Temporal Partitioning of the Yeast Cellular Network - Systems Biology of Metabolic and Signaling Networks - page 324

of the yeast oscillatory phenotype (Murray et al. 2007 ). This revealed that the entire
biochemical network or reactome self-organises into two distinctive regions:
oxidative and reductive. Transcription during the oxidative phase was almost
exclusively focused on biosynthesis, and a clear temporal programme was initiated
starting with nucleotide biosynthesis and ending with acetate metabolism. This
programme spanned 10-12 min, and metabolical, physiological and morphological
processes (e.g. sulphate uptake (Fig. 12.2 ), amino acid biosynthesis, S-phase, etc.)
occurred 10-14 min after the peak in their transcriptional activity. Statistical
analysis of transcription factor binding targets and the reconstruction of a yeast
protein-protein interaction map (from multiple sources) implicated the temporal
construction and destruction of a transcription factor complex (Fig. 12.3b ), com-
prising Cbf1p (centromere binding factor), Met4p, Met28p, Met31p and Met32p
(methionine regulation factors) and Gcn4p (general control protein involved in
nitrogen catabolite repression); these processes orchestrated the majority of tran-
scriptional changes of sulphate uptake and amino acid synthesis, and by targeting
other transcription factors. The gene targets of the complex formed by these
transcription factors produced the largest amplitude oscillation of the measured
transcripts, and their metabolic products GSH, H 2 S and S -adenosylmethionine also
oscillated. Network analysis of the oscillatory transcription factor network (com-
prising some 33 transcriptional regulators) indicated that this works by temporally
spacing gene transcription via the formation of multiple input feed-forward genetic
circuits.
Out of phase with oxidative phase transcription, a much larger group of
transcripts (~80 % of the most oscillatory transcripts) showed a peak production
in the reductive phase. Many transcripts encode for proteins that encode mitochon-
drial assembly, respiration, carbohydrate catabolism and the stress response
proteins. Moreover, many of these processes are shown to occur 14-18 min later,
indicating that transcription actually precedes subsequent activity of the gene
products. It was clear from the annotation and statistical analysis of network
structure that this group of transcriptional regulatory proteins controlling the
expression of these processes are the most highly interconnected nodes in the
yeast network. These have complex regulatory patterns and are key components
of the differentiation responses in yeast (for pseudohyphae and spore formation).
However, the regulation of this system, whose targets show a strong oscillatory
pattern and peak in the reductive phase, remained largely enigmatic.
Recently, new computational analyses aimed at elucidating a more global
regulatory system that modulate these transcripts have implicated energetic state
(ATP:ADP ratio) as a key factor (Machn´ and Murray 2012 ). When transcript data

Fig. 12.3 (continued) respective phenotype of biosynthetic. The heat maps were ordered
according to the phase angle ( ) of the measurements' peak production. Activities of the tran-

scriptional regulators were derived from the statistical analysis of their target transcripts and a ball-
and-stick network representation of the transcriptional regulators (b). See Fig. 12.2 for a guide to
the network. The shaded box indicates a transcription factor complex comprising of Cbf1p and the
Met transcriptional regulators (Murray et al. 2007 )

ϕ