J. Biochem. 110, 232-236 (1991)

Xenobiotic Responsive Element in the 5'-Upstream Region of the Human P-450c Gene1 Momoyo Kubota,* Kazuhiro Sogawa,* Yoshiko Kaizu,' Takuji Sawaya,** Junko Watanabe,"' Kaname Kawajiri,"* Osamu Gotoh,*" and Yoshiaki Fujii-Kuriyama* 'Department of Chemistry, Faculty of Science, Tohoku University, Aobayama, Sendai, Miyagi 980; "Hokkaido Industrial Technology Center, 379 Kikyo, Hakodate, Hokkaido 041; and "'Department of Biochemistry, Saitama Cancer Center Research Institute, Ina-machi, Saitama 362 Received for publication, March 5, 1991

The nucleotide sequence of the 5 -upstream region up to about —4.1 kb of the human P-460c gene was determined. Two kinds of repetitive sequences were located; one was the Alu sequence which was inserted at three positions (-3127 to - 3 0 3 8 , - 3 0 1 7 to - 2 7 7 0 , and —2167 to —1851), and the other was the SINE-R element located just upstream of the most distal Alu sequences. The region other than the two repeated sequences showed an overall similarity of 70% to that of the rat P-450c gene. Survey of XRE or its homologues, responsible for the inducible expression of the rat P-450c gene, revealed eight XRE core sequences in this region of the human P-460c gene. Three of them were carried in the Alu sequences. A fusion gene which was constructed by ligating the upstream region of the human P-460c gene to the chloramphenicol acetyltransferase (CAT) gene expressed the CAT activity in response to the inducer, methylcholanthrene, when transfected into Hepa-1 cells. Stepwise decrease in CAT activity in three regions was observed as the 5 -upstream sequence containing XRE motifs was removed. However, the XRE core sequence in the Alu sequences seemed inactive, because elimination of the three elements in the Alu sequences did not affect the expressed CAT activity. In accordance with this observation, competition experiments using gel mobility shift assay showed that XRE core sequences in the Alu sequences could not compete with the XRE sequence for the inducer-bound receptor. Other XRE motifs than those in the Alu sequences appeared active and were conserved at similar positions to their counterparts in the rat P-450c gene.

A variety of chemical substances in the environment are oxidatively detoxified by members of the P-450 family (1, 2). Of these drug-metabolizing P-450s, P-450c [P-450IA1 (3) ] mainly oxidizes polycyclic aromatic hydrocarbons such as benzo[a]pyrene and 3-methylcholanthrene (MC) (4, 5). This enzyme is expressed ubiquitously in various organs at a very low level but is induced more than 20-fold by treatment with inducers, including MC and 2,3,7,8-tetrachlorodibenzodioxin (TCDD). The transcriptional response of the P-450c gene to xenobiotic inducers was observed not only in animal organs but also in several established cell lines. Using Hepa-1 cells (6-9), a cell line derived from mouse hepatoma, as recipient cells, transfection expertments with fusion genes consisting of various lengths of the 5'-upstream sequence of the rat P-450c gene and chloramphenicol acetyltransferase (CAT) gene led us and others to find XRE (xenobiotic responsive element) as the sequence element responsible for inducible activation of the gene (10-12). This motif is found at least 5 times in the 5'-upstream region of the rat P-450c gene, and has the common core sequence (T)CACGC (20). -== ;

We previously reported that the 5'-upstream region of the human P-450c gene conferred inducibility to the subordinate gene in response to xenobiotic inducers (13). In this paper, we surveyed the XRE motifs in the 5'-upstream region of the human P-450c gene, and found an additional XRE in proximity to the transcription-initiation site. Including this XRE, all the functional XREs were positionally conserved between the human and rat P-450c genes, MATERIALS AND METHODS Materials—Restriction enzymes, T4 DNA ligase, Klenow fragment (the large fragment of Escherichia coli DNA polymerase I) and Bal31 exonuclease were obtained from Takara Shuzo, (Kyoto). [a--32P]dCTP (-3,000 Ci/ mmol) and [y- 32 P]ATP (~ 6,000 Ci/mmol) were purchased from Amersham International. DNA Sequence Analysis—Restriction fragments were inserted into appropriate vectors, Ml3mp 18 and Ml3mp 19 at the polylinker site, then sequenced by the dideoxy method (14). All sequencing was performed in both direc-

