang, Long Xu and

Peitang Huang

Laboratory of Protein Engineering, Beijing Institute of Biotechnology, China

The poxvirus and zinc finger (POZ) domain, formerly

termed the broad complex, tramtrack and bric-a-brac

(BTB) domain, was initially characterized in the

Drosophila proteins broad complex, tramtrack and

bric-a-brac [1] It is  120 amino acids long and

usu-ally exists in a few transcriptional repression complexes

[2] The POZ domain is highly conserved from yeast to

humans, and is involved in many critical cellular

pro-cesses such as development [3,4], oncogenesis [5,6],

apoptosis [7] and ion channel activity [8]

More than 200 proteins have been found in

associa-tion with the POZ domain [9], and they are usually

grouped according to their distinct C-terminal

struc-tures, such as the zinc finger motif, basic zipper motif,

actin-binding repeats, kech domains and ion channel

motifs [10] Proteins containing the POZ domain and

zinc finger motif are termed POZ-ZF or POK proteins Via the POZ domain, many POK proteins can recruit transcriptional co-repressors such as nuclear co-repres-sor (N-CoR), silencing mediator of retinoic acid, thyroid hormone receptor (also known as N-CoR2), mSin3A and histone deacetylases to the target gene promoter regions, thereby decreasing these gene tran-scriptional activities [11–14]

Currently,  60 POK genes have been identified in the human genome [15] Many of them, such as PLZF, BCL-6, Zbtb7 and HIC1, are involved in development, differentiation and oncogenesis [2] Pokemon, the POK erythroid myeloid ontogenic factor, was previously known by several names (LRF, OCZF and FBI-1) and was originally identified as a protein that binds specifi-cally to the inducer of short transcripts (IST) element

Keywords

DNA decoy; element; mutation; Pokemon;

promoter

Correspondence

P Huang, Laboratory of Protein Engineering,

Beijing Institute of Biotechnology,

Beijing 100071, China

Fax ⁄ Tel: +86 10 6381 0272

E-mail: amms832@126.com

(Received 15 October 2007, revised 12

February 2008, accepted 18 February 2008)

doi:10.1111/j.1742-4658.2008.06344.x

Pokemon, the POK erythroid myeloid ontogenic factor, not only regulates the expression of many genes, but also plays an important role in cell tumorigenesis To investigate the molecular mechanism regulating expres-sion of the Pokemon gene in humans, its 5¢-upstream region was cloned and analyzed Transient analysis revealed that the Pokemon promoter is constitutive Deletion analysis and a DNA decoy assay indicated that the NEG-U and NEG-D elements were involved in negative regulation of the Pokemon promoter, whereas the POS-D element was mainly responsible for its strong activity Electrophoretic mobility shift assays suggested that the NEG-U, NEG-D and POS-D elements were specifically bound by the nuclear extract from A549 cells in vitro Mutation analysis demonstrated that cooperation of the NEG-U and NEG-D elements led to negative regu-lation of the Pokemon promoter Moreover, the NEG-U and NEG-D ele-ments needed to be an appropriate distance apart in the Pokemon promoter in order to cooperate Taken together, our results elucidate the mechanism underlying the regulation of Pokemon gene transcription, and also define a novel regulatory sequence that may be used to decrease expression of the Pokemon gene in cancer gene therapy

Abbreviations

BTB, broad complex, tramtrack and bric-a-brac domain; EMSA, electrophoretic mobility shift assay; Pokemon, POK erythroid myeloid ontogenic factor; POZ, poxvirus and zinc finger domain; SRE, sterol regulatory element; SREBP, SRE-binding protein.

