Multiple Basic Helix-Loop-Helix Proteins Regulate Expression of the ENO1 Gene of Saccharomyces cerevisiae

Plasmid-containing Escherichia coli DH5α cells (Invitrogen, Carlsbad, CA) were grown in LB-Amp medium (10% [wt/vol] Bacto tryptone, 5% [wt/vol] yeast extract, 10% [wt/vol] NaCl, and 50 μg/ml ampicillin) at 37°C. Plasmid-containing E. coli BL21(DE3)/pLysS cells (Novagen, Madison, WI) were grown at 37°C and 25°C in LB-Amp medium supplemented with 50 μg/ml chloramphenicol.

The S. cerevisiae strains used in this study were BY4742 (MATα his3-Δ1 leu2-Δ0 lys2-Δ0 ura3-Δ0), BY4741 (MATahis3-Δ1 leu2-Δ0 met15-Δ0 ura3-Δ0), and isogenic strains containing ino2Δ, ino4Δ, pho4Δ, cbf1Δ, sgc1Δ, rtg1Δ, rtg3Δ, hms1Δ, and ygr290wΔ alleles (22, 81). Yeast cultures were grown at 30°C in a complete synthetic medium lacking inositol, choline, KH₂PO₄, and uracil (for reporter plasmid selection) (38). Where indicated, 75 μM inositol (I+) and/or 1 mM choline (C+) was added. Low-P_i medium contained 0.22 mM KH₂PO₄ and 20 mM KCl, and high-P_i medium contained 11 mM KH₂PO₄.

Plasmid YEp357R-ENO1 contains 720 bp of the sequence upstream of the ENO1 ORF and the first codon fused in frame to the lacZ reporter gene in YEp357R (56). This 720-bp region was previously shown to contain all of the regulatory elements necessary for ENO1 expression (79). YEp357R is a multicopy episomal plasmid with a URA3 selectable marker (56). This fusion plasmid was constructed by first amplifying 1,000 bp of the ENO1 promoter from S. cerevisiae genomic DNA (Invitrogen, Carlsbad, CA), using primers ENO1 F and ENO1 R (Table 1). The 1,000-bp PCR product was cloned into pGEM-T (Promega, Madison, WI) and sequenced, and then the ORF-proximal 720-bp sequence was excised by digestion with EcoRI and inserted into YEp357R.

Plasmids that complemented the cbf1Δ, sgc1Δ, ino2Δ, and ino4Δ mutant alleles were constructed by cloning each ORF and promoter into pRS315. Plasmid pRS315-CBFI was constructed by amplifying a 1,556-bp fragment from S. cerevisiae genomic DNA (Invitrogen, Carlsbad, CA), using primers CBF1 F′ (position −500) and CBF1 R′ (position +1056) (Table 1). The 1,556-bp PCR product was cloned into pGEM-T (Promega, Madison, WI), sequenced, excised by digestion with NotI and XbaI, and ligated into pRS315. Likewise, pRS315-SGC1 was made using primers SGC1 F′ (−1000) and SGC1 R′ (+1426) (Table 1), which amplified the SGC1 ORF, 1,000 bp of upstream sequence, and 550 bp of downstream sequence. The 2,426-bp PCR product was cloned into pGEM-T, sequenced, excised by digestion with BamHI and HindIII, and inserted into pRS315. Plasmids pRS315-INO2 and pRS315-INO4 were constructed previously (K. R. Gardenour and J. M. Lopes, unpublished data). Briefly, pRS315-INO2 was constructed by inserting a 2.4-kb SalI/ClaI fragment (containing 500 bp of promoter, the INO2 ORF, and 400 bp of 3′-untranslated region [3′UTR]) from YCp50-INO2 (17) into pRS200 (pRS200-INO2) and subsequently cloning a SalI/PstI fragment from pRS200-INO2 into pRS315. Likewise, pRS315-INO4 was constructed by inserting a 2.4-kb SacII/SalI fragment (containing 500 bp of promoter, the INO4 ORF, and 400 bp of 3′UTR) from YCp50-INO4-496 (64) into pRS315.

