www.bba-direct.com

Biochimica et Biophysica Ac

Identification of the functional elements in the promoter region of

human DNA topoisomerase IIIh gene

Young Hoon Cho, Jee Young Park, Sang Youp Han, In Kwon Chung*

Department of Biology, Molecular Aging Research Center, and Protein Network Research Center, College of Science, Yonsei University,

134 Shinchon-dong, Seodaemun-gu, Seoul 120-749, South Korea

Received 1 October 2003; accepted 3 August 2004

Available online 20 August 2004

Abstract

In this study, we have isolated and characterized the promoter region of the human DNA topoisomerase IIIh (hTOP3b) gene. The 5VRACE assay showed a short exon 1 encoding only the 35-bp untranslated region and suggested the presence of multiple transcription

initiation sites. The hTOP3b gene promoter lacks a canonical TATA box or initiation element and is moderately high in GC content. Transient

expression of a luciferase reporter gene under the control of serially deleted 5V-flanking sequence identified an activator element between

�141 and �119 upstream of the transcription initiation site and a second regulatory element between �91 and �71. On the basis of scanning

mutations of triple nucleotides, we demonstrated that a 5VGGAACC3Velement between �117 and �112 plays a critical role in the up-

regulation of the basal transcription activity. Changing the 5VGGAACC3Vsequence leads to markedly reduced promoter activity. Gel mobility

shift assays revealed that the 5VGGAACC3Velement is required for DNA binding by the transcription factor complex. These observations lead

to the conclusion that the positive regulatory region including the 5VGGAACC3Vcore element is essential for efficient expression of the

hTOP3b gene as well as for the binding of as yet unidentified regulatory factor(s).

D 2004 Elsevier B.V. All rights reserved.

Keywords: Topoisomerase IIIh; Human; Gene expression; Promoter; Transcription; Transcription factor

DNA topoisomerases are nuclear enzymes that are able

to break and reseal the sugar-phosphate backbone bonds of

DNA, thereby adjust the topological states of DNA [1].

Eukaryotic type I topoisomerases catalyze the removal of

both positive and negative supercoils by transiently break-

ing one strand of the DNA double helix for strand passage

[2]. These enzymes are divided into two subfamilies, IA and

IB, based on structural and mechanistic differences [3].

Eukaryotic topoisomerase III belongs to the type IA

subfamily of topoisomerases, which link covalently to the

5V-end of the cleaved DNA via a phosphotyrosine linkage.

Unlike single TOP3 gene in prokaryotic and lower

eukaryotic cells, two TOP3 isozymes encoded by different

genes have been identified in mammals [4–7]. The hTOP3a

0167-4781/$ - see front matter D 2004 Elsevier B.V. All rights reserved.

doi:10.1016/j.bbaexp.2004.08.001

* Corresponding author. Tel.: +82 2 2123 2660; fax: +82 2 312 8660.

E-mail address: topoviro@yonsei.ac.kr (I.K. Chung).

gene is located at 17p11.2–12 and has two potential start

codons for the synthesis of two proteins [8]. A larger protein

possesses a mitochondria-targeting signal at its N-terminus,

which can direct the protein to mitochondria. Overexpres-

sion of a truncated form of hTOP3a was found to inhibit

spontaneous and radiation-induced apoptosis upon trans-

fection into ataxia telangiectasia cells [9], and targeted

disruption of the mTOP3a gene revealed that this gene is

essential in early embryogenesis [10]. The hTOP3b gene

was initially identified in the immunoglobulin E locus

located at 22q11–12 [11]. In contrast to the embryonic

lethality of TOP3a knockout mice, mice lacking TOP3bdevelop normally to maturity without apparent defects but

have a shorter life span than their wild-type littermates [12].

These results suggest that TOP3a and h isozymes cannot

fully substitute for each other, despite their similar

enzymatic characteristics. The two human TOP3 isozymes

show distinct tissue specificities, the a form being predom-

ta 1679 (2004) 272–278

Y.H. Cho et al. / Biochimica et Biophysica Acta 1679 (2004) 272–278 273

inantly expressed in the testis and the h form in the thymus

[7]. In order to understand the mechanisms governing

expression of TOP3 isozymes, we previously identified and

characterized the functions of YY1 and USF-binding

elements that are conserved in both human and mouse

TOP3a promoters [13–15]. In this report, we describe the

isolation and initial characterization of the 5V-flankingsequence for the hTOP3b gene.

