Crop biotech
Biofortification of crops
Why phytate, provitamin A conversion and the food matrix decide whether a biofortified crop changes anyone's nutritional status — and why Golden Rice stalled for reasons that were never biochemical.
Biofortification means breeding or engineering a staple crop to carry more of a micronutrient, on the argument that people who are deficient already eat the staple every day. The logic is sound and the delivery problem is genuinely smaller than for supplements or fortified processed foods. But the intervention is nutritional, not agronomic, and it succeeds or fails on a distinction that concentration figures hide: how much of the nutrient the eater actually absorbs.
Concentration is not bioavailability
Cereal and legume seeds store phosphorus as phytate — myo-inositol hexakisphosphate — concentrated in the aleurone layer and the germ. Phytate is a strong chelator of divalent cations, and iron and zinc bound to it in the gut are largely unabsorbed. This is why the phytate-to-zinc molar ratio, rather than zinc content, is the usual predictor of zinc status from a diet. It also sets up an unpleasant geometry: the mineral and its inhibitor sit in the same tissue, so milling that removes phytate removes the mineral too.
Low-phytate mutants exist and do raise absorption, but phytate is the seed’s phosphorus reserve, and the mutants commonly show poor germination, reduced seed vigour and lower field emergence. The alternative approach works around it rather than through it — overexpressing ferritin to add an iron sink, and raising nicotianamine to improve loading and, in principle, uptake — and combining that with the ordinary dietary levers, ascorbate promoting non-haem iron absorption and polyphenols inhibiting it.
Provitamin A has its own arithmetic. β-carotene is not vitamin A; it must be cleaved and reduced to retinol, and the efficiency of that conversion is not fixed. It depends on the food matrix, on how much fat accompanies the meal, and on the individual’s vitamin A status. A figure for micrograms of β-carotene per gram of grain therefore does not translate to retinol activity without knowing what the meal looks like. Carotenoids also degrade during storage and cooking, so the number that matters is the one at the point of eating.
What actually worked, and what did not
The clearest evidence comes from conventional breeding rather than transgenics: orange-fleshed sweet potato, selected for high β-carotene, has been tested in controlled introduction trials in sub-Saharan Africa and improved vitamin A status in children — helped by the fact that the trait is visible, the crop is already familiar, and the change is a variety swap. Conventionally bred high-zinc wheat and high-iron pearl millet followed a similar route.
Golden Rice is the instructive contrast. Its biochemistry has been settled for years: a phytoene synthase and a bacterial carotene desaturase reconstitute β-carotene accumulation in endosperm, and the second-generation event raised it well above the original. What blocked it was regulatory process, litigation and the ordinary agronomic requirement that the trait be introgressed into locally adapted, locally acceptable varieties — a Philippine approval for commercial propagation in 2021 was followed by a court ruling in 2024 that vacated the biosafety permits. Nothing in that sequence is a question about carotenoids.