on the HIV-1 genome [16] It was first termed the

fac-tor binding to IST-1 (FBI-1) and is encoded by the

Zbtb7 gene Pokemon not only regulates the HIV-1

Tat transactivation process [17,18], but is also involved

in human and murine adipogenesis [19] It acts as a

transcription factor and regulates the expression of

many gene-encoding proteins such as extracellular

matrix collagen types I, II, IX, X and XI, fibronectin,

elastin, human cartilage oligomeric matrix protein

[20,21], ARF tumor suppressor [22], and the c-fos and

c-myc oncoproteins [23] This activity is due to its

capacity to bind to the consensus sequence within the

promoters of these target genes Furthermore, Pokemon

can regulate the expression of other genes via an

inter-action between its POZ domain and other important

transcription factors such as Sp-1 and the p65 subunit

of NF-jB or IjB [24,25]

Pokemon is also a repressor of the ARF tumor

sup-pressor gene and is a central regulator in oncogenesis

Overexpression of the Pokemon gene can decrease

expression of the ARF gene, which in turn results in

p53 degradation and oncogenic transformation

Con-versely, depletion of the Pokemon gene both inhibited

oncogene-mediated cellular transformation and

induced cell senescence and apoptosis [22,26]

There-fore, Pokemon plays a crucial role in cell tumorigenesis

and may be a potential therapeutic target for human

cancer therapy

Although considerable work has been done to

eluci-date the biological functions of the Pokemon, its

regu-lation mechanism has not been reported In order to

identify the elements that regulate Pokemon gene

expression, we cloned and characterized the Pokemon

promoter Our results suggest a role for two strongly

negative elements and one positive element in

regula-tion of the Pokemon gene However, the two negative

elements could not individually exhibit the negative

regulatory activity; they required mutual cooperation

with each other in order to negatively regulate the

Pokemon promoter In conclusion, our studies are the

first to elucidate transcriptional regulation mechanism

of the Pokemon gene, and this will be beneficial for

gene therapy in cancer

Results

Cloning of the 5¢-upstream region of the

Pokemon gene

To identify the regulatory sequences that control

expression of the Pokemon gene, a 2204-bp section of

the 5¢-upstream region of the Pokemon gene was

cloned by PCR using human genomic DNA as the

template Figure 1 shows the nucleotide sequence of the 2204-bp promoter region and a short stretch of the transcription region The translation start site was des-ignated as +1, and the transcribed region was shaded Some reports showed that the Pokemon gene can be expressed in different cell lines and different human tis-sues [25,27]; later reports also confirmed these results [22,26] To examine whether the Pokemon promoter can drive reporter gene expression in a similar manner, the 2204-bp promoter linked to the luciferase reporter gene was used in transient transfection studies with dif-ferent cell lines Luciferase assays showed that the Pokemon promoter could direct luciferase expression

in HeLa, A549, DU145, Jurkat and HepG2 cells, whereas the pGL3-basic construct could not (Fig 2); this suggests that the Pokemon promoter can drive reporter gene expression in different cell lines, which was in agreement with the expression patterns of the Pokemon gene [22,25–27] Because Pokemon is also highly expressed in lung and prostate carcinomas, A549 and DU145 cells were used to study the regula-tion mechanisms of the Pokemon gene

Computer analysis of putative transcription factor-binding sites

For a rough understanding of the regulation of the Pokemon gene, the 2204-bp section of its 5¢-upstream region was analyzed for putative cis-elements in the TRANSFAC 7.0 database (http://www.generegulation com/pub/databases.html) [28] After scanning the TRANSFAC 7.0 database, we found that the putative TATA and CCAAT sequences are absent in the upstream region of the Pokemon gene; however, some transcription factor-binding sequences, including Sp1, AP-1, AP-2, PU.1, Hb, CBF-1, GATA-1 elements and p53-binding sites are present in the promoter (Fig 1), implying their potential roles in the regulation of the Pokemon gene

Deletion analysis of the Pokemon promoter

To broadly determine the main regulatory regions in the Pokemon promoter, we created five 5¢-deletion con-structs; the activities of these deletion constructs were measured in A549 and DU145 cells As shown in Fig 3A, luciferase activity was markedly reduced when the region from )837 to )560 was deleted, but was dramatically increased when the region from )560 to )233 was deleted These results demonstrated the pos-sible presence of some potential positive elements in the region from)837 to )560 and negative elements in the region from)560 to )233