A search for homologous sequences in the GenBank

database revealed that a genomic sequence encoding

hTOP3h is located within the human immunoglobulin Egene locus [11]. The hTOP3b gene spans a genomic

region of ~29 kb and expresses three alternatively spliced

transcripts [7]. These transcripts have the same 5V-endsequence but contain alternatively spliced 3V-end sequen-

ces. A short exon 1 encoded only the 5V-untranslatedregion and was separated from the second exon by a 6.9-

kb intron. To understand the regulatory mechanism

responsible for regulating hTOP3b expression, the 5V-flanking region of the hTOP3b gene was cloned by PCR

amplification using a primer annealing to the upstream

sequence of hTOP3b cDNA (5V-GAGCAACCAGGAA-GATGGAGTTGCC-3V) and a primer derived from the 5V-end sequence of the first intron (5V-CCTGTGAGACTGG-TACAGCAGATC-3V) (see location of primers in Fig. 1A).

A 2.1-kb DNA fragment was amplified from HeLa cell

Fig. 1. Nucleotide sequence of the 5Vflanking region of the hTOP3b gene. (A) Sche

the hTOP3b gene. Transcription initiation site and the ATG initiation codon are d

lines denote introns and flanking regions. The coding region is drawn as closed b

amplification are shown. Gene-specific primers (Sp-1, Sp-2, Sp-3) used for 5VRtranscription initiation site (designated as +1), which is marked with an arrow, an

boxed, and the potential binding sites in the positive regulatory region are underl

genomic DNA, and the nucleotide sequence of the cloned

fragment showed a perfect match to the published genomic

sequence [11].

The transcription initiation site of the hTOP3b gene was

determined by 5V-RACE using a 5V RACE kit (Roche

Molecular Biochemicals). The cDNA strand was generated

by reverse transcription of total RNA isolated from HeLa

cells using a gene-specific primer-1 (5V-GATGAGAGA-GATGTCTAAATCACCGT-3V) complementary to the cod-

ing sequence of the hTOP3b cDNA and was subjected

to dATP-tailing. The dA-tailed cDNA was amplified by

PCR using a gene-specific primer-2 (5V-GGCATTACA-GATGTCTGTGTCCG-3V) and the oligo dT-anchor primer.

The obtained cDNAwas further amplified by a second PCR

using nested gene-specific primer-3 (5V-AGCCACAGCAC-GATGTAGTCGC-3V) and the PCR anchor primer. The final

PCR products with approximately 550 bp were cloned and

sequenced. Examination of 17 independent clones revealed

several extended products with varying length, suggesting

that transcription was initiated at multiple sites. The most

abundant transcript is being initiated from the adenine

residue located 132 bp upstream of the start ATG codon

(Fig. 1B). Accordingly, this base was designated hereafter as

+1 bp. Examination of the 5V-flanking sequence revealed

several notable sequence motifs for the binding of tran-

scription factors but a lack of consensus TATA box or

matic representation of exon–intron organization in the 5V-flanking region ofesignated as +1 and +133, respectively. Boxes represent exons, while thin

oxes. The promoter region sequences and the positions of primers for PCR

ACE are indicated. (B) Bases are numbered with respect to the major

d vertical lines indicate minor initiation sites. The ATG initiation codon is

ined. Location of the intron separating exons 1 and 2 is indicated.

Y.H. Cho et al. / Biochimica et Biophysica Acta 1679 (2004) 272–278274

initiation element. The CpG dinucleotide occurs approx-

imately every 9 bp in the first exon and the 200-bp region

immediately upstream of the transcription initiation site,

while CpG occurs every 50 to 100 bp on average in the

major fraction of the mammalian genome (Fig. 1B). [16,17].

Thus, methylation of these regions might be involved in

regulation of hTOP3b gene expression.