1

This work was supported in part by Grants-in-Aid for Scientific Research from the Ministry of Education, Science and Culture of

.•

Japan and by a fund from Mitsubishi Foundation. Abbreviations: XRE, xenobiotic responsive element; MC, 3-methylcholanthrene; CAT, chloramphenicol acetyltransferase.

Construction of Pfosmids-Construction of the original P-450c-CAT chimeric plasmid, phMC7.6, was described previously (23). For the construction of phMC4.2 and 232

'

..

. _,

- j o . . -

r J.I_

• • i

J. Biochem.

XRE in Human P-450c Gene phMC2729, the original plasmid was partially digested with Stul and the Stul site at about — 5.4 kb was changed to a Xhol site with synthetic Xhol linker after filling in with Klenow fragment. The plasmid was linearized with Xhol then digested with Bal31 exonuclease. The digested plasmid was ligated with Clal linker to make intermediate plasmids. The intermediate plasmids were cut with Clal and allowed to be ligated intramolecularly to produce phMC4.2 and phMC2729. For the construction of other 5'-deletion plasmids, phMC7.6 was completely digested with Stul, and then digested with Bo/31 exonuclease. The digested plasmid was ligated with Clal linker to make intermediate plasmids. The intermediate plasmids were cut with Clal and allowed to be ligated intramolecularly. Cell Culture and DNA Transfection—Hepa-1 cells (15) were maintained in Dulbecco's modified essential medium supplemented with 10% bovine fetal calf serum and 0.35% glucose. Transfection was performed by the calcium phosphate precipitation method as described (8). P-450c-CAT chimeric plasmids (5//g) were mixed with lug of a plasmid (pXGH5) which contained the structural gene for human growth hormone downstream of SV40 promoter (16), and used for the transfection. The resultant secreted growth hormone was determined by radioimmunoassay (17) and used as an internal control for efficiency of the transfection. The CAT activity was assayed by the method of Gorman et al. (17). The acetylated forms of [ u C]chloramphenicol were separated by thin-layer chromatography, and quantified by scintillation spectrometry after detection by autoradiography as described previously (17). Gel Mobility Shift Assay—Gel mobility shift assay was performed as described (6). An oligonucleotide of XKE1 (6), whose sequence is

5'- GATCCCTCCAGGCTCTTCTCACGCAACTCC GGAGGTCCGAGAAGAGTGCGTTGAGGCTAG -5' was used as a probe after being labeled at the 5' ends with [y-"P]ATP by T4 polynucleotide kinase. The specific activity of the probe was 1.5 X10' cpm/pmol. RESULTS AND DISCUSSION 5' - Upstream Sequence of the Human P-450c Gene—The nucleotide sequence of the 5'-upstream region up to —4.1 kb of human P-450c gene was determined as shown in Fig. 1. There are two types of repetitive sequences in the region. One is the Alu sequence located at three positions ( — 3127 to - 3 0 3 8 , - 3 0 1 7 to -2770, and - 2 1 6 7 to -1851). The most upstream one is a half part of the Alu sequence and the most downstream is inserted in the opposite direction. The other repetitive sequence is the SINE-R element, which is located at the position of — 3128. This sequence is a kind of nonviral retroposon whose repeating unit is about 630 bp in length and which is present at 4,000 to 5,000 copies per haploid human genome (18). A highly GC-rich subregion "a" of the element (18) was tandemly repeated 16 times as shown in Fig. 1. The nucleotide sequence of the other region showed an overall similarity of 70% to that of the rat P-450c gene, suggesting that the repetitive sequences were inserted in the upstream region of the human P-450c gene after divergence between human and rat. We previously reported that a cis-acting element termed XRE, is responsible for inducible activation of the rat Vol. 110, No. 2, 1991