To determine the regulatory region in the Pokemon

promoter more accurately, we performed further

5¢-deletion analysis with the regions from )837 to

)560 and )560 to )233 Five 5¢-deletion constructs

were constructed for the region from )837 to )560

and used in transient transfection studies As shown in Fig 3B, only the deletion from )580 to )560 resulted

in a moderate reduction in luciferase activity; it reduced the luciferase activity of A549 cells by 4.6-fold and that of DU145 cells by 4.2-fold compared with

Fig 1 Nucleotide sequence of the 5¢-upstream region of the Pokemon gene The upstream region of the Pokemon gene containing the pro-moter and a short stretch of the transcribed region is shown The nucleotides are numbered on the left, with the translation start site desig-nated as +1 The translation start site is indicated by an arrowhead The transcribed region is shaded The POS-D, NEG-U and NEG-D elements are boxed The putative cis-elements are underlined The TRANSFAC database was used to identify putative cis-elements in the 5¢-upstream region of the Pokemon gene.

F-580, suggesting the presence of potential positive

element(s) in this region Six 5¢-deletion constructs

were constructed for the region from )560 to )233,

and their activities were measured in A549 and DU145

cells Figure 3C shows that deletion of the region from

)560 to )542 resulted in a remarkable increase in

luciferase activity, by 25-fold in A549 cells and

24.6-fold in DU145 cells compared with F-560,

sug-gesting the presence of a strong negative element in

this region This negative element was termed NEG-U

The F-233 construct still directed reporter gene

expression to a great degree, and some essential

ele-ments might be responsible for this property Further

deletion analysis showed that the region from )83 to

)71 was responsible for the strong activity of the

0

1

2

3

4

5

6

7

8

9

HeLa A549 DU145 Jurkat HepG2

Fig 2 The 2204-bp section of the Pokemon promoter can drive

luciferase gene expression in different cell lines Different cell lines

were transfected with the 2204-bp section of the Pokemon promoter

construct or the pGL3-basic construct Solid bars represent the

2204-bp stretch showing Pokemon promoter construct activity,

and open bars represent pGL3-basic construct activity The values

are the mean ± SE for three independent experiments performed in

triplicate and are normalized to Renilla luciferase activity.

A

B

C

D

LUC activity

LUC activity

LUC activity

LUC activity

Fig 3 5¢-Deletion analysis of the Pokemon promoter

Progres-sively truncated fragments of the upstream region of the Pokemon

gene were inserted into the pGL3-basic vector and their ability to

activate transcription of the luciferase gene was assessed in A549

and DU145 cells The values are the mean ± SE for three

indepen-dent experiments performed in triplicate and are normalized to

Renilla luciferase activity (A) Rough characterization of the

Poke-mon promoter using the larger, gradually truncated fragment from

)2220 to )233 (B) Refined analysis of the Pokemon promoter

using the smaller, progressively truncated fragment from )837 to

)560 (C) Refined analysis of the Pokemon promoter using the

smaller, progressively truncated fragment from )560 to )233.

(D) Refined analysis of the Pokemon promoter using the smaller,

progressively truncated fragment from )233 to )71.

Pokemon promoter When this region was removed,

luciferase activity was decreased by 25.6-fold in the

A549 cells and 23.4-fold in the DU145 cells compared

with F-83, indicating that the region from )83 to )71

is necessary for strong expression of the Pokemon

pro-moter in both A549 and DU145 cells (Fig 3D); we

termed this positive element as POS-D

Importance of the POS-D element for the strong activity of the Pokemon promoter

Because the POS-D element plays an important role in the strong activity of the Pokemon promoter, it may

be the target of some transcriptional factors To deter-mine the presence of binding sites for transcriptional factors in this element, we performed electrophoretic mobility shift assays (EMSAs) with A549 cell nuclear extract Figure 4A shows the formation of complexes when wild-type double POS-D was used as a probe and incubated with the nuclear extracts The specificity

of the complexes was confirmed by incubation with a 50-fold excess of unlabeled wild-type double POS-D However, mutant POS-D did not compete with the labeled wild-type probe, suggesting that the POS-D element is specifically recognized by nuclear proteins from A549 cells

