High-resolution timing of cell cycle-regulated gene expression

Abstract

The eukaryotic cell division cycle depends on an intricate sequence of transcriptional events. Using an algorithm based on maximum-entropy deconvolution, and expression data from a highly synchronized yeast culture, we have timed the peaks of expression of transcriptionally regulated cell cycle genes to an accuracy of 2 min (approximately equal to 1% of the cell cycle time). The set of 1,129 cell cycle-regulated genes was identified by a comprehensive analysis encompassing all available cell cycle yeast data sets. Our results reveal distinct subphases of the cell cycle undetectable by morphological observation, as well as the precise timeline of macromolecular complex assembly during key cell cycle events.

Fig. 1.

Characteristic temporal profile of cell cycle-regulated genes. Presented are the raw expression data (14) of several transcripts that peak at different times: cyclins (CLN3, CLN2, CLB1, and PCL9); a Cdc28p inhibitor (SWE1); a positive regulator of mitotic exit (SPO12); and an endochitinase, crucial for cell separation after mitosis (CTS1). y axis, normalized expression (arbitrary units); x axis, time in minutes.

Fig. 2.

Comparison of expression patterns of key cell cycle-regulated genes in slowly growing [YMC (14), Left] vs. rapidly growing [cdc15, alpha, and cdc28 (1, 5)] yeast cultures. (Upper) G₂ cyclin CLB1, mitotic transcription factor SWI5, and G₁ cyclin CLN3. (Lower) Mitotic cyclin PCL9 and MCM subunit MCM3. Environment-dependent G₁ phase is ≈10 times longer in YMC than found in previous studies (cdc15, alpha, and cdc28 synchronization), allowing the expression of key cell cycle genes to be timed with higher accuracy.

Fig. 3.

Expression profiles measured for the whole culture differ considerably from single-cell mRNA profiles. Imperfect synchronization of cells results in broadening of expression profiles measured in the culture (Right), compared with the respective single-cell profiles (Left). The original single-cell expression profiles (A, C, and E) can be reconstructed from the observed profiles (B, D, and F, respectively) by using deconvolution. This method can also be applied when the single-cell expression profile is complex (C and E) rather than just one short-lived pulse (A).

Fig. 4.

Transcriptional program of the yeast cell cycle. (A and B) Proteins involved in DNA replication initiation, color-coded according to timing of their expression. The prereplication complex (pre-RC) undergoes several changes before an elongation complex (EC), capable of initiating DNA synthesis, is formed (24). The order of expression (A) agrees with the order in which gene products are needed (B). In A, dotted outlines denote non-CCTR genes. Note the two groups of MCM subunits, each containing one nuclear localization signal (NLS). In B, solid outlines denote primary expression peak; dashed outlines denote secondary (lower scoring) peak. The only exception to just-in-time transcription is ORC1. (C) Timing of CCTR complexes. DSE, daughter-cell-specific expression program; APC act, APC activation; SPB sat, spindle pole body satellite formation; SPB sep, spindle pole body duplication and separation. (D) Peaks of selected CCTR genes. (E) Phases and subphases of the cell cycle. Note new prereplicative G₁ (P) phase. (F) Histogram of expression peaks of CCTR genes. (G) Peaks of transcripts regulated by selected cell cycle transcription factors (11, 20, 21, 29, 30). Note the differences in expression of MBF and SBF targets. (H) Histograms of peaks of CCTR genes involved in selected cell cycle functions. Compare peaks of predicted Cdc28p targets (26) vs. peaks of CDC28 (D).