Chemical-genetic profile analysis of five inhibitory compounds in yeast

Abstract

Background: Chemical-genetic profiling of inhibitory compounds can lead to identification of their modes of action. These profiles can help elucidate the complex interactions between small bioactive compounds and the cell machinery, and explain putative gene function(s).

Results: Colony size reduction was used to investigate the chemical-genetic profile of cycloheximide, 3-amino-1,2,4-triazole, paromomycin, streptomycin and neomycin in the yeast Saccharomyces cerevisiae. These compounds target the process of protein biosynthesis. More than 70,000 strains were analyzed from the array of gene deletion mutant yeast strains. As expected, the overall profiles of the tested compounds were similar, with deletions for genes involved in protein biosynthesis being the major category followed by metabolism. This implies that novel genes involved in protein biosynthesis could be identified from these profiles. Further investigations were carried out to assess the activity of three profiled genes in the process of protein biosynthesis using relative fitness of double mutants and other genetic assays.

Conclusion: Chemical-genetic profiles provide insight into the molecular mechanism(s) of the examined compounds by elucidating their potential primary and secondary cellular target sites. Our follow-up investigations into the activity of three profiled genes in the process of protein biosynthesis provided further evidence concerning the usefulness of chemical-genetic analyses for annotating gene functions. We termed these genes TAE2, TAE3 and TAE4 for translation associated elements 2-4.

Figure 1

Clustering of drug sensitive gene deletion mutants. The haploid non-essential yeast gene deletion array was subjected to sub-inhibitory concentrations of five inhibitory compounds. Colony size reduction was used to detect sensitivity. (A) Drug sensitive yeast gene deletion mutants were clustered according to the cellular processes in which their deleted genes participated. The overall distributions of gene functions were comparable for different treatments with protein biosynthesis as a major group for all treatments. (B) Chemical profiles were clustered according to drug sensitivities to two or more drugs. Hierarchical clustering of mutants is illustrated using complete linkage. Absolute correlation coefficient (centered) is used for comparability and displayed in Java TreeView. Several regions of interest (a-e) are enlarged. The cellular processes of the deleted genes are color-coded. On the basis of sensitivity profiles, paromomycin is grouped with neomycin. Cycloheximide is grouped with 3-AT, which then merges with streptomycin. Sensitivity indexes of the gene deletion mutants are shown as high to low (light to dark red). (C) Sensitivity overlaps for gene deletion mutants to different drug treatments. The number of gene deletion mutants with a particular sensitivity, for example paromomycin (P) alone (89), paromomycin and 3-AT (17) and paromomycin, 3-AT and neomycin (3), are indicated. (D) The overlapping drug sensitive yeast gene deletion mutants are clustered according to the cellular processes in which their deleted genes participate. No significant enrichment for protein biosynthesis genes among overlapping sensitive strains was observed. The number of sensitive strains is presented on the z-axis. C: cycloheximide; P: paromomycin; A: 3-AT; N: neomycin; and S: streptomycin. The sensitivity overlaps between P and N, C and 3-AT, C and S, and 3-AT and S were significant with P-values ≤ 5 × 10^-14. Other overlaps are significant with P-values of ≤ 0.029.

Figure 2

Strain sensitivity to different translation-inhibitory drugs. Wild type (WT) or gene deletion mutant strains (yploo9cΔ, yil137cΔ, ypl183w-aΔ, ydr056cCΔ and yjr111cΔ) were serially diluted to 10^-3 to 10^-6 and spotted on solid medium with sub-inhibitory concentrations of cycloheximide, paromomycin, 3-AT, streptomycin and neomycin as indicated, or without drugs (control). The plates were incubated at 30°C for 1-2 days. Deletion of ypl009c confers increased sensitivity to cycloheximide; yil137c and ypl183w-a to 3-AT, ydr056c to streptomycin and neomycin, and yjr111c to streptomycin.

Figure 3

Synthetic genetic interaction analysis for TAE2, TAE3 and TAE4 with translation related genes. There are 72 interactions that represent synthetic genetic interactions for three query genes TAE2, TAE3 and TAE4, with 59 different translation genes. Genes are represented as nodes (circles) and interactions are represented as edges (lines). The interacting genes are further divided into eight functional categories. There are a number of shared interactions that highlight the interconnectivity of the network. The nodes are coloured according to functional groups. Black edges represent synthetic sick (aggravating) interactions, and the six pink thick edges represent synthetically rescue (alleviating) interactions.

Figure 4

Overexpression of TAE2 and TAE4 suppresses the sensitivity of numerous translation genes to drug treatments. Overexpression of TAE2 and TAE4 suppresses the phenotype of a number of translation gene deletion strains against neomycin and/or streptomycin treatments. Genes are represented as nodes (circles) and interactions are represented as edges (lines). The interacting genes are divided into functional categories and colored accordingly. (A) TAE2 over-expression rescued 20 gene deletions with a variety of functions. (B) TAE4 over-expression rescued 18 gene deletions, the majority of which are 40 S subunit proteins (nine genes) or function as translation-associated RNA processing proteins (five genes). Blue letters represent genes that are rescued by the overexpression of both TAE2 and TAE4.

Figure 5

Characterization of TAE2, TAE3 and TAE4 deletions. (A) Total protein synthesis was measured using [³⁵S] methionine incorporation in wild type, tae2Δ, tae3Δ and tae4Δ strains. The average count for [³⁵S] methionine incorporation for wild type was 11,356,073 (± 1,400,000) counts, which is set to 100%. On average, in the absence of Tae2p, Tae3p and Tae4p, [³⁵S] methionine incorporation was reduced by approximately 30, 14 and 10%, respectively. (B) The efficiency of protein synthesis was measured using an inducible β-galactosidase reporter construct (p416). The average β-galactosidase activity for wild type was 7.5 (± 0.6) units, which is set to 100%. The β-galactosidase activity was measured after 4 h induction. Deletion of TAE2, TAE3 and TAE4 limited the expression of β-galactosidase to 13, 21 and 17% of that in wild type, respectively. (C) Deletion of TAE2, TAE3 and TAE4 resulted in increased levels of β-galactosidase from lacZ reporters with different premature stop codons (pUKC817 and pUKC818). The activity of β-galactosidase was determined by normalizing the activity of the mutant (pUKC817 and pUKC818) to the control (pUKC815). pUKC815 is the background construct without a premature stop codon and is used as a control. Bars represent standard deviations for the means. (D) Ribosome profile analysis of yeast deletion strains tae3Δ and tae4Δ compared to wild type. Deletion of TAE3 decreased the levels of polysomes and increased free 60 S subunits. Deletion of TAE4 caused an increase in free 60 S subunits and a slight decrease in larger polysomes. Each experiment was repeated a minimum of three times. Ratios of free 60S:40 S were calculated from the areas under the curves.