Temporal Partitioning of the Yeast Cellular Network - Systems Biology of Metabolic and Signaling Networks - page 326

on the five consensus clusters. The programme consisted of cell structural and
metabolic changes and as it is cyclical we arbitrarily define the start point as the
minimum first derivative of residual dissolved oxygen concentration (Fig. 12.1a ).
Three distinct clusters were observed in the oxidative phase, and two were observed
in the reductive phase. Functionally, the first common oxidative phase cluster
(A) consisted mainly of structural function, e.g. nucleolar synthesis, cytoplasmic
ribosome biogenesis and RNA polymerase I and III. The next common cluster
(AB) was enriched in cytoplasmic ribosomal proteins and translation. The final
oxidative phase cluster comprised mostly of anabolic reactions (B), such as amino
acid biosynthesis. The first reductive phase cluster (C) was enriched mitochondrial
ribosomal transcripts, i.e. coding for a structural component. The final common
reductive phase cluster (D) was primarily a metabolic cluster and comprised of
many catabolic pathways and the genes involved in redox regulation and the
general stress response. Therefore, both the oxidative and reductive phases start
with a structural bias and finish with a metabolic bias. The period differences
observed during the two oscillations (40 min and 300 min) are thought to be due
to differences in growth conditions and strain, and they led an interesting context-
dependent development of temporal gene expression, where the reductive phase
clusters C and D were co-expressed in the 40-min oscillation and the oxidative
phase clusters A, AB and B were co-expressed during the oxidative phase.
Strikingly, when these clusters were compared to a compendium of 1,327
microarray hybridisations (McCord et al. 2007 ), the oxidative and reductive
phase clusters were differentially expressed in the majority of the array
hybridisations (Fig. 12.4b ). This analysis indicates that most of the experiments
carried out in yeast effect a global hitherto unknown regulation that is epitomised
by the respiratory oscillation. Further analysis supported a role for the general
growth rate response where genes up-regulated during rapid growth corresponded
with oxidative phase genes (A, AB and B) and those down-regulated corresponded
with reductive phase genes (C and D) (Brauer et al. 2008 ; Slavov and Botstein
2011 ; Machn´ and Murray 2012 ). Moreover, the environmental stress response
(Gasch and Werner-Washburne 2002 ) tended to correlate negatively with the
growth rate response. These analyses support a much more basic mechanism for
gene regulation that switches the system from the “growth” to “stress” responses.
Indeed, it also makes it difficult to define what the “growth” or “stress” responses
are as both these conditions alternate in the stable, unperturbed environment that is
afforded by continuous culture. Implicit in this regulation is an oscillation in ATP:
ADP ratio (Akiyama and Tsurugi 2003 ; Lloyd and Murray 2007 ; Machn´ and
Murray 2012 ) as this is a major determinant of growth rate and is intricately
interwoven with the redox state of the cell and the membrane status of the
mitochondria. The appropriate chemical and membrane potentials must be
maintained to generate ATP (Chance and Williams 1955 ; Murray et al. 2011 ).
Maximum ATP availability occurs at the beginning of the oxidative phase and
minimum ATP availability occurs at the beginning of the reductive phase. There-
fore, we argue that the “growth” and “stress” responses are better described as a
global switch between catabolic and anabolic processes. The oscillation in