Annual Scientific Report 2015

Beltrao Group
Evolution of Cellular Networks
Our group is interested in understanding how novel cellular functions arise
and diverge during evolution. We study the molecular sources of phenotypic
novelties, exploring how genetic variability introduced at the DNA level is
propagated through protein structures and interaction networks to give rise to
phenotypic variability.
Within the broad scope of this evolutionary problem, we
focus on two areas: the function and evolution of posttranslational
regulatory networks, and the evolution
of genetic and chemical-genetic interactions. Looking
beyond evolutionary process, we also seek to understand
the genomic differences between individuals and
improve our capacity to devise therapeutic strategies.
In collaboration with mass-spectrometry groups, we
develop a resource of experimentally derived, posttranslational
modifications (PTMs) for different
species in order to study the evolutionary dynamics
and functional importance of post-translational
regulatory networks. We use these data to create novel
computational methods to predict PTM function and
regulatory interactions. Our goal is to gain insights into
the relationship between genetic variation and changes
in PTM interactions and function.
Changes in cellular interaction networks underpin
variation in cellular responses and sensitivity to
environmental perturbations or small molecules.
As we model and study the evolution of cellular
interaction networks, we begin to see how different
individuals or species diverge in their response to
drugs. Understanding this relationship will enable us to
develop methods to predict how genetic changes result
in specific sensitivity to drug combinations.
Major achievements
In 2015 we published a study on the conservation
and structural properties of phosphosites in X. laevis
(Johnson et al., 2015). We collected phosphoproteomics
data for X. laevis and analysed the conservation of these
phosphosites across 13 other species. We found that the
degree of conservation of phosphosites and putative
kinase–protein interactions is predictive of functionally
relevant sites and interactions. We then used protein
homology modelling to show that phosphosites tend
to be in regions of proteins that are conformationally
variable. This suggests that some of these sites might
exert their function by controlling protein conformation.
We believe that such studies of PTM function might,
in future, lead to the engineering of PTM regulation by
rational design.
Understanding how protein kinases identify their
targets substrates is an open question in cell signalling.
We developed a new approach that combines protein
phosphorylation with interaction networks to predict
what are the sequence determinants for kinase
recognition. We used this method to predict the
specificity of hundreds of human kinases (Wagih et
al., 2016).
Changes in external conditions are sensed by the cell,
which needs to trigger a response in order to adapt to
the new environment. This adaptation can take many
forms, such as changing protein post-translational
modifications, their expression and interaction patterns.
It is expected that the degree of functional association
between pairs of genes will depend greatly on the
environment. To study this, we measured the fitness, in
different conditions, of yeast cells missing combinations
of pairs of genes (Martin et al., 2015). In total we
measured around 250 000 conditional gene-gene
interactions. This large-scale endeavour showed that
that a large fraction of genetic interactions are only
observed under specific environmental conditions.
Using these data, we were able to identify and validate
novel condition-specific roles for several yeast genes.
Future plans
In 2016 we will continue our studies of the function and
evolution of cellular networks with a strong emphasis
on phospho-regulatory networks. Human cells have
on the order of 500 kinases that are used by the cell
to react to different stimuli and to reach decisions on
what changes need to occur to cellular state. We seek to
understand how these kinase-signalling networks are
used, in different environmental conditions, to define
specific cellular responses. To study this we have been
developing approaches to study cell signalling states
using comparative phosphoproteomics. We are also
interested in understanding the structural properties
that define protein-kinase specificity and continue to
develop ways to combine different types of information
to predict specificity from sequence. We aim to study
the evolution of kinase specificity using these methods.
We are also using some of the tools we have developed in
applied research into protein kinases.
131
2015 EMBL-EBI Annual Scientific Report