SNAT2 ERE-Containing Promoter Region Is Activated by 17β-Estradiol.

To determine whether the putative ERE found in the SNAT2 promoter region was functional, we performed assays using SNAT2 promoter constructs and the luciferase reporter gene. We observed (Fig. 1A) that 1 × 10⁻⁸ M 17β-estradiol increased luciferase activity approximately twofold when using base pairs −1,872 to +265 of the SNAT2 promoter region. To corroborate that loss of the ERE site between bases −1,486 and −1,471 diminished the effect of the hormone, we used a series of unidirectional 5′ deletion constructs generated from the complete sequence. The 17β-estradiol–induced luciferase activity was reduced by ∼63% and 75% with the constructs encoding base pairs −872 to +265 and −336 to +265, respectively. The remaining activity was not generated by a direct interaction of ER-α with a specific ERE because there are no other putative EREs in the SNAT2 promoter. Deletion of base pairs −1,872 to −192 resulted in the loss of 17β-estradiol–induced activity.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Functional analysis of the ERE site in the rat SNAT2 promoter. (A) Effect of 1 × 10⁻⁸ M 17β-estradiol on SNAT2 promoter activity. HeLa cells were cotransfected with each promoter–reporter construct, the ER-α expression vector, and pRL-TK and incubated in the presence of 1 × 10⁻⁸ M 17β-estradiol or vehicle (ethanol). Luciferase activities were measured 24 h after transfection. The results are the mean ± SE of three independent experiments per construct. (B) Effect of the ER-α-sh plasmid on SNAT2 promoter activity. HeLa cells were cotransfected with the ER-α-sh plasmid or empty vector (Superior), the SNAT2 promoter–reporter construct (pSNAT2), or empty vector (pGL3b), the ER-α expression vector, and pRL-TK and incubated in the presence of 1 × 10⁻⁸ M 17β-estradiol or vehicle (ethanol). Luciferase activities were measured 24 h after transfection. The results are the mean ± SD of three independent experiments per construct. *P < 0.05; **P < 0.01 (significant difference compared with their respective control).

SNAT2 Promoter Activation by 17β-Estradiol Is ER-α–Dependent.

To determine whether the SNAT2 promoter activation mediated by 17β-estradiol is ER-α–dependent, we used an ER-α–sh plasmid to knock down the ER-α protein. As shown in Fig. 1B, ER-α knockdown prevented SNAT2 promoter activation by 17β-estradiol. There was no reduction in SNAT2 promoter activity induced by 17β-estradiol in control cells transfected with an empty plasmid. These results corroborate the importance of ER-α regulation of the SNAT2 promoter in response to 17β-estradiol and suggest that the remaining activity observed in the promoter constructs containing base pairs −872 to +265 and −336 to +265 could be ER-α–dependent.

ER-α Binds to the ERE in the SNAT2 Promoter in Vitro.