Because the POS-D element is responsible for the strong activity of the Pokemon promoter, we specu-lated whether mutation of the POS-D element would result in a decrease in the activity of the promoter We mutated a 9-bp section of the POS-D element in the F-233 construct (MF-233) and transfected MF-233 and F-233 into A549 and DU145 cells, respectively Luciferase assays showed that MF-233 displayed lower luciferase activity than F-233 (Fig 4B) In addition,

we also examined the function of the POS-D element

by using the DNA decoy technique Our results showed that introduction of the POS-D decoy could efficiently suppress the F-233 activity, whereas the mutant POS-D decoy could not (Fig 4C) All these results suggest that POS-D is an essential regulatory element that is responsible for the strong activity of the Pokemon promoter

Role of the NEG-U element in the negative regulation of the Pokemon promoter 5¢-Deletion analysis showed that the NEG-U element was involved in the negative regulation of the Pokemon

A

B

C

Fig 4 The POS-D element is necessary for strong activity of the Pokemon promoter (A) EMSA was performed with 32 P-labeled POS-D element in the absence or presence of the wild-type POS-D element or mutant POS-D element at the molar excess indicated above each lane (B) Activities of F-233 and MF-233 in A549 and DU145 cells (C) Activities of F-233 and varying amounts of the decoy oligonucleotides in A549 and DU145 cells WP-decoy cates the wild-type POS-D oligonucleotide, while MP-decoy indi-cates the mutant POS-D oligonucleotide The values are the mean ± SE for three independent experiments performed in tripli-cate and are normalized to Renilla luciferase activity.

promoter To examine whether the NEG-U element

shows nuclear protein-binding activity, we synthesized

double NEG-U and mutant double NEG-U elements,

and then performed EMSAs with A549 cell nuclear

extracts As shown in Fig 5A, specific complexes were

observed with the labeled wild-type probe; moreover,

250-fold excess of the unlabeled wild-type probe

almost entirely eliminated complex formation, whereas

250-fold excess of the unlabeled mutant probe did not

Interestingly, we found that the mutated element can

compete with the wild-type NEG-U element to some

extent; this suggests that the corresponding nuclear

factor may bind to the region between the mutation

site and the marginal sequence in the mutant probe

However, our decoy analysis showed that the MNEG-U

decoy had almost no effect on the activity of F-560,

indicating that the mutated competitor had only weak

nonspecific binding capacity for proteins in the A549

nuclear extract (Fig 5B)

Because the NEG-U element lent a strong negative

character to the Pokemon promoter, we speculated

whether it could also decrease the activity of the SV40

promoter NEG-U and mutant NEG-U elements were

cloned into the KpnI⁄ XhoI sites of the pGL3-control

plasmid in both the normal and reverse orientations,

and the resultant constructs were used in transient

transfection studies Interestingly, as shown in Fig 5C,

both normally and reversely oriented NEG-U elements

increased the activity of the SV40 promoter, whereas

the mutant element did not Therefore, the NEG-U

element could exhibit the negative regulatory function

only in a special DNA context

Role of the NEG-D element in the negative

regulation of the Pokemon promoter

The NEG-U element alone cannot negatively regulate

the function of the Pokemon promoter, therefore, we

proposed that it might interact with other downstream regulatory elements to exhibit negative activity To accurately locate the region that can cooperate with the NEG-U element, we performed 3¢-deletion analysis

in the region from )560 to )88 All the deletion con-structs of this region contained the region between)88 and )17 but different internal deletion fragments As

