The effect of cell geometry on polarization in budding yeast

Michael Trogdon, Brian Drawert, Carlos Gomez, Samhita P. Banavar, Tau-Mu Yi, Otger Campàs, Linda R. PetzoldPLOS Computational Biology 2018 · DOI · OpenReview

TL;DR

This paper investigates how cell shape influences the localization of key proteins during yeast mating, a crucial process for cell growth and reproduction. Using computational models and experiments, the authors demonstrate that cell geometry can significantly destabilize protein polarization, causing it to drift away from the cell's tip. This finding highlights the underappreciated role of cell shape in biological pattern formation and suggests that cells may employ additional mechanisms to maintain polarization during growth.

What Problem Does This Paper Address?

Polarization, the process by which cells establish distinct functional regions, is fundamental to many biological processes, including cell division, migration, and differentiation. In budding yeast, mating involves a complex cascade of events leading to the formation of a mating projection, guided by the localized accumulation of specific proteins on the cell membrane. While numerous mathematical models exist to describe this polarization process, they often simplify the cell's geometry, typically using basic shapes like spheres or lines. This simplification overlooks a critical factor: the actual, often complex and dynamic, shape of the cell. The authors identify a gap in our understanding of how realistic cell geometries, particularly the tip-shaped structures formed during mating projection growth, affect the dynamics of protein polarization. They hypothesize that cell geometry might play a more significant role than previously appreciated, potentially leading to instability in the polarization site and impacting the accuracy of existing models.

Methods

The study employs a combination of computational modeling and experimental validation.

  • Computational Modeling: The core of the computational work involves simulating spatial stochastic models of protein polarization using the Reaction-Diffusion Master Equation (RDME) formalism. This approach allows for the modeling of chemical reactions and diffusion within a spatial domain, capturing the inherent randomness (stochasticity) of molecular interactions within cells.
  • PyURDME: The authors utilize their in-house software, PyURDME, which is capable of simulating RDME on complex, 3D, and time-dependent geometries. This is crucial for modeling the dynamic shapes of yeast cells during mating projection growth.
  • Geometries: Computational meshes representing realistic yeast cell shapes, including spheres and various tip-shaped projections, were generated based on a separate model of cell wall mechanics.
  • Models of Polarization: Two main classes of models were investigated:
  • A detailed, mechanistic model of Cdc42 polarization (S1 Model), adapted from existing literature.
  • A simplified model of Cdc42 polarization (S2 Model), known to rely on stochastic effects for polarization.
  • A model of polarisome (actin) formation (S3 Model) was also incorporated to investigate the interplay between actin dynamics and Cdc42 polarization.
  • Simulation Approach: Simulations were run for extended periods (up to 100,000 seconds) with multiple realizations to capture the stochastic nature of the system and assess the stability and location of the polarization cap.
  • Experimental Validation: To test the computational predictions, the researchers designed an experiment using a cla4∆ yeast strain, which is known to exhibit elongated projections.
  • Experimental Setup: Cells were treated with pheromone to induce mating projection growth and then with latrunculin A (LAT-A) to depolymerize actin cables. This treatment isolates the Cdc42 polarization pathway, allowing for the observation of its behavior in different cell shapes.
  • Observation: The localization of a GFP-tagged reporter for active Cdc42 (Ste20-GFP) was tracked over time in spherical, short-projection, and long-projection cells. Stability of the polarization cap was assessed by observing whether its location changed after actin depolymerization.

Datasets

Not explicitly discussed as distinct datasets for training or evaluation in the traditional machine learning sense. The "datasets" are generated through the computational simulations and experimental observations described in the "Methods" section. The computational meshes representing cell geometries are derived from a separate cell wall mechanics model [34].