To identify the regulatory elements in the 5V-flankingsequence of the hTOP3b gene, DNA fragments with nested

5V-deletions of the 5V-flanking sequence were placed

upstream of the luciferase reporter gene (see deletion map

of each plasmid in Fig. 2A). For a directional cloning, DNA

fragments were prepared by PCR reactions using the

appropriate synthetic oligonucleotides and inserted into the

pGL2-basic vector (Promega). These plasmids were tran-

siently transfected into HeLa cells, and the luciferase

activities were measured from the cell lysates (Fig. 2A).

Since hTOP3h was functionally expressed in HeLa cells

[18] and its mRNAwas detected in 5VRACE experiment, we

chose HeLa cells to analyze the promoter activity. The

results showed that deletions between �1752 and �142 had

no significant effect on luciferase activity. However, the

deletion of the region between �141 and �119 resulted in

about 85% decrease in reporter gene activity as compared

Fig. 2. Transient expression analysis of the hTOP3b promoter. (A) Deletion const

upstream of the luciferase reporter gene and transiently transfected into HeLa cells

marked. Amounts of cell lysates employed for the luciferase activity assay were no

each construct was expressed as a percentage of that of �1752/+13 construct. (B) H

�118/+13 construct and h-galactosidase expression plasmid. The results are the m

duplicate.

with that observed for the 141/+13 construct. These results

suggest that a positive regulatory element(s) is located

between �141 and �119, and that this region is required for

a high level expression of the hTOP3b gene. Transfection of

the further deleted construct �70/+13 resulted in a lower

level of luciferase activity as compared with that of the �91/

+13 construct. This suggests the presence of a second

positive regulatory element(s) between �91 and �71. To

examine whether the regulatory region between �141 and

�119 is functionally conserved between human and mouse,

the �141/+13 and �118/+13 constructs were transfected

into mouse NIH3T3 cells, and the luciferase activities were

measured (Fig. 2B). The results indicate that the positive

regulatory region is indispensable for the high level

expression in mouse cells, suggesting that this region could

be conserved in both human and mouse genes, and that

mammalian TOP3b genes may possess a common mecha-

nism of transcriptional regulation.

To determine precisely the boundaries of the sequence

element required for the high level of gene expression in the

hTOP3b promoter, the promoter sequences were mutated by

site-directed mutagenesis. Each construct contained triple

nucleotide transitions in the region between �141 and �119

(Fig. 3A). Their promoter activities were examined by

ructs containing different lengths of the hTOP3b promoter were subcloned

. Relative positions of the 5Vends of deleted promoter in each construct are

rmalized to the h-galactosidase activity, and the relative luciferase assay of

eLa and mouse NIH3T3 cell lines were cotransfected with the �141/+13 or

ean and standard deviation of two independent transfections performed in

Fig. 3. Mutational analysis of the promoter region between �141 and �119. (A) DNA sequences of the wild-type and mutant constructs. M1–M12 constructs

contain the same sequence except for the three nucleotide substitutions. (B) The promoter activity of each construct was measured by transfection into HeLa

cells. The relative luciferase assay of each construct was expressed as a percentage of that of �141/+13 construct. The results are the mean and standard

deviation of two independent transfections performed in duplicate.

Y.H. Cho et al. / Biochimica et Biophysica Acta 1679 (2004) 272–278 275

transient transfection of HeLa cells. As shown in Fig. 3B,

mutations of the M1–M8 sites had no significant effect on

luciferase activity. However, mutation of the M9 (GGA) or

M10 site (ACC) resulted in about 82% decrease in

luciferase activity when compared with that of the �141/

+13 construct. M11 or M12 mutation showed high levels of

luciferase expression. These results indicate that the M9 and

M10 sequences play a crucial role as activator elements in

the efficient expression of the hTOP3b gene. Although the

�118/+13 construct contained wild-type sequences at the

M9 and M10 sites, its promoter activity was markedly

reduced as compared with that of �141/+13. This suggests

that the sequence immediately upstream of the M9 and M10

sites could also contribute to high level of the hTOP3b gene

expression.

