Structure of gene flanking regions and functional analysis of sequences upstream of the rat hsp70.1 stress gene

Biochimica et Biophysica Acta 1625 (2003) 77 – 87 www.bba-direct.com

Structure of gene flanking regions and functional analysis of sequences $ upstream of the rat hsp70.1 stress gene Anna Fiszer-Kierzkowska, Aleksandra Wysocka, Micha» Jarza˛b, Katarzyna Lisowska *, Zdzis»aw Krawczyk Department of Tumor Biology, Center of Oncology, Maria Sk lodowska Curie Memorial Institute, Wybrzez˙e Armii Krajowej 15, 44-100 Gliwice, Poland

Received 8 January 2002; received in revised form 4 September 2002; accepted 6 November 2002

Abstract We present structural and comparative analysis of the flanking regions of the rat hsp70.1 stress gene. Several repetitive sequences, microsatellites and short interspersed repetitive elements (SINEs) were found, as well as a significant gap in the 3V UTR, as compared to the orthologous mouse gene. We also show that the complex microsatellite region composed of partially overlapping inverted repeat and long homopurine – homopyrimidine sequence, which is localized 1.8 kbp upstream of the transcription start site, is capable to adopt non-B DNA structures (an H-DNA and a cruciform structure) in vitro. Functional analysis performed with the use of various fragments of the 5Vend flanking regions ligated to the chloramphenicol acetyltransferase (CAT) reporter gene revealed a crucial role of cooperation between heat shock element (HSE) regulatory sequences, while none of the three HSEs alone is able to drive efficient heat induced transcription of the reporter gene. We also found that the microsatellite region does not influence transcription by itself, however, it abolishes the effect of the adjacent putative silencing element. To our knowledge, this is a first extensive structural and functional analysis of the promoter region of the mammalian heat inducible hsp70i gene localized distally to the hsp70-related spermatid-specific gene in the major histocompatibility complex III. D 2002 Elsevier Science B.V. All rights reserved. Keywords: Heat shock gene; hsp70 family; Transcription; Promoter sequence; MHC III; Non-B DNA structure

1. Introduction Mammalian hsp70 family contains two to four hsp70i (inducible) genes [1]. In human, rat and mouse two duplicated, major histocompatibility complex (MHC) III-linked hsp70i genes co-localize with the spermatid-specific hsp70related gene (see Fig. 1). Two additional hsp70i genes called

Abbreviations: CAT, chloramphenicol acetyltransferase; DMEM, Dulbecco’s modified Eagle’s medium; HAM, nutrient mixture; HSE, heat shock element; HSF, heat shock factor; HSP, heat shock protein; hsp70, gene encoding 70-kDa HSP; hsp70i, inducible hsp70; MHC, major histocompatibility complex; SINE, short interspersed repetitive element; tsp, transcription start point; UTR, untranslated region $ Nucleotide sequence of the rat hsp70.1 gene and its flanking regions has been deposited in the EMBL Data Bank, Ac X74271. * Corresponding author. Tel.: +48-32-278-9669; fax: +48-32-231-3512. E-mail address: [email protected] (K. Lisowska).

hsp70B and hsp70BV were found in human genome, and they also originated by ancestral duplication. The nucleotide sequences of both MHC III-linked hsp70i genes are highly similar and they code for virtually identical chaperone proteins. A telomeric gene called hsp70A/ hsp70.1 in human [2] (hsp70.3 in mouse [3] and hsp70.2 in rat [4]) is positioned in head-to-head orientation with the spermatid-specific hsc70t gene [4 – 6]. The transcription initiation sites of these two types of genes are only about 600 bp apart. A centromeric MHC III-linked hsp70i gene, called hsp70.2 in human [6], hsp70.1 in mice [7] and hsp70.1 in rat [4,8 – 10], is separated from the telomeric hsp70i gene by approximately 10 kb and is transcribed in the same direction (Fig. 1). The functional significance of the duplicated hsp70i locus coding for identical protein was partially elucidated in the very elegant study employing knock-out mice model, which showed differential expression of both MHC III-linked genes during ontogenesis [11]. These observations together with the results of other studies

0167-4781/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 4 7 8 1 ( 0 2 ) 0 0 5 9 2 - 4

78

A. Fiszer-Kierzkowska et al. / Biochimica et Biophysica Acta 1625 (2003) 77–87

Fig. 1. Structure of the hsp70.1 gene locus. (A) Localization of three hsp70 genes within the MHC III complex of rat, mouse and human. Arrows indicate direction of transcription. ‘MS’ indicates microsatellite region; ‘C2’ and ‘Bf’, genes of the complement complex; ‘TNF’, TNF genes. The telomeric hsp70i gene is called hsp70.1 in rat [8,4], hsp70.1 in mouse [7] and hsp70.2 in human [6]. The centromeric hsp70i gene is called hsp70.2 in rat [4], hsp70.3 [3] in mouse and hsp70.1 in human [2]. (B) Structure of the region containing the rat hsp70.1 gene and its flanking sequences. Boxes marked ID and B2 indicate localization of these SINE sequences found around the gene (the 5V-end ID sequence represents class 4 ID elements and the 3V-end one class 3 [60]). The arrowhead indicates the site of the suspected deletion in the 3VUTR (see text for details).

