Morales, Romero, Méndez, Segura, González and Moenne (University of Santiago of Chile) report in the International Journal of Molecular Sciences (2026, 27(19):8827) the third chapter of a careful metal-by-metal dissection of how the green macroalga Ulva compressa handles the three metals it is known to accumulate from contaminated seawater. Earlier work from the same laboratory, using transmission electron microscopy with EDXS, showed copper is stored as electron-dense CuS nanoparticles in cytoplasm and chloroplast, and cadmium as CdO and/or fatty-acid-bound deposits. Zinc, it turns out, follows neither pattern: the authors’ experiments argue it is not precipitated as ZnS, and that excess Zn is instead extruded back into the water bound to glutathione (GSH).
The mechanism evidence
Cultured thalli from Cachagua, Chile (identity confirmed by chloroplast markers at 98–100% nucleotide similarity) were grown in synthetic seawater with 10 µM Zn, with or without the H₂S donor NaHS (50–200 µM) or the sulfur acceptor hypotaurine. Intracellular Zn rose from a basal 23 µg g⁻¹ dry tissue to 42 µg g⁻¹ at day 5 with Zn alone — far above cadmium’s ~2.9 µg g⁻¹ under identical conditions in the earlier study, and below copper’s ~70 µg g⁻¹. Against expectation from the Cu and Cd results, the sulfur donor decreased intracellular Zn: 50 µM NaHS drove it to 16 µg g⁻¹ at day 5, and 200 µM to 16 µg g⁻¹ at day 7 — with one caveat the authors themselves gloss over, namely that 100 µM NaHS essentially tracked the control (40 µg g⁻¹ at day 5 versus 42), so “NaHS decreases Zn” is true mainly at the concentration extremes. None of the candidate ligands tracked the metal: GSH (~95 µg g⁻¹ baseline) rose only at 200 µM NaHS (to 390 µg g⁻¹ at day 1 and 585 µg g⁻¹ at day 3); phytochelatin profiles shifted without a coherent pattern; and metallothionein transcripts were induced strongly but inconsistently — MT1.4 peaked at 204-fold with Zn alone, MT2 only with Zn + 100 µM NaHS (58-fold at day 5), and MT3 stayed silent throughout.
The decisive experiment was an extrusion assay: after five days in 10 µM Zn, thalli were moved to Zn-free medium, and the medium was sampled. Zinc appeared in the water alongside GSH, with no phytochelatins detected, and — unlike cadmium, whose co-release with GSH and PCs is amplified by NaHS — sulfur-donor treatment did not increase Zn or GSH release at all. The authors’ explanation for the metal-specific chemistry is solubility: CuS is far more insoluble than CdS, which is far more insoluble than ZnS, so sulfide precipitation simply does not trap Zn the way it traps Cu. A neat molecular signature supports the distinct handling: at the same 10 µM exposure, Cu and Cd both switch on MT1.2 (80-fold and 600-fold respectively), while Zn’s dominant MT is MT1.4 (125-fold with NaHS; 204-fold without) — the MT isoform looks like a metal fingerprint.
What this means for phytoremediation and bioremediation
The practical lesson concerns sulfur-based priming. H₂S and thiol/sulfur donors are an active research theme as low-dose priming agents that raise metal accumulation and stress tolerance in plants, and the lab’s earlier Cd result suggested they could enhance algal metal capture too. This paper shows the leverage is metal-specific: priming that helps with cadmium actively hurts zinc retention in the same organism. For algal bioremediation of Zn-rich effluents — aquaculture discharge, galvanising runoff, acid-mine-drainage plumes reaching coastal waters — an H₂S-based amendment strategy borrowed from the Cd literature could double the problem.
The second implication is about where the metal ends up. Land hyperaccumulators succeed by keeping metal: our records for Noccaea caerulescens reach 40,000 µg g⁻¹ Zn in shoots, and Arabidopsis halleri sequesters Zn in leaf vacuoles — the /metals/zinc/ hub tracks that accumulation-first lineage. Ulva takes the opposite tack, pumping Zn back out complexed with GSH. That makes it a water-cleaning organism rather than a phytomining one: the recovery unit would have to harvest the effluent, not the biomass, and the 42 µg g⁻¹ tissue ceiling measured here is two to three orders of magnitude below the hyperaccumulation threshold. GSH-mediated extrusion is also a candidate explanation for why Ulva-based treatments can lower dissolved Zn while building modest tissue concentrations — the metal is being moved, not merely absorbed.
The genuine risks, stated plainly
This is one species, one Chilean population, lab culture at a single mild exposure (10 µM Zn), triplicate flasks, transcript-level MT data with the MT proteins never measured. The hypotaurine result — a sulfur acceptor also lowering intracellular Zn — is left unexplained. And the paper has internal problems a peer reviewer would flag. The Results text gives the medium Zn series as 107 → 218 nmol L⁻¹ (day 0→1) and the medium GSH series as 107 → 218 µmol L⁻¹ (day 1→2) — the identical number pair in different units on shifted days, which an independent read of the Figure 5 plot suggests is a copy error for Zn rather than a coincidence. The Discussion also contradicts itself once on whether cadmium release responds to NaHS, cites the wrong reference for the Cu release result, and gives a Cu paper’s year as 2014 where the volume and DOI say 2024. Finally, the “co-release” stoichiometry is loose: released GSH exceeds released Zn by roughly three orders of magnitude on a molar basis, and the two curves visibly decouple at day 3 (GSH collapses to 22 µmol L⁻¹ while Zn “remains stable”) without comment. The qualitative conclusion — Zn is exported with GSH, not stored as sulfide — survives these wobbles; the precise release numbers should be treated as approximate.
Provenance: gold OA (CC-BY) full text read directly; every load-bearing number above (23 → 42 µg g⁻¹ tissue Zn; 16 µg g⁻¹ under NaHS; GSH 95 → 585 µg g⁻¹; MT1.4 204-fold; medium Zn/GSH release values; 70 and 2.9 µg g⁻¹ Cu/Cd comparators) was verified verbatim by three independent cross-vendor reviewers (Claude, GPT and Kimi) who re-retrieved the primary source separately, with zero inter-vendor discrepancies; the text-vs-figure issue on Figure 5, the Cd-release contradiction and the citation errors are the paper’s own and are reported, not smoothed over. Methods and limits at /methodology/.