nucleation

LIT-853: A heritable iron memory enables decision-making in Escherichia coli

3 min read

Properties4
StatusActive — read 2026-10-09 (NOTE-657); worth reading as an experimental case of a non-genetic, multigenerational state in single bacteria: a cell's swarming potential is shared by its descendants for about four generations and gone by seven, and it tracks the cell's iron status as reported by a Fur-repressed promoter (Spearman r = 0.71 between reporter signal and swarm diameter over 96 sorted cells; 0.15 and 0.05 in the controls). The state is not a bistable switch: an ODE of three swarming classes fits the lineage data only when each founding cell holds its class for a fixed delay (about three or four generations) before switching at constant rates. The "memory", "decision-making" and "conditioning" of the title and text are figurative; the iron link rests on a promoter reporter and on correlation, and the model fixes five parameters on two time points.
Tagsnatural-sciences, complex-systems
Cited as “paper” byNOTE-657 — Bhattacharyya et al. 2023, heritable iron memory in E. coli
Cited as “source” byTHEORY-1209 — In genetically identical E. coli, a cell's swarming potential is a transient inherited state: its descendants share it for about four generations and lose it by about seven, and it tracks the cell's iron status rather than a bistable switch, THEORY-179 — In Bhattacharyya et al.'s single-cell swarm assays of E. coli MG1655, the daughters of one mother swarmed alike at the fourth generation sampled and varied again at the seventh; swarm size correlated with a plasmid fepA-promoter reporter of Fur derepression by low iron (Spearman r = 0.71), and an iron chelator or added iron shifted swarming and held the shared state longer; a three-class switching model fitted to two time points matched the lineages only with a delay of three to four generations before switching, which the authors read as ruling out a bistable switch

Graph View

Souvik Bhattacharyya, Nabin Bhattarai, Dylan M. Pfannenstiel, Brady Wilkins, Abhyudai Singh and Rasika M. Harshey (2023), Proceedings of the National Academy of Sciences 120(48):e2309082120 — DOI-10.1073/pnas.2309082120; preprint bioRxiv 10.1101/2023.05.19.541523 (“Iron Memory in E. coli”)

Key takeaways

  • Swarming potential is heterogeneous among genetically identical planktonic cells, and inherited for a few generations. Swarms started from single cells vary far more (coefficient of variation) than swarms from 100 or 10,000 cells. Sixteen daughters of one mother, grown four generations and plated singly, swarm alike; by seven generations the spread among siblings is back. Cells taken from a swarm edge are uniform good swarmers and keep that for four generations too, while cells from hard agar, where E. coli does not swarm, are as variable as planktonic ones.
  • The variable is iron status. Of environmental perturbations (medium, salt, pH, aeration, agar) and plasmid-expressed swarming genes, only fepA and fur (iron uptake and its repressor) cut the variability. Iron chelation (DFO) shifts cells to better swarming and holds that to the seventh generation; added iron or fepA overexpression shifts them to worse swarming and holds that to the twelfth.
  • Measured through Fur. A plasmid sfGFP reporter driven by the Fur-repressed fepA promoter (bright = low iron) correlates with the sorted mother’s swarm diameter (Spearman r = 0.7067, p = 8.66 × 10⁻¹⁶, n ≈ 96), not in a constitutive-promoter control (r = 0.15) or in ΔfepA (r = 0.05). Daughters of five sorted mothers per strain follow the mother’s reporter level at four generations, not at seven. The same reporter correlates with biofilm (more iron, more biofilm) and with survival at about half-MIC kanamycin and chloramphenicol (less iron, better survival), each at one generation only.
  • Not memoryless, and not bistable. Three swarming classes (XS < 35 mm, M 35–65 mm, L > 65 mm; 52/35/13% at steady state) switch along XS ⇄ M ⇄ L at constant rates, but each founding cell holds its class for a fixed delay first (τ ≈ 3 generations for XS and M, ≈ 4 for L). The authors report that constant-rate switching alone does not fit, and themselves rule out bistable and toggle switches and epigenetic marks because the memory is too short.

Standing in the record

Filed on 2026-10-09 at the owner’s request, with no stated context, and read on its own merits the same day (NOTE-657). The reading is the source of THEORY-179, which states the finding with the scope the evidence gives it. It sits beside the record’s other reading of history-dependent state in E. coli, Irreversibility in bacterial regulatory networks, as a contrast: there a self-maintaining attractor of the regulatory network, here a transient state that decays on a timescale of generations.

Backlinks