Figure 1.
A standardized reporter library uncovers splicing regulatory information encoded within introns.
A) Overview of the reporter approach for studying splicing mediated gene expression regulation. Intron insertion cassettes were constructed in-vitro, each comprised of a selection marker (URA3), a constitutive promoter, the first 195 nucleotides (nt) of the YFP gene, and one of 240 native S. cerevisiae introns followed by an additional 60 nt of the YFP gene. Each insertion cassette was transformed into the genome of a master strain which contained a promoter-less YFP gene, thus creating an in-vivo intron-reporter yeast library (YiFP). Culture growth and YFP expression levels of each variant in the library were monitored using a micro-plate reader. B) All strains in the YiFP reporter library grew similarly. C) However, each intron conferred unique YFP expression levels. D) Each strain's average expression levels, YiFP(i), were compared to that of an intron-less reference strain, YFP(wt), to get an assessment of “splicing efficiency”. YiFP strains whose YFP levels did not pass the detection limit were considered as “non-spliced” (marked with circles). Error bars represent ±1 SD from four independent experiments. E) Splicing efficiency of ribosomal and non-ribosomal protein genes (RPGs) is distributed in a similar manner. F) Analysis of YFP expression at the single cell level (using automated microscopic imaging) validated splicing efficiency measurements of spliced introns (inset graph, r = 0.94), and enabled assessment of splicing efficiency noise in a population. Noise is represented by the (squared) YFP expression coefficient of variation (CV2), i.e. the variance (σ2) normalized by the squared mean YFP expression (μ2), for each intron strain as determined using microscopic imaging analysis. Gene names of introns that presented noise higher (red) or lower (blue) than normal are indicated (outliers of linear regression; p<0.1). G) Splicing efficiency in a synthetic context is robust to environmental change. Yeast were grown in several stress conditions (Amino acid starvation, Rapamycin, 1M KCl) known to affect the splicing machinery. Error bars represent ±1 SD from three independent experiments.
Figure 2.
Sequence motifs function “out of context”.
A) Motifs associated with splicing efficiency were revealed by comparing intron sequences of high and low expressing YiFP strains. Examples of five enhancers (enrichment p<3.4e-4) and silencers (p<4e-4) are shown. B) The novel motifs were enriched in proximity to intron ends. Enrichment of motifs in introns compared to a randomized/permutated version of the motifs that maintain their properties (blue line, Materials and methods) is presented in respect to distance from 5′- or 3′ Splice Site (SS) (top and bottom, respectively). Positions significantly enriched or deprived of motifs are marked in red and green, respectively. C) Reporter YFP expression is decreased upon exchange of an enhancer to a silencer motif in two independent intron strains. Mean YFP expression was calculated for triplicates of the two intron library strains (YGL076C- p = 0.02, and YDL064W- p = 0.2). Average expression of each mutated motif strain is shown in comparison to that of the natural intron harboring YiFP strain (100%). Numbers inset in bars indicate the mean YFP expression level (non-relative).
Figure 3.
The exonic context of introns is a major regulatory determinant of gene expression.
A) Profiles of the correlation of GC content (top) and local mRNA folding energy (bottom) around the 5′SS and 3′SS with YiFP expression levels identified these features as determinants affecting gene expression. Sliding window sizes are 50 nt for GC content and 40 nt for folding energy. B) Ten introns were inserted into a stronger folding location in the YFP to test this feature's effect on gene expression. C) Introns inserted into a location within the YFP reporter with stronger folding (blue) confer lower expression levels. Averages of three independent experiments are presented (paired t test p = 8.3e-03).
Figure 4.
Modeling intron features uncovers design principles and allows the prediction of gene expression in a synthetic system.
A) Sequence based predictor of gene expression assembly process: In every iteration the feature contributing the highest correlation to the reporter expression measurements was added. The first eight features and their description are presented. B) Bar diagram of the predictor's cumulative correlation with expression levels of YiFP variants as a function of the number of added features. C) A predictor function based on 3, 13, or 38 features was able to explain 49%, 77% and 90% of gene expression variation, respectively. (for 13 features: p<2.2e-16; empirical p<5e-03); D) Cross validation of the predictor assembly method using training and test sets, with 80% and 20% of introns respectively, demonstrated a predictive power of 50% (for >15 features: 0.37<r<0.5; p<3.6e-02). E) A new predictor assembled using strains with introns inserted to several locations in the YFP maintains 80% of the model's predictive power (r = 0.38; p = 0.036), suggesting that although some of the regulatory splicing information is not located in intronic regions, our methodology is able to predict intron regulation under several exon contexts.