Plasmids were made to contain dominant-negative mutants of INO2 and INO4. The expressed mutants were capable of dimerization with other bHLH proteins but inhibited their binding to DNA because they either contained mutations in the DNA-binding basic charged domain (ino2-R13L and ino4-R13L) or completely lacked the basic charged domain (ino4-BRD). The ino4-BRD mutant was created by PCR. The region upstream of the basic charged domain (including 500 bp of the INO4 promoter) was amplified using primers INO4F (containing a BamHI site) and INO′4R (Table 1) to yield a 650-bp product. The region downstream of the basic charged domain (including the INO4 3′UTR) was amplified using primers INO′4F and INO4R (containing a KpnI site) (Table 1) to yield an 800-bp product. The two PCR products were annealed (primers INO′4R and INO′ include a 30-bp overlap that deletes the basic charged region), extended, reamplified using primers INO4F and INO4R (to yield a 1,420-bp product), and cloned into pGEM-T. The insert was sequenced, excised by digestion with BamHI and KpnI, and inserted into pRS315 to yield pRS315-INO4-BRD. Plasmids pRS315-INO2-R13L and pRS315-INO4-R13L were constructed previously by site-directed mutagenesis (Gardenour and Lopes, unpublished data). Briefly, YCp50-INO2 was mutagenized using a QuikChange XL site-directed mutagenesis kit (Stratagene, La Jolla, CA), using primers INO2 R13L 5′ SD and INO2 R13L 3′ SD (Table 1). The INO2-R13L mutant allele was sequentially cloned into pRS200 and pRS315 as described above. Plasmid pRS315-INO4 was directly mutagenized using a QuikChange XL site-directed mutagenesis kit and primers INO4 R13L 5′ SD and INO4 R13L 3′ SD (Table 1).

Hemagglutinin (HA)-tagged derivatives of CBF1, SGC1, INO2, and INO4 to be used for ChIP assays were either generated or purchased. A YCp50-CBF1-HA construct was created by mutational PCR. PCR was used to replace the second and third codons of CBF1 with a BglII site. To do this, a PCR using primers CBF1A and CBF1-HA 3′ (Table 1) yielded a 500-bp product containing the CBF1 promoter and the new BglII site. A second PCR with the primers CBF1B and CBF1-HA 5′ (Table 1) yielded a 1,520-bp product containing the CBF1 ORF and the new BglII site. The PCR products were digested with BglII and ligated to create a product which contained the CBF1 promoter and ORF with the BglII site. The ligated fragment was used for another round of PCR with the primers CBF1A and CBF1B, resulting in a 2,056-bp product. This PCR fragment was cloned into pGEM-T. A 120-bp BglII fragment containing three tandem copies of the HA epitope was isolated from pSM492 (8) and inserted into the pGEM-T derivative partially digested with BglII. A SalI-BamHI fragment was isolated from the pGEM-T derivative and cloned into YCp50. The YCp50-CBF1-HA construct was confirmed by DNA sequencing. The YCp50-INO2-HA and YCp50-INO4-HA plasmids have been described previously (17, 64). A strain containing an HA-tagged SGC1 gene was purchased from Open Biosystems (Huntsville, AL).

Three E boxes in the ENO1 promoter (positioned at −460, −656, and −704) were mutagenized using a QuikChange XL site-directed mutagenesis kit and the pGEM-T-ENO1 promoter derivative described above. The E primer set (Table 1) was used to create three single E box mutants. The single mutants were used to create the three possible combinations of double mutants and the triple mutant. The mutant ENO1 promoters were cloned into YEp357R as described above.

