ARTICLE   Open Access    

Compositional characteristics of red clover (Trifolium pratense) seeds and supercritical CO2 extracted seed oil as potential sources of bioactive compounds

  • Authors contributed equally: Ying Zhou, Ye Tian

More Information
  • Plant seeds from the Fabaceae (Leguminosae) family are commonly edible. However, little has been done to study the phytochemicals of red clover (Trifolium pratense) seeds. Our study aims to obtain comprehensive and novel findings on red clover seeds and supercritical fluid extraction (SFE)-extracted oil, with the purpose of exploring their potential as a new source of functional ingredients for food and health care products. In our study, red clover seed oil was extracted by supercritical CO2. Forty-four phytochemical compounds were preliminarily identified in red clover seeds and the extracted oil by UPLC-ESI-MS/MS metabolomics method. These compounds mainly belong to lipids, phenolic compounds, terpenoids and phytosterols. Red clover seeds contain fatty acids (4,676.1 mg/100 g dried seeds) and bioactive components such as phenolic compounds (228.4 mg/100 g) and tocopherols (94.9 mg/100 g). In red clover seed oil, unsaturated fatty acids are over 83% and are rich in linoleic acid (54.7 g/100 g oil) and oleic acid (14.0 g/100 g oil). These findings provide important guidance for introducing red clover seed oil into pharmaceutical products or as functional foods.
  • Grapevines are among the most widely grown and economically important fruit crops globally. Grapes are used not only for wine making and juice, but also are consumed fresh and as dried fruit[1]. Additionally, grapes have been increasingly recognized as an important source of resveratrol (trans-3, 5,4'-trihydroxystilbene), a non-flavonoid stilbenoid polyphenol that in grapevine may act as a phytoalexin. In humans, it has been widely reported that dietary resveratrol has beneficial impacts on various aspects of health[2, 3]. Because of the potential value of resveratrol both to grapevine physiology and human medicine, resveratrol biosynthesis and its regulation has become an important avenue of research. Similar to other stilbenoids, resveratrol synthesis utilizes key enzymes of the phenylpropanoid pathway including phenylalanine ammonia lyase (PAL), cinnamate 4-hydroxylase (C4H), and 4-coumarate-CoA ligase (4CL). In the final steps, stilbene synthase (STS), a type II polyketide synthase, produces trans-resveratrol from p-coumaroyl-CoA and malonyl-CoA, while chalcone synthase (CHS) synthesizes flavonoids from the same substrates[4, 5]. Moreover, trans-resveratrol is a precursor for other stilbenoids such as cis-resveratrol, trans-piceid, cis-piceid, ε-viniferin and δ-viniferin[6]. It has been reported that stilbenoid biosynthesis pathways are targets of artificial selection during grapevine domestication[7] and resveratrol accumulates in various structures in response to both biotic and abiotic stresses[812]. This stress-related resveratrol synthesis is mediated, at least partialy, through the regulation of members of the STS gene family. Various transcription factors (TFs) participating in regulating STS genes in grapevine have been reported. For instance, MYB14 and MYB15[13, 14] and WRKY24[15] directly bind to the promoters of specific STS genes to activate transcription. VvWRKY8 physically interacts with VvMYB14 to repress VvSTS15/21 expression[16], whereas VqERF114 from Vitis quinquangularis accession 'Danfeng-2' promotes expression of four STS genes by interacting with VqMYB35 and binding directly to cis-elements in their promoters[17]. Aided by the release of the first V. vinifera reference genome assembly[18], genomic and transcriptional studies have revealed some of the main molecular mechanisms involved in fruit ripening[1924] and stilbenoid accumulation[8, 25] in various grapevine cultivars. Recently, it has been reported that a root restriction treatment greatly promoted the accumulation of trans-resveratrol, phenolic acid, flavonol and anthocyanin in 'Summer Black' (Vitis vinifera × Vitis labrusca) berry development during ripening[12]. However, most of studies mainly focus on a certain grape variety, not to investigate potential distinctions in resveratrol biosynthesis among different Vitis genotypes. In this study, we analyzed the resveratrol content in seven grapevine accessions and three berry structures, at three stages of fruit development. We found that the fruits of two wild, Chinese grapevines, Vitis amurensis 'Tonghua-3' and Vitis davidii 'Tangwei' showed significant difference in resveratrol content during development. These were targeted for transcriptional profiling to gain insight into the molecular aspects underlying this difference. This work provides a theoretical basis for subsequent systematic studies of genes participating in resveratrol biosynthesis and their regulation. Further, the results should be useful in the development of grapevine cultivars exploiting the genetic resources of wild grapevines. For each of the seven cultivars, we analyzed resveratrol content in the skin, pulp, and seed at three stages of development: Green hard (G), véraison (V), and ripe (R) (Table 1). In general, we observed the highest accumulation in skins at the R stage (0.43−2.99 µg g−1 FW). Lesser amounts were found in the pulp (0.03−0.36 µg g−1 FW) and seed (0.05−0.40 µg g−1 FW) at R, and in the skin at the G (0.12−0.34 µg g−1 FW) or V stages (0.17−1.49 µg g−1 FW). In all three fruit structures, trans-resveratrol showed an increasing trend with development, and this was most obvious in the skin. It is worth noting that trans-resveratrol was not detectable in the skin of 'Tangwei' at the G or V stage, but had accumulated to 2.42 µg g−1 FW by the R stage. The highest amount of extractable trans-resveratrol (2.99 µg g−1 FW) was found in 'Tonghua-3' skin at the R stage.
    Table 1.  Resveratrol concentrations in the skin, pulp and seed of berries from different grapevine genotypes at green hard, véraison and ripe stages.
    StructuresSpeciesAccessions or cultivarsContent of trans-resveratrol (μg g−1 FW)
    Green hardVéraisonRipe
    SkinV. davidiiTangweindnd2.415 ± 0.220
    V. amurensisTonghua-30.216 ± 0.0410.656 ± 0.0432.988 ± 0.221
    Shuangyou0.233 ± 0.0620.313 ± 0.0172.882 ± 0.052
    V. amurensis × V. ViniferaBeibinghong0.336 ± 0.0761.486 ± 0.1771.665 ± 0.100
    V. viniferaRed Global0.252 ± 0.0510.458 ± 0.0571.050 ± 0.129
