As more research is conducted on the potential health benefits of saffron value-added metabolites, the demand for these compounds is likely to increase. Finding alternative sources of these metabolites could help ensure a stable supply for the market. While historically the stigmas of some other crocuses like C. cartwrightianus, the known ancestor of saffron [35, 36], were used as wild saffron, only C. sativus stigmas contain the most considerable amount of these metabolites. Since previous investigations confirmed the presence of value-added metabolites in saffron tepals, tepals of other crocuses are a potential source worth exploring. In the present work, using the targeted metabolomics and transcriptomics approaches, four yellow-tepal crocus species were investigated: C. flavus subsp. dissectus, C. graveolens, C. korolkowii and C. chrysanthus.
By examining the saffron apocarotenoid pathway, we could gain a deeper understanding of the genes involved in producing saffron-like metabolites. The presence of various types of crocins was proved in all four species. However, major crocins of C. sativus (t-C-2, t-C-3, t-C-4) were only detected in C. chrysanthus and C. korolkowii (Fig. 3), which is also supported by our transcriptomic analysis (Table 1). We analyzed three key enzymes of the apocarotenoid pathway, CCDs, ALDHs, and UGTs. We observed that CCD2, the enzyme catalyzing the first committed step of crocin biosynthesis, is present in the tepals from all of the four yellow-tepal crocuses. We also identified transcripts for several ALDH homologs including ALDH2B4, ALDH2C4, ALDH3I1, ALDH5F1, ALDH6B2, ALDH7B4 (Fig. 4A, Table 1).
The presence of CCD2, ALDH and crocin-related UGTs transcripts was confirmed in tepals of C. flavus subsp. sissectus and C. graveolens, despite no saffron-like crocins being detected in their LC-PDA analysis.
The absence of detected stigma-like crocins in LC-PDA analysis prompted us to explore a combination of three different hypotheses to understand the underlying reason. The post-harvest degradation of crocins could be considered as the first explanation for the absence of crocins. Since all transcripts of the crocin pathway genes were observed in C. graveolens, crocins might have become degraded subsequently. The effect of environmental factors on crocins were investigated by HPLC-DAD-MS [37] where light and temperature were introduced as the main elements that could degrade crocins after only one week. Due to the detection of other types of crocins in all four species, as well as their detection in saffron stigmas, this hypothesis is not likely.
We considered the possibility that other compounds and CCD enzymes, rather than those directly involved in the synthesis of saffron-like crocins, may play a more important role in yellow tepal color. Besides crocins, there are various apocarotenoids such as β-cyclocytral and β-ionone which are CCD cleavage products that provide specific aromas and colors ranging from yellow to red in fruits and flowers [38]. CCD2 belongs to the larger family of CCD enzymes which also includes CCD1, CCD4, CCD7, and CCD8 [39]. As it is shown in our phylogenetic analysis (Fig. 4A), candidate genes for CsCCD and CCDs of other crocin-producing plants clustered separately while BdCCD1 clustered in CsCCD. CsCCDs clustered into two groups led by CCD1 and CCD4. The role of CCD1 and CCD4 in providing unique aromas and colors through carotenoid degradation was demonstrated in different plants [40]. For instance, in Medicago truncatula CCD1 is responsible for yellow color [41, 42], while in petals of chrysanthemum, CmCCD4a results in colorless compounds in white flowers [43]. In contrast, the reddish color in citrus fruits is attributed to CitCCD4 [44]. Moreover, previous studies validated the presence of CCD1a-, CCD1b- and CCD4a/b-encoding genes in the stigmas of C. sativus and in tepals of C. chrysanthus and C. korolkowii [45]. CsCCD1 and CsCCD4 convert β-carotene into β-ionone and β-cyclocitral [11] that play significant roles in the synthesis of aroma and flavor.
In our study, CCD1 and CCD4 transcripts were also observed in all the studied yellow-tepal species. The presence of these transcripts could contribute to their yellow color along with the presence of crocins (Fig. 4). What is evident is that crocins, although present, are found in low abundance on a dry weight basis compared to saffron stigmas. However, it is essential to consider them as potential commercial sources due to their significantly higher mass compared to saffron stigmas.
The third and the most probable hypothesis refers to the glycosylation of plant secondary metabolites that is done by UGTs displaying the well-known PSPG (plant secondary-product-glycosyltransferase) box [34]. While more than a hundred UGT-encoding genes exist in each plant genome [46], the saffron-related ones were identified recently [2] that are shown in phylogenetic tree (Fig. 4A). While UGT74AD1 (also named GL2) allows the formation of crocins with one and two glucose molecules, UGT91P3 does not use crocetin as a substrate and generate crocins with more than two glucose molecules [33]. Studies demonstrated the expression patterns of the six identified genes (UGT91K2, UGT91K3, UGT91P3, UGT91P4, UGT91P5 and UGT91P6) of the UGT91 subfamily in several tissues of C. sativus [33]. While only UGT91P3 showed higher expression levels in the stigma, higher expression levels of UGT91K3 were detected in leaves [33], suggesting the involvement of the latter in the synthesis of other compounds rather than crocins, which are absent in the leaves. As it is shown in the phylogenetic tree, these genes are grouped together (Fig. 4A). UGT91K2, UGT91K3 and UGT91P3 had expression patterns similar to UGT74AD1[33]. Based on the Blast analysis of the crocin-related UGTs, we observed a higher identity percentage among UGT74AD1 (GL2), UGT91K2, UGT91K3, UGT91P3, UGT91P4, UGT91P5 and UGT91P in C. korolkowii (Table 1). It could potentially confirm the observed variety of crocins at the metabolome level. In contrast, a lower number of these genes with a high identity percentage were found in C. chrysanthus and the decreasing trend continued in C. graveolens and reached to almost non-significant ones in C. flavus subsp. dissectus. This observation could also suggest that these species are increasingly distantly related to saffron.