B

Competitor Nuclear protein

Free probe

Complex

0 0 50× 250× 50× 250×

- + + + + +

C

SV40-Promoter LUC

SV40-Promoter LUC

LUC

LUC

LUC NEG-U

SV40-Promoter MNEG-U

SV40-Promoter MNEG-U

SV40-Promoter NEG-U

0 50 100 150 MU-I

MU-F WU-I WU-F W

LUC activity

A549 DU145

F-560 (ng) 300 300 300 300 300 WU-decoy (µg) 0 1 2 0 0 MU-decoy (µg) 0 0 0 1 2

0 1 2 3 4 5 6

A549 DU145

Fig 5 The NEG-U element is involved in the negative regulation of

the Pokemon promoter (A) EMSA was performed with 32 P-labeled

NEG-U element in the absence or presence of the wild-type NEG-U

element or mutant NEG-U element at the molar excess indicated

above each lane (B) Activities of F-560 and varying amounts of the

decoy oligonucleotide in A549 and DU145 cells WU-decoy

indi-cates the wild-type NEG-U oligonucleotide, wherreas MU-decoy

indicates the mutant NEG-U oligonucleotide (C) The left-hand panel

shows the different chimeric constructs used to test the effect of

the NEG-U element on the SV40 promoter The right-hand panel

shows the results of luciferase activity assays for different chimeric

constructs in A549 cells and DU145 cells The values are the

mean ± SE for three independent experiments performed in

tripli-cate and are normalized to Renilla luciferase activity.

shown in Fig 6A, deleting the region from )127

to )88 resulted in a significant increase in luciferase

activity compared with F-560, whereas deleting the

region from )107 to )88 resulted in a slight increase

in luciferase activity, indicating that the region from

)127 to )107 is also involved in the negative

regula-tion of the Pokemon promoter; we termed this region

NEG-D In addition, progressive deletion of the region

from )156 to )473 resulted in only slight changes in

luciferase activity compared with T-127, further

con-firming the importance of the NEG-D element In

order to fully examine the function of the NEG-D

ele-ment, we performed EMSA and NEG-D decoy analy-sis EMSA showed that the NEG-D element could be bound specifically by the nuclear extract from A549 cells (Fig 6B) Decoy analysis demonstrated that the NEG-D decoy could increase the activity of F-560; a similar observation was made with regard to the NEG-U decoy-treated cells (Fig 6C) These results indicate that the NEG-D element is also necessary for negative regulation of the Pokemon promoter

Because the NEG-D element also lent a strong nega-tive character to the Pokemon promoter, we speculated whether it might decrease the activity of the SV40

Fig 6 The NEG-D element is involved in the negative regulation of the Pokemon promoter (A) The left-hand panel shows 3¢-deletion con-structs in the region between )560 and )88 of the Pokemon promoter All these constructs contained the region between )88 and )17 of the Pokemon promoter but had different internal deletion fragments The right-hand panel shows the results of the luciferase activity assays

of the 3¢-deletion constructs in A549 and DU145 cells (B) EMSA was performed with 32

P-labeled NEG-D element in the absence or pres-ence of the wild-type NEG-D element or mutant NEG-D element at the molar excess indicated above each lane (C) Activities of F-560 and varying amounts of the decoy oligonucleotide in A549 and DU145 cells WD-decoy indicates the wild-type NEG-D oligonucleotide, whereas MD-decoy indicates the mutant NEG-D oligonucleotide (D) The left-hand panel shows the different chimeric constructs used to test the effect of the NEG-D element on the SV40 promoter The right-hand panel shows the results of the luciferase activity assays for different chi-meric constructs in A549 and DU145 cells The values are the mean ± SE for three independent experiments performed in triplicate and are normalized to Renilla luciferase activity.

