Riboflavin in phloem sap helps wild tomato combat whiteflies

Abstract

The whitefly Bemisia tabaci feeds on vegetable and ornamental crops including chilli, okra, potato, tobacco and tomato, causing global annual losses of up to billions of US dollars (Sani et al., 2020). As phloem-feeding insects, B. tabaci feed on host plants by inserting their mouthpart, the stylet, directly into the phloem. During that process, they inject RNA and protein that suppress the plant immune system. Additionally, they can transfer viruses such as the tomato yellow leaf curl virus (Jones, 2003). Whitefly is a rapidly adapting pest that has evolved resistance to even the most potent chemical treatments (Barman et al., 2022). Sustainable whitefly management requires integrating multiple defence layers—including genetic resistance—into comprehensive integrated pest management strategies. Some wild relatives of cultivated tomato are resistant to whitefly damage. Identifying the mechanisms and genes underlying whitefly resistance in wild tomato can equip the vegetable breeding industry with genetic tools to introgress this trait into cultivated varieties.

Most of these resistance mechanisms in wild tomato are based on the production of specialised metabolites in glandular trichomes that are bioactive against B. tabaci (Kortbeek et al., 2021). Lissy-Anne Denkers, first author of the highlighted publication, did a research internship in Petra Bleeker's lab at the University of Amsterdam, in which she worked on this resistance mechanism. This sparked her interest in tomato and the role of specialised metabolites in biotic interactions. She was fascinated by the tomato–whitefly interaction as a research system because it presents complex yet tangible questions that can be approached from multiple disciplines, including plant physiology, biochemistry, insect physiology and behaviour, ecotoxicology and broader ecology. Although the different tomato species with their large variation in appearance, chemistry and resistances were what captured her initial interest, she could not help but develop some secret appreciation for whiteflies, especially the clumsily walking first instar nymphs.

For her PhD project, Denkers analysed a defence mechanism that was independent of trichome repellents. The wild tomato accession Solanum chmielewskii was found to be susceptible to adult whiteflies; however, the whiteflies deposited fewer eggs on its leaves, and the development of juvenile whitefly stages was hampered (de Almeida et al., 2023). The team hypothesised that this resistance in S. chmielewskii might be due to something in the phloem that specifically targets young stages of whiteflies, the nymphs, which rely on constant feeding from the phloem (Denkers et al., 2025).

Whiteflies deposit their eggs on host plant leaves, then the first instar nymph hatches and searches for a feeding site towards the leaf vasculature. There, it probes the tissue to access phloem sap before progressing through three additional nymphal instar stages (Figure 1a). The authors quantified the nymph developmental stages on four S. chmielewskii accessions and on a susceptible tomato cultivar after infestation. Two accessions, LA1028 and LA1840, had fewer eggs deposited and a lower number of nymphs (Figure 1b). To assess whether phloem sap was the causal factor, the authors used LA1840 in grafting experiments. When LA1840 served as a rootstock for the susceptible tomato cultivar shoot, both egg deposition and nymph abundance were reduced, suggesting that the resistance was mediated by a vasculature-mobile factor originating from LA1840.

To identify the metabolite responsible for the resistance, the authors used inbred lines from crosses of LA1840 with a susceptible parent and selected one susceptible and one resistant inbred line for further analysis. To analyse non-volatile compounds in the leaf material, they used ultra-high-performance liquid chromatography-quadrupole time of flight (UPLC-qTOF). UPLC-qTOF is a liquid chromatography mass spec platform (LC-MS) that can analyse a complex mix of unknown compounds. First, the LC separates the different compounds in the sample, based on physical properties like size, polarity and affinity to the column. The MS then fragments molecules by bombardment with electrons. The resulting fragments are accelerated in the qTOF, and for each ion, the mass-over charge can be determined. This provides a particularly accurate mass determination that helps the identification of the molecule. In total, this approach identified 792 features.

The UPLC-qTOF analysis generated numerous features of unknown identity, with a pronounced imbalance between the number of samples and the number of features per sample. To address this, the authors applied a random forest (RF) machine learning approach to identify metabolites associated with resistance. RF is well suited for sparse, high-dimensional datasets with low sample-to-feature ratios, common in multiomics datasets (Kortbeek et al., 2021). This analysis identified 41 metabolic features as significant discriminators between susceptible and resistant lines. The top candidate was 21-fold more abundant in LA1840 than in the susceptible parental line and was identified as riboflavin.

When riboflavin was supplied to susceptible plants via a hydroponic growth solution, whitefly nymphal development on the leaves was impaired, reproducing the resistance phenotype observed in LA1840. Conversely, treatment of LA1840 plants with a riboflavin synthase inhibitor increased susceptibility, confirming riboflavin as the phloem sap metabolite responsible for resistance against B. tabaci.

The authors propose two possibilities for why riboflavin is a deterrent. First, it might directly inhibit nymphal development when ingested above a threshold concentration, as was demonstrated for aphids (Nakabachi & Ishikawa, 1999). Alternatively, it might act indirectly by triggering other plant defence mechanisms, such as the induction of pathogenesis-related gene expression, as reported in Arabidopsis and tobacco (Dong & Beer, 2000).