The Identification of a Cis-element and a Trans-acting Factor Involved in the Response to Polyamines and Polyamine Analogues in the Regulation of the Human Spermidine/Spermine N 1 -Acetyltransferase Gene Transcription*

The superinduction of spermidine/spermine N 1 acetyltransferase (SSAT) gene has been associated with a cytotoxic response to a new class of antineoplastic polyamine analogues. The initial mechanism of SSAT superinduction is an increase in transcription in response to analogue exposure. This increased transcription appears to be modulated through the association between a nuclear protein factor and a cis-element described here as the polyamine-responsive element (PRE). The PRE was identified as a 9-base pair sequence, 5 * -TATGACTAA-3 * , in the context of a 31-base pair stretch from –1522 to –1492 base pairs with respect to the SSAT transcriptional start site. This element binds a nuclear factor from polyamine analogue-respon-sive cells, but not from polyamine analogue-insensitive cells. The labeled PRE was used to clone and identify the transcription factor, Nrf-2, that binds constitutively to the PRE sequence. Although the PRE sequence shares homology to the originally identified Nrf-2 recognition sequence, the two sequences are not identical. The Nrf-2 transcription factor appears only to be present in cell types that are capable of expressing high amounts of SSAT. The results of these studies suggest that Nrf-2, bound to the PRE, plays an important regulatory role of expression of the human SSAT gene.

The requirement for polyamines in growth and development is absolute in eukaryotic cells (1). Although this requirement has been well established, some of the precise molecular functions of the polyamines have only recently been elucidated. Several newly synthesized polyamine analogues have been examined for their potential as antitumor agents (2)(3)(4). These compounds have been designed to alter the regulation of polyamine metabolism and interfere with the normal functions of polyamines in tumor cells. Some of these compounds appear to act in an association with an ability to superinduce the expression of spermidine/spermine N 1 -acetyltransferase (SSAT), 1 the rate-limiting enzyme in polyamine catabolism (5)(6)(7). The su-perinduction of SSAT has been associated with a phenotypespecific sensitivity to some of the antitumor polyamine analogues, and there are several reports demonstrating superinduction of SSAT and tumor cell toxicity (8 -13). Therefore, there is considerable interest in elucidating the mechanisms that differentiate sensitive tumors from analogue-insensitive tumors and normal tissue.
Control of the analogue-mediated superinduction of SSAT is initiated at the level of transcription (14,15). The -fold induction of SSAT transcription by analogue exposure is relatively small (2-7-fold) compared with the ultimate induction of enzyme activity; however, this increased transcription is necessary for the downstream events, which can result in Ͼ1000-fold increases in SSAT activity (8,16,17). Here we identify a cis-element in the 5Ј regulatory region of the human SSAT and define it as a polyamine-responsive element (PRE). Further, electrophoretic mobility shift assay (EMSA) results demonstrate that the PRE is bound by a cell type-specific DNAbinding protein that appears to exist only in cells that are capable of transcribing the SSAT gene (17,18). We used the labeled PRE in an expression library screening strategy to clone and identify the protein(s) responsible for the cell typespecific binding activity and subsequent transcriptional activation. The results of the experiments presented below suggest that the natural polyamines and some analogues can play a regulatory role in gene expression through modulation of the interaction between cis-elements and trans-acting factors.