showing differential expression of both genes in response to different stressors [12 –15] suggest that these two genes posses distinct regulatory mechanisms and each may respond to different stimuli. The hsp70i genes show a highly differentiated expression pattern, from being strictly inducible in certain tissues to being constitutively active in various tumors (reviewed in Refs. [16,17]). Interestingly, an altered expression of the hsp70i genes seems to be involved in the changes of cell phenotype and in the etiology and progression of some diseases, e.g. autoimmune, neurodegenerative or cancer. The general inducing mechanism of hsp70i gene expression, the interaction of heat shock factor(s) (HSF) with the heat shock element (HSE) sequences, is relatively well recognized (reviewed in Ref. [18]). Much less is known about other regulatory interactions, e.g. with c-myc [19], B-myb [20], c-myb [21], p53 [22,23], E1A [24,25] in trans or MAR [26,27] in cis. This stresses the need for more extensive functional and comparative investigation of DNA regions flanking the hsp70i genes. Of the mammalian hsp70i genes, the human hsp70.1 (telomeric, MHC III-linked) has been most extensively studied since its cloning in 1985 [2,28 – 30]. Despite that, it appeared recently that its structure is more complex than suspected. An intron was detected at the 5V end of the gene and it was found that transcription can be initiated from two different sites and both promoter regions have distinct regulatory elements [31]. Other human and rodent hsp70i genes were less extensively studied. So far, only limited functional studies of the rat hsp70.2 gene (telomeric, MHC

III-linked, [32]) and the orthologous mouse hsp70.3 gene [3] have been performed. Interestingly, some silencing activities were found in the region between telomeric hsp70i gene and the adjacent spermatid-specific gene [32,33]. In contrast, neither the mouse nor the human promoter of the centromeric MHC III-linked hsp70i gene has been a subject of extensive functional analysis. Also, very limited information is available about the regulation of expression of the orthologous rat hsp70.1 gene. In our earlier work we determined the most important promoter region for heat inducibility of the gene [8], while Konishi et al. [34] reported that only the second, more distal HSE element is engaged in DNA – protein interactions under heat shock conditions. We also found a complex microsatellite sequence localized 1.8 kbp upstream of the transcription start site that is potentially able to adopt non-B DNA structure(s) in vitro [35]. Similar sequences, with a potential to adopt cruciform or H-DNA structure, occur frequently in the 5Vflanking regions and UTRs of genes, and in some cases were shown to influence transcription efficiency [36 – 39]. Their effect on transcription may probably depend on such factors as length of the sequence, perfection of its symmetry, distance from the gene and possible interactions with specific proteins recognizing either primary or secondary DNA structure. Here we present the results of structural and functional studies of the 5Vend flanking region of the rat hsp70.1 gene. Special attention has been paid to elucidate whether a complex microsatellite sequence found upstream of the transcription start site can adopt non-B DNA structure and

A. Fiszer-Kierzkowska et al. / Biochimica et Biophysica Acta 1625 (2003) 77–87

affect expression of the gene. In addition, we present comparative analysis of the gene flanking sequences of the hsp70.1 and homologous, MHC III-linked hsp70i genes from mouse and human.

2. Materials and methods 2.1. Plasmid constructs Plasmids p68/1.0, p68/700, p68/600, p68/450 and p68/ 300 contain different fragments of the rat hsp70.1 gene upstream sequence [from 1924 to 869 region, relative to the transcription start point (tsp)], cloned into pUC19 vector (see Fig. 3A). Plasmids used for transient transfections contain different DNA fragments from between 1924 and + 85, cloned into pBLCAT6 vector bearing the bacterial chloramphenicol acetyltransferase (CAT) reporter gene [40]. For further details of plasmid content and construction, see Table 1 and Fig. 6. 2.2. S1 nuclease cleavage and mapping of the cleavage site Plasmid DNA was incubated in S1 buffer (final concentration: 30 mM sodium acetate, pH 4.5, 50 mM zinc chloride, and 5% glycerol) for 30 min on ice. S1 nuclease (Pharmacia) was diluted in buffer containing 10 mM sodium acetate, pH 5.0, 1 mM dithiothreitol, 0.1 mM zinc chloride and 0.001% Triton X-100). Different concentrations of the nuclease were added to the DNA samples and reactions were carried out on ice for 10 min. Reactions were stopped by the addition of Tris base and EDTA to the final concentrations of 90 and 45 mM, respectively. For mapping of the cleavage site the S1-treated DNA was phenol-extracted and

79

ethanol-precipitated, then digested with appropriate restriction endonucleases. 2.3. Construction of deletion plasmids Plasmid p68/300 was digested with S1 nuclease and cleavage products were separated electophoretically. DNA band corresponding to the linear form of plasmid, containing products of complete (specific) digestion by S1 nuclease, was isolated from agarose gel, blunted with Klenow enzyme and ligated. Twenty-one clones were successfully sequenced and their sequence was compared to that of p68/ 300 in order to determine size and localization of the deletion introduced by S1. 2.4. Cell culture and transient transfection Rat hepatoma FTO cells were maintained in DMEM/ HAM medium supplemented with 10% foetal bovine serum (ICN). For transient transfections cells were seeded into 100-mm petri dishes (6  105 cells per dish) and grown in 10 ml of medium to 50 – 60% confluency. Transfections were performed by the DEAE-dextran method using a modification of a published procedure (Maniatis). Briefly, cells were overlaid with 0.65 ml of mixture containing 2 Ag of DNA and 0.5 mg/ml of DEAE-dextran in TBS*(P) buffer (25 mM Tris – Cl, 137 mM NaCl, 5 mM KCl, 0.7 mM CaCl2, 0.5 mM MgCl2 and 0.3 mM Na2HPO4) and incubated at room temperature for 30 min. Then, cells were washed with TBS*(P) and overlaid with 10% solution of DMSO in DMEM/HAM medium for 2 min. After washing and addition of medium with serum, cells were grown for 48 h at 37 jC (controls) or for 43 h, then heat shocked at 42.5 jC for 45 min and allowed to recover for 4 h.

