Tanacetum cinerariifolium, Schultz Bip., Schultz Bip.

Li, Jinjin, Hu, Hao, Chen, Yu, Xie, Jing, Li, Jiawen, Zeng, Tuo, Wang, Manqun, Luo, Jing, Zheng, Riru, Jongsma, Maarten A. & Wang, Caiyun, 2021, Tissue specificity of (E) - β-farnesene and germacrene D accumulation in pyrethrum flowers, Phytochemistry (112768) 187, pp. 1-11 : 2-5

2.1. Morphological and anatomical analysis of T. cinerariifolium flowers

The T. cinerariifolium flower head is supported by a long peduncle rising from the base of the plant, and composed of an outer ring of white ray florets and a yellow flower heart densely populated by disk florets in the center of the receptacle ( Fig. 1A ). Both disk and ray florets have ribbed achenes, which are seated on the slightly convex receptacle ( Fig. 1A ). Each disk floret has a yellow tubular corolla opening at the top into five lobes ( Fig. 1B and C ). SEM observation of the longitudinal sections of involucre bracts located at the outer side of the receptacle showed that the joining site constituted mostly of parenchyma cells and vascular bundles ( Fig. 1D and E ). During flower development, the peduncle below the flower head coarsened with the disintegration of the parenchyma cells in the pith to support the heavier flower head ( Fig. S1 ). First signs of disintegration of parenchyma cells in the pith were visible in the upper flower peduncle of S2 flowers ( Fig. 1F and G ).

2.2. Localization of terpenes in young flowers

In our previous work ( Li et al., 2019), we reported that flowering T. cinerariifolium plants emitted a blend of five terpenes and one green leaf volatile in changing ratios as the flowers matured. The ratios in this blend appeared to control ladybird beetle attraction and aphid repellence. To further investigate the location and production of these terpenes, we performed hexane extraction and GC-MS analysis on selected early stage flower tissues (S0–S2) showing a significant transition in terpene emission composition and quantity. Focusing on terpenes with the terpene-specific mass 93 shows that fewer compounds were in peduncle extracts compared to the flower head ( Table 1 , Fig. 2A ). The principal volatile terpene constituents of the flower head were a set of isomeric sesquiterpenes with very similar mass spectra ( Table 1 ). (E)-β- farnesene (EβF) and germacrene D (GD) were the predominant molecules, accounting for 68% of total terpenes (34% EβF and 34% GD) in the S0 flower head and 87% of total terpenes (80% EβF and 7.6% GD) in the upper flower peduncle ( Table 1 ). To our surprise, these two dominant terpene volatiles EβF and GD were not present in common storage site trichomes, so this raised the question how these terpenes were then distributed inside the flower tissues. In fact, we found that approximately 88% of the total amount of GD was concentrated inside the S2 flower head, while in contrast 71% of EβF was stored in the flower peduncle (4 cm upper and 4 cm lower peduncle together) with less trichomes ( Fig. 2B ). Table 1 shows how during flower development, the amount of EβF and GD in the flower head decreased relative to the strongly accumulating of pyrethrins, but maintained a constant level in the flower peduncle. Also in terms of specific weight in the flower head, EβF (257 ng per mg fresh tissue) and GD (251 ng per mg fresh tissue) similarly decreased by a factor two during S0 – S2 flower development ( Fig. 2B ).

2.3. Detailed EβF and GD emission and distribution patterns during flower development

The early stage flowers contained similar levels of EβF and GD in the flower head ( Fig. 2B ), but in the peduncle EβF levels were ~10-fold higher than GD. From our earlier work ( Li et al., 2019), we were interested to understand how the change in EβF/GD emission ratio was achieved during flower development, with GD becoming the dominant volatile in the blend ( Fig. 3A ). In a first step, we analyzed the kinetics of EβF and GD emissions in the S0-2 flowering plants, and found that EβF emission levels peaked at 10 a.m., whereas GD emissions peaked 4 h earlier at 6 a.m. and were significantly higher than EβF at night ( Fig. 3B ). The differences in emissions we presumed to be partly due to differences in storage tissues and gene expression patterns.

To analyze the temporal distribution of these two sesquiterpenes in the flower and peduncle in detail, we further analyzed hexane extracts of flower heads and flower peduncles from a wider range of developmental stages (S1–S5). Whole plant EβF and GD emissions both increased during flower development ( Fig. 3A ), but the content per fresh weight of EβF and GD stored in the flower head decreased and remained constant in the flower peduncle ( Fig. 3C–D ). On a per flower basis, the content of EβF and GD in the flower head peaked at stage 2 ( Table S1 ). The majority of this stored amount is never emitted as only a small fraction (around 1%) is measured to be released into the environment during a flower lifetime per day ( Fig. 3A ). 100–200 ng /h EβF or GD was released from S0-2 whole flowering plant (15–20 flowers) in a 24 h collection period, which means about 120–310 ng EβF or GD was released from a single flower per day. Yet, we found nearly 18 μg EβF and 28 μg GD to be stored in a single S2 flower head ( Table S1 ) ( Li et al., 2019).