To assay β-galactosidase (β-Gal) activity, yeast strains were grown in 5 ml of appropriate medium to mid-log phase (60 to 80 Klett units) and pelleted by centrifugation at 5,000 × g for 10 min. The cell pellet was suspended in 200 μl of β-Gal assay buffer (20% glycerol, 0.1 M Tris-HCl [pH 8.0], 1 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride [PMSF]), transferred to a sterile 1.5-ml microcentrifuge tube, and stored at −80°C overnight. Cells were thawed on ice, and 100 μl of glass beads (0.45-mm diameter) was added. Cells were lysed by use of a vortex mixer six times for 15 s each. Cellular debris was removed by centrifugation at 14,000 rpm for 15 min at 4°C. The supernatant was transferred to another 1.5-ml microcentrifuge tube. To assay β-Gal activity, reaction mixtures were set up with 20 μl of cell extract and 80 μl of β-Gal assay buffer and incubated for 5 min at 28°C. The reaction was initiated by the addition of 40 μl of ONPG (o-nitrophenyl-β-d-galactopyranoside; 4 mg/ml). The absorbance of the reaction was measured by determining the optical density at 420 nm at 12-second intervals for a total of 30 min. Protein concentration was determined by the Bio-Rad Protein Assay (Bio-Rad, Rockville Center, NY). Both the β-Gal activity reactions and the protein concentration reactions were quantified using Soft_maxPro software and a Versa_max tunable microplate reader (Molecular Devices, Sunnyvale, CA). Units of β-Gal activity are given as A₄₂₀/min/mg total protein × 1,000.

Yeast cell cultures (200 ml) were grown in I−C− medium at 30°C to mid-log phase (60 to 80 Klett units). Formaldehyde was added to a 1% final concentration, followed by a 30-min incubation at 30°C. Glycine was added to 125 mM, and the mixture was incubated for an additional 5 min. Cells were pelleted at 1,500 × g for 5 min, and pellets were washed twice with 700 ml of 1× phosphate-buffered saline (0.43 mM Na₂HPO₄, 0.14 mM KH₂PO₄, 13.7 mM NaCl, and 0.27 mM KCl) and once with 15 ml of bead-beater lysis buffer (50 mM HEPES-KOH, pH 7.5, 10 mM MgCl₂, 150 mM KCl, 0.1 mM EDTA, 10% glycerol, 0.1% NP-40, 1 mM dithiothreitol, 1 mM sodium metabisulfite, 0.2 mM PMSF, 1 mM benzamidine, and 1 μg/ml pepstatin). The cell pellet was weighed and resuspended in 2.5× bead-beater lysis buffer. One milliliter of the resuspended cells was added to 1 ml of 0.45-mm glass beads. Cells were lysed in a mini-Beadbeater 8 with four 1-minute pulses at the highest setting (with cells being placed on ice for 2 min between pulses). The extract was recovered by pouring the bead-extract slurry into a 6-ml syringe fitted with a 25-gauge, 5/8-inch-long needle. The syringe was washed with 0.75 ml of bead-beater lysis buffer. The extract was sonicated three times for 30 seconds each, using a model 100 Sonic Dismembrator with a Branson 250 microtip sonicator (Fisher Scientific, Pittsburgh, PA) at 50% duty cycle with a power setting of 5, with cells being placed on ice for 2 min between pulses. The extract was cleared of debris twice in a microcentrifuge at full speed at 4°C for 5 min. Samples were fractionated in an agarose gel to ensure that DNA was sheared to a size range from 500 to 2,000 bp. The protein concentration was determined using the Bio-Rad Protein Assay, and 750 ng of extract was diluted with IP buffer (25 mM HEPES-KOH, pH 7.5, 150 mM KCl, 1 mM EDTA, 12.5 mM MgCl₂, 0.1% NP-40, 1 mM sodium metabisulfite, 0.2 mM PMSF, 1 mM benzamidine, and 1 μg/ml pepstatin) to a final reaction volume of 500 μl. Mouse anti-HA (clone 12CA5; Boehringer Mannheim) was added (to 2.5 ng/μl) and incubated overnight at 4°C. The antibody-protein-DNA complexes were recovered with protein G beads (equilibrated in IP buffer) by incubation at 4°C for 1 to 2 h. The beads were washed four times for 15 min each with 1 ml of IP buffer at 4°C, 100 μl of IP elution buffer (50 mM Tris-HCl, pH 8.0, 10 mM EDTA, 1% sodium dodecyl sulfate) was added, and the samples were incubated at 65°C for 30 min. Samples were fractionated by centrifugation, and 80 μl of the supernatant was recovered. The elution step was repeated with 50 μl of IP elution buffer, and 50 μl of the supernatant was recovered and combined with the first eluate. Seventy microliters of each eluate was incubated overnight at 65°C to reverse cross-linking. Seventy microliters of Tris-EDTA (TE), pH 7.4, 1 μl of 20-mg/ml glycogen, and proteinase K (final concentration, 100 μg/ml) were added and incubated at 37°C for 2 hours. Samples were extracted with phenol-chloroform, and the organic phase was reextracted with 100 μl of TE. Sodium acetate was added to 0.3 M, and 2 volumes of 100% ethanol was added. DNAs were precipitated at −20°C for 1 h, collected by centrifugation, washed with 70% ethanol, and dried. DNAs were resuspended in 25 μl of TE with 100 μg/ml RNase A and incubated at 37°C for 30 min.