Table 1 Structure of the plasmids Plasmid name

Coordinates of the hsp70.1 sequence

p68/1.0 p68/300 p68/600 p68/700 p68/450 p150/CAT6 p250/CAT6 p350/CAT6 p350(DHSE1)/CAT6 p550/CAT6 p550(DHSE1)/CAT6 p950/CAT6

1924 to 869 1924 to 1611 1924 to 1322 1611 to 869 1322 to 869 64 to + 85 167 to + 85 269 to + 85 269 to 167 and 64 to + 85 478 to + 85 478 to 167 and 64 to + 85 869 to + 85

p950(DHSE1 + 2)/CAT6 p150 – 950/CAT6 p300 – 950/CAT6 p450 – 950/CAT6 p700 – 950/CAT6 p1.0 – 960/CAT6

869 to 1024 to 1180 to 1322 to 1611 to 1924 to

280 and + 85 + 85 + 85 + 85 + 85

64 to + 85

Features MS, ID, and 600 bp nonrepetitive sequence MS MS, ID ID and 600 bp nonrepetitive sequence 450 bp nonrepetitive sequence Sp1, TATA, 85 bp 5VUTR HSE1, Sp1, TATA, 85 bp 5VUTR HSE2, HSE1, Sp1, TATAS, 85 bp 5VUTR as p350/CAT6 except DHSE1 HSE3, HSE2, HSE1, Sp1, TATA, 85 bp 5VUTR as p550/CAT6 except DHSE1 400 bp nonrepetitive sequence, HSE3, HSE2, HSE1, Sp1, TATA, 85 bp 5VUTR as p950/CAT6 except DHSE1 and 2 as p950/CAT6 + 150 bp nonrepetitive sequence as p950 + 300 bp nonrepetitive sequence as p950 + 450 bp nonrepetitive sequence as p950/CAT5 + 450 bp nonrepetitive sequence + ID as p950/CAT6 + 450 bp nonrepetitive sequence + ID + MS

Coordinates shown in the second column concern the hsp70.1 sequence cloned in each plasmid; they are indicated relative to the transcription start point of the gene. Third column contains information concerning specific features of the cloned DNA fragment, e.g. cis regulatory element or repetitive sequences.

80

A. Fiszer-Kierzkowska et al. / Biochimica et Biophysica Acta 1625 (2003) 77–87

Fig. 2. Nucleotide sequence of the regions flanking the rat hsp70.1 gene coding unit. The sequence of the coding region of the gene was published earlier [8] and is not shown here. Inverted repeat and the homopurine/homopyrimidine sequence are boxed. TG repeats are underlined (dashed line). AT-rich sequences are shown in italics, ID sequences are double underlined, and the B2 sequence is double overlined. Direct repeats flanking the ID sequence are shown in bold. ‘A’ indicates heat shock elements (HSE); ‘B’, CAAT-box; ‘C’, Sp1 sequences; ‘D’, inverted CAAT-box; ‘E’, TATA-box; ‘F’, polyadenylation signal; arrowhead, site of putative deletion. The sequence shown here is deposited in EMBL Data Bank (acc. no. X74271).

A. Fiszer-Kierzkowska et al. / Biochimica et Biophysica Acta 1625 (2003) 77–87

2.5. CAT assay

3. Results

Cell culture extracts were prepared using Reporter Lysis Buffer (Promega) according to the manufacturer’s instructions. Protein concentration in extracts was determined by Bradford method [41]. CAT assay was performed as a modification of procedure described by others (Maniatis). Briefly, 100 Al of cell extract containing 25 Ag of protein was added to 100 Al of reaction mixture containing 0.25 Tris –HCl (pH 7.8), 1 mM EDTA, 4.5 Al of 100 mM acetylCoA (Sigma) and 2.5 Al of [14C] chloramphenicol (2.5 ACi/ ml; Amersham). Samples were incubated in 37jC for 4 h. The reaction products were separated by thin layer chromatography. To quantitate CAT activity, radioactive spots representing acetylated and nonacetylated forms of chloramphenicol were cut out from the plate and their activity was measured in a scintillation counter. Counts were converted to the percentage of acetylated chloramphenicol as described by others [42]. Data for each experimental point were gathered in quadruplicate.

3.1. Structure of the hsp70.1 gene flanking regions

81

Fig. 2 shows DNA sequence of the regions flanking the coding unit of rat hsp70.1 gene with distinguished repetitive sequences. The most complex of these sequences is the microsatellite sequence localized upstream of the tsp of the gene. The core region of this sequence exhibits some unique structural properties, being composed of a long inverted repeat and a homopurine/homopyrimidine (Pu/Py) mirror repeat. In close proximity to the microsatellite sequence we detected three separate DNA regions highly enriched in AT base pairs and poly(A) tracts (Fig. 2). One of the AT-rich regions constitutes a part of the ID sequence (Fig. 2), a rodent-specific short interspersed repetitive element (SINE) [44]. We also found a second ID element located 128 bp downstream of the polyadenylation signal. Another SINE sequence, the B2 repeat, was identified 608 bp downstream of the poly(A) signal on the complementary strand. We also

Fig. 3. S1 endonuclease sensitivity of plasmids containing various fragments of the DNA region located upstream from the rat hsp70.1 gene. (A) Structure of plasmids. The EcoRI – PstI fragment located between positions 1924 and 869 (relative to tsp) and its shorter subfragments were cloned into pUC19. Drawing of the 5Vgene flanking region shows restriction sites used for cloning. Dark grey box represents microsatellite sequence, light grey box indicates ID repeat. (B) S1 nuclease assay. Plasmids described in (A) and control pUC19 plasmid were treated with S1 nuclease. Lanes marked ‘C’ contain native plasmid (control); lanes ‘L’, plasmid cleaved with the restriction enzyme, which serve as a marker for lanes ‘S1’, indicating position of linear plasmid; lanes ‘S1’, plasmid treated with S1 nuclease. Supercoiled plasmids with non-B DNA structures consisting partially of single-stranded fragments are cleaved by S1 nuclease to the OC form and subsequently to the linear form (depending on enzyme concentration and duration of reaction). Plasmids devoid of non-B DNA structures remain intact.