Because it has been established that ER-α activated by 17β-estradiol binds to the ERE sites of target genes, the next step was to corroborate whether ER-α could bind to the ERE site found in the SNAT2 promoter region. As shown in Fig. 2A, lane 2, electrophoretic mobility-shift assay (EMSA) analysis of the SNAT2 ERE site using extracts from HeLa cells transfected with ER-α indicated binding with a retarded band. Mutation of the SNAT2 ERE site (Table S1) abolished the formation of this complex (Fig. 2A, lane 3), suggesting that this site binds to ER-α. As a positive control, we used the ERE consensus sequence, which produced a very intense retarded band (Fig. 2A, lane 4). Interestingly, the ERE consensus band traveled further than the SNAT2 ERE complex in the gel, suggesting that different cofactors might bind to each complex. To demonstrate whether ER-α was bound to the SNAT2 ERE sequence, we used an anti–ER-α antibody. Fig. 2A, lane 5, demonstrates that preincubation with the antibody abolished the formation of the complex, possibly by blocking the ER-α DNA-binding site. The formation of a super-shifted band was observed (Fig. 2A, lane 7) with the consensus ERE sequence, indicating that the proteins involved in the formation of the complex might vary between the SNAT2 ERE and the ERE consensus sequence. In fact, liquid chromatography-elevated energy mass spectrometry (LC-MS^E) analysis confirmed the difference between both complexes (see below). The SNAT2 ERE sequence competed with the ERE consensus sequence for ER-α binding because the addition of a 200-fold molar excess of unlabeled (cold) SNAT2 ERE sequence nearly obliterated binding to the ERE consensus sequence (Fig. 2C).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Binding of ER-α to the ERE in the SNAT2 promoter in vitro. Double-stranded DNA oligonucleotides containing the SNAT2 ERE site (ERE SNAT2), the SNAT2 ERE mutated site (ERE SNAT2 m), or the ERE consensus sequence (ERE cons) were incubated with HeLa nuclear extracts from cells previously transfected with ER-α. The DNA sequences of the oligonucleotides used are indicated in Table S1. (A) Localization of the ERE sequence in the SNAT2 promoter region between the bases −1,486 and −1,471 upstream of the transcriptional start site. (B) EMSA analysis of HeLa nuclear extracts previously transfected with ER-α using labeled oligonucleotides containing the wild-type SNAT2 ERE site (lane 2), the SNAT2 ERE mutated site (lane 3), or the ERE consensus site (lane 4). The supershift assay was performed by preincubating 1 µg of anti–ER-α antibody with the HeLa nuclear extracts from cells previously transfected with ER-α. Probes for the labeled SNAT2 ERE site (lane 5), the SNAT2 ERE mutated site (lane 6), or the consensus ERE sequence (lane 7) were added. (C) Competitive assays using the labeled ERE consensus sequence and an excess of unlabeled SNAT2 ERE sequence (SNAT2 ERE cold) at 50-, 100-, 150-, and 200-fold molar excess are shown in lanes 3, 4, 5, and 6, respectively. FP, free probe.

ERα Binds to the SNAT2 ERE Site in the Rat SNAT2 Promoter in Vivo.

The functionality of the SNAT2 ERE site was also demonstrated in vivo. The increase of 17β-estradiol levels during gestation favored the binding of ER-α to the SNAT2 ERE site in the mammary gland as assessed by chromatin immunoprecipitation (ChIP) assays, demonstrating an amplification product on days 5, 14, and 20 of gestation (Fig. 3 A and B). Quantitative PCR was performed to quantify the binding of ER-α to the SNAT2 ERE site in the rat SNAT2 promoter. There was a progressive increase in binding during the gestation period, which corresponded to the increase in serum estradiol levels in the rat (Fig. 3 B and C, respectively). Virgin control rats in the diestrus period have very low levels of the hormone, and no ER-α binding was observed. A high level of hormone is needed to activate SNAT2 transcription, and therefore SNAT2 is up-regulated exclusively during the gestation period and possibly during the proestrus phase of the female rat. Additionally, the use of different concentrations of 17β-estradiol (from 1 × 10⁻⁶ to 1 × 10⁻¹¹ M) demonstrated that the SNAT2 promoter requires a higher concentration of hormone to induce transcription compared with the consensus ERE sequence (Fig. 3D). For instance, whereas 10⁻¹¹ and 10⁻¹⁰ M 17β-estradiol robustly activated the consensus ERE construct, the SNAT2 ERE required higher hormone concentrations to be activated to the same degree.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

ER-α binds to the SNAT2 ERE site in the rat SNAT2 promoter in vivo. (A) ChIP showing ER-α binding. ChIP assays were conducted on mammary glands from virgin control rats and 5-, 14-, or 20-d pregnant rats. Soluble chromatin was immunoprecipitated with 4 µg of anti–ER-α antibody or negative control IgG and subjected to PCR with a specific primer to the SNAT2 ERE site (Table S1). (B) Real-time PCR analysis of amplified products corresponding to ER-α binding to the SNAT2 ERE. This graph is representative of three independent experiments (n = 5 rats per group). (C) Serum 17β-estradiol levels in virgin control rats and 5-, 14-, and 20-d pregnant rats (n = 5 rats per group). (D) Effect of different concentrations of 17β-estradiol (1 × 10⁻⁶, 1 × 10⁻⁷, 1 × 10⁻⁸, 1 × 10⁻⁹, 1 × 10⁻¹⁰, and 1 × 10⁻¹¹ M) on the ERE consensus promoter–reporter construct and the SNAT2 promoter reporter construct activity. HeLa cells were cotransfected with each promoter–reporter construct, the ER-α expression vector, and pRL-TK and incubated in the presence of 17β-estradiol or vehicle (ethanol). Luciferase activities were measured 24 h after transfection. The results are the mean ± SD of three independent experiments per construct and are expressed as percentage (%) of luciferase activity increase compared with the vehicle (ethanol). Different letters differ: a > b > c. P < 0.05.