Since the wild-type sequences at the M9 and M10 sites

function as activator elements for the efficient expression of

hTOP3b gene, we tested the ability of the 30-bp region

extending from �132 to �103 to interact with nuclear

proteins by gel mobility shift assay (Fig. 4A). As shown in

Fig. 4B, a single major complex was formed with nuclear

extracts from HeLa cells. The specificity of this complex for

the sequence was shown by a competition experiment. The

complex was completely abolished by competition with 50-

fold molar excess of an unlabeled wild-type probe, whereas

the same molar excess of a nonspecific DNA failed to

compete. When two mutated oligonucleotides, M9 and

M10, were used as competitors, the complex was not

significantly depleted, indicating that the GGAACC ele-

ment is essential for DNA binding by the transcription

factor complex. Computer analysis of the positive regu-

latory elements using the TRANSFAC 4.0 database

predicted the presence of a putative RFX1-binding site at

the GGAACC core element [19]. Moreover, members of the

Ets transcription factors including Ets-1- and Ets-related

Elk-1 are known to bind to DNA sequences containing a

GGAA core motif [20]. To identify the nuclear proteins

involved in the formation of the protein–DNA complex, the

Fig. 4. Gel mobility shift assays of nuclear protein factors for the regulatory elements. (A) DNA sequence of the wild-type oligonucleotide (WT), mutant

oligonucleotides (M9 and M10), and oligonucleotides containing AP2, RFX1, and Ets-1 consensus binding sites. (B) The 30-bp radiolabeled WT duplex

oligonucleotide (1 ng) was incubated without (lane 1) or with 4.8 Ag of HeLa cell nuclear extracts (lane 2). Competition was performed with 50-fold molar

excesses of unlabeled duplex oligonucleotides as depicted on the top of each lane. NS is the unlabeled 30-bp nonspecific duplex oligonucleotide. Bound and

free represent protein–DNA complex and free probe, respectively. (C) The 30-bp radiolabeled duplex probe was incubated with 4.8 Ag of HeLa cell nuclear

extracts in the binding mixture containing antibody against RFX1, Ets-1 or Elk-1.

Y.H. Cho et al. / Biochimica et Biophysica Acta 1679 (2004) 272–278276

duplex oligonucleotides containing the binding consensus

sequences of RFX1, Ets-1, and AP2 were prepared (Fig.

4A) and used as competitors [21–23]. In the competition

experiments shown in Fig. 4B, the complex was not

affected by any of the competitor oligonucleotides used.

Since the nuclear extract from HeLa cells contains RFX1

and Ets-1 proteins [21,24], this observation cannot be

attributed to the absence of the factors. To further confirm

nuclear factor in the protein–DNA complex, we performed

a supershift analysis of the complex by utilizing antibody

against RFX1, Ets-1 or Elk-1. The results from this assay

revealed that preincubation with each antibody had no

effect on the complex (Fig. 4C). In conjunction with

reporter gene assays, these observations suggest that the

GGAACC element (M9 and M10 sites) contributes

significantly to high level expression of the hTOP3b gene

and is essential for the binding of as yet unidentified

regulatory factor.

We next performed a mutagenesis study to determine the

regulatory element in the region between �91 and �71

(Fig. 5A). The plasmids containing triple nucleotide

transitions were transfeced into HeLa cells, and the

luciferase activities were measured from the cell lysates.

As shown in Fig. 5B, mutations at M13, M14, M17, M18,

and M19 sites resulted in about 80% to 90% decrease in

luciferase activity when compared with that of the �91/+13

construct, whereas M15 and M16 mutations reduced the

promoter activity by 50% and 69%, respectively. To

determine the true effect of the element on the expression

of hTOP3b gene, the mutated M13 and M14 sequences

were placed back into the context of the �141/+13

sequence, and their promoter activities were examined

(Fig. 5C). Mutations of the M13 and M14 showed 16%

and 18% reduction in the luciferase activity, respectively,

when compared with that of the �141/+13 construct. These

results indicate that the sequence element between �91 and

�71 plays a role as a second activator in the hTOP3b gene

expression.