EXPERIMENTAL PROCEDURES
Chemicals-N 1 ,N 11 -Bis(ethyl)norspermine (BENSpm) was supplied by Parke-Davis Pharmaceuticals (Ann Arbor, MI). 2-Difluoromethylornithine (DFMO) was obtained as a gift from Marion-Merrell-Dow Research Institute (Cincinnati, OH). The radionucleotides ([␣-32 P]dATP, [␣-32 P]dCTP, and [␥-32 P]ATP) and the Thermo Sequenase radiolabeled terminator cycle sequencing kit were purchased from Amersham Pharmacia Biotech. Restriction and DNA modifying enzymes and Lipofectin reagent were purchased from Life Technologies, Inc. and New England Biolabs (Beverly, MA). The luciferase assay system was purchased from Promega (Madison, WI), and the Gal-XE chemiluminescent reporter gene assay system was purchased from ICN Pharmaceuticals (Cosa Mesa, CA). Oligo(dT)-cellulose was purchased from Boehringer Mannheim. The cDNA synthesis system was from Promega (Madison, WI). The gt11 vector and packaging extract were purchased from Stratagene (La Jolla, CA). The TA cloning kit was purchased from Invitrogen (Carlsbad, CA), and the Lambda Midi kit was from Qiagen (Santa Clarita, CA). Other chemicals were purchased from Sigma, Boehringer Mannheim, or J.T. Baker. All oligonucleotides used in PCR and EMSA were synthesized by Life Technologies, Inc.
Transient Transfection Assays-For transient transfection, 4 ϫ 10 5 H157 cells were seeded in a 35-mm diameter culture dish and cultured in RPMI 1640 medium containing 5 mM DFMO for 48 h. The Lipofectinmediated transfections were performed with 1.5 g of luciferase reporter plasmid DNA and 0.4 g of control plasmid pSV-␤-gal (Promega) according to the manufacturer's protocol. After a 5-h incubation, the DNA-Lipofectin complex-containing medium was replaced by RPMI 1640 medium containing 5 mM DFMO. Forty-eight hours after transfection, the cells were exposed to 10 M BENSpm as described under "Results." The cells were then prepared for luciferase activity measurements as per the instructions of the manufacturer (Promega). To account for variations in transfectional efficiency, the luciferase activity was normalized to ␤-galactosidase activity.
Cell Culture and Preparation of Nuclear Extracts-Human lung cancer cell lines H157, A549, and H82 were cultured in RPMI 1640 medium containing 9% calf serum, 100 units/ml penicillin, 100 units/ml streptomycin, and nuclear extracts were prepared from H157, A549, and H82 cells as previously reported (9). Frozen aliquots of the nuclear proteins were stored at Ϫ70°C for future use. The protein concentration of nuclear extracts was determined by the methods of Bradford (20).
EMSA-EMSA was performed as described previously (21). Briefly, DNA-binding reactions were carried out in a buffer (25 l final volume) containing 1 g of nuclear protein, 14 mM HEPES, pH7.9, 84 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, 0.3 mM MgCl 2 , 0.4 mM phenylmethylsulfonyl fluoride, 1 g of poly(dI⅐dC), and 10% glycerol. The reaction mixture was incubated at 0°C for 15 min, after which time the radiolabeled probes (10,000 -20,000 cpm) were added and the incubation was continued at room temperature for another 30 min. For competition experiments, the unlabeled double or single strand DNA was added to the binding reaction mixture before the addition of the radiolabeled probe. DNA-protein complexes were resolved on 5% polyacrylamide gel (37.5:1 acrylamide/bisacrylamide) with 0.5ϫ TBE buffer, using 150 V, at 4°C.
DNase I Protection Assay-DNase I protection assays were performed according to the method previously described (22). Briefly, F1 DNA fragment excised from pCRII/F1 (Fig. 1B) with HindIII ϩ XhoI was 3Ј end labeled by Klenow DNA polymerase. Binding reaction was carried out in the EMSA buffer described above (25 l final volume), containing 10 g of nuclear protein. After a preincubation of the reaction mixture on ice for 15 min, the radiolabeled probe (20,000 cpm) was added and the reaction was incubated at room temperature for another 30 min. Then 0.0625 units of DNase I in a total volume of 25 l (50 mM CaCl 2 , 10 mM MgCl 2 ) was added. The mixture was incubated for 2 min at room temperature, and digestion was stopped by adding 50 l of 0.2 M NaCl, 30 mM EDTA, 1% SDS, and 100 g/ml yeast tRNA. After phenol extraction and ethanol precipitation, the sample was run on an 8% urea-polyacrylamide gel with a Maxam-Gilbert G reaction of the same restriction fragment as a marker.