PARP1, Ku70, and GAPDH Bind to the SNAT2 Promoter ERE Sequence with ER-α.

We used a technique based on a desthiobiotin–triethylene glycol (TEG) oligo and streptavidin beads to precipitate the ER-α–ERE complex and identify the other proteins bound to the SNAT2 promoter region following activation by 17β-estradiol. For this assay, double-stranded DNA oligonucleotides containing the SNAT2 ERE site (ERE SNAT2 oligo) and the ERE consensus sequence (ERE cons oligo) were incubated with nuclear extracts from HeLa cells transfected with ER-α. HeLa nuclear extracts from untransfected cells were used as negative controls for each oligo. The ER-α–ERE complex was precipitated with streptavidin beads and eluted with 5 mM biotin (Fig. 4A). LC-MS^E (Table S2) and Western blot analysis (Fig. 4B) revealed the presence of poly(ADP-ribose) polymerase 1 (PARP1), Lupus Ku autoantigen protein p86 (Ku80), Ku70, and topoisomerase IIβ with ER-α on the ERE consensus oligo and PARP1, Ku70, and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as well as topoisomerase I and IIβ with ER-α on the SNAT2 ERE oligo. Although PARP1 bound equally well to both ERE sequences, the intensity of some of the bands bound to the SNAT2 ERE oligo was lower than to the ERE consensus oligo, particularly Ku70 and topoisomerase IIβ, possibly because ERα has a lower affinity for the SNAT2 ERE than the consensus ERE. Interestingly, topoisomerase I and GAPDH bind exclusively to the SNAT2 ERE site. Binding of these proteins was dependent on ER-α binding because most of them did not precipitate in the absence of ER-α (Fig. 4B).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

ER-α–ERE complex purification. (A) ER-α–ERE consensus and ER-α–SNAT2 ERE complexes were precipitated by using desthiobiotin–TEG oligos for the ERE consensus and the SNAT2 ERE sites (Table S1). The desthiobiotin–TEG oligos (2 µg) were incubated with HeLa nuclear extracts or HeLa nuclear extracts from cells previously transfected with ER-α, poly(dI-dC) (2 µg) and 1 × 10⁻⁸ M 17β-estradiol at 4 °C overnight in a rotation device. The volume was adjusted to 400 µL with binding buffer [50 mM Tris⋅HCl, pH 7.5, 5 mM MgCl₂, 2.5 mM EDTA, 2.5 mM DTT, 250 mM NaCl, 0.25 µg/µL poly(dI-dC), 20% glycerol]. Each sample was incubated with streptavidin magnetic beads for 2 h at 4 °C and washed three times with the same binding buffer in a rotation device. After the washes, the precipitated oligo-complex was eluted with 5 mM biotin and run on an SDS/PAGE for Western blot analysis. (B) Occupancy of Topo IIβ, Topo I, PARP1, Ku80, Ku70, ER-α, and GAPDH on the ERE consensus and SNAT2 ERE sequences. HeLa nuclear extracts without ER-α transfection were used as negative controls for each oligo. (C) Representative immunoblot of the knockdown sh-plasmids (PARP1, Ku70, and GAPDH) efficiency. (D) Effect of the PARP1, Ku70 and GAPDH sh-plasmids on ERE consensus activity. (E) Effect of the PARP1, Ku70, and GAPDH sh-plasmids on SNAT2 promoter activity. HeLa cells were cotransfected with PARP1, Ku70, or GAPDH sh-plasmids, with or without the ER-α expression vector, each promoter–reporter construct, and pRL-TK and incubated in the presence of 1 × 10⁻⁸M 17β-estradiol or vehicle (ethanol). Luciferase activities were measured 24 h after transfection. The results are the mean ± SD of three independent experiments per construct. Different letters differ: a > b > c. P < 0.05.