The percentage of identical amino acid residues and

conservative changes between hTOP3a and hTOP3h is

50%. However, comparison of the promoter sequence of the

hTOP3a gene with the hTOP3b gene revealed no signifi-

Fig. 5. Mutational analysis of the promoter region between �91 and �71. (A) DNA sequences of the wild-type and mutant constructs. M13–M19 constructs

contain the same sequence except for the three nucleotide substitutions. (B) The promoter activity of each construct was measured by transfection into HeLa

cells. The relative luciferase assay of each construct was expressed as a percentage of that of �91/+13 construct. The results are the mean and standard

deviation of two independent transfections performed in duplicate. (C) The mutated M13 and M14 sequences were placed back into the context of the �141/

+13 sequence. The relative luciferase assay of each construct was expressed as a percentage of that of �141/+13 wild-type (WT) construct.

Y.H. Cho et al. / Biochimica et Biophysica Acta 1679 (2004) 272–278 277

cant sequence similarity. We have previously described that

the positive regulatory elements containing the YY1- and

USF-binding sites are important for efficient expression of

the hTOP3a promoter [13]. However, these binding sites

were not observed in the hTOP3b promoter. On the

contrary, the positive regulatory region including the

GGAACC element of the hTOP3b promoter was not found

in the hTOP3a promoter. Although hTOP3a and hTOP3hproteins show very similar enzymatic characteristics

[6,25,26], these findings suggest that the expression of the

two genes could not be coordinated. Although we isolated

and initially characterized the promotor region of the

hTOP3b gene in this work, further study will be required

to identify the nuclear protein factors specific for binding to

the regulatory elements and to elucidate the physiological

roles of other embedded elements that regulate the basal and

tissue-specific expression of the hTOP3b gene.

Acknowledgements

This work was supported by grant from the Korea

Science and Engineering Foundation through the Protein

Network Research Center.

References

[1] J.C. Wang, DNA topoisomerases, Annu. Rev. Biochem. 65 (1996)

635–692.

Y.H. Cho et al. / Biochimica et Biophysica Acta 1679 (2004) 272–278278

[2] J.J. Champoux, in: N.R. Cozzarelli, J.C. Wang (Eds.), DNATopology

and its Biological Effects, Cold Spring Harbor Laboratory Press,

Plainview, NY, 1990, pp. 217–242.

[3] M.R. Redinbo, L. Stewart, P. Kuhn, J.J. Champoux, W.G.J. Hol,

Crystal structures of human topoisomerase I in covalent and

noncovalent complexes with DNA, Science 279 (1998) 1504–1513.

[4] R. Hanai, P.R. Caron, J.C. Wang, Human TOP3: a single-copy gene

encoding DNA topoisomerase III, Proc. Natl. Acad. Sci. U. S. A. 93

(1996) 3653–3657.

[5] T. Seki, M. Seki, T. Katada, T. Enomoto, Isolation of a cDNA

encoding mouse DNA topoisomerase III which is highly expressed at

the mRNA level in the testis, Biochim. Biophys. Acta 1396 (1998)

127–131.

[6] T. Seki, M. Seki, R. Onodera, T. Katada, T. Enomoto, Cloning of

cDNA encoding a novel mouse DNA topoisomerase IIIh possessing

negatively supercoiled DNA relaxing activity, whose message is highly

expressed in the testis, J. Biol. Chem. 273 (1998) 28553–28556.

[7] S.W. Ng, Y. Liu, K.T. Hasselblatt, S.C. Mok, R.S. Berkowitz, A new

human topoisomerase III that interacts with SGS1 protein, Nucleic

Acids Res. 27 (1999) 993–1000.

[8] Y. Wang, Y.L. Lyu, J.C. Wang, Dual localization of human DNA

topoisomerase IIIa to mitochondria and nucleus, Proc. Natl. Acad.

Sci. U. S. A. 99 (2002) 12114–12119.

[9] E. Fritz, S.H. Elsea, P.I. Patel, M.S. Meyn, Overexpression of a

truncated human topoisomerase III partially corrects multiple aspects

of the ataxia-telangiectasia phenotype, Proc. Natl. Acad. Sci. U. S. A.

94 (1997) 4538–4542.