Immunoprecipitated DNA and input DNA were analyzed by real-time quantitative PCR using an Mx3000P QPCR thermocycler and MxProQPCR software (Stratagene, La Jolla, CA). Specific primers (Table 1) flanking 60 to 90 base pairs of each ENO1 E box, the INO1 promoter, and the TCM1 promoter were designed using Primer3 software (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi ). Primer and template DNA concentrations were optimized, and amplification reactions with SYBR green were carried out for 1 cycle of 50°C for 2 min and 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 55°C for 1 min. The data were calibrated using the TCM1 ChIP signal and normalized to the input DNA.

It was known that ENO1 expression is regulated by the Sgc1p bHLH protein and Gcr1p (49, 58, 68). Here we tested if other bHLH proteins also regulate ENO1 expression by using an ENO1-lacZ reporter. The ENO1-lacZ reporter was assayed in a wild-type and nine isogenic bHLH knockout strains. Since inositol, choline, and phosphate concentrations affect the functions of different bHLH proteins (Ino2p:Ino4p and Pho4p, respectively), ENO1-lacZ expression was tested under different growth conditions. We used the following four growth conditions: I−C− low-P_i medium, I−C− high-P_i medium, I+C+ low-P_i medium, and I+C+ high-P_i medium. In the case of the ino2Δ and ino4Δ strains, the I−C− medium contained 10 μM inositol, which is required for growth of these mutant strains but still yields derepressed expression of their target genes (1, 17, 24).

As expected, the SGC1 gene was required for maximal expression of ENO1-lacZ under all four growth conditions (Fig. 2A) (58, 68). In addition, the data showed that ENO1-lacZ expression in the wild-type strain was repressed in the presence of inositol-choline regardless of the phosphate concentration (Fig. 2A). ENO1-lacZ expression was reduced 51% and 21% by inositol-choline in low- and high-phosphate media, respectively (Fig. 2A). While these levels of repression are relatively modest, they are within the range observed for several well-characterized inositol-choline-responsive genes (14, 55, 69, 71). However, inositol-choline repression was not observed in most other bHLH knockout strains (Fig. 2A). While this would be expected for the ino2Δ and ino4Δ strains, it was unexpected for the other bHLH knockout strains (Fig. 2A).

In addition, this experiment revealed that every yeast bHLH protein was required for maximal ENO1-lacZ expression, although the ino2Δ, pho4Δ, rtg1Δ, cbf1Δ, and hms1Δ strains showed the most dramatic effects, which resembled the effect of an sgc1Δ mutant. In this study, we focus on the Ino2p, Ino4p, Cbf1p, and Sgc1p bHLH proteins because the first three are associated with the response to inositol-choline (33, 67; Y. He, A. Shetty, and J. M. Lopes, unpublished results) and Sgc1p is well established as being required for maximal expression of ENO1 (58, 68).

A complementation test was performed to confirm that the decreased ENO1-lacZ expression (Fig. 2A) was due to the bHLH knockout alleles. The mutant strains were transformed with pRS315-based plasmids carrying the INO2, INO4, CBF1, or SGC1 gene under the control of its own promoter. The plasmids carrying the INO2, CBF1, and SGC1 genes partially restored ENO1-lacZ expression to the wild-type levels of expression (Fig. 2B). It is not unusual that these genes did not yield completely restored expression. However, the INO4 plasmid did not appear to complement the defect in the ino4Δ strain, even though it did complement the inositol auxotrophy and defective INO1-lacZ expression in an ino4Δ strain (He et al., unpublished data). The reasons for this are not clear.