The activity of the luciferase reporter plasmid encoding the consensus ERE sequence in cells cotransfected with ER-α and stimulated with 17β-estradiol decreased significantly when cotransfected with PARP1 and Ku70 sh-plasmids (Fig. 4C). Interestingly, the luciferase reporter plasmid encoding the SNAT2 ERE sequence and the ER-α expression vector showed a significant decrease in activity in response to 17β-estradiol when GAPDH, PARP1, or Ku70 was knocked down (Fig. 4D).

GAPDH Accumulates in the Nucleus, Binds in Vivo to the SNAT2 Promoter of Mammary Epithelium During Gestation, and Is Specific for SNAT2 ERE Activation.

GAPDH is a metabolic enzyme involved in glycolysis, gluconeogenesis, and glyceroneogenesis. Intriguingly, most of the amino acids transported by SNAT2 are considered gluconeogenic substrates. Feeding dams a high-protein diet (30%) that contains an excess of gluconeogenic amino acids significantly increased the expression of cytoplasmic GAPDH in the rat mammary gland during gestation on day 20 compared with those fed an adequate-protein (20%) diet (Fig. 5 A and C). In addition, there was a progressive increase in the nuclear abundance of GAPDH during gestation in rat mammary tissue, reaching a peak on day 20 of gestation as shown by Western blot and immunohistochemical analysis (Fig. 5 B and D–G). Moreover, ChIP assays demonstrated that GAPDH binds in vivo to the SNAT2 ERE in rat mammary gland during gestation, especially on days 14 and 20, whereas virgin control rats lack this binding (Fig. 5H). All these data show that GAPDH is translocated to the nucleus during gestation to actively maintain SNAT2 gene expression. To determine the specificity of GAPDH binding to the SNAT2 ERE, we performed a precipitation assay using single-nucleotide mutations and flanking regions similar to the SNAT2 consensus sequence. Our results clearly demonstrated that GAPDH binding to the SNAT2 ERE is exclusively dependent on the palindromic sequence 5′-AGGTAATACAGACTT-3′; 5′-AGGTCATACAGACTT-3′ does not bind. Likewise, GAPDH binds to the 5′ flanking region of the palindromic sequence 5′-AAGA AGGTAA TAC AGACTT TTCT-3′ but not to 5′-TATC AGGTAA TAC AGACTT TTCT-3′ (bold type letters indicate the nucleotide or nucleotides that were changed in the sequence) (Fig. 5I).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

The expression of GAPDH in rat mammary glands. (A and B) A representative immunoblot of GAPDH, β-tubulin, and RNA polymerase II (Pol II) in cytoplasmic extract (A) and GAPDH, β-tubulin, and Pol II in the nuclear extract of mammary glands from rats fed 20% or 30% dietary protein (B). (C and D) Western blot densitometric analysis of cytoplasmic GAPDH/tubulin (C) and nuclear GAPDH/RNA polymerase II of the mammary glands from rats fed 20% or 30% dietary protein (D). The results are the mean ± SD of three different blots. (E) GAPDH immunohistochemistry in rat mammary gland during gestation on day 20. (F and G) Negative control (F) and GAPDH (G). The arrows indicate the presence of GAPDH in the nucleus of mammary epithelium cells. [Scale bars: 50 µm (E and F); 10 µm (G).] (H) ChIP showing GAPDH binding to the SNAT2 ERE. ChIP assays were conducted on mammary glands from virgin control rats and 5-, 14-, and 20-d pregnant rats. Soluble chromatin was immunoprecipitated with 4 µg of anti-GAPDH antibody or negative control IgG and subjected to PCR with a specific primer to the SNAT2 ERE site (Table S1). Results shown are representative of three independent experiments (n = 5 rats per group). Different letters differ: a > b > c. P < 0.05. (I) A representative immunoblot of GAPDH precipitation in SNAT2 ERE and SNAT2 ERE mutated sequences using biotinylated oligos incubated with ER-α transfected nuclear extracts from HeLa cells.