82

A. Fiszer-Kierzkowska et al. / Biochimica et Biophysica Acta 1625 (2003) 77–87

found a triple tandem repeat of 33 nt within the 5VUTR and a 239-bp gap within the 3VUTR as compared to the orthologous mouse region (Figs. 1B and 2). 3.2. Structure of the Pu/Py region The microsatellite region found about 1.8 kbp upstream of the tsp of the gene is composed of a (GA)6CAG(TC)6 inverted repeat and a partially overlapping (TC)24T mirror repeat. This core sequence is flanked by imperfect (TG)9 and (TG)11 repeats. Theoretically, both the long inverted repeat and the Pu/Py mirror repeat are able to form non-B DNA structures such as a cruciform or a triple-stranded HDNA, respectively. Our preliminary results indicated that the microsatellite sequence cloned into the pUC19 vector (plasmid p68/600) was susceptible to S1 nuclease which digests single-stranded DNA regions that accompany different non-B DNA structures [35]. We also observed a DNA polymerase blockage at the beginning of the (TC)24 tract in the sequencing reaction with the lower DNA strand as template (not shown). Here we map more precisely the region of S1 nuclease cleavage and attempt to define the nature of the non-B DNA structure. The series of plasmids was constructed (see Fig. 3A and Materials and methods) that contained either all repetitive sequences present in analyzed region, or some of them, or only the adjacent nonrepetitive sequences. These plasmids were treated with S1 nuclease and specific cleavage was observed in p68/1.0, p68/600 and p68/300, each containing the microsatellite region (Fig. 3B). Plasmid p68/450, devoid of any repetitive sequences, was not susceptible to S1 digestion. Plasmid p68/700 was very slightly cleaved by S1, probably due to the lower melting temperature within the long poly(A) tract present in its insert. In order to map the cleavage site, S1treated p68/300 plasmid was further digested with different restriction enzymes (Fig. 4A). Two main cleavage sites could be deduced from the restriction pattern, the first within the CAG sequence (the core sequence of the inverted repeat (GA)6CAG(TC)6) and the second in the center of the homopurine – homopyrimidine mirror repeat (TC)24T (Fig. 4B and C). This suggested that under our experimental conditions, part of the plasmid molecules adopted a cruciform structure while other part formed an H-DNA. To test this hypothesis we re-ligated the S1-treated p68/300 plasmid and sequenced 21 of the clones obtained. In four clones we observed huge deletions encompassing the whole microsatellite region and some neighboring sequences which can probably be ascribed to the S1 nuclease overdigestion and unspecific activity on double-stranded DNA (Fig. 5, deletions marked ‘‘A’’). Fifteen clones had deletions suggesting the formation of an H-DNA structure (marked ‘‘C’’). Fourteen of them had deletions ranging from 8 to 31 bp localized within the Pu/Py sequence. In one clone the deletion started in the middle of the Pu/Py sequence and proceeded far into the neighboring nonrepetitive sequence which can be explained similarly to the ‘‘A’’-type deletion. Only two

Fig. 4. Low resolution mapping of the S1 nuclease sensitive sites in p68/ 300 plasmid. (A) Samples of supercoiled p68/300 plasmid were treated with S1 nuclease, then cleaved by restriction enzymes (lanes 1 – 4). Large restriction fragments containing vector sequences were cut out off the picture. Sizes of reference marker (lanes ‘M’) are indicated on the right. Sizes of the restriction fragments were defined using ONE-D Scan software (Scananalytics); they correspond well to the theoretical cleavage products that may be obtained when one fraction of plasmid particles adopts H-DNA structure, while the other adopts cruciform structure. Diffuse shape of the bands may reflect differential activity of S1 nuclease (different plasmid particles are digested more or less intensively). (B) Theoretical sizes of the cleavage products. ‘S1A’ indicates most probable S1 cleavage site in the case of cruciform structure formation; ‘S1B’, site of cleavage in the case of H-DNA. Gray box at the top drawing represents insert in p68/300; line represents vector sequences. (C) Localization of S1A and S1B sites within the nucleotide sequence of the MS region.

clones had deletions that can be interpreted as the result of cruciform cleavage by S1 (marked ‘‘B’’). This result, together with the results of restriction mapping of the S1 cleaved plasmids, shows that the equilibrium between two possible non-B DNA structures may vary depending probably on slight differences in the experimental conditions (e.g. buffer concentration and temperature). We also tested

