Literature

Australia's strongest selenium hyperaccumulator stores its selenium as selenocystathionine — a metabolic pathway the standard hyperaccumulator model does not predict

Harvey, van der Ent, Kirby, Virtue, Kidman, Lovett, Erskine and Harris (Wageningen University, the Universities of Queensland and Adelaide, CSIRO and the University of Saskatchewan) report in Plant Physiology and Biochemistry (2026, 232:111091, open access, CC BY) what is in effect a correction to their own 2020 speciation model of Neptunia amplexicaulis, the “selenium weed” — an herbaceous legume endemic to the seleniferous Richmond district of north-west Queensland and one of the strongest Se hyperaccumulators known. Combining synchrotron X-ray fluorescence microscopy, spatial and bulk X-ray absorption spectroscopy, and LC-tandem MS on live, hydrated plants dosed with selenate or selenite, the team asked a deceptively simple question: what chemical species is the selenium actually stored as?

The answer is not the textbook one

The canonical model of Se hyperaccumulation, built on North American Astragalus spp. and Stanleya pinnata (both in our database), runs Se through selenocysteine methyltransferase (SMT) into methylselenocysteine (MeSeCys) — a non-protein seleno-amino acid that detoxifies Se by keeping it out of proteins. N. amplexicaulis does not do this. The dominant water-extractable organic Se species the authors identify is selenocystathionine (SeCT), detected at m/z ≈ 271.019 ([C₇H₁₄N₂O₄Se + H]⁺, ⁸⁰Se) with γ-glutamyl-SeCT and fragment ions at m/z ≈ 400.069, 181.971, 166.959 and 135.966. SeCT is the Se-analogue of cystathionine, made by cystathionine-γ-synthase from selenocysteine and O-phosphohomoserine, and it sits at the rate-limiting step toward selenomethionine and volatilised dimethyl selenide (DMSe) — meaning this species diverts Se into storage before the volatilization branchpoint, rather than into the MeSeCys detour.

That matters because the group’s own 2020 Metallomics paper had fitted this plant’s Se as MeSeCys + SeMet. The authors now say that fit was an artefact of silica-dried sample material and the simple unavailability of a SeCT analytical standard — a cautionary tale for the whole speciation literature: identifications are only as good as the standards and preparation methods behind them.

A quarter of the inventory is elemental — and invisible to chemistry

The spatial XANES fits are the number that should stay with anyone planning to phytoextract or agromine selenium. Of total plant Se, organic selenoether forms dominate (~65% in selenate-dosed, ~68% in selenite-dosed plants), but elemental Se⁰ makes up 25.0 ± 1.0% and 28.7 ± 0.9% respectively, with residual selenate 9.4% vs 3.1%. The elemental fraction concentrates in and on the roots — which run visibly red — and the authors attribute it at least partly to bacterial reduction, while noting XANES-detected Se⁰ could also be metabolised from selenocysteine. Two implications follow directly:

  • Mass balances lie. Elemental Se⁰ would not register in standard LC speciation or water extraction — in this study water extracts recovered only ~50% of total plant Se (563 of 1,037 µg g⁻¹). Any field trial that quantifies “recovered selenium” by extractable species alone will undercount the plant’s true uptake by as much as the Se⁰ fraction.
  • The biggest stock is below ground. Peak tissue concentrations were in roots — fine roots reaching 16,246 µg Se g⁻¹ in the 60 mg kg⁻¹ selenate soil treatment (the text says “up to 16,300”; Table 1 prints 16,246), against 4,726 in young leaves — and old leaves carried the lowest Se of any tissue. Harvest-and-remove economics for this species have to price in root recovery, not just leaf cutting. The famous 13,600 µg g⁻¹ young-leaf record for the species remains the 2020 hydroponic result, not this paper’s.

A second spatial detail: in selenate-dosed plants, mature-leaf tips held ~29.1% selenate versus ~14.3% at the leaf base — a twofold tip-to-base gradient the authors link to volatilization at the leaf margin. Volatilization-driven Se removal (export as DMSe gas) is the classic argument for in-place phytoremediation of seleniferous drainage; here it appears active but operating on a plant that has already committed most of its Se to non-volatile storage.

The sister species runs the “wrong” pathway too — in reverse

The comparative non-accumulator Neptunia heliophila — sympatric, but with ~5× lower tissue Se (208 vs 1,037 µg g⁻¹) and toxicity symptoms by ~50 µM — produced the mirror-image surprise: LC-MS/MS evidence of MeSeCys, the hyperaccumulator detox compound, in a non-accumulator. The authors are appropriately careful: the spectra were noisy, the putative MeSeCys primary ion (m/z 183.9873) did not show the correct Se-isotope pattern, and the identification rests on the ammonia-loss fragment at m/z 166.9586. If it holds, the SMT pathway is not the hyperaccumulation switch per se — N. heliophila has the metabolite without the tolerance, and N. amplexicaulis has extreme tolerance without the metabolite. That undermines a tidy one-enzyme story and pushes the interesting biology toward flux control, sequestration location, and the elemental-Se branch instead.

What it means for the three use cases

  • Biofortification / Se products: the value of a Se crop is the identity of the molecule. SeCT is a different, water-extractable product with predictable Se content — not the selenomethionine of standard supplements or the MeSeCys of the cancer literature. Which compound a hyperaccumulator actually makes has to precede any commercial claim. See our selenium hub.
  • Phytoremediation: a quarter-to-a-third elemental Se⁰ inventory is a stabilizing end product — Se⁰ is far less mobile and bioavailable than selenate — which suits containment of seleniferous soils and drainage, but it complicates claims of “removal” and any downstream recovery chemistry that assumes organic species.
  • Conservation: the species is a narrow endemic of one Queensland district; the same seleniferous grasslands that make it a scientific oddity make its habitat genuinely rare, and our database lists it as Not Evaluated by the IUCN.

The honest risks

First, the tolerance question: the paper cites Freeman et al. (2010) as suggesting mixed SeCT/MeSeCys accumulation is associated with lower Se tolerance than pure MeSeCys accumulation — an apparent tension with N. amplexicaulis’s extreme SeCT-dominated accumulation that the authors flag but do not resolve. The SeCT-storage strategy should not be assumed benign in other genetic backgrounds. Second, the caveats are substantial and stated: no certified SeCT standard existed, the XANES “organic” fraction is a SeMet-model proxy (SeMet and MeSeCys are spectrally indistinguishable at this resolution), water extraction recovered only ~50% of Se, the mass spectra were noisy, and Table 1 rests on one to two specimens per treatment in spiked glasshouse soils and hydroponics. This is a speciation study — one species complex, one population’s seed line, controlled conditions — not a field trial. The correction of the 2020 result is exactly how science should work; treat today’s identifications with the same standing and the same scepticism.


Source: open-access full text (CC BY) read in full via the Wageningen eDepot accepted-manuscript mirror; every load-bearing number verified verbatim by three independent reviewers plus an independent re-retrieval of the primary source (eDepot, WUR OAI-PMH, Crossref, OpenAlex, Europe PMC); retrieval and cross-check notes per /methodology/.

Primary source: https://doi.org/10.1016/j.plaphy.2026.111091

← All news & analysis