Morales, Romero, Méndez, Segura, González and Moenne report in the International Journal of Molecular Sciences (2026, vol. 27, art. 8827) the third installment of a careful comparative project in the marine green alga Ulva compressa: how does this cosmopolitan, metal-tolerant seaweed handle a surplus of a physiologically essential element — zinc — when confronted with it at marine-relevant concentrations? The alga was collected at Cachagua, central Chile (April–November 2025), cleaned and sonicated free of epiphytes, and cultured in synthetic seawater at 14 °C with 10 µM Zn, either alone or with the H₂S-donor NaHS (50, 100 or 200 µM) or the H₂S scavenger hypotaurine (500 µM), in triplicate, for up to seven days. Tissue was EDTA-washed before digestion so that only intracellular Zn was counted.
The backdrop is what makes the result interesting. This same laboratory had previously shown, by transmission electron microscopy with EDX spectroscopy, that U. compressa challenged with copper accumulates Cu as electron-dense CuS nanoparticles in cytoplasm and chloroplast (~70 µg Cu g⁻¹ dry tissue at 10 µM), and that cadmium behaves differently again — only ~2.9 µg Cd g⁻¹ taken up at 10 µM, stored as CdO and/or fatty-acid-associated forms, with CdS appearing only when a sulfur donor is supplied. Metal sulfide formation looked like a general detoxification script. Zinc, it turns out, does not read that script.
The numbers
With 10 µM Zn alone, intracellular Zn climbed from a basal 23 µg g⁻¹ dry tissue to 34 at day 3 and 42 at day 5, holding steady to day 7 — the highest steady-state burden of the three metals tested, yet still ordinary background-tissue territory for a terrestrial plant, nowhere near the ~300 µg g⁻¹ phytotoxicity line the authors cite, let alone the 10,000 µg g⁻¹ of a Brassicaceae hyperaccumulator. Adding the sulfur donor NaHS did not build ZnS: intracellular Zn fell under 200 µM NaHS, to 16 µg g⁻¹ at day 7 (and dipped to 16 under 50 µM NaHS at day 5 before recovering). The H₂S scavenger hypotaurine pushed intracellular Zn down too, from 39 to 22 µg g⁻¹ at day 5 — a result the authors themselves say needs further study before it can be interpreted.
Crucially, the classical thiol defence system did not track any of this. Glutathione (95 µg g⁻¹ at baseline) was raised only by the highest NaHS dose — to 585 µg g⁻¹ at day 3 — while Zn was falling; phytochelatins moved inconsistently (PC4, the most abundant at baseline, 128 µg g⁻¹, actually decreased to 38 under 200 µM NaHS, though PC2 rose to 192 at day 7 under 100 µM); metallothionein transcripts were massively induced by Zn alone — MT1.4 up to 204-fold, MT1.3 to 71-fold, MT1.1 peaking at 36-fold only at day 7 — yet none of these responses paralleled the NaHS-driven loss of intracellular Zn. Where did the Zn go? Into the medium. Algae transferred to Zn-free seawater released Zn (107 → 218 nmol L⁻¹ over the first day) along with glutathione (0 → 218 µmol L⁻¹ by day 2); phytochelatins were never detected in the medium, and NaHS did not enhance the release. The authors’ reading: excess Zn is mainly extruded from the alga complexed with GSH, not precipitated as ZnS — consistent with the sulfide-solubility ladder (CuS ≪ CdS ≪ ZnS) that makes ZnS formation the least favourable of the three.
Why a phytoremediation site cares about an alga
The direct application to soil remediation is nil, and we flag that plainly: this is one marine alga, one laboratory, a low µM exposure, and tissue burdens two orders of magnitude below hyperaccumulator territory. The value is epistemic. Hyperaccumulator physiology still argues, decade after decade, about what ligands actually hold the metal in the leaf — for zinc in Noccaea caerulescens, our /species/noccaea-caerulescens-zn/ record (Zn to 40,000 µg g⁻¹), X-ray absorption work has pointed to histidine and organic-acid coordination rather than thiol ligands, while phytochelatin and metallothionein responses are repeatedly found not to scale with the stored metal. The Ulva experiment is a clean, controlled demonstration of exactly that decoupling: thiol pools, PC profiles and MT induction can all be measured, all be large, and still carry essentially no information about where the metal is or how it leaves. Anyone interpreting metal–sulfur co-localisation from electron microscopy as evidence of metal sulfide should note that Zn and Cu, in the same cell type under the same protocol, reached opposite fates that only direct speciation could have predicted. For the Zn side of our element hubs, see /metals/zinc/.
There is also an under-appreciated process pointer here. The paper reminds us that in Brassica napus and transgenic Arabidopsis, glutathione feeding increases Zn loading into the xylem and Zn accumulation in shoots. If a Zn–GSH complex is the currency of both xylem loading and algal efflux, then GSH metabolism sits at a genuine fork in the road for Zn phytoextraction/agromining: manipulating it could either drive Zn into harvestable shoot tissue or leak it back out of the system. The Ulva result is the leak side of that fork made visible.
The sober ledger
Bound this, because the paper’s own evidence demands it. The title claim is an inference, not a measurement. This paper contains AAS, HPLC and qPCR only — no TEM-EDXS, no XAS; the CuS nanoparticles it leans on for contrast were imaged in a different publication. “Probably not accumulated as ZnS” is honest hedging, and the discussion concedes Zn could equally sit with nicotianamine, citrate or ferritin, none of which were assayed. The medium data have a copy-paste smell: the medium Zn series (107 → 218 nmol L⁻¹) and the medium GSH series (0 → 107 → 218 µmol L⁻¹) reuse identical numerals in different units, and by day 3 the putative co-release uncouples — GSH collapses to 22 µmol L⁻¹ while Zn holds at 218 nmol L⁻¹ — a wobble the authors do not discuss. The time-course and hypotaurine experiments disagree quietly on the control: 42 vs 39 µg g⁻¹ at day 5, unremarked. The abstract’s “Zn mainly decreased” flattens a dose-dependent picture in which 100 µM NaHS barely moved the metal, and one discussion sentence attributes the NaHS effect to increased activity of GSH-synthesising enzymes that this study never measured. Small bibliography errors (a ref-18 year typo; a misattributed citation) suggest the manuscript moved fast. None of this overturns the core result — the decoupling of Zn from every thiol marker measured is robust within the design — but the ZnS negative should be filed as “unlikely on solubility grounds, plausible given efflux, unproven by speciation.”
Provenance: open-access full text (PMC13665299, Int. J. Mol. Sci. 27(19):8827, CC-BY; published 2 October 2026 per the PMC record) retrieved and read in full; every load-bearing number verified against the primary source by three independent reviewers (Claude, GPT, Kimi) with zero claim-level discrepancies; paper-internal limitations (control-value mismatch between experiments, duplicated release-curve numerals, day-3 efflux uncoupling, unassayed enzyme-activity claim, inference-without-speciation) reported above rather than smoothed over. Methods and limits at /methodology/.