A. Fiszer-Kierzkowska et al. / Biochimica et Biophysica Acta 1625 (2003) 77–87

83

Fig. 5. Size and localization of deletions introduced in p68/300 plasmid by S1 nuclease. A sample of p68/300 plasmid was treated with the S1 nuclease. Cleavage products were separated by agarose gel electrophoresis; linearized plasmid DNA was isolated from gel, filled with Klenow enzyme to blunt possible overhanging ends and ligated. Twenty-one of the obtained clones were sequenced. Upper line shows the DNA sequence within the microsatellite region (positions from 1792 to 1691 relative to the tsp). Lines below indicate fragments of the sequence that were deleted by S1 digestion. Four types of deletion were repeated more than once (indicated by the number on the left side of the line). Deletions marked by dashed lines are located within (TC)24 sequence and due to the repetitive nature of that sequence, their precise localization cannot be established. For description of the deletions marked as A, B or C, see text.

the susceptibility of the deletion clones to the subsequent S1 nuclease cleavage. As expected, the clones with minor deletions (TC)4 to (TC)8 were still cleaved by S1 while clones with larger deletions were much more resistant (not shown). Surprisingly, clones with huge deletions encompassing the whole microsatellite region were also cleaved by S1, although very slightly. We think that this was due to the high AT content (over 80%) in the remaining part of the insert. 3.3. Functional analysis of the hsp70.1 gene 5V flanking sequences We made several constructs containing various fragments of the 5V sequence of the hsp70.1 gene fused with a CAT reporter gene (see Fig. 6). The first series of constructs was made to study the basic heat shock regulatory elements of the gene, the HSEs. Plasmids contained each of three HSEs a l o n e ( p 2 5 0 / C AT 6 , p 3 5 0 ( DH S E 1 ) / C AT 6 , p 9 5 0 (DHSE1 + 2)/CAT6), combinations of two HSEs (p350/ CAT6, p550(DHSE1)/CAT6), all three HSEs (p550/CAT6) or only proximal promoter sequences without any HSE (p150/CAT6). These plasmids were transiently transfected to FTO cells and CAT activity was analyzed in extracts of control and heat shocked cells (Fig. 6B). No CAT activity was observed in case of the plasmid p150/CAT6 containing only proximal promoter sequences encompassing TATAbox and Sp1 sequence, but without any HSE. Significant constitutive as well as heat inducible CAT activities (12% and 30% of acetylated chloramphenicol, respectively) were achieved with the construct p350/CAT6 containing both proximal HSEs. Addition of the third HSE (construct p550/CAT6) caused only a slight, insignificant increase in both types of CAT activity. Interestingly, we found that none of HSEs alone is able to drive inducible transcription of the reporter gene. A further question was whether the more upstream sequences, with special emphasis on the repetitive motifs,

have any impact on transcription efficiency. Two constructs were made containing, in addition to the region with three HSEs, a further upstream nonrepetitive sequences of 391 bp (p950/CAT6) and 844 bp (p450 –950/CAT6; compare with plasmid p68/450, Fig. 3A). Next plasmid p700– 950/CAT6 contained additionally the poly(A) tract of the ID sequence (compare with p68/700), while p1.0 –950/CAT6 contained both the microsatellite sequence and the poly(A) tract. Surprisingly, plasmid p950/CAT6, which differs from p550/CAT6 by addition of 391 bp of upstream region devoid of any evident regulatory elements, exhibited much higher CAT activity (75% versus 35%). Plasmids p950/CAT6 and p1.0 – 950/CAT6 showed comparable activity of the reporter gene, although it seemed that the presence of the microsatellite sequence in the latter construct caused a decrease of transcription in control cells. This is closer to the behavior of the strictly inducible promoter of the rat hsp70.1 gene in vivo. Surprisingly, in the case of the intermediate constructs (p450 – 950/CAT6 and p700 – 950/CAT6) a significant decrease in the CAT activity was observed, both in control and heat shocked cells. This suggests the possible presence of a silencing element in this DNA region, whose negative influence on transcription can be reversed by the presence of more distal regulatory elements (plasmid p1.0– 950/CAT6). Searching for a negative regulatory element we shortened the p450 – 950/CAT6 plasmid (the shorter of two constructs with diminished transcriptional efficiency) by 142 and 298 bp from the 5Vend. Both thus constructed plasmids (p300 – 950/CAT6 and p150 –950/CAT6) showed diminished heat inducibility, which suggests that the silencing activity is localized within the 155 bp of the HincII – PstI fragment (see Fig. 6B). 3.4. Interspecies comparison of noncoding and flanking DNA sequences of the hsp70.1 gene In search for evolutionary conserved repetitive motifs we compared the gene-flanking sequences of three orthologous

84

A. Fiszer-Kierzkowska et al. / Biochimica et Biophysica Acta 1625 (2003) 77–87

Fig. 6. In vitro assay of the promoter activity. Left panel: Schematic representation of the plasmids used for transfections. Different fragments of the promoter and further upstream region of the rat hsp70.1 gene were cloned in pBLCAT6 [40] vector. Drawing of the structure of analyzed DNA region (upper row) shows restriction sites used for subcloning. Right panel: Data from transfection experiments. Plasmids were transiently transfected to the rat FTO cells and CAT activity was determined in control (white bars) and heat shocked (gray bars) cells. CAT activity is shown as percentage of acetylated chloramphenicol. Each value is expressed as a mean (with the standard deviation) of four transfections. (A) Experiments performed to evaluate function of the proximal promoter region. (B) Functional analysis of the far upstream DNA region and repetitive sequences therein.