    Thompson seedless0.120 ± 0.0251.770 ± 0.0320.431 ± 0.006
    V. vinifera × V. labruscaJumeigui0.122 ± 0.0160.170 ± 0.0210.708 ± 0.135
    PulpV. davidiiTangwei0.062 ± 0.0060.088 ± 0.009nd
    V. amurensisTonghua-30.151 ± 0.0660.324 ±0.1040.032 ± 0.004
    Shuangyou0.053 ± 0.0080.126 ± 0.0440.041 ± 0.017
    V. amurensis × V. ViniferaBeibinghong0.057 ± 0.0140.495 ± 0.0680.087 ± 0.021
    V. viniferaRed Global0.059 ± 0.0180.159 ± 0.0130.027 ± 0.004
    Thompson seedless0.112 ± 0.0160.059 ± 0.020nd
    V. vinifera × V. labruscaJumeigui0.072 ± 0.0100.063 ± 0.0170.359 ± 0.023
    SeedV. davidiiTangwei0.096 ± 0.0140.169 ± 0.0280.049 ± 0.006
    V. amurensisTonghua-30.044 ± 0.0040.221 ± 0.0240.113 ± 0.027
    Shuangyound0.063 ± 0.0210.116 ± 0.017
    V. amurensis × V. ViniferaBeibinghongnd0.077 ± 0.0030.400 ± 0.098
    V. viniferaRed Global0.035 ± 0.0230.142 ± 0.0360.199 ± 0.009
    Thompson seedless
    V. vinifera × V. labruscaJumeigui0.077 ± 0.0250.017 ± 0.0040.284 ± 0.021
    'nd' indicates not detected in samples, and '−' shows no samples are collected due to abortion.
     | Show Table
    DownLoad: CSV
    To gain insight into gene expression patterns influencing resveratrol biosynthesis in 'Tangwei' and 'Tonghua-3', we profiled the transcriptomes of developing berries at G, V, and R stages, using sequencing libraries representing three biological replicates from each cultivar and stage. A total of 142.49 Gb clean data were obtained with an average of 7.92 Gb per replicate, with average base Q30 > 92.5%. Depending on the sample, between 80.47%−88.86% of reads aligned to the V. vinifera reference genome (Supplemental Table S1), and of these, 78.18%−86.66% mapped to unique positions. After transcript assembly, a total of 23,649 and 23,557 unigenes were identified as expressed in 'Tangwei' and 'Tonghua-3', respectively. Additionally, 1,751 novel transcripts were identified (Supplemental Table S2), and among these, 1,443 could be assigned a potential function by homology. Interestingly, the total number of expressed genes gradually decreased from the G to R stage in 'Tangwei', but increased in 'Tonghua-3'. About 80% of the annotated genes showed fragments per kilobase of transcript per million fragments mapped (FPKM) values > 0.5 in all samples, and of these genes, about 40% showed FPKM values between 10 and 100 (Fig. 1a). Correlation coefficients and principal component analysis of the samples based on FPKM indicated that the biological replicates for each cultivar and stage showed similar properties, indicating that the transcriptome data was reliable for further analyses (Fig. 1b & c).
    Figure 1.  Properties of transcriptome data of 'Tangwei' (TW) and 'Tonghua-3' (TH) berry at green hard (G), véraison (V), and ripe (R) stages. (a) Total numbers of expressed genes with fragments per kilobase of transcript per million fragments mapped (FPKM) values; (b) Heatmap of the sample correlation analysis; (c) Principal component analysis (PCA) showing clustering pattern among TW and TH at G, V and R samples based on global gene expression profiles.
    By comparing the transcriptomes of 'Tangwei' and 'Tonghua-3' at the G, V and R stages, we identified 6,770, 3,353 and 6,699 differentially expressed genes (DEGs), respectively (Fig. 2a). Of these genes, 1,134 were differentially expressed between the two cultivars at all three stages (Fig. 2b). We also compared transcriptional profiles between two adjacent developmental stages (G vs V; V vs R) for each cultivar. Between G and V, we identified 1,761 DEGs that were up-regulated and 2,691 DEGs that were down-regulated in 'Tangwei', and 1,836 and 1,154 DEGs that were up-regulated or down-regulated, respectively, in 'Tonghua-3'. Between V and R, a total of 1,761 DEGs were up-regulated and 1,122 DEGs were down-regulated in 'Tangwei', whereas 2,774 DEGs and 1,287 were up-regulated or down-regulated, respectively, in 'Tonghua-3' (Fig. 2c). Among the 16,822 DEGs between the two cultivars at G, V, and R (Fig. 2a), a total of 4,570, 2,284 and 4,597 had gene ontology (GO) annotations and could be further classified to over 60 functional subcategories. The most significantly represented GO terms between the two cultivars at all three stages were response to metabolic process, catalytic activity, binding, cellular process, single-organism process, cell, cell part and biological regulation (Fig. 2d).
    Figure 2.  Analysis of differentially expressed genes (DEGs) at the green hard (G), véraison (V), and ripe (R) stages in 'Tangwei' (TW) and 'Tonghua-3' (TH). (a) Number of DEGs and (b) numbers of overlapping DEGs between 'Tangwei' and 'Tonghua-3' at G, V and R; (c) Overlap among DEGs between G and V, and V and R, for 'Tangwei' and 'Tonghua-3'; (d) Gene ontology (GO) functional categorization of DEGs.
    We also identified 57 Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways that were enriched, of which 32, 28, and 31 were enriched at the G, V, and R stages, respectively. Seven of the KEGG pathways were enriched at all three developmental stages: photosynthesis-antenna proteins (ko00196); glycine, serine and threonine metabolism (ko00260); glycolysis/gluconeogenesis (ko00010); carbon metabolism (ko01200); fatty acid degradation (ko00071); cysteine and methionine metabolism (ko00270); and valine, leucine and isoleucine degradation (ko00280) (Supplemental Tables S3S5). Furthermore, we found that the predominant KEGG pathways were distinct for each developmental stage. For example, phenylpropanoid biosynthesis (ko00940) was enriched only at the R stage. Overall, the GO and KEGG pathway enrichment analysis showed that the DEGs in 'Tangwei' and 'Tonghua-3' were enriched for multiple biological processes during the three stages of fruit development. We then analyzed the expression of genes with potential functions in resveratrol and flavonoid biosynthesis between the two cultivars and three developmental stages (Fig. 3 and Supplemental Table S6). We identified 30 STSs, 13 PALs, two C4Hs and nine 4CLs that were differentially expressed during at least one of the stages of fruit development between 'Tangwei' and 'Tonghua-3'. Interestingly, all of the STS genes showed increasing expression with development in both 'Tangwei' and 'Tonghua-3'. In addition, the expression levels of STS, C4H and 4CL genes at V and R were significantly higher in 'Tonghua-3' than in 'Tangwei'. Moreover, 25 RESVERATROL GLUCOSYLTRANSFERASE (RSGT), 27 LACCASE (LAC) and 21 O-METHYLTRANSFERASE (OMT) DEGs were identified, and most of these showed relatively high expression at the G and V stages in 'Tangwei' or R in 'Tonghua-3'. It is worth noting that the expression of the DEGs related to flavonoid biosynthesis, including CHS, FLAVONOL SYNTHASE (FLS), FLAVONOID 3′-HYDROXYLASE (F3'H), DIHYDROFLAVONOL 4-REDUCTASE (DFR), ANTHOCYANIDIN REDUCTASE (ANR) and LEUCOANTHOCYANIDIN REDUCTASE (LAR) were generally higher in 'Tangwei' than in 'Tonghua-3'at G stage.