To reveal the spatial-distribution of stored EβF and GD in more detail, we next dissected the flower head into four parts (upper disk floret; lower disk floret (ovary); bract; receptacle; flower stages S1–S5), extracted the samples in hexane, and analyzed by GC-MS. EβF and GD concentrations stored in the disk flower (upper and lower parts) both decreased during flower development, but the amount of GD in the disk flower and bract was 2–3 times higher than EβF ( Fig. 4 ). We interpret the decrease in GD and EβF to result from volatile release when the corolla of the disk floret opens although the increase in flower weight will also play a role in the decrease. The much higher GD content of the upper disk floret (GD, 1181 ng /mg fresh tissue; EβF, 478 ng /mg fresh tissue) most likely caused the higher ratio of GD/EβF during flower development ( Fig. 4 ).

2.4. Terpene volatiles in the flower stigma and corolla

As shown above, with the opening of the disk flower, EβF and GD stored in the disk floret were released. The higher GD emission is correlated with a higher stored content in the upper disk floret. We performed a NADI staining which shows that in closed S1 and S2 disk flowers purple stained terpene oil (EβF and/or GD based on content analysis) was present in the young two-lobed stigma of the pistil in the form of two channels/cavities/secretory ducts ( Fig. 5A–C and S 3 ). Following the flower bud development, more blue-violet stained oils (mainly pyrethrins, nearly 66%–80% of the stained terpenes) began to emerge in the ovary wall of S1 disk floret, and this proceeded further in the S3 flower (Table S2, Fig. 5A and D ). In open S3 and S4 flowers, we found NADI-stained terpenes (mainly EβF and GD, 70–80% of the stained terpene oil droplets) were also pushed upward in the two lobes of the stigma and five lips of the corolla in the opened disk floret with the elongation of the style (Table S2, Fig. 5B–E and S 3 ). We dissected the stigma and corolla of the opened disk florets of S4 and S5 flowers and then did hexane extraction for GC-MS analysis. We found stigma extracts to contain ten times higher concentrations of EβF and GD (EβF, 169 ± 73 ng /mg fresh weight; GD, 257 ± 126 ng /mg) compared to the corolla (EβF, 17.8 ± 4.9 ng /mg; GD, 38.6 ± 10.9 ng /mg) ( Fig. 6 ). We, therefore, propose a scenario that when the corolla opens up, the style and stigma elongates and terpenes are much promoted upwards the terminal ends of the two separating stigma lobes resulting in a release into the air, presumably with the purpose of promoting pollination.

2.5. EβF stored in the flower peduncle

EβF was much more dominant and nearly pure in the flower peduncle extract ( Fig. 2A ) and shown earlier to be mostly produced in cortex cells neighboring vascular bundles ( Li et al., 2019). Here we looked into the question of the fate of the oil droplets produced early on during later development. Cross sections of young upper flower peduncles showed that parenchymatous cortex with chloroplasts consisting of 6–7 layers and that lamellar collenchyma is discontinuous under the epidermis ( Fig. S1 ). The peduncle has 13–15 open collateral type vascular bundles surrounded by schlerenchyma fibres. There were secretory cavities next to the schlerenchyma fibres that were especially abundant in the upper peduncle ( Fig. S2 ). These secretory cavities developed parallel to the vascular system and may consist of elongated cavities extending into the cortex cells (cortical canals).

In T. cinerariifolium, NADI staining of cross sections of younger flower peduncles (S0–S1 stages) showed up positive for purple-violet terpene oil droplets that largely accumulated inside cortex cells surrounding the vascular bundles ( Fig. 7A ) sometimes also filling parenchyma cells of the pith ( Fig. 7B ). When flower peduncles develop to the S2 stage, the pith in the upper peduncle disintegrates. NADI coupled with aniline blue staining revealed the presence of terpene oil droplets in the intercellular space between the two vascular bundles at that stage ( Fig. 7C and D ).

Moreover, we also observed very large mature secretory cavities close to the vascular bundles filled with purple-violet terpene oil ( Fig. 8A and B ). Strings of purple (beaded) staining were also found in the cortex cells of longitudinal sections of the flower peduncles ( Fig. 8C and D ). These purple-violet stained oil droplets consist mainly of EβF based on: (i) our previous gene expression study ( Li et al., 2019), (ii) the chemical content of peduncles dominated by EβF ( Fig. 2A ), and (iii) the correspondence of the violet color with pure EβF stained by NADI reagent ( Fig. 8A ). It may be that these droplets (dominated by EβF) are on a path towards the very large secretory cavities around the vascular bundles to be stored and possibly transported from there.

Figures

Fig. 1. Morphological and anatomical structures of stage 2 (S2) Tanacetum cinerariifolium flowers. (A) Longitudinal section of an S2 flower with peduncle attached. The arrow points to a cavity in the center of the pith that will become larger when the flower matures. The inset pictures show cross sections of the upper and lower peduncle (4 cm individually). (B) A closed disk flower. (C–G) SEM images of the T. cinerariifolium S2 flower organs and tissues. (C) The corollas of closed disk flowers. (D and E) Different magnifications of bract tissue, with an arrow pointing to the longitudinal serpentine vascular system located in the joining site with receptacle. (F and G) Cross sections of the flower peduncle and a detailed view of the disintegrating pith tissue.

