Glycan processing in the Golgi -- optimal information coding and constraints on cisternal number and enzyme specificity

Alkesh Yadav, Quentin Vagne, Pierre Sens, Garud Iyengar, Madan Rao

TL;DR

This paper uses information theory to explain why the Golgi apparatus is organized into multiple compartments (cisternae). By modeling glycan synthesis as a molecular coding process, the authors demonstrate that multiple cisternae and a "Goldilocks" level of enzyme specificity are essential for cells to accurately produce the complex glycan signatures required for biological recognition.

What Problem Does This Paper Address?

In biology, glycans (sugar chains attached to proteins) act as a "molecular code" on the cell surface, signaling cell identity and state. These glycans are synthesized in the Golgi apparatus, a series of stacked compartments called cisternae.

A long-standing question in cell biology is: Why is the Golgi divided into multiple compartments instead of being one big vat?

The authors tackle this by viewing the Golgi as a chemical assembly line. They investigate the trade-offs between the number of cisternae ($N_C$), the number of enzymes ($N_E$), and enzyme specificity ($\sigma$). The challenge is to understand how these physical constraints limit a cell's ability to produce a specific, complex "target" distribution of glycans with high fidelity.

Methods

The authors developed a quantitative framework combining chemical kinetics with information theory:

  • The Model: They represent glycan synthesis as a sequence of enzymatic reactions (Michaelis-Menten kinetics) coupled with transport between cisternae.
  • Fidelity Metric: To measure how well the Golgi "computes" the right output, they use Kullback-Leibler (KL) divergence. This quantifies the error between the synthesized glycan distribution and a target distribution.
  • Complexity Metric: They define the complexity of a glycan profile by fitting it to a Gaussian Mixture Model (GMM). The number of Gaussian components ($k$) serves as a proxy for complexity.
  • Optimization: They formulated a constrained optimization problem to find the enzyme reaction rates and specificities that minimize the KL divergence for a given number of cisternae. They utilized Sequential Quadratic Programming (SQP) to solve these non-convex problems.

Datasets

The study uses real-world biological data to ground its theoretical model:

  • MSMS Data: Mass Spectrometry (MSMS) profiles of glycans from human T-cells, human neutrophils, and mouse neutrophils.
  • KB2005 Model: Parameters for enzymatic reaction rates were derived from established models of N-glycan synthesis pathways, which include tens of thousands of possible chemical reactions.

Key Insights and Ideas

  • The Necessity of Compartments: To synthesize complex glycan distributions (like those found in real T-cells), a single compartment is insufficient. Multiple cisternae allow for "precise enzyme partitioning," effectively creating a multi-stage assembly line that reduces "noise" (the production of unwanted intermediate glycans).
  • Optimal "Sloppiness": There is an optimal level of enzyme specificity. If enzymes are too specific, they can't handle the variety of precursors; if they are too "sloppy" (promiscuous), the output becomes a blurry mess. The optimal specificity increases as the target distribution becomes more complex.
  • Robustness through Architecture: Increasing the number of cisternae makes the system less sensitive to the exact tuning of enzyme specificity. In other words, a more complex Golgi structure allows for "sloppier" enzymes while maintaining high-fidelity output.
  • Diversity vs. Fidelity: While high specificity is needed to match a specific code (fidelity), low specificity combined with many cisternae can actually help generate a diverse repertoire of glycans, a strategy potentially used by plants and algae.

Key Results

  • Complexity Saturation: Real cell glycan profiles are highly complex; the log-likelihood of GMM fits for T-cells only saturates at around 20 components, indicating a high-information-density code.
  • Trade-off Curves: The normalized KL divergence is a convex function of enzyme specificity ($\sigma$). For a fixed number of enzymes, increasing $N_C$ from 1 to 7 significantly reduces the minimum achievable error (fidelity).
  • Cisternal Constraints: For typical mammalian cells ($N_E \approx 10-20$), the model suggests an optimal cisternal number of 3 to 8, which aligns remarkably well with observed biological values.

Limitations and Future Work

  • Mean-Field Assumption: The model is "mean-field," meaning it ignores stochastic noise resulting from low copy numbers of molecules. In real cells, the small volumes of cisternae might make stochasticity a significant factor.
  • Linear Pathways: The current analysis focuses on a line-graph of reactions. Real glycan synthesis involves complex, branched "tree-like" reaction networks.
  • Dynamic Assembly: The authors suggest that future research should combine these chemical processing constraints with the physical dynamics of the Golgi (fission, fusion, and transport) to understand how the organelle maintains its structure while performing these complex computations.

Summary updated

Abstract

Many proteins that undergo sequential enzymatic modification in the Golgi cisternae are displayed at the plasma membrane as cell identity markers. The modified proteins, called glycans, represent a molecular code. The fidelity of this glycan code is measured by how accurately the glycan synthesis machinery realises the desired target glycan distribution for a particular cell type and niche. In this paper, we quantitatively analyse the tradeoffs between the number of cisternae and the number and specificity of enzymes, in order to synthesize a prescribed target glycan distribution of a certain complexity. We find that to synthesize complex distributions, such as those observed in real cells, one needs to have multiple cisternae and precise enzyme partitioning in the Golgi. Additionally, for fixed number of enzymes and cisternae, there is an optimal level of specificity of enzymes that achieves the target distribution with high fidelity. Our results show how the complexity of the target glycan distribution places functional constraints on the Golgi cisternal number and enzyme specificity.
BibTeX
@inproceedings{yadav2022glycan,
  author = {Alkesh Yadav and Quentin Vagne and Pierre Sens and Garud Iyengar and Madan Rao},
  title = {{Glycan processing in the Golgi -- optimal information coding and constraints on cisternal number and enzyme specificity}},
  booktitle = {eLife},
  year = {2022},
  doi = {10.7554/elife.76757},
}

Loading related content...

Figure