(centromeric) MHC III-linked hsp70i genes—the human hsp70.2 gene [6] (acc. no. AF134726), the mouse hsp70.1 gene [10] (acc. no. AF109906) and the rat hsp70.1 gene [7] (acc. no. X74271). We observed similar distribution of SINEs in the 5V flanking regions of all three genes. As in the case of the rat hsp70.1 gene three SINE sequences were found upstream of the mouse hsp70.1 gene (PB1D7, PB1D10, MMB1F, beginning 1.07, 1.56 and 1.76 kbp upstream of the ATG codon, respectively). Also, in the region preceding the human hsp70.2 gene, three Alu sequences were found at distances of 1.4, 1.65 and 1.76 kbp upstream of the ATG codon. As all these are speciesspecific sequences, it seems that SINE sequences evolved in

this region of the mammalian genome independently, after diversification to the distinct species. Searching for conserved microsatellite sequences in the mouse hsp70.1 gene, we found a long tract (CC(TC)9 CTCCC(TC)11TT(TC)9(GT)19) composed of Pu/Py sequences devoid of perfect mirror symmetry and a (GT)19 repeat. This sequence, with partial similarity to the microsatellite sequence found in rat is, however, localized on the opposite DNA strand, what probably excludes a common evolutionary origin. In human there are no simple repeats neither in the region corresponding to the position of microsatellite sequence in the rat hsp70.1 gene nor in the whole 3.0-kbp region upstream of the ATG codon.

A. Fiszer-Kierzkowska et al. / Biochimica et Biophysica Acta 1625 (2003) 77–87

We also compared the nucleotide sequence of the untranslated regions (UTRs) of the rat, mouse and human genes. The 5VUTRs showed about 80% identity between mouse and rat hsp70.1 genes, and about 50% identity between rat hsp70.1 and human hsp70.2 genes (see also Ref. [4]). We found a 33-nt triple tandem repeat (TTR) in the 5VUTR of the rat hsp70.1 gene (Fig. 2), which is very similar to the TTR described previously in the mouse hsp70.1 gene [7]. A database search revealed that such a TTR is also present in the 5V UTR of the second (telomeric) MHC III-linked hsp70i gene of the mouse (hsp70.3 gene [3]) and in the orthologous rat gene (hsp70.2 gene [4]) but is absent from any other hsp70 gene. The functional meaning of this TTR in hsp70 genes is unknown. The nucleotide sequences of the 3V UTR of the rat hsp70.1 and human hsp70.2 genes show 70% identity. Interestingly, the nucleotide sequence of the rat gene is evidently interrupted by a gap of 239 bp (Fig. 2) and the identity between the DNA regions which flank the gap is 84%. The presence of the microsatellite sequence (AT)17 flanking the 5Vend of this 239-bp sequence in the mouse gene suggests a possible deletion/insertion event, and according to the mammalian phylogenetic tree, it seems that this diversity could have been caused by insertion rather than by deletion.

4. Discussion Functional analysis performed in this study revealed several interesting features concerning differential effect of various fragments of the promoter region on the level of heat inducible expression of the rat hsp70.1 gene. Promoter region of the gene contains 3 HSE elements, all localized within the approximately 450 bp DNA fragment adjacent to the transcription unit of the hsp70.1 gene. We found that functional cooperation between these HSE sequences is required for efficient heat inducible transcription. None of HSEs alone, even the HSE1 with almost perfect match with the consensus sequence, is able to activate transcription of the reporter gene. Our data show that although HSE elements are indispensable for heat shock activation of the hsp70.1 promoter, there are DNA sequences localized more upstream, which can significantly affect transcription efficiency. This upstream fragment, of approximately 1.5 kbp, was found to contain regions which can exert either stimulatory or inhibitory effect. We observed a very significant enhancement of heat inducible transcription when the promoter region containing three HSE sequences was extended to approximately 900 bp (up to 869 PstI restriction site); see Fig. 6B. In contrast, DNA sequences localized between positions 869 and 1024 (relative to the tsp) significantly reduced the inducibility of the rat hsp70.1 gene promoter. This silencing effect could be, however, abolished when upstream sequen-

85

ces containing the microsatellite region were present in the construct. This suggests possible cis – trans interactions at various sites found within the 1.5-kbp distal promoter region containing three HSEs. These interactions can probably affect transcription machinery and/or stabilize/destabilize other crucial interactions. As a relatively large DNA fragment has to be analyzed in search for cis regulatory elements involved in modulation of HSE-HSF dependent activation of the hsp70i genes, this would need a separate study. It seems, however, that even the present preliminary data may be useful, e.g. for construction of the vectors containing inducible promoters. It has to be noted that the differential effect exerted by various fragments of the promoter region was already observed by others in the case of telomeric hsp70i genes. The 224-bp DNA fragment of the rat hsp70.2 gene, adjacent to the transcription initiation site, was sufficient to govern heat inducible expression of the CAT reporter gene; however, the induction level was rather low (approximately threefold) and some activity was observed at physiological temperature. Quite unexpectedly, a longer fragment of the promoter region (of 387 bp) activated the reporter gene in a constitutive way and the expression level was independent of heat shock [32]. Similar studies of the mouse hsp70.3 gene revealed that a 2.3-kb fragment of the 5Vend sequences was at least three times more effective as a heat inducible promoter than a 292-bp fragment when assayed in HeLa cells. However, the same DNA fragments showed comparable promoter activities in mouse L cells [3]. Taken together, our data and those of others, it seems evident that cis elements other than HSEs could be involved in the regulation of hsp70i gene expression. It may well be that such elements are responsible for frequently observed tissue-specific differences in the induction level of the hsp70i genes. It is also possible that such elements can play a role in differential expression of two of a MHC III-linked hsp70i genes. Differential expression in response to various conditions such as hypertonicity [13], ischemia/reperfusion [14], exposition to H2O2 and NaCl [12] or during differentiation of hematopoietic cells [15] was already communicated. It remains to be established whether DNA sequences localized upstream the HSEs have any functional meaning in vivo. The aim of our study was also to find out whether a complex microsatellite region identified far upstream from the beginning of the hsp70.1 gene transcription unit can affect expression of the gene. Potential cruciform and HDNA forming sequences are relatively frequent in genomes of higher eucaryotes. They are localized predominantly in the 5V flanking regions and UTRs of genes but are almost excluded from coding sequences [43,44]. Both types of structures are suspected to play some functional roles, e.g. in transcription, recombination or DNA replication. Cruciform forming sequences may affect transcription by differential DNA – protein interactions depending on the DNA conformation (dsDNA versus cruciform) [45]. The effect of H-