    Figure 3.  Expression of differentially expressed genes (DEGs) associated with phenylalanine metabolism. TW, 'Tangwei'; TH, 'Tonghua-3'. PAL, PHENYLALANINE AMMONIA LYASE; C4H, CINNAMATE 4-HYDROXYLASE; 4CL, 4-COUMARATE-COA LIGASE; STS, STILBENE SYNTHASE; RSGT, RESVERATROL GLUCOSYLTRANSFERASE; OMT, O-METHYLTRANSFERASE; LAC, LACCASE; CHS, CHALCONE SYNTHASE; CHI, CHALCONE ISOMERASE; F3H, FLAVANONE 3-HYDROXYLASE; FLS, FLAVONOL SYNTHASE; F3'H, FLAVONOID 3′-HYDROXYLASE; DFR, DIHYDROFLAVONOL 4-REDUCTASE; LAR, LEUCOANTHOCYANIDIN REDUCTASE; ANR, ANTHOCYANIDIN REDUCTASE; LDOX, LEUCOANTHOCYANIDIN DIOXYGENASE; UFGT, UDP-GLUCOSE: FLAVONOID 3-O-GLUCOSYLTRANSFERASE.
    Among all DEGs identified in this study, 757 encoded potential TFs, and these represented 57 TF families. The most highly represented of these were the AP2/ERF, bHLH, NAC, WRKY, bZIP, HB-HD-ZIP and MYB families with a total of 76 DEGs (Fig. 4a). We found that the number of downregulated TF genes was greater than upregulated TF genes at G and V, and 48 were differentially expressed between the two cultivars at all three stages (Fig. 4b). Several members of the ERF, MYB, WRKY and bHLH families showed a strong increase in expression at the R stage (Fig. 4c). In addition, most of the TF genes showed > 2-fold higher expression in 'Tonghua-3' than in 'Tangwei' at the R stage. In particular, a few members, such as ERF11 (VIT_07s0141g00690), MYB105 (VIT_01s0026g02600), and WRKY70 (VIT_13s0067g03140), showed > 100-fold higher expression in 'Tonghua-3' (Fig. 4d).
    Figure 4.  Differentially expressed transcription factor (TF) genes. (a) The number of differentially expressed genes (DEGs) in different TF families; (b) Number of differentially expressed TF genes, numbers of overlapping differentially expressed TF genes, and (c) categorization of expression fold change (FC) for members of eight TF families between 'Tangwei' and 'Tonghua-3' at green hard (G), véraison (V), and ripe (R) stages; (d) Heatmap expression profiles of the three most strongly differentially expressed TF genes from each of eight TF families.
    We constructed a gene co-expression network using the weighted gene co-expression network analysis (WGCNA) package, which uses a systems biology approach focused on understanding networks rather than individual genes. In the network, 17 distinct modules (hereafter referred to by color as portrayed in Fig. 5a), with module sizes ranging from 91 (antiquewhite4) to 1,917 (magenta) were identified (Supplemental Table S7). Of these, three modules (ivory, orange and blue) were significantly correlated with resveratrol content, cultivar ('Tonghua-3'), and developmental stage (R). The blue module showed the strongest correlation with resveratrol content (cor = 0.6, p-value = 0.008) (Fig. 5b). KEGG enrichment analysis was carried out to further analyze the genes in these three modules. Genes in the ivory module were significantly enriched for phenylalanine metabolism (ko00360), stilbenoid, diarylheptanoid and gingerol biosynthesis (ko00945), and flavonoid biosynthesis (ko00941), whereas the most highly enriched terms of the blue and orange modules were plant-pathogen interaction (ko04626), plant hormone signal transduction (ko04075) and circadian rhythm-plant (ko04712) (Supplemental Fig. S1). Additionally, a total of 36 genes encoding TFs including in 15 ERFs, 10 WRKYs, six bHLHs, two MYBs, one MADs-box, one HSF and one TRY were identified as co-expressed with one or more STSs in these three modules (Fig. 5c and Supplemental Table S8), suggesting that these TFs may participate in the STS regulatory network.
    Figure 5.  Results of weighted gene co-expression network analysis (WGCNA). (a) Hierarchical clustering tree indicating co-expression modules; (b) Module-trait relationship. Each row represents a module eigengene, and each column represents a trait. The corresponding correlation and p-value are indicated within each module. Res, resveratrol; TW, 'Tangwei'; TH, 'Tonghua-3'; (c) Transcription factors and stilbene synthase gene co-expression networks in the orange, blue and ivory modules.
    To assess the reliability of the RNA-seq data, 12 genes determined to be differentially expressed by RNA-seq were randomly selected for analysis of expression via real-time quantitative PCR (RT-qPCR). This set comprised two PALs, two 4CLs, two STSs, two WRKYs, two LACs, one OMT, and MYB14. In general, these RT-qPCR results strongly confirmed the RNA-seq-derived expression patterns during fruit development in the two cultivars. The correlation coefficients between RT-qPCR and RNA-seq were > 0.6, except for LAC (VIT_02s0154g00080) (Fig. 6).
    Figure 6.  Comparison of the expression patterns of 12 randomly selected differentially expressed genes by RT-qPCR (real-time quantitative PCR) and RNA-seq. R-values are correlation coefficients between RT-qPCR and RNA-seq. FPKM, fragments per kilobase of transcript per million fragments mapped; TW, 'Tangwei'; TH, 'Tonghua-3'; G, green hard; V, véraison; R, ripe.
