Comparative metabolomics of Tilia platyphyllos Scop. bracts during phenological development Author Szűcs, Zsolt * & University of Debrecen, Department of Botany, Division of Pharmacognosy, H- 4010 Debrecen, Egyetem tér 1, Hungary Author Cziáky, Zoltán Author Kiss-Szikszai, Attila Author Sinka, László Author Vasas, Gábor Author Gonda, Sándor text Phytochemistry 2019 2019-11-30 167 1 11 http://dx.doi.org/10.1016/j.phytochem.2019.112084 journal article 10.1016/j.phytochem.2019.112084 1873-3700 10483578 2.1. Identified specialized metabolites of Tilia platyphyllos bracts Bracts of various phenological stages were shown to contain various phenolic compounds, based on the identification of LC-ESI-MS spectra ( Table 1 ). Several flavonoids, including kaempferol, quercetin and luteolin glycosides, as well as catechin derivatives, procyanidins, coumarin derivatives and quinic acid derivatives were identified according to reference spectra (see section 4.6.), along with some non-phenolic compounds. As specialized metabolites were only tentatively identified, description of novel compounds was not attempted. The compounds found were previously found in inflorescences of Tilia sp. consisting of the bract and the flowers. Procyanidin oligomers, catechin and epicatechin were found in Tiliae flos ( Karioti et al., 2014 ), and flavonoid hexosides, deoxyhexosides were described from inflorescences of T. americana ( Aguirre-Hernández et al., 2010 ; Pérez-Ortega et al., 2008 ). Specifically regarding T. platyphyllos , kaempferol and quercetin monoglycosides, diglycosides as well as catechin, epicatechin and procyanidin B2 have been described ( Jabeur et al., 2017 ). However, it is unclear how much of these compounds come from the botanical flowers, as only Toker et al. (2001) specifically examined the bract, providing a comparison to flowers and leaves. They found that the bract has a similar composition to that of the leaves ( Toker et al., 2001 ). 2.2. Seasonal variation of the Tilia platyphyllos bract metabolome The overall changes in the T. platyphyllos bract metabolome change are plotted as a PCA score plot ( Fig. S1 ). PC1 and PC4 were chosen as they were most affected by time ( p = 8.12E- 24, p = 3.95E- 4, ANOVA ). It is obvious that severe changes occur in the metabolome between days 0–32, before flowering, especially during the major growth phase of the bract (days 0–21) ( Fig. 1 , Figs. S1–4 ). In the young organ (days 0–21), the metabolome is extremely different from the metabolome during the flowering and later developmental stages. Later on, a relative stability can be observed during flowering, followed by a slow characteristic change during fruit growth, and rapid changes from the onset of senescence (days 72–112). As Fig. S1 only covers a small proportion of variance of the whole change (9.68% and 5.85%, respectively), it is better to examine the change kinetics of the major groups, as detailed in 2.4. The significance of the effects of time and tree number was also studied using ANOVA models for each compound separately ( Table S1 ). After Bonferroni correction (n = 504), 241 features turned out to be significantly ( p < 9.92E- 5) affected by time (47.82% of features). Of these, 202 (40.07%) were highly significant ( p < 1.98E- 6). The difference among trees was only significant for 6.34% of all metabolites studied (n = 32), while the interaction between the two experimental factors was only significant in the case of a single feature (0.19%). A set of 15 metabolites (2.97%) showed significant differences ( p < 9.92E- 5) between trees and sampling times as well. Studies dealing with the seasonal variability of specialized metabolites in plant organs usually have much lower resolution than the current study ( Dai et al., 2015 ; Vagiri et al., 2015 ; Valares Masa et al., 2016 ; Yang et al., 2015 ; Zhang et al., 2016 ), which is not enough to adequately detect several trends. For example, in the case of Salminen et al. (2004) and Tuominen and Salminen (2017) , phenomena would have been missed with a typical once-a-month sampling. Using the untargeted metabolomics approach complemented quantification of a set of compounds after calibration ( Cosmulescu et al., 2014 ), giving much more insight ( Dai et al., 2015 ). The metabolome of young tissues was also shown to be substantially different from that of the older organs in leaves of Quercus sp. ( Salminen et al., 2004 ), in Geranium sylvaticum ( Tuominen and Salminen, 2017 ) and in Cistus ladanifer ( Valares Masa et al., 2016 ) .