Source-linked AI summary
Localization of protein aggregation in Escherichia coli is governed by diffusion and nucleoid macromolecular crowding effect
Anne-Sophie Coquel, Jean-Pascal Jacob, Maël Primet, Alice Demarez, Mariella Dimiccoli, Thomas Julou, Lionel Moisan, Ariel B. Lindner, Hugues Berry
TL;DR
How aging-related protein aggregates localize in E. coli, and whether their movement is active or passive, remained unresolved. Using live-cell tracking and 3D individual-based modeling, the study found purely diffusive motion and that diffusion, aggregation, and nucleoid hindrance reproduced salient localization observations.
Problem
The mechanisms localizing aging-related protein aggregates in E. coli remained controversial, particularly whether their transport is active or passive.
Source: AUTHOR SUMMARY · p. 3, INTRODUCTION · p. 6Method
The study combined live-cell single-particle tracking with 3D individual-based modeling to assess aggregate movement and localization.
Source: DISCUSSION · pp. 17–18Results
Aggregate movements were purely diffusive, with coefficients of the order of 500 nm^2/s depending on size; diffusion, aggregation, and nucleoid hindrance reproduced salient localization observations.
Source: DISCUSSION · pp. 17–18Takeaways & Limitations
The findings support polar aggregate localization in E. coli resulting from passive diffusion-aggregation coupled with spatially non-homogeneous macromolecular crowding.
Source: DISCUSSION · p. 18Computation-time limits inherent to individual-based models restrict simulated times, motivating a mean-field approach to reach times spanning multiple cell generations.
Source: DISCUSSION · p. 18
Abstract
from arXiv · showhide
Aggregates of misfolded proteins are a hallmark of many age-related diseases. Recently, they have been linked to aging of Escherichia coli (E. coli) where protein aggregates accumulate at the old pole region of the aging bacterium. Because of the potential of E. coli as a model organism, elucidating aging and protein aggregation in this bacterium may pave the way to significant advances in our global understanding of aging. A first obstacle along this path is to decipher the mechanisms by which protein aggregates are targeted to specific intercellular locations. Here, using an integrated approach based on individual-based modeling, time-lapse fluorescence microscopy and automated image analysis, we show that the movement of aging-related protein aggregates in E. coli is purely diffusive (Brownian). Using single-particle tracking of protein aggregates in live E. coli cells, we estimated the average size and diffusion constant of the aggregates. Our results evidence that the aggregates passively diffuse within the cell, with diffusion constants that depend on their size in agreement with the Stokes-Einstein law. However, the aggregate displacements along the cell long axis are confined to a region that roughly corresponds to the nucleoid-free space in the cell pole, thus confirming the importance of increased macromolecular crowding in the nucleoids. We thus used 3d individual-based modeling to show that these three ingredients (diffusion, aggregation and diffusion hindrance in the nucleoids) are sufficient and necessary to reproduce the available experimental data on aggregate localization in the cells. Taken together, our results strongly support the hypothesis that the localization of aging-related protein aggregates in the poles of E. coli results from the coupling of passive diffusion- aggregation with spatially non-homogeneous macromolecular crowding. They further support the importance of "soft" intracellular structuring (based on macromolecular crowding) in diffusion-based protein localization in E. coli.
AUTHOR SUMMARY
An integrated experimental and modeling strategy indicates that misfolded-protein aggregates accumulate at polar and mid-cell positions in E. coli through diffusive movement and nucleoid occlusion. Increased macromolecular crowding in the nucleoids is central to this localization.
AUTHOR SUMMARY
- Diffusion combined with nucleoid occlusion underlies aggregate accumulation at polar and mid-cell positions in E. coli.The results indicate that increased macromolecular crowding in the nucleoids contributes to this localization.Source: AUTHOR SUMMARY · p. 3
- An integrated strategy combining experimental and modeling approaches was devised to identify mechanisms governing aggregate localization.Source: AUTHOR SUMMARY · p. 3
- Because bacteria lack internal organelles and vesicle-based sorting systems, protein localization relies on alternative mechanisms that often remain unclear.In E. coli, localization of misfolded-protein aggregates at cell poles or the center has been linked to aging, but its mechanisms remain controversial.Source: AUTHOR SUMMARY · p. 3
INTRODUCTION
This section examines how aging-related protein aggregates localize in E. coli, focusing on passive diffusion, aggregation, nucleoid crowding, and their roles in the observed spatial distribution.
