Balancing transcriptional interference and initiation on the GAL7 promoter of Saccharomyces cerevisiae

Abstract

Transcriptional termination of the GAL10 gene in Saccharomyces cerevisiae depends on the efficiency of polyadenylation. Either cis mutations in the poly(A) signal or trans mutations of mRNA 3' end cleavage factors result in GAL10 read-through transcripts into the adjacent GAL7 gene and inactivation (occlusion) of the GAL7 promoter. Herein, we present a molecular explanation of this transcriptional interference phenomenon. In vivo footprinting data reveal that GAL7 promoter occlusion is associated with the displacement of Gal4p transcription factors from the promoter. Interestingly, overexpression of Gal4p restores promoter occupancy, activates GAL7 expression, and rescues growth on the otherwise toxic galactose substrate. Our data therefore demonstrate a precise balance between transcriptional interference and initiation.

Figure 1

Diagram of the GAL cluster and the GAL7 promoter. (A) Diagram showing the arrangement of the structural GAL genes. GAL10 and GAL7 are contained within YC10-7. Deletions of the GAL10 poly(A) site result in Δ-55 and Δ-75, respectively. YC10-7 gives rise to monocistronic GAL10 and GAL7 mRNA, respectively (dotted lines below), and is viable on gal medium (Gal+). Δ-55 and Δ-75 give rise to bicistronic mRNA, which does not produce GAL7 mRNA, and are therefore gal sensitive (Gal−). Black boxes represent Gal4p-binding sites. (B) Schematic of the region 260 bp upstream of the GAL7 transcription initiation site (+1). The TATA box and the two Gal4p binding sites are shown (dark boxes); the distances between these elements are indicated below. Also shown are the positions of primers used for in vivo footprinting (P1–P4). (C) Sequences of the two Gal4p binding sites (GAL4-1 and GAL4-2). The center of the roughly symmetrical 17-bp sites is indicated by a vertical, dashed line. The base pairs are numbered sequentially from the center; the central base pair has been assigned the number 0. The protected guanines of the highly conserved outer triplet are in bold and indicated by horizontal bars. They are shown in uppercase and lowercase, depending on the degree of protection. Also indicated are the primers (P1–P4), which were used to visualize the particular strand.

Figure 2

Overexpression of Gal4p rescues Gal7p expression. (A) Overexpression of Gal4p restores growth of Δ-55 on gal + ethidium bromide medium. Strains, indicated below in the key diagram, were grown on YP/dextrose and YP-gal + ethidium bromide for 4 days at 30°C. Cells were streaked onto YP-gal + ethidium bromide first then onto YP/dextrose. (B) Northern blot of total RNA from gal-induced cells. (Upper) GAL10-7 and GAL7 mRNA was detected with a GAL7-specific probe. Gal4p was overexpressed to different levels in each strain as shown on the top. Lanes 1, 4, and 7, Gal4p at physiological levels; lanes 2, 5, and 8, cells transformed with Gal4p expressed from the autonomous replication sequence plasmid pRJR197; lanes 3, 6, and 9, cells transformed with Gal4p expressed from the 2-μ plasmid pRJR216. Crosshybridization to rRNA is indicated on the side. (Lower) An actin loading control (ACT1). (C) Northern blot of poly(A)⁺ mRNA. GAL transcripts were detected with a probe recognizing GAL10 and GAL7. Cryptic GAL10 poly(A) sites (cryptic pA sites), formed by Δ-55, are indicated by empty arrowheads on the side. (D) Western blot of Gal7p. Yeast extracts from strains, indicated on the top, were separated by SDS/10% PAGE. The gel was stained with Coomassie brilliant blue after the transfer and serves as a loading control (Coomassie). The blot was probed with Gal7p antibody. Gal4p was overexpressed to different levels in each strain as shown on the top. Lanes 1 and 4, Gal4p at physiological levels; lanes 2 and 5, cells transformed with Gal4p expressed from the autonomous replication sequence plasmid; lanes 3 and 6, cells transformed with Gal4p expressed from the 2-μ plasmid.

Figure 3

In vivo footprinting of Gal4p sites on the upper strand. (A) Primer 1: both binding sites GAL4-1 and GAL4-2 are detected and are indicated by black vertical bars. DMS protections are denoted by filled circles on the right side; the respective protected guanines are shown on the left. The hypermethylated G 4 in GAL4-2 is indicated by a filled square. Lanes 1 and 2, YC10-7 in the absence and presence of extra Gal4p, respectively; lanes 3 and 4, Δ-55 in the absence and presence of extra Gal4p, respectively; lane 5, in vitro (vt) control (purified DNA treated with DMS in vitro). (B) Primer 3: detection of GAL4-2. Symbols used are the same as in A. Lane 1, footprint on DNA from the Gal4⁻ strain JPY9; lane 2, in vitro (vt) control; lanes 3 and 4, Δ-55 in the presence and absence of Gal4p, respectively; lanes 5 and 6, YC10-7 in the presence and absence of Gal4p, respectively. (C) Quantitation of the GAL4-1 site shown in Fig. 1A. Shown are residues G −7, −6, and 4 (no changes were observed at position 4). The gel was scanned in a PhosphorImager. Lanes were equalized relative to a “neutral” guanine outside the footprinted region; values were calculated relative to the corresponding residue in the in vitro lane.

Figure 4

In vivo footprinting of Gal4p sites on the lower strand. (A) Primer 4: both binding sites GAL4-2 and GAL4-1 are detected. Symbols used are the same as in Fig. 3A. The empty circle in GAL4-2 indicates that residue G 7 is unchanged in all lanes. Protected guanines are shown on the left side. Lane 1, in vitro (vt) control; lanes 2 and 3, Δ-55 in the presence and absence of extra Gal4p, respectively; lanes 4 and 5, YC10-7 in the presence and absence of Gal4p, respectively. (B) Primer 2: detection of GAL4-1. Lane 1, in vitro (vt) control; lanes 2 and 3, YC10-7 in the absence and presence of extra Gal4p, respectively; lanes 4 and 5, Δ-55 in the absence and presence of extra Gal4p, respectively. Note that lane 5 is more than 2-fold overrepresented (relative to lane 4). (C) Quantitation of the GAL4-1 site shown in B. Shown are residues G 7, 6, and −1 (no changes were observed at position −1). Quantitation was performed as described for Fig. 3C.

Figure 5

The ts mRNA 3′ end CFs affect GAL10 transcription termination. (A) Northern blot of total RNA from gal-induced ts strains (indicated on the top). Strains were grown at the permissive temperature (26°C), induced with gal for 1 h, and then shifted to the nonpermissive temperature (37°C) for 45 min. Because mRNA levels are overall reduced at 37°C, twice the amount of RNA was loaded for all samples at the restrictive temperature (except for Δ-55, where equal amounts were loaded; lanes 1 and 2). GAL transcripts were detected with a probe recognizing GAL10 and GAL7. The filter was prehybridized with total RNA (10 μg/ml hybridization solution) extracted from cells grown in glucose to reduce crosshybridization to rRNA. (B) Northern blot of total RNA from the hpr1 Δgal10-7 strain, transformed with GAL gene plasmids as indicated (lanes 1–4) versus wt (strain N222; lanes 5 and 6).