Our metabolome results provide evidence explaining why we still find crocins in yellow tepals, but these differ significantly from those found in the stigmas. The main reason behind these differences could be attributed to variations in the substrate affinity and catalytic properties of the putative crocetin UGTs. Notably, all the detected crocins in the tepals exhibited higher polarity compared to those present in saffron stigmas, leading to their earlier appearance in the LC chromatogram. This higher polarity suggests a greater degree of glycosylation. Investigating glucose availability in the tissues, as glucose is used for glycosylating the crocetin backbone, could be a compelling aspect to explore further. This might shed light on the discrepancy between tepals and stigmas in terms of crocin composition and abundance. This is in line with another study on some yellow-tepal crocuses showing that the crocins in tepals possess higher glycosylation levels in comparison to the crocins in C. sativus stigmas [6]. While crocins in saffron stigmas contain up to five and six glucose molecules [47, 48], crocins in yellow-tepal crocuses include up to eight sugar molecules [6]. In fact, these highly glycosylated crocins demonstrate differences in retention times indicating different arrangements of the glucose molecules on the ends of the crocetin [49]. Moreover, previous studies showed that these crocins are also present in the stigmas of spring crocuses and because of their absence in the stigma of autumn crocuses could be considered a distinguishing factor between spring and autumn crocuses [50].
The apocarotenoid HTCC is also glycosylated by UGTs to produce picrocrocin, the precursor of safranal. UGT709G1 catalyzes the HTCC glucosyltransferase reaction [2] Fig. 4). Not HTCC, neither picrocrocin nor safranal were detected in yellow-tepals samples (Fig. 3). Nevertheless, at the transcriptome level, the UGT709G1 was identified in almost all samples. While the final enzyme in safranal production is still unknown [51, 52], we could only check HTCC glucosyltransferase that eventually produces picrocrocin. Considering the unknown safranal pathway, several possibilities could be contemplated to explain the absence of safranal despite the presence of UGT709G1. Furthermore, it has been reported that the presence of picrocrocin and safranal in stigmas is greatly influenced by the developmental stage of the flowers [14]. Conducting a time-course analysis of tepals could potentially shed light on the ability of crocus tepals to produce these monoterpenoids. Alternatively, the substrates of UGT709G1 might differ in a species- and/or tissue-specific manner, leading to variations in the production of certain compounds.
To provide a comprehensive view of the presence of saffron value-added metabolites in yellow-tepal crocuses, the presence of saffron stigma-like kaempferols was also investigated. Our results confirmed the presence in C. korolkowii of two kaempferol derivatives also identified in C. sativus stigmas. Alongside with kaempferol glycosides, which are the sole flavonoids in saffron stigmas, a higher diversity of flavonoids, such as anthocyanins, quercetin- and flavone-related compounds (luteolin, tricin, acacetin, apigenin, and scutellarein) are found in saffron tepals. [53, 54]. The detection of multiple compounds with maximum UV-Vis absorption values at 320 nm in all tepal samples, supports the presence of a higher diversity of flavonoids in the yellow-tepal crocuses. Analysis of C. chrysanthus tepals has proved the existence of kaempferol 3-O glucoside, [55]. Other flavonoids such as luteolin-derivatives in C. corsicus and C. minimus, and tricin in C. heuffelianus and C. korolkowii have been reported [56]. In our study, we found flavonoid glucosyltransferase transcripts with the highest to the lowest E-value and identity percentage in C. korolkowii, C. chrysanthus, C. graveolens and C. flavus subsp. dissectus, respectively.
Ultimately, tepals of C. korolkowii with three proven types of saffron-like crocins and two kaempferol-derivatives showed the highest similarity to saffron stigmas. It is important to highlight that additional crocin species were found in yellow tepals which despite differing from the ones in saffron, support the metabolic capability of these tissues and species to synthesize crocins. The identified crocins differ in the decorations of the crocetin backbone, likely displaying high glycosylation levels. Although the crocins found in the tepals are different from those in saffron stigmas, they could still be a potential source of pigments that meet industrial demands. The identification of additional CCD transcripts in the yellow crocuses may also be involved in the synthesis of non-crocin pigments, which could also be of commercial interest. Moreover, our results support the presence of diverse flavonoid in tepals, which could be exploited as a source of natural products for nutritional and medicinal purposes.