The bHLH proteins regulate transcription by forming protein dimers, most of which recognize DNA sequences called E boxes (CANNTG) (53, 63). There are five potential E boxes in the ENO1 promoter region present in the ENO1-lacZ construct used in this study. It was previously reported that sequences required for ENO1 expression localize to the −724-to-−353 region, thereby eliminating two ORF-proximal E boxes (79). Therefore, we focused on the three ORF-distal E boxes (E-460 [E1], E-656 [E2], and E-704 [E3]; distances are relative to the start codon). We mutated each E box by site-directed mutagenesis and assayed the strains for β-galactosidase activity (Fig. 3). All three E boxes were found to play a role in ENO1-lacZ expression. Mutating the E1 box had the biggest impact, reducing expression 21.2-fold (average for the four growth conditions), while mutating the E2 and E3 boxes reduced expression 4.6- and 2.8-fold, respectively (Fig. 3). Deleting the E2 element, alone or in combination with E1 and E3, completely eliminated repression in response to inositol and choline (Fig. 3). However, deletion of E1 and/or E3 eliminated repression in response to inositol and choline only in high-P_i medium (Fig. 3).

To determine if there is any synergy between the E boxes, every double mutant and triple mutant combination was created, and the effect of each pairwise combination of elements was determined. The triple mutant was almost completely devoid of UAS activity, suggesting that the three E boxes constitute all of the required positive regulatory elements (Fig. 3). To define the contribution of each E box to activation, double E box mutants were compared to the triple E box mutant. For example, to calculate the E1 contribution in I−C− low-P_i medium, the activity in the E2+E3 mutant was divided by the activity in the E1+E2+E3 triple mutant (5.7-fold activation) (Table 2). To calculate the combined activation from the E1 and E2 elements, the activity in the E3 mutant was divided by the activity in the E1+E2+E3 triple mutant (24.3-fold activation) (Table 2). The ratio of the observed activation from two elements to the product of the activation levels from the individual elements represents the synergy factor. A synergy factor of 1.0 indicates additive activation (i.e., no synergy) (41). The data showed that most E box relationships yielded synergy factors in the range of 0.8 to 1.8 under all growth conditions, showing the absence of synergy between the three E boxes.

To define the bHLH protein-E-box interactions, we used two approaches, namely, an epistatic analysis and ChIP. The dual approach was expected to provide corroborating evidence, and the epistatic analysis could additionally provide information about indirect regulation. For the epistatic analysis, the three E box mutant promoters were transformed into a wild-type strain and the nine bHLH knockout strains. Cells were grown in I−C− low-P_i and I+C+ low-P_i media.

The results suggest that Sgc1p functions exclusively through the E1 element, since deleting the SGC1 gene had no effect on expression from the E1 mutant promoter but affected expression from the E2 and E3 mutants (Fig. 4). Likewise, Cbf1p functions through the E2 element (Fig. 4). The data suggest that Ino2p functions through the E3 element because deleting the INO2 gene had virtually no effect on expression from the E3 mutant promoter. The data also suggest that Ino4p functions through the E1 and E3 boxes (Fig. 4). While deleting the INO4 gene affected expression from all three E box mutants, the most dramatic effect was with the E2 mutant. Thus, while Ino4p may function through all three E boxes, it is more likely that it functions through the E1 and E3 elements. It is curious that the inositol-choline response is mediated through the E2 element (Fig. 3 and 4), which does not appear to bind the Ino2p-Ino4p dimer but instead interacts with Cbf1p. As for the other five bHLH proteins, it was not possible to assign them to specific E boxes since the bHLH knockout alleles did not affect expression from any of the E box mutants. The simplest explanation is that the other five bHLH proteins may function cooperatively through multiple sites.