    Grapevines are among the most important horticultural crops worldwide[26], and recently have been the focus of studies on the biosynthesis of resveratrol. Resveratrol content has previously been found to vary depending on cultivar as well as environmental stresses[27]. In a study of 120 grape germplasm cultivars during two consecutive years, the extractable amounts of resveratrol in berry skin were significantly higher in seeded cultivars than in seedless ones, and were higher in both berry skin and seeds in wine grapes relative to table grapes[28]. Moreover, it was reported that total resveratrol content constantly increased from véraison to complete maturity, and ultraviolet-C (UV-C) irradiation significantly stimulated the accumulation of resveratrol of berry during six different development stages in 'Beihong' (V. vinifera × V. amurensis)[9]. Intriguingly, a recent study reported that bud sport could lead to earlier accumulation of trans-resveratrol in the grape berries of 'Summer Black' and its bud sport 'Nantaihutezao' from the véraison to ripe stages[29]. In the present study, resveratrol concentrations in seven accessions were determined by high performance liquid chromatography (HPLC) in the seed, pulp and skin at three developmental stages (G, V and R). Resveratrol content was higher in berry skins than in pulp or seeds, and were higher in the wild Chinese accessions compared with the domesticated cultivars. The highest resveratrol content (2.99 µg g−1 FW) was found in berry skins of 'Tonghua-3' at the R stage (Table 1). This is consistent with a recent study of 50 wild Chinese accessions and 45 cultivars, which reported that resveratrol was significantly higher in berry skins than in leaves[30]. However, we did not detect trans-resveratrol in the skins of 'Tangwei' during the G or V stages (Table 1). To explore the reason for the difference in resveratrol content between 'Tangwei' and 'Tonghua-3', as well as the regulation mechanism of resveratrol synthesis and accumulation during berry development, we used transcriptional profiling to compare gene expression between these two accessions at the G, V, and R stages. After sequence read alignment and transcript assembly, 23,649 and 23,557 unigenes were documented in 'Tangwei' and 'Tonghua-3', respectively. As anticipated, due to the small number of structures sampled, this was less than that (26,346) annotated in the V. vinifera reference genome[18]. Depending on the sample, 80.47%−88.86% of sequence reads aligned to a single genomic location (Supplemental Table S1); this is similar to the alignment rate of 85% observed in a previous study of berry development in Vitis vinifera[19]. Additionally, 1751 novel transcripts were excavated (Supplemental Table S2) after being compared with the V. vinifera reference genome annotation information[18, 31]. A similar result was also reported in a previous study when transcriptome analysis was performed to explore the underlying mechanism of cold stress between Chinese wild Vitis amurensis and Vitis vinifera[32]. We speculate that these novel transcripts are potentially attributable to unfinished V. vinifera reference genome sequence (For example: quality and depth of sequencing) or species-specific difference between Vitis vinifera and other Vitis. In our study, the distribution of genes based on expression level revealed an inverse trend from G, V to R between 'Tangwei' and 'Tonghua-3' (Fig. 1). Furthermore, analysis of DEGs suggested that various cellular processes including metabolic process and catalytic activity were altered between the two cultivars at all three stages (Fig. 2 and Supplemental Table S3S5). These results are consistent with a previous report that a large number of DEGs and 100 functional subcategories were identified in 'Tonghua-3' grape berries after exposure to UV-C radiation[8]. Resveratrol biosynthesis in grapevine is dependent on the function of STSs, which compete with the flavonoid branch in the phenylalanine metabolic pathway. Among the DEGs detected in this investigation, genes directly involved in the resveratrol synthesis pathway, STSs, C4Hs and 4CLs, were expressed to significantly higher levels in 'Tonghua-3' than in 'Tangwei' during V and R. On the other hand, DEGs representing the flavonoid biosynthesis pathway were upregulated in 'Tangwei', but downregulated in 'Tonghua-3' (Fig. 3 and Supplemental Table S6). These expression differences may contribute to the difference in resveratrol content between the two cultivars at these stages. We note that 'Tangwei' and 'Tonghua-3' are from two highly diverged species with different genetic backgrounds. There might be some unknown genetic differences between the two genomes, resulting in more than 60 functional subcategories being enriched (Fig. 2d) and the expression levels of genes with putative roles in resveratrol biosynthesis being significantly higher in 'Tonghua-3' than in 'Tangwei' during V and R (Fig. 3). A previous proteomic study also reported that the expression profiles of several enzymes in the phenylalanine metabolism pathway showed significant differences between V. quinquangularis accession 'Danfeng-2' and V. vinifera cv. 'Cabernet Sauvignon' at the véraison and ripening stages[33]. In addition, genes such as RSGT, OMT and LAC involved in the production of derivatized products of resveratrol were mostly present at the G and V stages of 'Tangwei', potentially resulting in limited resveratrol accumulation. However, we found that most of these also revealed relatively high expression at R in 'Tonghua-3' (Fig. 3). Despite this situation, which does not seem to be conducive for the accumulation of resveratrol, it still showed the highest content (Table 1). It has been reported that overexpression of two grapevine peroxidase VlPRX21 and VlPRX35 genes from Vitis labruscana in Arabidopsis may be involved in regulating stilbene synthesis[34], and a VqBGH40a belonging to β-glycoside hydrolase family 1 in Chinese wild Vitis quinquangularis can hydrolyze trans-piceid to enhance trans-resveratrol content[35]. However, most studies mainly focus on several TFs that participate in regulation of STS gene expression, including ERFs, MYBs and WRKYs[13, 15, 17]. For example, VvWRKY18 activated the transcription of VvSTS1 and VvSTS2 by directly binding the W-box elements within the specific promoters and resulting in the enhancement of stilbene phytoalexin biosynthesis[36]. VqWRKY53 promotes expression of VqSTS32 and VqSTS41 through participation in a transcriptional regulatory complex with the R2R3-MYB TFs VqMYB14 and VqMYB15[37]. VqMYB154 can activate VqSTS9/32/42 expression by directly binding to the L5-box and AC-box motifs in their promoters to improve the accumulation of stilbenes[38]. In this study, we found a total of 757 TF-encoding genes among the DEGs, including representatives of the MYB, AP2/ERF, bHLH, NAC, WRKY, bZIP and HB-HD-ZIP families. The most populous family was MYB, representing 76 DEGs at G, V and R between 'Tangwei' and 'Tonghua-3' (Fig. 4). A recent report indicated that MYB14, MYB15 and MYB13, a third uncharacterized member of Subgroup 2 (S2), could bind to 30 out of 47 STS family genes. Moreover, all three MYBs could also bind to several PAL, C4H and 4CL genes, in addition to shikimate pathway genes, the WRKY03 