Celosia argentea is already on this site’s manganese roster for a reason: field surveys around a manganese-mining area of Guangxi, South China, recorded leaf Mn concentrations as high as 20,228 mg kg⁻¹ (Liu et al. 2014, 2018, as cited in the paper under discussion). What that number does not tell you is where inside the plant the metal sits, in what chemical form, and what it costs the tissue that stores it. A new hydroponic study by Zhang, Jiang and colleagues in BMC Plant Biology (2026, 26:852, CC BY-NC-ND) maps all three — and the map has a gradient.
A humped uptake curve, not a linear one
Seedlings were grown in modified Hoagland solution with Mn added as MnCl₂·4H₂O at 0 (CK), 50, 100, 200 and 300 mg L⁻¹ for 30 days. Tissue concentrations rose monotonically: at 150 mg L⁻¹ Mn, roots held 3,685 mg kg⁻¹, stems 5,605, young leaves 11,543 and mature leaves 14,238 mg kg⁻¹ — mature leaves out-concentrated every other tissue at every dose. Translocation factors were extreme even by hyperaccumulator standards, 6.45 at 50 mg L⁻¹ and 8.52 at 150 mg L⁻¹, reflecting how xylem-mobile Mn²⁺ is (the authors point to IRT1 and NRAMP transporters, with storage in leaf-cell vacuoles and chloroplasts, citing Castaings et al. 2016). But the bioconcentration factor peaked at 365.35 at 50 mg L⁻¹ and fell significantly to 233.82 by 150 mg L⁻¹ — the plant’s uptake-and-concentration machinery saturates, and past a moderate dose each additional unit of external Mn buys a smaller fraction of a unit inside the plant. That is the non-obvious number in this paper for anyone sizing a phytoextraction or phytomining operation: removal rate per plant is humped across the contamination gradient, and the optimum sits at moderate, not maximal, soil Mn.
The damage follows the transpiration stream
SEM and TEM work was done at CK versus 100 mg L⁻¹, and it localises the stress with unusual precision. The leaf tip — the terminus of the transpiration stream, and per the tissue data the metal’s preferred endpoint — showed deformed, compressed guard cells and mostly closed stomata, folded and thinned cell walls, plasmolysis, collapsed chloroplast endomembrane systems and ruptured mitochondria. Cells in the middle of the same leaf, at the same external concentration, kept intact walls, ordered grana and healthy mitochondria. Gas exchange tracked the anatomy: at 300 mg L⁻¹, transpiration fell from 8.43 to 4.8 mmol m⁻² s⁻¹ in the leaf middle (~43%) and from 3.63 to 1.75 at the tip (~52%), with tip stomatal conductance down 67.8%. The authors read this as an adaptive throttle — stomatal closure limits transpiration-driven Mn delivery — an interpretation consistent with the data but not demonstrated causally.
The chemistry that pays for the map
Stepwise chemical-extraction of leaf Mn showed the detox strategy directly: NaCl-extractable (protein-bound) plus water-soluble forms together made up more than 70% of leaf Mn at all doses, and as stress rose the balance shifted toward the protein-bound pool — at leaf tips that fraction exceeded 40%, higher than in the middle of the leaf. Chelation to proteins (the pattern seen in Mn hyperaccumulators such as Phytolacca and Macleaya) immobilises the ion exactly where the metal arrives hardest. Behind the chemistry sits an antioxidant system that responds non-monotonically: at 50 mg L⁻¹ the enzymes actually fell (SOD −19.28%, POD −8.76%, CAT −30.05%) while chlorophyll rose 13% to a peak of 3.24 mg kg⁻¹ — a low-dose stimulation echoing the Cd hormesis we reported in this species in July — whereas at 300 mg L⁻¹ ascorbate peroxidase surged up to +369% and glutathione reached 5.381× control, alongside MDA up 121.23%: the ascorbate–glutathione cycle working at its ceiling while lipid peroxidation still climbs.
What it means — and the honest bounds
For phytoremediation of Mn, the practical reading is twofold. Opportunity: with TF ≈ 6.5–8.5 and mature-leaf concentrations above 14,000 mg kg⁻¹, the harvestable shoot is where the metal is, so leaf removal is metal removal; the protein-bound speciation also means the Mn in the biomass is relatively immobilised, a handling plus. Risk: the humped BAF says a very hot site will underperform per plant relative to a moderately contaminated one — extraction efficiency must be planned against the actual soil concentration, not assumed to scale with it — and the tip-cell damage at 100 mg L⁻¹ plus rising MDA at 300 mg L⁻¹ mark a genuine physiological ceiling below the concentrations of some mine-waste leachates. The bounds matter as much as the mechanism: this is one accession from a Guangxi remediation base, hydroponic, 30 days, with tissue Mn, TF and BAF measured only up to 150 mg L⁻¹ even though the stress series ran to 300 — so the decline in BAF is documented across its first two points, not the full dose range, and chlorophyll, enzyme and MDA responses are physiology endpoints, not field extraction rates.
Source: Zhang, Jiang, You, Li, Lu, Chen & Jiang (2026), BMC Plant Biology 26:852, DOI 10.1186/s12870-026-08664-x (CC BY-NC-ND, published 2026-04-09). Every load-bearing number was verified verbatim against the full text by three independent reviewers via two independent retrieval paths (publisher page and Europe PMC full-text XML); no disagreements on values — only noted caveats: the 20,228 mg kg⁻¹ field record is cited by the authors to Liu et al. 2014 and 2018 jointly, and the stomatal-throttle reading is the authors’ interpretation rather than a demonstrated mechanism. See /methodology/ for how we source and check analysis pieces.