A dynamic model of proteome changes reveals new roles for transcript alteration in yeast

Abstract

The transcriptome and proteome change dynamically as cells respond to environmental stress; however, prior proteomic studies reported poor correlation between mRNA and protein, rendering their relationships unclear. To address this, we combined high mass accuracy mass spectrometry with isobaric tagging to quantify dynamic changes in ~2500 Saccharomyces cerevisiae proteins, in biological triplicate and with paired mRNA samples, as cells acclimated to high osmolarity. Surprisingly, while transcript induction correlated extremely well with protein increase, transcript reduction produced little to no change in the corresponding proteins. We constructed a mathematical model of dynamic protein changes and propose that the lack of protein reduction is explained by cell-division arrest, while transcript reduction supports redistribution of translational machinery. Furthermore, the transient 'burst' of mRNA induction after stress serves to accelerate change in the corresponding protein levels. We identified several classes of post-transcriptional regulation, but show that most of the variance in protein changes is explained by mRNA. Our results present a picture of the coordinated physiological responses at the levels of mRNA, protein, protein-synthetic capacity, and cellular growth.

Figure 1

Experimental workflow and mass spectrometry identification summary. (A) Yeast cells were grown to mid-log phase and exposed to 0.7 M NaCl; culture volumes were removed at 30, 60, 90, 120, and 240 min after stress, as well as from unstressed cells (0 min), for microarray and quantitative MS proteomic analysis. Each proteomic sample was lysed, followed by protein extraction and enzyme digestion. Peptides were labeled with isobaric TMT and mixed in equal rations. The labeled mix was then subjected to an orthogonal first-dimension separation: SCX. Fractions were subsequently analyzed on an LTQ Orbitrap Velos mass spectrometer coupled with nano-RP HPLC. Biological replicates were performed in triplicate. Spectra were analyzed with in-house developed software. (B) Peptide and protein identifications across the three biological replicates (BR) are outlined in the above table with the overlap depicted in the Venn diagrams. Of the protein identifications, 60% overlap was observed across all biological replicates. On average, 81.5% of all identifications were quantifiable with an overlap of 55% across all biological replicates. Numbers in each colored circle represent the correspondingly colored sector in the Venn diagrams.

Figure 2

Correlation between mRNA and protein changes. (A) The average log₂ changes in mRNA (left) and protein (right) are shown for each gene, represented as rows, and time point shown as columns within each time course (triangles). Red and green colors indicate increased and decreased abundance, respectively, according to the key. The figure shows all transcripts whose change was statistically significant (FDR <0.05) and whose corresponding proteins were measured in triplicate, amounting to 408 transcripts that increased and 702 transcripts that decreased in abundance. (B) Regression of the maximum average log₂ changes in mRNA and protein for transcript–protein pairs shown in (A), where red indicates increased and green represents reduced transcripts. (C) R²-values (shaded according to the key in the center) were calculated as in (B) except that they compared mRNA with protein at each denoted time point for increased (left) and reduced (right) transcripts. Comparisons with the highest correlations are highlighted with asterisks. (D) Average log₂ change over time for all mRNAs (solid lines) and corresponding proteins (dashed lines).

Figure 3

Distribution of protein levels predicted from model simulations. Distribution of the minimum log₂ change over time (i.e. greatest log₂ reduction) for transcripts (red) that decreased in abundance, corresponding protein changes as measured (orange) or calculated using the original model (green), corresponding protein changes calculated under the assumption of a constant growth rate (blue), and protein changes calculated with the original model but predicted from mRNAs simulated without reduction during the time course (purple). Analysis is for 579 triplicated mRNA–protein pairs from the training set for which mRNA was reduced with statistical significance (FDR<0.05).

Figure 4

Translational profiles in wild-type and dot6Δtod6Δ cells responding to NaCl. Polysome profiles were measured as described in Materials and methods for (A) wild-type cells and (B) a mutant lacking the Dot6p/Tod6p transcriptional repressors. Absorbance at 260 nm across collected fractions is shown at 0, 5, 30, 45, and 90 min after NaCl treatment, relative to the starting baseline; 40S, 60S, and 80S monosome (M), and polysome (P) peaks are indicated. The monosome/polysome (M/P) ratio was calculated based on the trapezoidal area under the curve. Relative abundance of (C) ARX1 or (D) HSP104 transcript in polysome fractions was measured in wild-type (left) and dot6Δtod6Δ cells (right) before and at 30 min after NaCl treatment. Relative mRNA abundance was measured compared with a doped control mRNA and normalized to baseline abundance measured in the trough between the 2 and 3 polysome peaks (see Materials and methods). Plots are representative of biological duplicates.

Figure 5

Transient mRNA changes produce faster protein changes. (A) Representative changes in mRNA (solid) and protein (dashed) for two pairs with and without transient mRNA induction. (B) In all, 127 transcripts whose increase in abundance peaked at 30 min were binned based on the magnitude of transient burst (see Materials and methods). The percentage with the indicated acclimation times is shown by quartile. (C) Average log₂ change of measured mRNA, simulated mRNA without the transient burst, and corresponding calculations of protein change, for GPD1. (D) Percentage of proteins shown in (B) with different acclimation times, for proteins measured (‘M’) or calculated (‘C’) from original or simulated (‘S’) mRNAs.

Figure 6

Influence of PTR. (A, B) Average log₂ changes in each replicate time course for 77 transcripts with increased (red) and 61 transcripts with decreased (green) abundance (A) and corresponding proteins (B) subject to noise reduction. (C) Fraction of variation in protein change explained by mRNA change for significantly changing proteins without mRNA changes, with opposing mRNA changes, with changes greater than mRNA, with noise reduction, and in the GO ‘stress response’ category as example of very high correlation.

Figure 7

Estimated fraction of translating ribosomes made available due to transcript reduction. The fraction of 171 000 translating ribosomes before stress that becomes available due solely to transcript reduction (blue bars) was estimated as described in Materials and methods. The estimated maximum fraction of those ribosomes needed to translate new transcripts at each time point (gray bars) was based on the increase in mRNA abundance at that time and assumed 100% ribosome loading to pre-stress levels at each time point. Each bar represents the average and s.d. over the three biological replicates of measured mRNA abundance changes. Source data is available for this figure at www.nature.com/msb.