A network of multiple regulatory layers shapes gene expression in fission yeast

Abstract

Gene expression is controlled at multiple layers, and cells may integrate different regulatory steps for coherent production of proper protein levels. We applied various microarray-based approaches to determine key gene-expression intermediates in exponentially growing fission yeast, providing genome-wide data for translational profiles, mRNA steady-state levels, polyadenylation profiles, start-codon sequence context, mRNA half-lives, and RNA polymerase II occupancy. We uncovered widespread and unexpected relationships between distinct aspects of gene expression. Translation and polyadenylation are aligned on a global scale with both the lengths and levels of mRNAs: efficiently translated mRNAs have longer poly(A) tails and are shorter, more stable, and more efficiently transcribed on average. Transcription and translation may be independently but congruently optimized to streamline protein production. These rich data sets, all acquired under a standardized condition, reveal a substantial coordination between regulatory layers and provide a basis for a systems-level understanding of multilayered gene-expression programs.

Figure 1

Genome-Wide Translational Profiling (A) Polysome profile showing the absorbance of RNA separated by density on a sucrose gradient, reflecting the number of associated ribosomes. The peaks of the profile are labeled for the small and large ribosomal subunits (40S and 60S), the monosome (80S), and the numbers of associated ribosomes for polysomal RNA (2 to >8). RNA from 12 fractions equally spaced throughout the profile (bottom) was labeled and hybridized against a total RNA reference on microarrays containing all S. pombe genes. (B) Translation profiles for selected transcripts obtained by microarray analysis, showing the relative RNA amounts for a given transcript contained in each of the 12 fractions. Fractions associated with ribosomes are indicated. Different transcripts are color coded, and polysome profiles from three independent biological repeats are shown for rrk1 (RNase P K-RNA), rpl4101 (encoding ribosomal protein), htb1 (encoding histone H2B), and fba1 (encoding fructose-biphosphate aldolase). (C) Average translation profiles for selected groups of RNAs, plotted as in (B) for one experiment. All mRNAs, the 3505 high-confidence mRNAs with complete profiles in this experiment; Introns, 11 long introns included on the microarray; and Translation, 62 mRNAs associated with the GO terms “translational intiation,” “translational elongation,” or “translational termination.”

Figure 2

Inverse Correlation between Ribosome Density and ORF Length (A) Ribosome density plotted against ORF length for the 3598 high-confidence mRNAs. The inset graph shows moving averages (100 gene window) of ribosome density as a function of genes ranked by ORF length. The corresponding Spearman rank correlation is also shown. (B) Ribosome density plotted against ORF length as in (A) but showing only the 134 mRNAs encoding ribosomal proteins, along with corresponding Spearman rank correlation.

Figure 3

Correlations between ORF Length and Ribosome Density with Protein Level Moving averages (100 gene window) of ribosome density (black) and ORF length (red) as a function of 3265 genes ranked by protein level. The Spearman rank correlations between protein level and ribosome density (n = 3265) and between protein level and ORF length (n = 4434) are also shown.

Figure 4

Correlations between ORF Length and Ribosome Density with Poly(A) Tail Length Moving averages (100 gene window) of ribosome density (black) and ORF length (red) as a function of 2576 genes ranked by poly(A) tail length. The Spearman rank correlations between poly(A) tail length and ribosome density (n = 2576) and between poly(A) tail length and ORF length (n = 2714) are also shown.

Figure 5

Correlations between mRNA Level and Poly(A) Tail Length and Ribosome Occupancy (A) Moving averages (100 gene window) of poly(A) tail length as a function of 2688 genes ranked by mRNA level, along with corresponding Spearman rank correlation. (B) Average translation profiles of the mRNAs with the 500 highest (red) or 500 lowest (blue) levels plotted as in Figure 1B. (C) Moving averages (100 gene window) of ribosome occupancy as a function of 3567 genes ranked by mRNA level, along with corresponding Spearman rank correlation.

Figure 6

Correlations between mRNA Half-Lives and Other Gene Expression Properties Bar graphs showing the mean mRNA levels (A), ribosome occupancies (B), ribosome densities (C), poly(A) tail lengths (D), and Pol II occupancies (E) for two groups of mRNAs with either short (light gray) or long (dark gray) half-lives. The significance of the difference between the means from the two mRNA groups is given for each panel.

Figure 7

Correlations between Pol II Occupancy and Other Gene Expression Properties, and Relationships between All Studied Properties (A) Moving averages (100 gene window) of relative mRNA level as a function of 4724 genes ranked by Pol II occupancy, along with corresponding Spearman rank correlation. (B) Moving averages (100 gene window) of ribosome occupancy as a function of 3598 genes ranked by Pol II occupancy, along with corresponding Spearman rank correlation. (C) Moving averages (100 gene window) of poly(A) tail length as a function of 2713 genes ranked by Pol II occupancy, along with corresponding Spearman rank correlation. (D) Weighted association map summarizing the relationships between all aspects of gene expression analyzed here. The blue nodes represent the different data sets as labeled, black lines show significant positive correlations between the connected data sets, and red lines show significant inverse correlations. The weight of the lines reflects the absolute correlation value.