promoter NEG-D and mutant NEG-D elements were

also cloned into the KpnI⁄ XhoI sites of the

pGL3-con-trol plasmid in both the normal and reverse

orienta-tions, and the activities of the resultant construct were

assayed As shown in Fig 6D, both normal- and

reverse-oriented NEG-D elements increased the

activ-ity of the SV40 promoter, whereas the mutant element

did not; this is similar to the function of the NEG-U

element, suggesting that a single NEG-D element alone

cannot exhibit negative activity

The NEG-U element cooperates with the NEG-D

element to promote negative regulation of the

Pokemon promoter

To further characterize the effects of the NEG-U and

NEG-D elements on the negative regulation of the

Pokemon promoter, mutations of the NEG-U and

NEG-D elements, alone or in combination, were

cre-ated in F-560 and transiently transfected into A549

and DU145 cells As shown in Fig 7A, mutation of

the NEG-U element alone resulted in a significant

increase in luciferase activity; a similar result was also

observed in the construct that only harbored the

mutated NEG-D element Interestingly, mutations in

both sites also led to a remarkable increase in

lucifer-ase activity These results indicate that both the

NEG-U and NEG-D elements are essential for

negative regulation of the Pokemon promoter

To examine the impact of the length between the two elements on inter-region synergism, the effect of deletions in the intervening sequence was evaluated Our results showed that the inhibition of the synergis-tic activity of the NEG-U and NEG-D elements was almost abolished in D-254 and D-106 (Fig 7B), thus suggesting that the inhibition of synergism may require the NEG-U and NEG-D elements to be located at a certain appropriate distance from each other

To further determine the cooperation between the NEG-U and NEG-D elements, we performed DNA

A

B

C

Fig 7 The NEG-U and NEG-D elements are necessary for the

neg-ative regulation of the Pokemon promoter (A) The left-hand panel

shows the F-560 and mutant constructs The right-hand panel

shows the results of the luciferase activity assays for all the

con-structs in A549 and DU145 cells M-U indicates that the NEG-U

ele-ment was mutated in F-560, M-D indicates that the NEG-D

element was mutated in F-560 and M-B indicates that both the

NEG-U and NEG-D elements were mutated in F-560 (B) The

left-hand panel shows F-560 and different mutant constructs harboring

shorter intervening sequences (254 and 106 bp) between the

NEG-U and NEG-D elements The distance between the NEG-NEG-U and

NEG-D elements was 253 and 106 bp in D-254 and D-106,

respec-tively The right-hand panel shows the results of the luciferase

activity assays for different constructs in A549 and DU145 cells.

(C) The effects of 2 lg of the NEG-U decoy, 2 lg of the NEG-D

decoy and a combination of 1 lg of each decoy on the activity of

F-560 The open ellipse indicates the wild-type NEG-U element,

whereas the solid ellipse indicates the mutant NEG-U element; the

open triangle indicates the wild-type NEG-D element, whereas the

solid triangle indicates the mutant NEG-D element ‘U’ indicates

the NEG-U decoy, ‘D’ indicates the NEG-D decoy and ‘U+D’

indi-cates the combination of the decoys The values are the

mean ± SE for three independent experiments performed in

tripli-cate and are normalized to Renilla luciferase activity.