ChIP was used to define the direct regulators of ENO1 expression. The TCM1 promoter was used as a negative control, and the INO1 promoter served as a positive control for the binding of Ino2p and Ino4p. As expected, none of the bHLH proteins interacted with the TCM1 promoter, whereas Ino2p and Ino4p interacted with the INO1 promoter (Fig. 5). The data show that all four bHLH proteins bind the ENO1 promoter directly in vivo via different E boxes. Moreover, the ChIP results are in complete agreement with the results of the epistatic study (compare Fig. 4 and 5). Ino2p and Ino4p interacted with the E3 box, but Ino4p also interacted with the E1 element (Fig. 5). This is interesting because Ino4p does not homodimerize and usually requires Ino2p for binding to DNA (72). However, Ino4p has been shown to regulate some genes in the absence of Ino2p (67). Furthermore, Sgc1p also interacted with the E1 box, and while Sgc1p binds a consensus E box as a homodimer (68), it may bind the ENO1 E1 element as a heterodimer with Ino4p. Cbf1p was the only bHLH protein tested that interacted with the E2 box. This is curious since the response to inositol-choline was found to be mediated by the E2 box and inositol-choline regulation was typically believed to be effected by Ino2p-Ino4p (13, 23, 27, 63). However, a recent microarray study identified a set of genes that are repressed by inositol and do not bind the Ino2p-Ino4p dimer but are instead part of the unfolded-protein response regulon (33).

The ability to form multiple dimer combinations is a general property of bHLH proteins (29, 30, 57, 66). The results described above showed that Ino2p, Ino4p, Cbf1p, and Sgc1p regulate the expression of ENO1-lacZ and suggested that novel dimers (such as Ino4p-Sgc1p) may contribute to this regulation. We have also previously shown that Ino4p can dimerize with every bHLH protein in a yeast two-hybrid assay (62). To determine if partner selection plays a role in ENO1-lacZ expression in vivo, we used two ino4 alleles and one ino2 allele containing mutations in the basic region. In two of these mutants, ino4-R13L and ino2-R13L (Gardenour and Lopes, unpublished data), the amino acid in the 13th position of the basic region was changed from an R to an L. This position of the basic region is highly conserved throughout bHLH proteins and contacts a G nucleotide in the fourth position of the E box (CANNTG) (63). However, MyoD contains an L amino acid at the 13th position of the basic region and contacts a C nucleotide in the 4th position of the E box (9). Therefore, while the ino4-R13L and ino2-R13L mutants function as dominant-negative mutants with respect to INO1 expression, they are able to bind to E boxes that contain a C nucleotide at the fourth position (Gardenour and Lopes, unpublished data). In the other mutant, ino4-BRD, we deleted the 13-amino-acid basic region.

The ino4 and ino2 mutants (in pRS315) were transformed into strains harboring the ENO1-lacZ reporter, grown under the four conditions described above, and assayed for β-galactosidase activity. The data clearly show that the presence of the R13L mutant bHLH proteins affects the expression of ENO1-lacZ in several strains (compare Fig. 6 and 2A). The ino4-R13L and ino4-BRD mutants reduced expression in the wild-type strain, mostly under I−C− conditions, and significantly reduced ENO1-lacZ expression in the ino2Δ and cbf1Δ strains (compare Fig. 6A and 2A; data not shown). This suggests that in the ino2Δ and cbf1Δ strains, expression is dependent on a bHLH protein(s) (likely Sgc1p) that is inhibited from binding by the Ino4p-R13L (and Ino4p-BRD) mutant. However, in the ino4Δ and sgc1Δ strains, expression was either unchanged or increased slightly (compare Fig. 6A and 2A; data not shown). Likewise, the presence of the ino2-R13L mutant allele either had no effect on expression (cbf1Δ and sgc1Δ strains) or significantly increased expression (wild-type, ino2Δ, and ino4Δ strains) (compare Fig. 6B and 2A). Because of the number of bHLH proteins that affect ENO1-lacZ expression (Fig. 2A), there are numerous possibilities that could account for the increase in expression in the presence of the ino4-R13L and ino2R-13L mutants. For example, the R13L mutation could inhibit binding of a repressor protein or could recruit a different bHLH protein to the E1 site, which has a C nucleotide at the fourth position. Regardless of the explanation for the increased expression, the data support the conclusion that bHLH dimer formation affects ENO1-lacZ gene expression in vivo.