Key Insights and Ideas

  • Geometry is Not Just a Passive Container: The study strongly suggests that cell shape is not merely a passive boundary but actively influences the dynamics of intracellular processes like protein polarization. This challenges the common practice of simplifying geometries in modeling.
  • The "Drift" Phenomenon: A key finding is the tendency for the polarization cap to "drift" away from the tip in tip-shaped geometries, even when starting from a polarized state. This indicates an inherent instability introduced by the geometry itself.
  • Length Scale Matters: The stability of the polarization cap appears to be dependent on the relative sizes of the cap and the geometric features of the cell. Longer projections seem to stabilize the cap at the tip, possibly because the cap is smaller than the projection length, whereas shorter projections can lead to instability.
  • Actin Dynamics May Not Be Enough: Integrating actin dynamics into the models did not fully rescue the polarization cap from drifting away from the tip in tip-shaped geometries. This implies that the geometry-induced bias is a strong effect that may require more than just the known actin-based mechanisms to overcome.
  • Experimental Confirmation: The experimental results qualitatively align with the computational predictions, showing greater stability of the polarization cap in spherical and long projection cells compared to short projection cells.

Key Results

  • Tip-Shaped Geometries Induce Instability: For tip-shaped geometries, the active Cdc42 polarization cap showed a tendency to drift away from the tip over time, a phenomenon not observed in spherical geometries.
  • Geometry-Dependent Manifolds: The final resting positions of the polarization cap in tip-shaped geometries formed distinct "manifolds," suggesting that the geometry dictates stable, albeit off-tip, locations for polarization.
  • Longer Projections Stabilize Polarization: In significantly elongated projections, the polarization cap was observed to be more stable at the tip compared to shorter projections.
  • Experimental Stability Trends: Experiments confirmed that polarization was more stable in spherical and long projection cells than in short projection cells, with statistical significance observed for the differences.
  • Actin Dynamics Insufficient (in these models): Coupled models of Cdc42 and actin polarization showed that actin dynamics, as modeled, were largely unable to overcome the geometry-induced bias away from the tip.

Limitations and Future Work

  • Model Simplifications: The authors acknowledge that their models, while incorporating stochasticity and realistic geometries, are still simplifications of the complex biological reality. The exact mechanisms by which cells overcome geometry-induced polarization instability in vivo remain an open question.
  • Experimental Isolation: While efforts were made to isolate the Cdc42 pathway by depolymerizing actin, it's acknowledged that completely isolating the effect of geometry from other cellular processes is challenging.
  • Deterministic vs. Stochastic: While the study primarily uses stochastic simulations, the authors note that replicating these experiments in a deterministic setting would be valuable but computationally challenging for complex geometries.
  • Future Directions:
    • Investigating the role of other cellular components and mechanisms (e.g., cell wall constraints on actin cables) in stabilizing polarization.
  • Exploring more complex and dynamic cell shapes.
  • Developing more efficient computational methods for deterministic simulations on complex geometries.
  • Further experimental validation of specific predictions, such as the role of length scales.

Summary updated

Abstract

Author summary The ability of cells to spontaneously break symmetry and form patterns via networks of protein reactions is a well-studied, yet not fully understood, phenomenon. One particularly useful model system is the mating of yeast cells, where a localization of proteins on the membrane leads to actin cable formation, vesicle traffic, changes in material properties of the cell wall and ultimately to the growth of a mating projection. There is no shortage of mathematical models describing the mechanisms of polarization, yet they are often formulated as differential equations on simple domains. The effect of geometry on the dynamics of polarization has not been systematically explored. There is an interaction between the length scales of polarization and the features of the geometry of the cell that has been under-appreciated and may lead to novel biological insights.
BibTeX
@article{trogdon2018effect,
  author = {Michael Trogdon and Brian Drawert and Carlos Gomez and Samhita P. Banavar and Tau-Mu Yi and Otger Campàs and Linda R. Petzold},
  title = {{The effect of cell geometry on polarization in budding yeast}},
  journal = {PLOS Computational Biology},
  year = {2018},
  doi = {10.1371/journal.pcbi.1006241},
}
Figure