stilbenoid co-regulator and resveratrol-modifying gene[39]. VqbZIP1 from Vitis quinquangularis has been shown to promote the expression of VqSTS6, VqSTS16 and VqSTS20 by interacting with VqSnRK2.4 and VqSnRK2.6[40]. In the present study, we found that a gene encoding a bZIP-type TF (VIT_12s0034g00110) was down-regulated in 'Tangwei', but up-regulated in 'Tonghua-3', at G, V and R (Fig. 4). We also identified 36 TFs that were co-expressed with 17 STSs using WGCNA analysis, suggesting that these TFs may regulate STS gene expression (Fig. 5 and Supplemental Table S8). Among these, a STS (VIT_16s0100g00880) was together co-expressed with MYB14 (VIT_07s0005g03340) and WRKY24 (VIT_06s0004g07500) that had been identified as regulators of STS gene expression[13, 15]. A previous report also indicated that a bHLH TF (VIT_11s0016g02070) had a high level of co-expression with STSs and MYB14/15[15]. In the current study, six bHLH TFs were identified as being co-expressed with one or more STSs and MYB14 (Fig. 5 and Supplemental Table S8). However, further work needs to be done to determine the potential role of these TFs that could directly target STS genes or indirectly regulate stilbene biosynthesis by formation protein complexes with MYB or others. Taken together, these results identify a small group of TFs that may play important roles in resveratrol biosynthesis in grapevine. In summary, we documented the trans-resveratrol content of seven grapevine accessions by HPLC and performed transcriptional analysis of the grape berry in two accessions with distinct patterns of resveratrol accumulation during berry development. We found that the expression levels of genes with putative roles in resveratrol biosynthesis were significantly higher in 'Tonghua-3' than in 'Tangwei' during V and R, consistent with the difference in resveratrol accumulation between these accessions. Moreover, several genes encoding TFs including MYBs, WRKYs, ERFs, bHLHs and bZIPs were implicated as regulators of resveratrol biosynthesis. The results from this study provide insights into the mechanism of different resveratrol accumulation in various grapevine accessions. V. davidii 'Tangwei', V. amurensis × V. Vinifera 'Beibinghong'; V. amurensis 'Tonghua-3' and 'Shuangyou'; V. vinifera × V. labrusca 'Jumeigui'; V. vinifera 'Red Globe' and 'Thompson Seedless' were maintained in the grapevine germplasm resource at Northwest A&F University, Yangling, Shaanxi, China (34°20' N, 108°24' E). Fruit was collected at the G, V, and R stages, as judged by skin and seed color and soluble solid content. Each biological replicate comprised three fruit clusters randomly chosen from three plants at each stage. About 40−50 representative berries were separated into skin, pulp, and seed, and immediately frozen in liquid nitrogen and stored at −80 °C. Resveratrol extraction was carried out as previously reported[8]. Quantitative analysis of resveratrol content was done using a Waters 600E-2487 HPLC system (Waters Corporation, Milford, MA, USA) equipped with an Agilent ZORBAX SB-C18 column (5 µm, 4.6 × 250 mm). Resveratrol was identified by co-elution with a resveratrol standard, and quantified using a standard curve. Each sample was performed with three biological replicates. Three biological replicates of each stage (G, V and R) from whole berries of 'Tangwei' and 'Tonghua-3' were used for all RNA-Seq experiments. Total RNA was extracted from 18 samples using the E.Z.N.A. Plant RNA Kit (Omega Bio-tek, Norcross, GA, USA). For each sample, sequencing libraries were constructed from 1 μg RNA using the NEBNext UltraTM RNA Library Prep Kit for Illumina (New England Biolabs, Ipswich, MA, USA). The library preparations were sequenced on an Illumina HiSeq2500 platform (Illumina, San Diego, CA, USA) at Biomarker Technologies Co., Ltd. (Beijing, China). Raw sequence reads were filtered to remove low-quality reads, and then mapped to the V. vinifera 12X reference genome[18, 31] using TopHat v.2.1.0[41]. The mapped reads were assembled into transcript models using Stringtie v2.0.4[42]. Transcript abundance and gene expression levels were estimated as FPKM[43]. The formula is as follows:
    FPKM =cDNA FragmentsMapped Fragments(Millions)× Transcript Length (kb)
    Biological replicates were evaluated using Pearson's Correlation Coefficient[44] and principal component analysis. DEGs were identified using the DEGSeq R package v1.12.0[45]. A false discovery rate (FDR) threshold was used to adjust the raw P values for multiple testing[46]. Genes with a fold change of ≥ 2 and FDR < 0.05 were assigned as DEGs. GO and KEGG enrichment analyses of DEGs were performed using GOseq R packages v1.24.0[47] and KOBAS v2.0.12[48], respectively. Co-expression networks were constructed based on FPKM values ≥ 1 and coefficient of variation ≥ 0.5 using the WGCNA R package v1.47[49]. The adjacency matrix was generated with a soft thresholding power of 16. Then, a topological overlap matrix (TOM) was constructed using the adjacency matrix, and the dissimilarity TOM was used to construct the hierarchy dendrogram. Modules containing at least 30 genes were detected and merged using the Dynamic Tree Cut algorithm with a cutoff value of 0.25[50]. The co-expression networks were visualized using Cytoscape v3.7.2[51]. RT-qPCR was carried out using the SYBR Green Kit (Takara Biotechnology, Beijing, China) and the Step OnePlus Real-Time PCR System (Applied Biosystems, Foster, CA, USA). Gene-specific primers were designed using Primer Premier 5.0 software (PREMIER Biosoft International, Palo Alto, CA, USA). Cycling parameters were 95 °C for 30 s, 42 cycles of 95 °C for 5 s, and 60 °C for 30 s. The grapevine ACTIN1 (GenBank Accession no. AY680701) gene was used as an internal control. Each reaction was performed in triplicate for each of the three biological replicates. Relative expression levels of the selected genes were calculated using the 2−ΔΔCᴛ method[52]. Primer sequences are listed in Supplemental Table S9. This research was supported by the National Key Research and Development Program of China (2019YFD1001401) and the National Natural Science Foundation of China (31872071 and U1903107).
  • The authors declare that they have no conflict of interest.
  • Supplemental Table S1 External standards applied in quantification of phenolic compounds.
    Supplemental Table S2 ESI-QTOF analytical condition of non-targeted metabolomics.
  • [1]