INTRODUCTION
- E. coli aging is associated with functional asymmetry and loss of fitness, while protein aggregates accumulate at older poles and are linked to age-related fitness loss.Whether their localization results from active transport or passive diffusion remained unresolved amid conflicting reports of ATP-dependent stages.Source: INTRODUCTION · pp. 4, 6
Trajectory analysis of single protein aggregates
- The study tracked naturally forming IbpA-tethered aggregates in non-stressed E. coli using automated cell segmentation, focus detection, and sub-pixel trajectory tracking.Single foci of IbpA-YFP fusion proteins were monitored in live cells at 37°C in LB medium.Source: Trajectory analysis of single protein… · p. 6
Localization of protein aggregates is non-homogeneous along the cells
- Aggregates were predominantly localized at one of the two poles, with the remaining positions mainly near the cell middle; their polar locations were consistent with a three-dimensional bulk distribution rather than membrane tethering.The observed density near the pole boundary resembled bulk simulations and differed from the sharp boundary accumulation predicted for membrane-bound aggregates.Source: Localization of protein aggregates is… · pp. 7–8
Polar protein aggregates exhibit Brownian motion
- Aggregates’ short-time mean-squared displacement was approximately linear, supporting Brownian diffusion and excluding directed active transport.The estimated diffusion constants were Dx ≈ 5.1×10^-4 µm2/s and Dy ≈ 4.0×10^-4 µm2/s; aggregate mobility decreased with increasing size.Source: Polar protein aggregates exhibit… · pp. 9–10
Protein aggregates exhibit passive confinement within bacterial polar
- Fitted aggregate motion was restricted to a 650 nm long-axis region near the pole, while motion in the other dimensions approached the available cell space.The restricted region spans approximately 1/4 to 1/9 of total cell length, consistent with hindrance by nucleoids.Source: Protein aggregates exhibit passive… · pp. 11–12
- Fitted diffusion constants decreased approximately inversely with aggregate radius, consistent with the Stokes-Einstein relation.The fitted relation was D(r)=C0/r, with C0 = 47230 nm3/s.Source: Protein aggregates exhibit passive… · p. 12
reconstitutes the observed patterns
- Brownian aggregate motion, long-axis confinement, and nucleoid crowding together reproduce the observed polar and mid-cell localization pattern in simulations.The model’s predicted distribution was qualitatively similar to experiment and remained robust across detection thresholds and initial protein locations.Source: Polar protein aggregates exhibit… · p. 10, Protein aggregates exhibit passive… · p. 12, reconstitutes the observed patterns · pp. 13–14
- In non-stressed cells, 1.2% had more than one aggregate; simulations with 100 proteins likewise mostly produced one detectable aggregate.The simulations retained a higher probability of one aggregate than two even at the end of the modeled period.Source: reconstitutes the observed patterns · p. 15
- With heat shock and a low detection threshold, simulations showed at least four aggregates per cell, while higher thresholds yielded mostly two aggregates by the end.At a threshold of 1,750 aggregates, around 60% of simulations displayed two aggregates, with the rest roughly evenly split between one and three; this matched reported heat-shock observations.Source: reconstitutes the observed patterns · p. 15
- Passive Brownian motion, aggregation, and increased nucleoid crowding suffice to reproduce observed aggregate distributions, without active or directed movement.The simulations support the experimental conclusion that aggregate movement is driven by passive diffusion.Source: reconstitutes the observed patterns · pp. 15–16
DISCUSSION
Protein aggregates in E. coli move by size-dependent Brownian diffusion, while nucleoid-associated crowding confines their movement and helps explain their localization. Modeling supports diffusion, aggregation, and nucleoid hindrance as sufficient to reproduce key observed patterns.
DISCUSSION
- Aggregate movements were purely diffusive, with diffusion coefficients around 500 nm^2/s that varied with size and followed Stokes–Einstein scaling through estimated radii of 250–270 nm.The scaling is consistent with D~1/r.Source: DISCUSSION · p. 17
- Unlike some reported protein probes, aggregate size–diffusion estimates followed Stokes–Einstein scaling, which the authors interpret as evidence of limited interactions with other cellular components.The cited studies reported strong deviations for some fusion proteins and a deviation for GFP pentamers.Source: DISCUSSION · p. 17
- Aggregates were confined by the cell membrane in the section plane and, along the long axis, to nucleoid-free polar regions; modeling showed diffusion, aggregation, and nucleoid hindrance reproduced observed distributions.Simulations matched distributions in non-stressed and heat-shock conditions.Source: DISCUSSION · p. 18
- Simulations explained differences in the number of distinct aggregates per cell as resulting from differences in the total number of aggregation-prone misfolded proteins.This explanation was reported alongside simulated distributions resembling those observed in non-stressed and heat-shock conditions.Source: DISCUSSION · p. 18
- To model several cell generations and track larger aggregates along a lineage, the authors propose deriving a mean-field diffusion-coagulation model with position-dependent diffusion properties.They identify individual-based model computation time as the limitation motivating this approach.Source: DISCUSSION · p. 18
- Nucleoid-driven soft intracellular structuring may also contribute to accurate, robust positioning of divisome proteins.The authors identify this as a broader implication of diffusion-based protein localization in E. coli.Source: DISCUSSION · p. 18
- Wild-type cells accumulated visible aggregates after >150 generations, and the authors propose that polar localization may extend beyond protein aggregates to other low-diffusive assemblies.The passage cites polar localization in large RNA-protein assemblies and fluorescent-protein assemblies, among other cases.Source: DISCUSSION · p. 19
MATERIALS AND METHODS
The study combined fluorescent IbpA-YFP imaging and automated aggregate tracking with a 3D individual-based model of diffusion, aggregation, and nucleoid crowding. A fitting procedure compared simulated and experimental mean-squared-displacement curves to estimate aggregate and cell parameters.