RNA Isolation-Total cellular RNAs from H157, A549, and H82 cells were extracted using the acid phenol-guanidine isothiocyanate method (23). Poly(A ϩ ) RNA was isolated using oligo(dT)-cellulose (Boehringer Mannheim) chromatography following the protocol provided by the manufacturer.
Construction and Screening of a A549 cDNA Expression Library-A549 cDNA was synthesized using the Promega cDNA synthesis system. Simply, oligo(dT) and avian myeloblastosis virus reverse transcriptase were used to prime the first strand of cDNA from 10 g of A549 poly(A ϩ ) RNA. Double-stranded cDNA was then synthesized us-ing DNA polymerase I. After size fractionation to remove small cDNA fragments (less than 400 bp) double-stranded cDNA were ligated with EcoRI adaptors on both ends and inserted into gt11 vector. The recombinant cDNA library was then packaged using Gigapack III packing extract (Stratagene) and amplified in Y1088 host strain.
The cDNA expression library was screened for clones encoding proteins that can bind to the polyamine-responsive element (PRE). The probe used for screening was generated starting with the two complementary, synthetic oligonucleotides containing the core PRE sequence (bold) shown below.

SEQUENCE 1
Five micrograms of each oligonucleotide were phosphorylated with unlabeled rATP and polynucleotide kinase in a 50-l reaction. The oligonucleotides were then mixed and heated at 65°C for 15 min. After cooling to 4°C within 1 h, T4 ligase was added and incubated for 16 h at 16°C. The resulting concatenated DNA had a mean length of ϳ 200 bp. Radiolabeled probe was prepared by fill-in reaction of concatenated PRE using [␣-32 P]dCTP and Klenow DNA polymerase.
Approximately 10 6 phage particles from the A549 cDNA expression library were plated for the primary screen. Filter lifts (Hybond-C Nitrocellulose filter, Amersham Pharmacia Biotech) were prehybridized in the binding buffer (50 mM NaCl, 1 mM dithiothreitol, 1 mM EDTA, 10 mM Tris-HCl, pH 7.5) with 5% (w/v) nonfat dry milk. Hybridization was performed 1 h in the binding buffer with 10 6 cpm/ml labeled concatenated probe and 5 g/ml poly(dI⅐dC). After hybridization, the filters were washed in binding buffer without probe. One positive clone was obtained after three rounds of screening from the expression library. The phage DNA from the positive clone was prepared using a Qiagen Lambda Midi kit, and the cDNA insert was sequenced using a Thermo Sequenase radiolabeled terminator cycle sequencing kit (Amersham Pharmacia Biotech).
Northern Blot Assay-Ten g of total RNA from H157, A549, or H82 cells were fractionated on a 1.5% agarose gel containing 6% formaldehyde and transferred to GeneScreen membrane (DuPont), which was hybridized with a random primer-labeled (24) Nrf-2 cDNA probe produced from PCR (25). Ribosomal RNA (28 S) was used as a loading control.
Southwestern Blotting-Thirty g of nuclear proteins were resolved in a 8% SDS-polyacrylamide gel and electrotransferred (20 V, 16 h) onto a Hybond ECL nitrocellulose filter (Amersham Pharmacia Biotech). The proteins were then renatured in the binding buffer (50 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, 10 mM HEPES, pH 7.9) containing guanidine hydrochloride in the concentrations of 6, 3, 1.5, 0.75, and 0.375 M, successively. The filter was blocked for 1 h in the binding buffer with 5% (w/v) nonfat dry milk powder, hybridized for 1 h in the binding buffer containing 10 6 cpm/ml radiolabeled probe and 5 g/ml poly(dI⅐dC), and washed 4 ϫ 5 min with the binding buffer. The filter was dried and exposed to Kodak X-Omat film.