86

A. Fiszer-Kierzkowska et al. / Biochimica et Biophysica Acta 1625 (2003) 77–87

DNA forming sequences on transcription efficiency is rather unpredictable without performing functional studies; in fact, positive [38,39,46 –49], negative [50,51] or neutral [52 – 54] or even variable [36,55] influence of Pu/Py sequences has been found. Other observations suggest that in some cases H-DNA may play the role of transcriptional insulator separating the gene from the influence of its chromosomal environment [56]. Looking at the results of the said studies, it seems that the way such sequences can affect transcription of a particular gene probably depends on such factors as the length of the tract, the perfection of its symmetry, its distance from the gene and even surrounding sequences context. Although there is still a controversy concerning the existence of H-DNA structures in vivo, crucial may be the differential interaction with cellular proteins showing affinity to either dsDNA or ssDNA or triplex structure formed within the same sequence [47,57 – 59]. It has been even proposed that a transcriptional mechanism functioning through interconversion between ds- and ssDNA (being part of H-DNA structure) is common, especially for growth-related genes [39]. In our study we found that hsp70i-linked Py-Pu-containing region had the potential to reverse the effect of putative silencing sequences. It remains to be found out what kind (if any) of non-B DNA structure may be formed in vivo by the microsatellite region identified by us and whether any DNA – protein interaction(s) influencing the transcription efficiency occur in this region.

Acknowledgements We thank Mrs. K. Klyszcz for technical assistance and B. Sarecka, M. Sci. for cooperation during preliminary transfection assays. We are grateful to Dr. A. Sochanik for critical reading of the manuscript. This work has been supported by a grant from the State Committee for Scientific Research (grant 6 P04A 036 15) to K.L., and by a grant from the Foundation for Polish Science (grant BIMOL 76/9) to Z.K.

References [1] E. Gunther, L. Walter, Experientia 50 (1994) 987 – 1001. [2] B. Wu, C. Hunt, R. Morimoto, Mol. Cell. Biol. 5 (1985) 330 – 341. [3] M.D. Perry, L. Aujame, S. Shtang, L.A. Moran, Gene 146 (1994) 273 – 278. [4] L. Walter, F. Rauh, E. Gunther, Immunogenetics 40 (1994) 325 – 330. [5] M. Matsumoto, H. Fujimoto, Biochem. Biophys. Res. Commun. 166 (1990) 43 – 49. [6] C.M. Milner, R.D. Campbell, Immunogenetics 32 (1990) 242 – 251. [7] C. Hunt, S. Calderwood, Gene 87 (1990) 199 – 204. [8] K. Lisowska, Z. Krawczyk, W. Widlak, P. Wolniczek, J. Wisniewski, Biochim. Biophys. Acta 1219 (1994) 64 – 72. [9] F.M. Longo, S. Wang, P. Narasimhan, J.S. Zhang, J. Chen, S.M. Massa, F.R. Sharp, J. Neurosci. Res. 36 (1993) 325 – 335.