Fig. 2. Analysis of terpene volatiles in flower head and peduncle (upper and lower) at early flowering stages (S0–S2). (A) Chromatogram profiles of T. cinerariifolium S1 flower head and upper peduncle hexane extracts analyzed by GC–MS and presented as total extracted ion current (EIC, m/z 93). Chromatogram profile of lower peduncle hexane extraction was not shown here due to the same peak patterns with the upper peduncle. For analysis conditions see text. List of the main terpene components, see Table 1 (B) Quantification of EβF and GD content in flower head and lower and upper peduncle at early flowering stages (S0–S2). Fresh flower tissues were extracted with hexane and analyzed by GC-MS. The GD and EβF concentrations were both estimated using an EβF standard reference curve. Error bars indicate ± SE of three biological replicates.

Fig. 3. EβF and GD accumulations in flowers and peduncles at different developmental stages. (A) Semi-quantitative comparison of GD and EβF emitted during different developmental stages (Pre-f, pre-flowering stage; S0-2, flowering plants with dominant S0–S2 flowers; S2-4, flowering plants with dominant S2–S4 flowers; S4-6, flowering plants with dominant S4–S6 flowers). Data are means ± SE of three biological replicates and expressed as peak area. The numbers indicate the ratio of EβF/GD emissions; the asterisks indicate the significance of the differences between EβF and GD (t-test: *, P ≤ 0.05; **, P ≤ 0.01).(B) Emission kinetics of EβF and GD in S0-2 stage T. cinerariifolium flowering plants are shown. The x-axis indicates the sampling time, and the y axis denotes the emission level of EβF and GD. Data are represented as means ± SD (n = 6). Numbers and asterisks are same as in A. (C) GC-MS analysis of the contents of GD and EβF in the T. cinerariifolium flower head at different developmental stages. Numbering of peduncles and flowers followed the developmental stages. Error bars indicate ± SD of three biological replicates. (D) GC-MS analysis of the contents of GD and EβF in T. cinerariifolium peduncles at different developmental stages. Error bars indicate ± SD of three biological replicates.

Fig. 4. EβF and GD accumulations in different parts of flower head at different developmental stages. Quantitated EβF and GD accumulations in different parts (upper disk floret, lower disk floret, receptacle, bract) of field flower following flower development. Fresh flower tissues were extracted by hexane and analyzed by GC-MS. GD concentration was evaluated by compared with EβF. Error bars indicate ± SD of three biological replicates.

Fig. 5. NADI staining for longitudinal sections of flower of T. cinerariifolium. (A) NADI staining for disk floret of S1 flower. (B) NADI staining of the upper section of a disk floret of S1 flower. (C) NADI staining of the upper section of a disk floret of an S2 flower showing the bilobed stained stigma on top of a short style. (D) NADI staining of longitudinal sections of S3 flower. (E) NADI staining of the upper sections of disk florets of S4 flower. Df, disk floret; Rec, receptacle; Sty, style; An, anther; Ac, anther cylinder; Sti, stigma; Co, corolla; GT, glandular trichome. Pure GD, EβF and pyrethrins were individually stained by NADI reagent and shown in A. The white arrows point to the positive NADI stained oils. The inset picture in D represents the top view of the flower.

Fig. 6. Chromatogram profiles of hexane extract of T. cinerariifolium stigma and corolla analyzed by GC–MS. For analysis conditions see text. GC-MS chromatogram of volatiles extracted from stigma and corolla are presented as total extracted ion current (EIC, m/z 93). Methyl laurate was used as an internal standard (IS). The concentration of EβF and GD were showed in the figure and presented as ng/mg fresh tissue. Localization of terpene oils in the closed and open disk floret was shown in the inset drawing pictures with purple (EβF and GD) and blue (pyrethrins) color.. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 7. NADI and aniline blue staining of cross sections of young T. cinerariifolium flower peduncles. (A and B) NADI staining of two cross sections of S0 and S1 upper flower peduncles. The positive NADI stained oil droplets are pointed out by white arrows. (C) Cross section of S2 flower peduncle stained by NADI reagent coupled with aniline blue staining observed under the bright light. (D) Cross section of the same S2 flower peduncles as in C observed under the UV light, callose in the sieve elements was stained by aniline blue and indicated by white arrowhead. The white arrows point to NADI stained oil droplets in C that appear to be outside the cell wall as indicated by red arrows in D. Ph, phloem; Xy, xylem; Co, cortex; Pi, pith; Se, sieve element.. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 8. NADI staining and aniline blue staining for cross and longitudinal sections of young T. cinerariifolium flower peduncles. (A) Bright field micrographs of the cross sections of S2 flower peduncle double-stained by NADI reagent and aniline blue. (B) Cross section of the same S2 flower peduncles as in A observed under the UV light. The arrows point to stained terpene oil droplets. (C and D) NADI staining of two longitudinal sections of flower peduncles. Co, cortex; Vb, vascular bundle. The purple-violet oil droplet in the inset of A is pure EβF standard stained by NADI reagent.. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)