RESULTS
Localization of the PRE in the SSAT Promoter-Initial experiments using deletion constructs covering 3500 bp upstream from the SSAT transcriptional start site revealed the region between Ϫ1553 and Ϫ1232 bp to contain a positive control element, which was induced approximately 3-fold by exposure to the bis(ethyl)polyamine analogues (26). To more precisely map the location of the PRE, this 321-bp region was divided into three smaller fragments. The three fragments of the human SSAT promoter sequence, F1 (Ϫ1562 to Ϫ1444 bp), F2 (Ϫ1463 to Ϫ1334 bp), and F3 (Ϫ1333 to Ϫ1238 bp), were produced by PCR reactions and subcloned into the pCRII vector for ease of large scale preparation. The wild type sequence of each fragment was confirmed by dideoxy sequencing. EcoRI restriction of the appropriate pCRII subclone produced the respective F1, F2, and F3 oligomers. Nuclear extracts were prepared from analogue-treated and untreated H157 and H82 cells, and these nuclear protein extracts were used with the fill-in labeled probes for shift analysis.
In the EMSA analysis, only the 119-bp F1 fragment (Ϫ1562 to Ϫ1444 bp) demonstrated a cell type-specific shift (Fig. 2). Importantly, the shifted band was only observed when nuclear extracts from H157 cells were used. This shift could be completely and specifically competed by excess cold F1 probe. It is important to note that the exposure to the analogue had no effect on the shift pattern, suggesting that the protein responsible for the shifting complex is constitutively expressed.
To further localize the cis-element responsible for H157 nuclear protein binding, DNase I protection analysis of the F1 fragment was performed in the presence of nuclear extracts from both the H157 and H82 cells. Consistent with the results of the EMSA studies, the H82 nuclear extracts provided no protection from DNase I digestion, suggesting that the specific trans-acting factors responsible for binding to this region are not expressed in the H82 cells (Fig. 3). By contrast, an area spanning a 31-bp region in F1 (bases Ϫ1522 to Ϫ1492) was protected when nuclear extracts from H157 cells were used (Fig. 3). The protection pattern observed suggested at least two contact points of DNA/protein interactions (Fig. 3).
A 9-bp Core Element in the SSAT Promoter Mediates the H157-specific Nuclear Protein-DNA Binding-To define the putative PRE, a 40-bp double-stranded oligomer (sequence 7b, Fig. 1A) was synthesized which spans the entire region of DNase I protection. That the nuclear protein from H157 could specifically shift this double-stranded oligomer was confirmed using the standard EMSA assay (Fig. 4). Using a series of deletions and mutations to the 7b sequence, a 9-base core region consisting of 5Ј-TATGACTAA-3Ј was identified (Table I).
Only two transversion substitutions in the core region maintain the ability to be recognized and shifted by the H157 nuclear proteins (Table I). These two double-stranded oligomers were also the only ones capable of competing with the binding of the H157 nuclear protein/F1 fragment shifted band. Combined, these results strongly suggest that this 9-bp core region contains the cognate sequence for nuclear protein binding. It should be noted that the lower band in Fig. 4 (A and B) represents a shift resulting from a single strand binding protein, which is not cell type-specific. The sequence in 7b probes responsible for this binding is 5Ј-TAAATTA-3Ј (data not shown). This band is not observed in the EMSA analysis using the core sequence without the 5Ј-TAAATTA-3Ј, sequence as represented by the F5 subclone (Fig. 5B). Also, F5 only competes with the cell type-specific upper band when the F1 probe is used in the shift-up assays (Fig. 5A).
The PRE Mediates Polyamine Analogue-induced Expression of Reporter Constructs Containing the SSAT Promoter Region-To determine whether the putative PRE could mediate polyamine analogue-induced transcription, a 40-bp region surrounding the core PRE was used to construct several luciferase reporter constructs. Two sets of constructs were made, one using the PRE in "ϩ" and "Ϫ" orientations upstream from the minimal promoter region of the SSAT gene (bp Ϫ93 to Ϫ1) and one set with the PRE in the "ϩ" and "-" orientations upstream from bp Ϫ798 to -1 (Fig. 1B). Initial transient transfection studies using these constructs in BENSpm-treated H157 cells demonstrated a small, but reproducible induction of luciferase expression (data not shown). However, the observed background of these experiments was consistently high. If the H157 cells were first treated with 5 mM DFMO to reduce their natural polyamine concentrations, the background levels of luciferase activity were consistently lower and the -fold induction was consistently greater (Fig. 6A). It should be noted that the observed increase in luciferase activity is consistent with the increase in transcription of the SSAT gene observed in run-off assays in response to polyamine analogue exposure (17,27). The results of these experiments clearly demonstrate the ability of the PRE to mediate an increased expression of the luciferase constructs in response to BENSpm treatment in a direction-and position-independent manner. It is also important to note that the natural polyamines, spermidine and spermine, induced the expression of the transfected constructs in a similar fashion to BENSpm (Fig. 6B).
