Basic HTML Version
Full Abstracts – SYSTEMS AND QUANTITATIVE BIOLOGY
162
112
The YAN Network is robust against YAN protein variation in the developing eye.
Nicolás Peláez
1,2,3
, Hiba Eltahir
1,3
, Alec Victorsen
3
, Kevin White
3,4
,
Ilaria Rebay
3,4
, Luís Amaral
1,2,3
, Richard Carthew
1,3
. 1) Molecular Biosciences, Northwestern University, Evanston, IL; 2) Howard Hughes Medical Institute;
3) Chicago Center for Systems Biology, IL; 4) Universitry of Chicago, Dept. Human Genetics / Dept. for Cancer Research, IL.
A central problem in systems biology is how molecular networks accurately regulate the transition of cells from a multipotent to a differentiated state. The
YAN molecular network controls this transition in the developing eye imaginal disc. YAN is an ETS-domain transcription factor that represses expression of
key genes such as Dpax2, which are necessary for differentiation. YAN switches from high to low activity as cells transition into photoreceptor or cone
differentiation. Modulation of Notch and EGFR signaling alters a variety of YAN regulators as part of this network during the multipotent to differentiated
state transition. To quantitatively study the YAN network more directly we used a combination of genetics, confocal microscopy, image processing and
computational modeling. We recombineered monomeric YFP to the C-terminus of YAN within a 33 kb genomic transgene. This YAN:YFP transgene fully
rescued loss of endogenous YAN. We measured YFP fluorescence within eye disc cells as a proxy for YAN protein abundance. From this we have made
several surprising discoveries. First, multipotent cells show remarkably large variation in YAN abundance, even between neighboring cells. Second,
differentiated cone cells also show high levels of variation. Moreover, the level of YAN:YFP in cone cells is highly comparable to levels in multipotent
cells. Third, changing the transgene copy number results in large changes in YAN:YFP abundance, but has little to no effect on the robustness of cell
differentiation. We provide potential mechanisms to explain these observations.
113
Quantitative insights into enhancer architecture of dorsal-ventral patterning in Drosophila.
Rupinder Sayal
1
, Jacqueline Dresch
2
, Irina Pushel
1
, David
Arnosti
1
. 1) Biochem & Molec Biol, Michigan State Univ, East Lansing, MI; 2) Mathematics, Michigan State Univ, East Lansing, MI.
Enhancers facilitate gene regulation by binding sequence-specific transcription factors to generate precise temporal and spatial control of transcription. We
are interested in understanding the enhancer function by utilizing thermodynamic models, which compute gene expression as a probability of binding of
transcription factors (TFs). We used the enhancer of rhomboid gene for our studies, which is activated by Dorsal and Twist proteins, and repressed by Snail
protein. We mutated each activator-binding site singly and measured its effect on gene expression, which showed little attenuation. Subsequently, we
mutated two activator binding sites in all possible combinations, which revealed a hitherto unknown pattern of soft spots and idiosyncrasies in this enhancer.
We used a global parameter estimation strategy (CMAES) to derive parameters for transcription factor scaling factors and cooperativities, as well as
quenching of activators by repressors. To investigate the distance-dependent functions of cooperativity and quenching, various types of mathematical
functions were applied. Since the in vivo binding affinity of sites is still unknown, we also tested the model using different thresholds for affinities of
binding sites. This allowed us to observe the effects of including or disregarding weaker binding sites. In all, approximately 100 variants of the model were
tested. To test the predictive power of parameters, we compared our model predictions with experimental data on 6 other evolutionary variants of this
enhancer from sequenced fly species. The parameters were also tested on enhancers of other genes regulated by the same set of transcription factors. To our
knowledge, this is the first systematic study based on a rigorous mutational analysis of a single enhancer for thermodynamic mathematical modeling. The
results illustrate the power of applying quantitative mathematical models to shed insight on biochemical mechanisms underlying enhancer function.
114
Increasing discriminative power in computational evaluation of the BMP activity distribution in wing disks.
Alexi Brooks
1,2
, Tara Brosnan
1,2
, Mohit
Bahel
2,3
, David Umulis
4
, Laurel Raftery
1,2
. 1) School of Life Sciences, University of Nevada, Las Vegas, Las Vegas, NV; 2) CBRC, MGH/Harvard Medical
School, Charlestown, MA; 3) New York University, New York, NY; 4) Department of Agriculture and Biological Engineering, Purdue University,
Lafayette, IL.
Bone morphogenetic proteins (BMPs) are morphogens in many tissues of vertebrates and invertebrates. Extracellular BMP activates intracellular R-Smads,
via receptor-kinase phosphorylation of R-Smad C-termini. For fly BMPs, Mad is the critical R-Smad. Nuclear levels of receptor-phosphorylated Mad (C-
phospho-Mad) are proportional to the level of BMP activity and resultant target gene expression. In the larval wing primordium, the distribution of C-
phospho-Mad is a gradient ranging from two high level peaks straddling the anterior/posterior compartment boundary to low levels near the edges of the
wing pouch. The characteristics of this gradient, through various downstream targets, define the locations of the veins of the adult wing. Recent
computational analyses have demonstrated that a linear representation of the posterior gradient may be characterized simply by exponential decay, but the
anterior gradient remains an open question. We have developed additional computational tools to quantitatively characterize the C-phospho-Mad gradient in
two dimensions, for a population of wing disks of the same genotype. To test for subtle alterations in BMP activity gradients, we express a potential
modulator from the Ras-MAP kinase pathway in the dorsal compartment, using Apterous-Gal4. Even in the presence of natural cell-to-cell and field-to-field
variability, we can now use reliable metrics of gradient length, amplitude, and shape throughout the wing pouch. Availability of such metrics for both
compartments allow us to discriminate between models for pathway cross-talk in BMP signaling.