    Akbaribazm M, Khazaei F, Naseri L, Pazhouhi M, Zamanian M, et al. 2021. Pharmacological and therapeutic properties of the Red Clover (Trifolium pratense L.): an overview of the new finding. Journal of Traditional Chinese Medicine 41(4):642−49

    doi: 10.19852/j.cnki.jtcm.20210604.001

    CrossRef   Google Scholar

    [2]

    Jones C, De Vega J, Lloyd D, Hegarty M, Ayling S, et al. 2020. Population structure and genetic diversity in red clover (Trifolium pratense L.) germplasm. Scientific Reports 10:8364

    doi: 10.1038/s41598-020-64989-z

    CrossRef   Google Scholar

    [3]

    Carswell A, Sánchez-Rodríguez AR, Saunders K, le Cocq K, Shaw R, et al. 2022. Combining targeted grass traits with red clover improves grassland performance and reduces need for nitrogen fertilisation. European Journal of Agronomy 133:126433

    doi: 10.1016/j.eja.2021.126433

    CrossRef   Google Scholar

    [4]

    McKenna P, Cannon N, Conway J, Dooley J. 2018. The use of red clover (Trifolium pratense) in soil fertility-building: a review. Field Crops Research 221:38−49

    doi: 10.1016/j.fcr.2018.02.006

    CrossRef   Google Scholar

    [5]

    EFSA ANS Panel. 2015. Risk assessment for peri- and post-menopausal women taking food supplements containing isolated isoflavones. EFSA Journal 13(10):4246

    doi: 10.2903/j.efsa.2015.4246

    CrossRef   Google Scholar

    [6]

    Kazlauskaite JA, Ivanauskas L, Bernatoniene J. 2021. Cyclodextrin-assisted extraction method as a green alternative to increase the isoflavone yield from Trifolium pratensis L. extract. Pharmaceutics 13:620

    doi: 10.3390/pharmaceutics13050620

    CrossRef   Google Scholar

    [7]

    Antonescu (Mintas) AI, Miere (Groza) F, Fritea L, Ganea M, Zdrinca M, et al. 2021. Perspectives on the combined effects of Ocimum basilicum and Trifolium pratense extracts in terms of phytochemical profile and pharmacological effects. Plants 10:1390

    doi: 10.3390/plants10071390

    CrossRef   Google Scholar

    [8]