decoy experiments using 2 lg of the NEG-U decoy,

2 lg of the NEG-D decoy, and a combination of 1 lg

each of the NEG-U and NEG-D decoys Our results

demonstrated that the F-560 activity is increased more

by the combination of the NEG-U and NEG-D decoys

than by the individual NEG-U and NEG-D decoys

(Fig 7C) These data demonstrate that the Pokemon

promoter can only be negatively regulated when the

NEG-U element cooperates with the NEG-D element

Discussion

Pokemon, a member of the POK protein family, plays

an important role in cell development, differentiation

and oncogenesis Abrogation of Pokemon often leads

to cell-cycle arrest and cellular senescence and

apopto-sis However, overexpression of Pokemon will lead to

reduced levels of the tumor suppressor gene ARF,

resulting in degradation of the wild-type nuclear p53

and oncogenic transformation [22,26] Although a

con-siderable amount of work has been done characterizing

the function of Pokemon, very little is known about

the mechanism that governs its expression In this

study, we performed deletion analysis, mutation

analy-sis, as well as decoy assays, and found that the NEG-U,

NEG-D and POS-D elements play important roles in

regulation of the Pokemon promoter; this helps us

understand the transcriptional mechanism of the

Pokemongene

In humans, the Pokemon gene localizes in syntenic

chromosomal regions (19p13.3), and is widely

expressed in adult tissues and cell lines [27] Reports

have shown that alternative splicing and alternative

promoters play important roles in the regulation of

some genes [29–31] Interestingly, our previous studies

also showed that the Pokemon transcripts could be

alternatively spliced, resulting in the formation of

mRNAs with four different 5¢-untranslated regions

Matching the nucleotide sequences of four first exons

to human genomic DNA showed that four alternative

first exons were located at )11 596, )10 224, )9109

and )17 bp upstream of the translation start site of

the Pokemon gene, suggesting that the Pokemon gene

could be regulated by four alternative promoters We

are currently performing deletion analysis and a DNA

decoy assay to study three other alternative promoters,

which will further provide better understanding of the

Pokemongene transcriptional mechanisms

From the TRANSFAC 7.0 database, we found some

putative regulatory elements in the Pokemon promoter,

including binding sites for Sp1, AP-1, AP-2 and

GATA-1 elements (Fig 1); however, deletion analysis

showed that the above-mentioned regulatory elements

cannot play decisive roles in the regulation of the Pokemon gene, suggesting the complexity of gene regu-lation Fortunately, we found that three regulatory ele-ments, namely, POS-D, NEG-U and NEG-D, play important roles in the regulation of the Pokemon gene

To determine whether these three elements are homo-logous with the regulatory sequences deposited in the database, we performed a BLAST search by using their sequences as queries in the TRANSFAC 7.0 database The results showed that none of them shared

a higher degree of homology with the reported regula-tory elements, thus indicating their novel roles Cur-rently, we are conducting yeast one-hybridization in order to isolate transcription factors that can interact with these novel regulatory elements; this will help further understand the regulatory mechanism of the Pokemon promoter

The DNA decoy technique, also referred to as the transcription factor decoy technique, involves the transfection of double-stranded oligodeoxynucleotides corresponding to the regulatory sequence into target cells; this results in the attenuation of authentic cis– trans interactions, leading to the removal of transcrip-tion factors from the endogenous regulatory element and suppression of the expression of the regulated genes Recently, some reports showed that the DNA decoy technique is a powerful tool for therapy related

to various diseases [32–34] In this experiment, we used the DNA decoy technique to successfully confirm the function of the POS-D, NEG-U and NEG-D elements, proving that the transcription factor decoy technique can be a powerful tool for the study of transcriptional regulation mechanisms However, we also found that the wild-type POS-D decoy cannot completely abolish reporter gene expression; this may occur for two rea-sons First, some DNA decoys may be degraded by endogenous nuclease Second, in the amounts used, the POS-D decoy cannot completely abolish the inter-action between the wild-type POS-D element and its corresponding transcription factor Although POS-D decoys cannot completely abolish the activity of the Pokemon promoter, they are still potential oligodeoxy-nucleotides that can be used to decrease the Pokemon gene expression in cancer gene therapy

The SV40 early promoter contains a TATA box, three copies of a 21-bp GC-rich repeat, and two copies

of a 72-bp repeat The 72-bp repeat acts as an enhan-cer to increase the activity of the SV40 promoter, whereas the 21-bp GC-rich repeat is the main recogni-tion signal for eukaryotic RNA polymerase II and is necessary for promoter activity [35] Deletion analysis and the decoy assay showed that the NEG-U and NEG-D elements were involved in the negative