Use of the PRE to Screen an Expression Library-The identification of the PRE sequence (5Ј-CGCTATGACTAAGCG-3Ј) (18) allowed the design of a ϳ200-bp concatemer to be used for screening of a gt11 expression library. Since the PRE se-quence was only shifted by nuclear extracts from the non-small cell lines, the library was constructed from the rapidly growing human adenocarcinoma line, A549. This cell line demonstrates increased transcription and SSAT enzyme activity in response to BENSpm (9,17). Most importantly for this study, A549 expresses a high amount of PRE nuclear binding factor as determined by EMSA analysis (Fig. 7).
Screening of the A549 expression library yielded a positive clone containing a 1.6-kilobase pair cDNA insert. Dideoxysequencing of the cDNA insert revealed 100% homology to an NF-E2-related transcription factor, Nrf-2 (25). The Nrf-2 transcription factor recognizes a 34-base pair sequence that contains two putative AP1 consensus sequences (25) (Fig. 8). These two AP1 binding sites contain a near consensus sequence of the PRE, both in the forward and reverse orientations.
As a first step to confirm that Nrf-2 was the protein responsible for the observed shift, we synthesized a 34-base pair Nrf-2 binding sequence to be used in EMSA competition assays. When the Nrf-2 binding element was used as the labeled probe, the core PRE, contained within an 18-bp oligomer homologous to -1515 to -1498 of the human SSAT gene (17), was fully capable of competing with the Nrf-2 sequence for binding of the H157 nuclear protein (Fig. 9A). The PRE concatemer also competed with the Nrf-2 probe for binding of the H157 nuclear protein (Fig. 9A). Similarly, when the 18-bp PRE-containing oligomer was labeled and used in the EMSA assay, the Nrf-2 oligomer successfully competed with the PRE for H157 nuclear protein binding (Fig. 9B). It should be noted that the H82 nuclear proteins were not capable of shifting either the PRE containing probe (Fig. 7) or the Nrf-2 containing probe (Fig.  9A). When A549 human adenocarcinoma cells, another cell line that is capable of inducing high SSAT expression (9), results were identical to those obtained with H157 nuclear extracts (data not shown).
To confirm that the core PRE sequence was required for the observed binding patterns, competition assays using representative mutant oligomers were performed. When the central 3 bases of the core region of the PRE sequence were mutated (probe 31, Fig. 1A), the mutant PRE was no longer capable of competing with the labeled Nrf-2 sequence for binding of the H157 nuclear extracts (Fig. 10). Identical results were observed when the same mutation was introduced into the concatenated probe (Fig. 10). These results are entirely consistent with the hypothesis that the Nrf-2 transcription factor requires the PRE core sequence for its recognition and the observed EMSA patterns.
Differential Expression of Nrf-2 mRNA in Human Lung Cancer Lines-The above results suggest that the ability of the lung cancer cell types to express SSAT may be related to the basal expression of the Nrf-2 transcription factor. Specifically, the exhibited EMSA shift patterns in H157 and A549 suggest that they, but not H82, express the Nrf-2 transcription factor. To determine if there was a difference between these cell types in their ability to express Nrf-2 at the message level, Northern blotting analysis was performed (Fig. 11A). Significant Nrf-2 message was detected only in the H157 and A549 cells, which express high levels of SSAT mRNA and protein in response to BENSpm treatment.
