3' Untranslated regions mediate transcriptional interference between convergent genes both locally and ectopically in Saccharomyces cerevisiae

Abstract

Paired sense and antisense (S/AS) genes located in cis represent a structural feature common to the genomes of both prokaryotes and eukaryotes, and produce partially complementary transcripts. We used published genome and transcriptome sequence data and found that over 20% of genes (645 pairs) in the budding yeast Saccharomyces cerevisiae genome are arranged in convergent pairs with overlapping 3'-UTRs. Using published microarray transcriptome data from the standard laboratory strain of S. cerevisiae, our analysis revealed that expression levels of convergent pairs are significantly negatively correlated across a broad range of environments. This implies an important role for convergent genes in the regulation of gene expression, which may compensate for the absence of RNA-dependent mechanisms such as micro RNAs in budding yeast. We selected four representative convergent gene pairs and used expression assays in wild type yeast and its genetically modified strains to explore the underlying patterns of gene expression. Results showed that convergent genes are reciprocally regulated in yeast populations and in single cells, whereby an increase in expression of one gene produces a decrease in the expression of the other, and vice-versa. Time course analysis of the cell cycle illustrated the functional significance of this relationship for the three pairs with relevant functional roles. Furthermore, a series of genetic modifications revealed that the 3'-UTR sequence plays an essential causal role in mediating transcriptional interference, which requires neither the sequence of the open reading frame nor the translation of fully functional proteins. More importantly, transcriptional interference persisted even when one of the convergent genes was expressed ectopically (in trans) and therefore does not depend on the cis arrangement of convergent genes; we conclude that the mechanism of transcriptional interference cannot be explained by the transcriptional collision model, which postulates a clash between simultaneous transcriptional processes occurring on opposite DNA strands.

Figure 1. RT-PCR expression assay of convergent gene pairs located in cis or in trans.

(A) Expression levels of four pairs of convergent genes with overlapping 3′-UTRs in wild type yeast, or when the upstream (group I) or the downstream (groups II–VII) genes were modified by: (I) inhibition of gene expression; (II) over-expression; (III) gene knock-out; (IV) frame-shift of the ORF; (V) re-location of the entire ORF and 3′-UTR; (VI) re-location of only the ORF; and (VII) re-location of only the 3′-UTR. The mean and standard deviation of expression level for three replicates are given in relative units compared to the wild type genes, which are assigned a value of 1.0. (B) Expression levels of nascent and mature messenger RNA for one of two convergent genes when its convergent partner's expression was altered.

Figure 2. Dual-luciferase reporter assay of protein levels for two convergent gene pairs.

Dual-luciferase reporter assay for the expression of the upstream genes of two pairs of convergent genes with overlapping 3′-UTRs. The reporter protein replaced the ORF of each upstream gene (A APT1 or B ADE1), while the downstream genes were modified: (I) wild type control with no modification; (II) over-expression of the entire gene; (III) ectopic over-expression of the ORF alone; and (IV) ectopic over-expression of the 3′-UTR alone. Mean and standard deviation based on three replicates are shown for mRNA expression level (upper panel) and protein abundance (lower panel), measured in relative units compared to the wild type control, which was assigned a value of 1.0.

Figure 3. Loss of anti-regulation in the absence of overlapping 3′-UTRs.

Expression levels of the upstream and downstream genes of two convergent gene pairs with their overlapping 3′-UTRs removed. Mean and standard deviation based on three replicates are given for the expression levels: (I) for a baseline control, assigned a value of 1.0; (II) with inhibition of the upstream gene; or (III) with over-expression of the downstream gene.

Figure 4. Expression patterns of four convergent gene pairs during the yeast cell cycle.

Expression of four pairs of yeast convergent genes with overlapping 3′-UTRs in cells from the wild type strain YL1A, the a-mating strain of YL1C (A ADE1/KIN3, B AXL2/REV7, C APT1/UNG1 and D SHM1/YPT10). Gene expression was assessed 10 minutes from the start of the cell-cycle and at 20 minutes intervals thereafter.

Figure 5. Cell-cycle expression of the gene pair ADE1 and KIN3, in convergent and tandem orientation.

(A) Three strains tested: (I) wild type, in which ADE1 and KIN3 were in convergent orientation; (II) tandem URA3, in which KIN3 was replaced with URA3, and ADE1 and URA3 were in tandem orientation; and (III) tandem KIN3, in which ADE1 and KIN3 were in tandem orientation. (B) Cell cycle expression of the three genes ADE1, KIN3 and URA3 in the wild type strain. (C) Cell cycle expression pattern of the genes ADE1 and KIN3 in tandem orientation. (D) Expression of the genes ADE1 and URA3 in the tandem URA3 strain.

Figure 6. Anti-regulation of a convergent pair in different nutritional environments.

Expression of the convergent gene pair (ADE1 and KIN3) with overlapping 3′-UTRs in two yeast strains cultured in complete (SC) or adenine barren (SC-A) medium. Strain D-K1 has ADE1 inhibited, while strain D-K2 has KIN3 over-expressed.

Figure 7. Phenotypic effect of anti-regulation between a convergent pair in a nutrient-limited environment.

Growth phenotype of the wild type strain (YL1C^WT) and seven strains (D-K1 to D-K7; see Table 2 ) genetically modified for the convergent gene pair ADE1 and KIN3. Strains were cultured in complete (SC) or adenine barren (SC-A) medium.