Camellia sinensis is one of the best-known aluminium accumulators in the plant kingdom: old leaves routinely reach several thousand µg Al g⁻¹ dry mass — Carr, Lombi, Küpper, McGrath and Wong measured 6,866 µg g⁻¹ in old leaves versus 380 µg g⁻¹ in young leaves in Agronomie back in 2003, an 18-fold gradient that qualified tea as a genuine, if unconventional, hyperaccumulator by any threshold reading. Fluoride rides along in the same leaf. Both elements form complexes in the tea apoplast and symplast, but the transporters that move those complexes across membranes had never been identified. Zhou, Yang, Long and colleagues at Hunan Agricultural University and the Yuelushan Laboratory now report in Plant Physiology and Biochemistry (239:111746, online September 2026) that the job belongs to two members of the aquaporin superfamily — the first time any Nodulin 26-like Intrinsic Protein has been shown to transport fluoride or an Al–F complex in any plant.
A division of labour, tissue by tissue
The two transporters are cleanly partitioned. CsNIP2 is expressed predominantly in roots and is enriched in phloem–epidermal tissues — the entry gate. CsNIP16 is expressed predominantly in stems and is abundant in xylem — the escalator. Both are plasma-membrane-localized, which is what makes them credible membrane-crossing machinery rather than intracellular sinks. The evidence chain is unusually complete for this class of claim: heterologous expression in yeast showed that both transporters facilitate Al–F uptake and significantly increase cellular sensitivity to Al and F — the standard yeast-toxicity test that separates genuine transport from correlation; antisense-oligonucleotide and VIGS knockdown in tea shoots impaired vascular translocation and leaf accumulation of the complexes; and overexpressing CsNIP2 in Arabidopsis promoted fluoride uptake and root-to-shoot transport, showing the transport function travels with the gene across species.
The throttle: a repressor and an accelerator on the same promoter
The part that should interest hyperaccumulation engineers most is the regulation. Al/F stress induces two transcription factors that act on the CsNIP2 promoter in opposite directions. CsMYB152, a root-specific repressor, inhibits CsNIP2 promoter activity — a brake whose stated job is to prevent excessive Al–F influx at the root. CsNST1, a cell-wall regulator, activates CsNIP2, promoting root-to-shoot translocation and “coupled detoxification”. In other words, tea does not just express a transporter; it expresses a transporter with a tunable promoter module — one regulator limiting uptake at the source, another pushing export once the metal is inside. That is exactly the architecture you would want to splice into a remediation or agromining plant if you could: uptake capacity and a governor, in the same cassette.
Why this matters beyond tea
Hyperaccumulators do not move raw metal ions; they move metal–ligand complexes through xylem — zinc and cadmium paired with organic acids in Noccaea caerulescens, arsenic shuttled as arsenite in Pteris vittata. The transporters behind those paired loads have been hunted for decades. The tea result adds a conceptually new class of carrier to the toolkit: aquaporin-family proteins that accept a charged metal–halide complex as substrate, rather than a bare ion or a neutral chelate. And it lands on a plant whose accumulation phenotype is extreme, well documented, and economically interesting — which makes the system practical to study. For phytoremediation, the immediate application is inverse: the authors themselves frame the breeding goal as tea varieties with reduced leaf Al and F, and the same knockdown toolkit they used for validation (asODN, VIGS) is a prototype for editing the module into low-accumulation crops.
The genuine risks and open edges
Three cautions belong in the same breath as the excitement. First, this is one layer of a known network, not the whole story. Tea fluoride transport was not a blank page before this paper — ALMT-type anion transporters (CsALMT6), FEX-family fluoride exporters and CLC-type F⁻/Cl⁻ transporters had already been characterised in tea. The NIPs add membrane handling of the Al–F complex; they do not replace the rest of the trafficking system, and the abstract does not quantify how much of total leaf Al or F load moves through them. Second, direction of travel matters. In a food crop, leaf F is the hazard, not the product: pushing CsNIP2 harder by editing out the MYB152 brake would raise uptake exactly as designed — and raise dietary fluoride with it. Any transfer of this module to other species inherits that trade-off; the transporter does not know whether its cargo is waste or ingredient. Third, the bounds: the article is closed access and its abstract reports no tissue concentrations, no flux values and no fold-changes — everything above about magnitude of transport is structurally supported but numerically unquantified here, and all of it is one species, one cultivar system. Whether the CsNIP2/CsNIP16 architecture exists in any of the metal hyperaccumulators in our database is an open, and very testable, question.
Source: Zhou, Yang, Long, Tan, Sun, Hu, Manan, Tan, Li, Zuo, Liu & Zhao (2026), Plant Physiology and Biochemistry 239:111746, DOI 10.1016/j.plaphy.2026.111746 (online 2026-09-21; October 2026 issue). The article is closed access; this piece is bounded to the peer-reviewed abstract and independently verifiable literature, and every mechanistic claim above is checkable against it. All seven load-bearing claims of the abstract — the two NIPs and their tissue partitioning; plasma-membrane localization; the yeast Al/F sensitivity result; the asODN/VIGS knockdown phenotype; the Arabidopsis overexpression result; and the opposing CsMYB152/CsNST1 regulation of CsNIP2 — were verified verbatim by three independent reviewers retrieving the Europe PMC, PubMed, Crossref, OpenAlex and Semantic Scholar records separately; the abstract contains no numeric quantities, so no numeric discrepancies were possible. The old-leaf Al value (6,866 µg g⁻¹) is from Carr et al. (2003), Agronomie 23:705–710, DOI 10.1051/agro:2003045, confirmed against the publisher record. Reviewer caveats reflected above: prior tea fluoride transporters (ALMT, FEX, CLC families) mean the NIP result is one layer of a network, and the novelty claim is precisely “first NIP mediators of Al–F complex transport”, not first fluoride transporters in tea. See /methodology/ for how we source and check analysis pieces.