Temporal Partitioning of the Yeast Cellular Network - Systems Biology of Metabolic and Signaling Networks - page 327

respiration serves as a mechanism to temporally partition these metabolic networks.
The importance of mitochondrial respiratory control in the generation of
temperature-compensated ultradian clock-driven (epigenetic) oscillatory metabo-
lism in a range of lower eukaryotic organisms (three yeasts and five protists species)
has been experimentally demonstrated (Edwards and Lloyd 1978 ; Lloyd and Rossi
1992 ; Murray et al. 2001 ).
Generally, there is a lack of annotation for many of the reductive phase
transcripts, which in part, may be due to the focus of research activity by the
yeast community on glucose repressed growth, where differentiation programmes
and respiratory catabolism are repressed. Therefore, we analysed the promoter
architectures, i.e. the regional sequence biases, in vitro and in vivo nucleosome
positioning, to reveal significant differences between each cluster (Machn´ and
Murray 2012 ). For example the promoters of the genes in cluster A (ribosomal) had
large nucleosome free regions that were enriched with AT containing motifs (well
known to have a lowered affinity for nucleosomes). Local nucleosome positions can
be classified by their “fuzziness”, i.e. as clearly defined positions or poorly defined.
Nucleosome position is dictated by the DNA sequence and enzymatically, e.g. as
the ATP-dependent remodelling complexes RSC and Isw2 complexes (Zhang
et al. 2011 ). It appears that these act differentially, where the RSC activates
(Lorch et al. 1998 ) and Isw2 represses transcription of genes (Vincent et al. 2008 )
based on their promoter configuration. In our analyses, we show the structural
clusters A and C have very well-defined nucleosomes while clusters B and D
have fuzzy nucleosome positions, presumably related to the development of struc-
ture to function between A to B and C to D. A caveat in in vivo nucleosome
positioning data is that the analyses were almost exclusively carried out in rapidly
growing exponential phase cultures, therefore the measured positioning may well
be context dependent. This model is supported by the strict ATP dependence of
in vitro promoter configurations (Zhang et al. 2011 ). Therefore, we proposed a
simple dual negative feedback loop involved in the regulation of catabolic and
anabolic processes, which would have a high potential to autonomously oscillate
(Fig. 12.5 ). When ATP availability is high anabolism is initiated by ATP-dependent
remodelling by the RSC on the anabolic and growth genes in cluster A, AB
and B. The protein products of these genes then produce cellular growth leading
to the down-regulation of these genes as ATP is consumed. In conjunction with this,
increased ATP availability early in the oxidative phase leads to Isw2 remodelling of
nucleosomes over the promoters of the catabolic genes. This leads to a repression of
the catabolic genes during the oxidative phase, leading to a decrease in expression.
Once ATP concentration declines late in the oxidative phase, the nucleosomes relax
and catabolic genes are transcribed. It is apparent that anything which stresses the
cell and causes a drop in ATP production (such as ROS damage of membranes) or
mutations that cause the balance between catabolism and anabolism to be altered,
will cause a rearrangement of chromatin structure, leading to stress and catabolic
transcript production (without the need for specific transcription factors). Indeed,
this model is supported in higher eukaryotes by previously reported in vivo and
in vitro ultradian oscillations in nucleosome remodelling in glucocorticoid and