Southwestern Analysis of Nrf-2 Expression in H157 Cells-Nrf-2 antibodies useful for supershift or Western analyses are not currently available. Therefore, to provide corroborative data that Nrf-2 is the protein binding to the PRE, we used labeled PRE as a probe to perform Southwestern analyses. Thirty g of H157 nuclear protein were separated by SDSpolyacrylamide gel electrophoresis and transferred to a nylon membrane, which was then probed with the labeled ϳ200-bp PRE concatemer (Fig. 11B). After an overnight exposure, one prominent band and one minor band were observed. The prom-inent band runs identically to the reported gel behavior of Nrf-2, running at ϳ99 kDa (25). The minor band at ϳ120 kDa is currently unidentified.

DISCUSSION
The rate-limiting enzyme in polyamine catabolism, SSAT, was the first enzyme in the polyamine metabolic pathway demonstrated to be positively regulated by the natural polyamines and the antineoplastic polyamine analogues (8,10,28). The increase in expression is regulated at multiple steps, from modest, but necessary increases in transcription, to the stabilization of the active protein (10, 16, 17, 28 -32). In the current work, a cis-element in the SSAT promoter has been identified, the PRE, which mediates the transcriptional induction of SSAT by the polyamine analogue, BENSpm. The increased transcription appears to be associated with a constitutively expressed nuclear protein, Nrf-2, which is expressed in the analogueresponsive cell types. The observed analogue-induced 2-3-fold increase in expression of the chimeric luciferase constructs are consistent with results of nuclear run-on experiments defining the transcriptional control of human SSAT (14,17). These results and the run-on experiments are in agreement with the hypothesis that once induced at the transcriptional level by the analogues, the majority of the observed increases in SSAT protein are a result of post-transcriptional mechanisms (33).
The initial transfection experiments using the SSAT promoter/luciferase chimeras yielded results with a relatively high background of luciferase expression. Although exposure to BENSpm consistently produced a significant induction of luciferase activity, the -fold increase was small. However, when  Ϫ Ϫ cells were first depleted of their natural polyamines by pretreating with DFMO, the -fold induction increased. These results suggest that the normal concentration of intracellular polyamines is sufficient to drive transcription of the SSAT promoter when a large number of the promoter regions are available, as is the case after transfection. However, once the intracellular concentrations of polyamines have been depleted by DFMO treatment, this background transcription is reduced. Additionally, the chimeric constructs may be missing negative regulatory regions, which may be necessary to keep basal transcription low in cells with normal intracellular concentrations of polyamines, and the observed high background may be due to an efficient promoter region in the sequence examined.
The effects of the polyamines and the analogues on SSAT expression is unique among the enzymes in the polyamine metabolic pathway (34). The other rate-limiting steps in polyamine metabolism, specifically ornithine decarboxylase and S-adenosylmethionine decarboxylase, are both negatively regulated by excess polyamines (35,36). However, this regulation appears to be limited to the post-transcriptional level (37,38). The polyamines have also been implicated in the regulation of expression of several other genes. The growth-related proto-oncogenes c-myc, c-fos, and c-jun have been demonstrated to be down-regulated at the transcriptional level by the depletion of the natural polyamines by DFMO treatment (39,40).
The current data are consistent with the hypothesis that Nrf-2 is the nuclear factor that binds to the human PRE in the EMSA studies. Although satisfactory antibodies are not currently available to perform supershift studies, the Southwestern analyses performed in the above studies provide solid supporting evidence that Nrf-2 is the trans-acting protein.
Although unlikely, we cannot presently exclude the possibility that a protein other than Nrf-2 binds to both the PRE and Nrf-2 elements and has identical electrophoretic mobility.