    Vlaisavljević S, Kaurinović B, Popović M, Vasiljević S. 2017. Profile of phenolic compounds in Trifolium pratense L. extracts at different growth stages and their biological activities. International Journal of Food Properties 20:3090−101

    doi: 10.1080/10942912.2016.1273235

    CrossRef   Google Scholar

    [9]

    Lee JS, Paje LA, Kim MJ, Jang SH, Kim JT, et al. 2021. Validation of an optimized HPLC–UV method for the quantification of formononetin and biochanin A in Trifolium pratense extract. Applied Biological Chemistry 64:57

    doi: 10.1186/s13765-021-00630-5

    CrossRef   Google Scholar

    [10]

    Prati S, Baravelli V, Fabbri D, Schwarzinger C, Brandolini V, et al. 2007. Composition and content of seed flavonoids in forage and grain legume crops. Journal of Separation Science 30:491−501

    doi: 10.1002/jssc.200600383

    CrossRef   Google Scholar

    [11]

    Çölgeçen H, Koca U, Büyükkartal HN. 2011. Use of red clover (Trifolium pratense L. ) seeds in human therapeutics. In Nuts and Seeds in Health and Disease Prevention, eds. Preedy VR, Watson RR, Patel VB. San Diego: Academic Press. pp. 975−80. https://doi.org/10.1016/b978-0-12-375688-6.10115-x

    [12]

    Sabudak T, Ozturk M, Goren AC, Kolak U, Topcu G. 2009. Fatty acids and other lipid composition of fiveTrifoliumspecies with antioxidant activity. Pharmaceutical Biology 47:137−41

    doi: 10.1080/13880200802439343

    CrossRef   Google Scholar

    [13]

    Kratovalieva S, Popsimonova G, Ivanovska S, Jankuloski L, Meglič V. 2012. Macedonian Genebank: Seed protein content of wild red clover (Trifolium pratense L.) accessions. Agriculturae Conspectus Scientificus 77(4):199−202

    Google Scholar

    [14]

    Ahmed IAM, Matthäus B, Özcan MM, Al Juhaimi F, Ghafoor K, et al. 2020. Determination of bioactive lipid and antioxidant activity of Onobrychis, Pimpinella, Trifolium, and Phleum spp. seed and oils. Journal of Oleo Science 69:1367−71

    doi: 10.5650/jos.ess20153

    CrossRef   Google Scholar

    [15]

    Rodway LA, Pauls SD, Pascoe CD, Aukema HM, Taylor CG, et al. 2023. Distinct effects of α-linolenic acid and docosahexaenoic acid on the expression of genes related to cholesterol metabolism and the response to infection in THP-1 monocytes and immune cells of obese humans. Biomedicine & Pharmacotherapy 159:114167

    doi: 10.1016/j.biopha.2022.114167

    CrossRef   Google Scholar

    [16]

    Vlaisavljevic S, Kaurinovic B, Popovic M, Djurendic-Brenesel M, Vasiljevic B, et al. 2014. Trifolium pratense L. as a potential natural antioxidant. Molecules 19:713−25

    doi: 10.3390/molecules19010713

    CrossRef   Google Scholar

    [17]

    Ahangari H, King JW, Ehsani A, Yousefi M. 2021. Supercritical fluid extraction of seed oils – A short review of current trends. Trends in Food Science & Technology 111:249−60

    doi: 10.1016/j.jpgs.2021.02.066

    CrossRef   Google Scholar

    [18]

    Wang W, Rao L, Wu X, Wang Y, Zhao L, et al. 2021. Supercritical carbon dioxide applications in food processing. Food Engineering Reviews 13:570−91

    doi: 10.1007/s12393-020-09270-9

    CrossRef   Google Scholar

    [19]

    Chiriac ER, Chiţescu CL, Sandru C, Geană EI, Lupoae M, et al. 2020. Comparative study of the bioactive properties and elemental composition of red clover (Trifolium pratense) and alfalfa (Medicago sativa) sprouts during germination. Applied Sciences 10:7249

    doi: 10.3390/app10207249

    CrossRef   Google Scholar

    [20]

    International Organisation of Standardization. 2009. Animal feeding stuffs -Determination of nitrogen content and calculation of crude protein content -Part 2: Block digestion/steam distillation method. ISO 5983-2.

    [21]

    Zhou Y, Tian Y, Beltrame G, Laaksonen O, Yang B. 2023. Ultrasonication-assisted enzymatic bioprocessing as a green method for valorizing oat hulls. Food Chemistry 426:136658

    doi: 10.1016/j.foodchem.2023.136658

    CrossRef   Google Scholar

    [22]

    Christie W, Han X. 2010. Lipid analysis: Isolation, separation, identification and lipidomic analysis. 4th edition. UK: The Oily Press.