[10] W. Li, J.C. Wang, Mammalian DNA topoisomerase IIIa is essential

in early embryogenesis, Proc. Natl. Acad. Sci. U. S. A. 95 (1998)

1010–1013.

[11] K. Kawasaki, S. Minoshima, E. Nakato, K. Shibuya, A. Shintani, J.L.

Schmeits, J. Wang, N. Shimizu, One-megabase sequence analysis of

the human immunoglobulin lambda gene locus, Genome Res. 7

(1997) 250–261.

[12] K.Y. Kwan, J.C. Wang, Mice lacking DNA topoisomerase IIIhdevelop to maturity but show a reduced mean life span, Proc. Natl.

Acad. Sci. U. S. A. 98 (2001) 5717–5721.

[13] J.C. Kim, J.B. Yoon, H.S. Koo, I.K. Chung, Cloning and character-

ization of the 5V-flanking region for the human topoisomerase III gene,

J. Biol. Chem. 273 (1998) 26130–26137.

[14] E.J. Park, S.H. Han, I.K. Chung, Regulation of mouse DNA

topoisomerase IIIa gene expression by YY1 and USF tran-

scription factors, Biochem. Biophys. Res. Commun. 283 (2001)

384–391.

[15] S.Y. Han, J.C. Kim, J.M. Suh, I.K. Chung, Cell type-dependent

regulation of human DNA topoisomerase IIIa gene expression by

upstream stimulatory factor 2, FEBS Lett. 505 (2001) 57–62.

[16] A. Bird, M. Taggart, M. Frommer, O.J. Miller, D. Macleod, A fraction

of the mouse genome that is derived from islands of nonmethylated,

CpG-rich DNA, Cell 40 (1985) 91–99.

[17] S.H. Cross, J.A. Charlton, X. Nan, A.P. Bird, Purification of CpG

islands using a methylated DNA binding column, Nat. Genet. 6

(1994) 236–244.

[18] M. Kobayashi, R. Hanai, M phase-specific association of human

topoisomerase IIIh with chromosomes, Biochem. Biophys. Res.

Commun. 287 (2001) 282–287.

[19] P. Emery, K. Strubin, K. Hofmann, P. Bucher, B. Mach, W. Reith, A

consensus motif in the RFX DNA binding domain and binding

domain mutants with altered specificity, Mol. Cell. Biol. 16 (1996)

4486–4494.

[20] B.J. Graves, J.N. Petersen, Specificity within the ets family of

transcription factors, Adv. Cancer Res. 75 (1998) 1–55.

[21] R. Castellano, C. Van Lint, V. Peri, E. Veithen, Y. Morel, R.

Costello, D. Olive, Y. Collette, Mechanisms regulating expression of

the tumor necrosis factor-related light gene. Role of calcium-

signaling pathway in the transcriptional control, J. Biol. Chem.

277 (2002) 42841–42851.

[22] M. Liu, B.H. Lee, M.B. Mathews, Involvement of RFX1 protein in the

regulation of the human proliferating cell nuclear antigen promoter,

J. Biol. Chem. 274 (1999) 15433–15439.

[23] P.R. Mertens, M.A. Alfonso-Jaume, K. Steinmann, D.H. Lovett, A

synergistic interaction of transcription factors AP2 and YB-1 regulates

gelatinase A enhancer-dependent transcription, J. Biol. Chem. 273

(1998) 32957–32965.

[24] A. Krehan, H. Ansuini, O. Bocher, S. Grein, U. Wirkner, W. Pyerin,

Transcription factors ets1, NF-kappa B, and Sp1 are major determi-

nants of the promoter activity of the human protein kinase CK2a

gene, J. Biol. Chem. 275 (2000) 18327–18336.

[25] H. Goulaouic, T. Roulon, O. Flamand, L. Grondard, F. Lavelle, J.F.

Riou, Purification and characterization of human DNA topoisomerase

IIIa, Nucleic Acids Res. 27 (1999) 2443–2450.

[26] N. Hotoda, R. Hanai, Characterization of recombinant human DNA

topoisomerase IIIalpha activity expressed in yeast, J. Biochem. 127

(2000) 1109–1113.