The current results are entirely consistent with past studies suggesting that H82 cells have low or no ability to express the SSAT gene (8,9). The lack of expression appears to result from the lack of measurable transcription of SSAT in H82, either basally or after analogue treatment (17). It is possible that one of the reasons for this inability is the low expression of Nrf-2 in H82 cells demonstrated here. The potential lack of this DNAbinding protein is also consistent with the observation that the H82 cells lack a DNase hypersensitivity site observed in the H157 cells that maps precisely to the region where the PRE has FIG. 8. Sequence of Nrf-2 and PRE elements. Note that the PRE has been found to function in a position-and orientation-independent manner. Consequently it is important to note that the PRE has multiple homologies to the Nrf-2 recognition sequence in both the forward and reverse orientations. been identified (17). Similarly, in the current study, DNase footprinting using H82 nuclear extracts indicates that H82 lacks the nuclear factor responsible for binding to the PRE. Moi et al. (25) demonstrated a widespread expression of Nrf-2 mRNA in multiple normal human tissues. Nrf-2 was found to be highly expressed in normal human lung. These results suggest that the H157 and A549 cells are more normal in their expression of Nrf-2 than H82, where Nrf-2 mRNA is nearly undetectable. Taken together, these data suggest that H82 and possibly other cell types may lack the necessary factors to express appreciable amounts of SSAT. Further, these data strongly support the importance of the PRE and Nrf-2 in playing a critical role in the polyamine analogue-induced expression of SSAT in sensitive cell types.
It should be noted that the binding of Nrf-2 to the cognate PRE does not appear to be mediated by the presence of the analogues. However, the increased expression of SSAT is only observed in the presence of analogue or natural polyamines, suggesting that the activity of Nrf-2 is altered in the presence of the cation. Basu et al. (41,42) have demonstrated profound changes in DNA conformation in the presence of the polyamine analogues. It has also been shown that DNA/protein interactions can be altered by the natural polyamines, as demonstrated by examining the effects of polyamines on restriction FIG. 9. A, competition EMSA analysis using the Nrf-2 binding element as the labeled probe. One g of nuclear extract from H157 cells (lanes 1-4) 11. A, expression of Nrf-2 mRNA in lung cancer cell lines. Ten g of total cellular RNA was used in each lane for Northern analysis. Lane 1, H82; lane 2, H157; lane 3, A549. Labeled Nrf-2 cDNA was used as the probe. B, Southwestern analysis of H157 nuclear proteins. Thirty g of H157 nuclear protein extract were separated by gel electrophoresis and transferred to nylon membrane, and the membrane was probed with the label PRE concatemer. The arrow indicates the position of the Nrf-2 protein.
These results suggest two possibilities for the involvement of Nrf-2 in SSAT expression. First, it is possible that Nrf-2 is modified in the presence of the polyamine analogue, thus activating transcription. Another possibility, potentially more likely, is that there is a partner protein that modulates the activity of Nrf-2 when cells are exposed to the analogue. Nrf-2 is known to have both an activation domain and a leucine zipper domain, which is typical among transcription factors that are modulated by protein/protein interactions (46). One family of proteins that has been demonstrated to associate with Nrf-2 and regulate its activity is the small Maf protein family (47)(48)(49). NF-E2 is a globin-specific transcription factor that recognizes the same consensus sequence as Nrf-2, and appears to be tissue-specific in its expression in erthythroid differentiation (50). By contrast, Nrf-2 appears to be more ubiquitous in its expression, but in all instances known thus far, it is modulated in a tissue-specific manner by the small Maf proteins (47). It has not yet been determined if the Maf family members play a role in expression of SSAT.
The mechanisms involved in SSAT expression will likely have significance beyond the regulation of the SSAT gene. Therefore, the polyamines may be expected to play a role in the regulation of other as of yet undefined genes. Since a number of polyamine analogues are under consideration for clinical trial, it is important to understand how such compounds might effect gene expression. If the action of the analogues is not limited solely to the regulation of the polyamine metabolic pathway, it will be important to understand how these compounds effect gene expression in order to avoid potential undesirable effects. Further, it has been demonstrated that overexpression of ODC is often associated with tumorigenesis and that polyamines are generally found in higher concentrations in tumor cells than in their normal counterparts (51,52). Therefore, it is possible that the increased polyamine concentration associated with transformed cells may be necessary for the expression of genes required for the neoplastic phenotype (53), perhaps by activation of Nrf-2.