Bacterial strains.
- MG1655 was modified to express C-terminal IbpA-YFP under the endogenous ibpA promoter, producing the MGAY strain, which was grown in half-salt LB at 37°C.The strain was based on sequenced wild-type E. coli MG1655 and used an improved YFP fused to IbpA.Source: Bacterial strains · p. 20
Fluorescence time-lapse microscopy setup.
- After recovery on an LB half-salt agarose pad, cells were imaged at 37°C by phase contrast and fluorescence at low- and high-frequency time scales.Low-frequency movies sampled images every 3 seconds for 5 minutes; high-frequency fluorescence images were sampled about every 0.60 seconds for 2 minutes.Source: Fluorescence time-lapse microscopy setup · p. 20
Image analysis and aggregate tracking.
- Cellst segmented cells and reconstructed lineages; customized software detected fluorescent aggregates, refined their coordinates to subpixel resolution, and aggtracker tracked them within cells.Spot localization was reported to reach one-tenth-pixel precision, including for very poorly contrasted spots.Source: Image analysis and aggregate tracking · pp. 20–21
Individual-based modeling of protein aggregation.
- The 3D model represented proteins as diffusing spheres, allowed overlapping proteins to aggregate, and made aggregate diffusion decrease with radius according to the Stokes–Einstein relation.Aggregation was treated as irreversible, consistent with the reported observation that focus fluorescence did not decay.Source: Individual-based modeling of protein… · pp. 22–23, §1.0 · p. 23
- The simulated cell was a reflective 3D cylinder with two nucleoid regions modeled as dense fields of immobile, impenetrable obstacles.The cylinder measured 1.0 µm in width and depth and 4.0 µm in length; simulations with spherical cell-end caps showed no significant differences.Source: Individual-based modeling of protein… · p. 22
Fitting procedure for the aggregate radius, diffusion constant and cell dimensions
- Simulated and experimental mean-squared-displacement curves were fit by least squares, using CMA-ES to optimize aggregate and cell-dimension parameters.The simulations used for fitting omitted nucleoids and aggregation; the experimental data were partitioned into five classes by initial aggregate fluorescence intensity.Source: Fitting procedure for the aggregate… · pp. 24–25
FIGURE LEGENDS
The legends describe aggregate localization and tracking, size-dependent diffusion, and simulations of aggregation under different cellular conditions. Supplementary figures extend the analyses and show aggregates forming at the old pole of aging wild-type cells.
FIGURE LEGENDS
- Aggregate positions were mapped along the cell axes, with 1,644 trajectory starts normalized to cell half-length and 9,242 pole-position points compared with synthetic distributions.Synthetic positions represented uniform sampling in the pole volume, on its membrane, or in the projected 2D semi-ellipse.Source: FIGURE LEGENDS · p. 34
- Corrected high-frequency mean squared displacements showed close-to-linear behavior during the first 10–15 seconds.The legends describe corrections for cell growth along x and the time-average y-coordinate.Source: FIGURE LEGENDS · pp. 34–35
- Diffusion constants decreased with aggregate radius, fitting D(r)=C0/r with C0 = 47.23×10^3 nm3/s.The figure also compares x- and y-direction fits and examines anomalous-diffusion exponents for the four smallest aggregate classes.Source: FIGURE LEGENDS · p. 35
- The individual-based model compared simulated and experimental aggregate positions at first detection and tracked aggregate counts under non-stressed and heat-shock conditions.Simulations varied detection thresholds; the legend specifies 100 proteins for non-stressed conditions and 7,000 for one heat-shock condition.Source: FIGURE LEGENDS · p. 35
Supplementary Figure Legends
- Supplementary simulations varied protein initialization inside or around nucleoids, while diffusion analyses also grouped trajectories by cell length.The nucleoid-initialized non-stressed simulations used 100 proteins, a detection threshold of 30, and aggregation probability pag=1; cell-length groups spanned L ≤ 3.4 µm to L ≥ 4.8 µm.Source: Supplementary Figure Legends · p. 36
- In aging wild-type cells followed for 32 hours, protein aggregates formed within the old-pole region.Cells were grown at 37°C in LB media in a microfluidic mother-machine device.Source: Supplementary Figure Legends · p. 36
TABLES
The table reports estimated aggregate radii and diffusion constants by initial median fluorescence, with values averaged across fits to MSDx and MSDy data.
TABLES
- The table compares estimated aggregate radii and diffusion constants across initial median fluorescence levels.The reported values are averages of fits to MSDx and MSDy data, with ± values shown.Source: Table 1 · p. 38, TABLES · p. 38
C Non-stressed conditions
The non-stressed-condition material labels detection probabilities for different aggregate counts and identifies cell length and x- and y-axis views.
C Non-stressed conditions
- The supplied panel labels detection probabilities for 1, 2, 3, and >4 aggregates and identifies cell length and x- and y-axis views.No plotted probability values or comparative findings are provided in the passages.Source: C Non-stressed conditions · pp. 42–45