[10] R. Mestril, S.H. Chi, M.R. Sayen, W.H. Dillmann, Biochem. J. 298 Pt 3 (1994) 561 – 569. [11] L. Huang, N.F. Mivechi, D. Moskophidis, Mol. Cell. Biol. 21 (2001) 8575 – 8591. [12] J.S. Lee, J.S. Seo, Exp. Mol. Med. 34 (2002) 131 – 136. [13] S.K. Woo, S.D. Lee, K.Y. Na, W.K. Park, H.M. Kwon, Mol. Cell. Biol. 22 (2002) 5753 – 5760. [14] Z. Akcetin, R. Pregla, D. Darmer, H. Heynemann, J. Haerting, H.J. Bromme, J. Holtz, Urol. Res. 27 (1999) 306 – 311. [15] S. Leppa, R. Kajanne, L. Arminen, L. Sistonen, J. Biol. Chem. 276 (2001) 31713 – 31719. [16] C. Jolly, R.I. Morimoto, J. Natl. Cancer Inst. 92 (2000) 1564 – 1572. [17] C. Sarto, P.A. Binz, P. Mocarelli, Electrophoresis 21 (2000) 1218 – 1226. [18] J.J. Cotto, R.I. Morimoto, Biochem. Soc. Symp. 64 (1999) 105 – 118. [19] T. Taira, Y. Negishi, F. Kihara, S.M. Iguchi-Ariga, H. Ariga, Biochim. Biophys. Acta 1130 (1992) 166 – 174. [20] H. Kamano, K.H. Klempnauer, Oncogene 14 (1997) 1223 – 1229. [21] C. Kanei-Ishii, T. Yasukawa, R.I. Morimoto, S. Ishii, J. Biol. Chem. 269 (1994) 15768 – 15775. [22] S.N. Agoff, J. Hou, D.I. Linzer, B. Wu, Science 259 (1993) 84 – 87. [23] J. Tanikawa, E. Ichikawa-Iwata, C. Kanei-Ishii, A. Nakai, S. Matsuzawa, J.C. Reed, S. Ishii, J. Biol. Chem. 275 (2000) 15578 – 15585. [24] S.N. Agoff, B. Wu, Oncogene 9 (1994) 3707 – 3711. [25] I.C. Taylor, R.E. Kingston, Mol. Cell. Biol. 10 (1990) 176 – 183. [26] E.M. Thompson, E. Christians, M.G. Stinnakre, J.P. Renard, Mol. Cell. Biol. 14 (1994) 4694 – 4703. [27] E.M. Thompson, E. Legouy, E. Christians, J.P. Renard, Development 121 (1995) 3425 – 3437. [28] C. Hunt, R.I. Morimoto, Proc. Natl. Acad. Sci. U. S. A. 82 (1985) 6455 – 6459. [29] B.J. Wu, G.T. Williams, R.I. Morimoto, Proc. Natl. Acad. Sci. U. S. A. 84 (1987) 2203 – 2207. [30] K. Abravaya, B. Phillips, R.I. Morimoto, Mol. Cell. Biol. 11 (1991) 586 – 592. [31] S. Shimizu, K. Nomura, M. Ujihara, H. Demura, Biochem. Biophys. Res. Commun. 257 (1999) 193 – 198. [32] H. Friedrich, L. Walter, E. Gunther, Biochim. Biophys. Acta 1395 (1998) 57 – 61. [33] T. Shimokawa, H. Fujimoto, Biochem. Biophys. Res. Commun. 221 (1996) 843 – 848. [34] T. Konishi, M. Nomoto, K. Shimizu, T. Abe, H. Itoh, J. Biochem. 118 (1995) 1021 – 1029. [35] K. Lisowska, T. Loch, A. Fiszer-Kierzkowska, D. Scieglinska, Z. Krawczyk, Acta Biochim. Pol. 44 (1997) 147 – 152. [36] Y. Kohwi, T. Kohwi-Shigematsu, Genes Dev. 5 (1991) 2547 – 2554. [37] E.H. Postel, Ann. N.Y. Acad. Sci. 660 (1992) 57 – 63. [38] M. Santra, K.G. Danielson, R.V. Iozzo, J. Biol. Chem. 269 (1994) 579 – 587. [39] A. Rustighi, M.A. Tessari, F. Vascotto, R. Sgarra, V. Giancotti, G. Manfioletti, Biochemistry 41 (2002) 1229 – 1240. [40] M. Boshart, M. Kluppel, A. Schmidt, G. Schutz, B. Luckow, Gene 110 (1992) 129. [41] M.M. Bradford, Anal. Biochem. 72 (1976) 248 – 254. [42] F. Pothier, M. Ouellet, J.P. Julien, S.L. Guerin, DNA Cell Biol. 11 (1992) 83 – 90. [43] G.P. Schroth, P.S. Ho, Nucleic Acids Res. 23 (1995) 1977 – 1983. [44] R. Cox, S.M. Mirkin, Proc. Natl. Acad. Sci. U. S. A. 94 (1997) 5237 – 5242. [45] C. Spiro, C.T. McMurray, J. Biol. Chem. 272 (1997) 33145 – 33152. [46] A.B. Firulli, D.C. Maibenco, A.J. Kinniburgh, Arch. Biochem. Biophys. 310 (1994) 236 – 242. [47] C. Mayfield, S. Ebbinghaus, J. Gee, D. Jones, B. Rodu, M. Squibb, D. Miller, J. Biol. Chem. 269 (1994) 18232 – 18238. [48] G. Raghu, S. Tevosian, S. Anant, K.N. Subramanian, D.L. George, S.M. Mirkin, Nucleic Acids Res. 22 (1994) 3271 – 3279.

A. Fiszer-Kierzkowska et al. / Biochimica et Biophysica Acta 1625 (2003) 77–87 [49] K. Zahedi, J.J. Bissler, A.E. Prada, J.A. Prada, A.E. Davis, J. Immunol. 162 (1999) 7249 – 7255. [50] S.K. Brahmachari, P.S. Sarkar, S. Raghavan, M. Narayan, A.K. Maiti, Gene 190 (1997) 17 – 26. [51] P.S. Sarkar, S.K. Brahmachari, Nucleic Acids Res. 20 (1992) 5713 – 5718. [52] G.S. Pahwa, L.J. Maher, M.A. Hollingsworth, J. Biol. Chem. 271 (1996) 26543 – 26546. [53] N.A. Becker, L.J. Maher, Nucleic Acids Res. 26 (1998) 1951 – 1958. [54] N.A. Becker, H.A. O’Neill, J.M. Zimmerman , L.J. Maher III, J. Biol. Chem. 275 (2000) 40218 – 40225.

87

[55] G. Xu, A.G. Goodridge, Arch. Biochem. Biophys. 363 (1999) 202 – 212. [56] D. Michel, G. Chatelain, Y. Herault, F. Harper, G. Brun, Cell. Mol. Biol. Res. 39 (1993) 131 – 140. [57] C. Mayfield, M. Squibb, D. Miller, Biochemistry 33 (1994) 3358 – 3363. [58] J.F. Reddoch, D.M. Miller, Biochemistry 34 (1995) 7659 – 7667. [59] G. Li, G.V. Tolstonog, P. Traub, DNA Cell Biol. 21 (2002) 163 – 188. [60] J. Kim, J.A. Martignetti, M.R. Shen, J. Brosius, P. Deininger, Proc. Natl. Acad. Sci. U. S. A. 91 (1994) 3607 – 3611.