regulation of the Pokemon gene However, our

gain-of-function experiment interestingly revealed that both

the NEG-U and NEG-D elements could increase the

activity of the SV40 promoter The negative function

of the NEG-U element is strictly dependent on the

NEG-D element When incorporated upstream of the

SV40 promoter, the NEG-U element may interact with

the 72-bp repeat enhancer to increase promoter

activ-ity In addition, there may be a similar reason why

NEG-D element could increase the activity of the

SV40 promoter Therefore, a DNA context in which

different regulatory elements exist also plays an

impor-tant role in gene regulation, and incorporating these

elements into new promoters may alter their original

functions

Eukaryotic gene expression is often controlled by

multiprotein transcriptional complexes that bind

differ-ent elemdiffer-ents in the 5¢-upstream regions of target genes

[36] Gallagher et al showed that the GATA-1 and

Oct-1 elements were required for the expression of the

gene encoding human a-hemoglobin-stabilizing protein

[37] Recently, Griffin et al showed that E-box and

sterol regulatory element (SRE) could mediate

syner-gistic activation of the fatty acid synthase promoter

[38] NEG-U and NEG-D elements were necessary and

sufficient for the negative regulation of the Pokemon

gene, but the NEG-U or NEG-D element alone could

not negatively affect gene expression, further

confirm-ing the importance of combinatorial control In this

study, we also found that mutations in both NEG-U

and NEG-D elements had the same effect as each

sin-gle mutation This is because negative regulation of the

Pokemon gene is strictly dependent on an interaction

between NEG-U and NEG-D elements Mutations in

both sites or mutation in a single site could abolish an

interaction between them Therefore, all mutated

con-structs displayed high luciferase activities

In many eukaryotic genes, transcription factors bind

to promoters located at sites distant from one another,

yet they act synergistically via DNA looping to

acti-vate transcription [39,40] The insulin gene promoter

contains three SREs and two E-boxes; two of the

SREs overlap with the E-boxes that can be bound by

the BETA2⁄ E47 protein Activation of the insulin

pro-moter by SRE-binding protein (SREBP-1c) was

mark-edly enhanced by the co-expression of BETA2⁄ E47

Synergistic activation by SREBP-1c and BETA2⁄ E47

was not mediated via SREs but via the E-boxes

Reducing the distance between the two E-boxes

abol-ished synergistic activation Therefore, the synergistic

action required the presence of two E-boxes separated

by an appropriate distance in a looped form,

presum-ably to form a DNA and SREBP-1c⁄ BETA2 ⁄ E47

complex [41] To determine whether the length between the NEG-U and NEG-D elements also plays an important role in the regulation of the Pokemon gene, the distance between them was reduced to 254 and

106 bp Our results showed that synergistic inhibition via the interaction between the NEG-U and NEG-D elements was almost abolished when the distance between the two elements was reduced, suggesting that synergistic inhibition also requires the regulatory ele-ments to be separated by a certain distance It is likely that the appropriate distance facilitates DNA looping structure formation and is the threshold distance for the interaction between the NEG-U and NEG-D ele-ments Deviation from the appropriate distance pre-vented the corresponding cis–trans complexes from acting synergistically with DNA looping to activate or suppress transcription

In conclusion, our studies are the first to elucidate the mechanism of the Pokemon gene transcription reg-ulation Future studies will focus on the identification

of proteins that can specifically bind to the NEG-U, NEG-D, and POS-D elements; this will provide a bet-ter understanding of the mechanisms of the Pokemon gene regulation

Experimental procedures Cells and cell culture

Human lung carcinoma A549 cells were grown in Ham’s F12K medium containing 10% fetal bovine serum (Invitro-gen Corp., Carlsbad, CA, USA), human prostate carci-noma DU145 cells and human tumor of cervix uteri HeLa cells were grown in Dulbecco’s modified Eagle’s medium containing 10% fetal bovine serum (Invitrogen), human hepatocyte carcinoma HepG2 cells and human acute T-cell leukemia cells were maintained in RPMI-1640 medium con-taining 10% fetal bovine serum (Invitrogen) All these cells

37C

Creation of deletion constructs of the upstream region of the Pokemon gene

The 2204-bp upstream region of the Pokemon gene, which spans the region)2220 to )17, was amplified by PCR from human blood genomic DNA by using the primers YU and

YD (Table 1) The PCR products were cloned into the pGEM-T easy vector (Promega, Madison, WI, USA) and sequenced; then, the 2204-bp promoter fragment was cloned into the BglII⁄ HindIII sites of the pGL-3 basic vec-tor (Promega), and the resultant construct was designated

as F-2220 The 5¢-deletion constructs with their endpoints