233 P-450c gene in response to inducers such as MC and TCDD (10). We searched for the core sequence, CACGC, of XRE in the 5'-upstream region and found eight of the sequences. Interestingly, three of them were located in the Alu sequences (-3068, -2116, and -1876). The Alu family contains subfamilies with the XRE core sequence around its position 20 (19, 20). The other XREs were positioned conservatively between human and rat genes with insertions of the two kinds of repetitive sequences (Fig. 1). In order to examine whether or not these XRE motifs possess a potential to activate the P-450c gene, we further analyzed the properties of the XREs. In particular, we investigated in more detail the possibility that the XRE cores in the Alu repeated sequences could activate the gene in response to xenobiotic inducers, because the Alu sequences are often inserted in the promoter region of various genes. The BTE sequence identified as a promoter element in the rat P-450c gene was also conserved immediately upstream of the TATA box of the human gene (21). Deletion Analysis of the P-450c-CAT Expression Plasmid—We have previously shown that the 5'-upstream sequence of the human P-450c gene is capable of activating the CAT gene used as a reporter gene in response to added inducers in a DNA transfection system of Hepa-1 cells (13). Various 5'-deletion mutants of the fusion gene were constructed and introduced into the cells. Figure 2 shows the structure and the CAT activity expressed from the mutated genes in the presence or absence of MC. The expressed CAT activity decreased progressively with elimination of three regions, —7.6 to —4.2 kb, —1262 to - 9 8 8 , and - 9 4 8 to - 5 3 2 . The deletion from - 1 2 6 2 to — 988 caused the largest decrease in the induced CAT activity. This region contained counterparts of XRE1 and XRE2 sequences which were found to be the most potent enhancers in the rat P-450c gene (6). The phMC948 plasmid which contained an XRE motif at —895 showed some inducibility of the CAT activity in response to the inducer, while the equivalent deletion mutant of the rat gene with the corresponding XRE sequence showed little response to the added inducer. The difference in the inducibility of the CAT activity between the two fusion genes may be due either to a slight variation in the sequence flanking the XRE sequence or to some other unknown reason. The mouse counterpart of this XRE sequence was reported to enhance a heterologous promoter in response to inducers suggesting that the activities of the individual XRE sequences may differ (22). In the rat gene, one of the XRE sequences was most remotely located at — 3586. The human most distant counterpart of the XRE might be located further upstream than the sequence determined (-4144), because the deletion of the - 7 . 6 to - 4 . 2 kb sequence caused a decrease in the expressed CAT activity by 50% and because a total of about 2 kb of the repetitive sequences was inserted in the upstream region to the human gene. The XRE cores in the Alu sequences did not seem to be effective on the gene expression, since the expression of the CAT activity was not affected by the deletion of the three sequences (phMC4.2, phMC2729, and phMCl784). To further investigate the enhancer activity of the XRE core sequences in the Alu sequences, we placed a fragment ( — 2238 to —1784) containing two XRE core sequences of the Alu sequences upstream or downstream of the CAT

234

M. Kubota et al.

-4019

&»;Tt#i»c»i*»M-"**t*tt'*»"**»t"e*—i**

-eAt6TCACCACtbSCa«— -CCCGCtCCCSCCTGCCWCTCCTSCTAGGC-TGT *ci*f***«»**tc«»»tt*tctttctetetcutcUfft*t»t*t*tcftc"c«**ci*t*t*t* > # • * g t t " " * * »

*

ACCTTTCTC-TCCAATCCUWGASACCttCCCGGT---TCACSCTtaTCTCCnCCATCTCAS " * •

-73 ac—KtCC«GC6A»a;