    [23]

    Klåvus A, Kokla M, Noerman S, Koistinen VM, Tuomainen M, et al. 2020. "Notame": workflow for non-targeted LC–MS metabolic profiling. Metabolites 10:135

    doi: 10.3390/metabo10040135

    CrossRef   Google Scholar

    [24]

    Wu Y, Chen Y, Lu Y, Hao H, Liu J, et al. 2020. Structural features, interaction with the gut microbiota and anti-tumor activity of oligosaccharides. RSC Advances 10:16339−48

    doi: 10.1039/d0ra00344a

    CrossRef   Google Scholar

    [25]

    Giese EC, Barbosa AM, Dekker RFH. 2011. Pathways to bioactive oligosaccharides: Biological functions and potential applications. In Handbook on Carbohydrate Polymers: Development, Properties and Applications, eds. Ito R, Matsuo Y. USA: Nova Science Publishers. pp. 279−309. https://doi.org/10.13140/2.1.2036.8323

    [26]

    Wei X, Fu X, Xiao M, Liu Z, Zhang L, et al. 2020. Dietary galactosyl and mannosyl carbohydrates: In-vitro assessment of prebiotic effects. Food Chemistry 329:127179

    doi: 10.1016/j.foodchem.2020.127179

    CrossRef   Google Scholar

    [27]

    Temerdashev ZA, Chubukina TK, Vinitskaya EA, Nagalevskii MV, Kiseleva NV. 2021. Assessment of the concentrations of isoflavonoids in red clover (Trifolium pratense L.) of the Fabaceae family using extraction by different methods. Journal of Analytical Chemistry 76:1071−82

    doi: 10.1134/s1061934821090112

    CrossRef   Google Scholar

    [28]

    Malca-Garcia GR, Zagal D, Graham J, Nikolić D, Friesen JB, et al. 2019. Dynamics of the isoflavone metabolome of traditional preparations of Trifolium pratense L. Journal of Ethnopharmacology 238:111865

    doi: 10.1016/j.jep.2019.111865

    CrossRef   Google Scholar

    [29]

    Shirvani A, Goli SAH, Shahedi M, Soleimanian-Zad S. 2016. Changes in nutritional value and application of thyme (Thymus vulgaris) essential oil on microbial and organoleptic markers of Persian clover (Trifolium resupinatum) sprouts. LWT - Food Science and Technology 67:14−21

    doi: 10.1016/j.lwt.2015.11.036

    CrossRef   Google Scholar

    [30]

    Innes JK, Calder PC. 2018. Omega-6 fatty acids and inflammation. Prostaglandins, Leukotrienes and Essential Fatty Acids 132:41−48

    doi: 10.1016/j.plefa.2018.03.004

    CrossRef   Google Scholar

    [31]

    Liu Q, Wu M, Zhang B, Shrestha P, Petrie J, et al. 2017. Genetic enhancement of palmitic acid accumulation in cotton seed oil through RNAi down-regulation of ghKAS2 encoding β-ketoacyl-ACP synthase II (KASII). Plant Biotechnology Journal 15:132−43

    doi: 10.1111/pbi.12598

    CrossRef   Google Scholar

    [32]

    Aksoz E, Korkut O, Aksit D, Gokbulut C. 2020. Vitamin E (α-, β+γ- and δ-tocopherol) levels in plant oils. Flavour and Fragrance Journal 35:504−10

    doi: 10.1002/ffj.3585

    CrossRef   Google Scholar

    [33]

    Grygier A, Chakradhari S, Ratusz K, Rudzińska M, Patel KS, et al. 2022. Seven underutilized species of the Fabaceae family with high potential for industrial application as alternative sources of oil and lipophilic bioactive compounds. Industrial Crops and Products 186:115251

    doi: 10.1016/j.indcrop.2022.115251

    CrossRef   Google Scholar

    [34]

    Kumar M, Zhang B, Potkule J, Sharma K, Radha, et al. 2023. Cottonseed oil: extraction, characterization, health benefits, safety profile, and application. Food Analytical Methods 16:266−80

    doi: 10.1007/s12161-022-02410-3

    CrossRef   Google Scholar

    [35]

    Knothe G, Razon LF, Madulid DA, Agoo EMG, de Castro MEG. 2016. Fatty acid profiles of some Fabaceae seed oils. Journal of the American Oil Chemists' Society 93:1007−11

    doi: 10.1007/s11746-016-2845-2

    CrossRef   Google Scholar

    [36]

    Doan LP, Nguyen TT, Pham MQ, Tran QT, Pham QL, et al. 2019. Extraction process, identification of fatty acids, tocopherols, sterols and phenolic constituents, and antioxidant evaluation of seed oils from five Fabaceae species. Processes 7:456

    doi: 10.3390/pr7070456

    CrossRef   Google Scholar

    [37]

    Buchbauer G, Jirovetz L, Nikiforov A. 1996. Comparative investigation of essential clover flower oils from Austria using gas chromatography–flame ionization detection, gas chromatography–mass spectrometry, and gas chromatography–olfactometry. Journal of Agricultural and Food Chemistry 44:1827−28

    doi: 10.1021/jf9506850

    CrossRef   Google Scholar

    [38]

    Chiriac ER, Chiţescu CL, Borda D, Lupoae M, Gird CE, et al. 2020. Comparison of the polyphenolic profile of Medicago sativa L. and Trifolium pratense L. sprouts in different germination stages using the UHPLC-Q exactive hybrid quadrupole orbitrap high-resolution mass spectrometry. Molecules 25:2321

    doi: 10.3390/molecules25102321

    CrossRef   Google Scholar

    [39]

    Akinmoladun AC, Olaleye MT, Komolafe K, Adetuyi AO, Akindahunsi AA. 2015. Effect of homopterocarpin, an isoflavonoid from Pterocarpus erinaceus, on indices of liver injury and oxidative stress in acetaminophen-provoked hepatotoxicity. Journal of Basic and Clinical Physiology and Pharmacology 26:555−62

    doi: 10.1515/jbcpp-2014-0095

    CrossRef   Google Scholar

    [40]

    Kaushal A, Sharma M, Navneet, Sharma M. 2020. Ethnomedicinal, phytochemical, therapeutic and pharmacological review of the genus Erythrina. International Journal of Botany Studies 5(6):642−48

    Google Scholar

    [41]

    Hu Q, Zhang J, Xing R, Yu N, Chen Y. 2022. Integration of lipidomics and metabolomics for the authentication of camellia oil by ultra-performance liquid chromatography quadrupole time-of-flight mass spectrometry coupled with chemometrics. Food Chemistry 373:131534

    doi: 10.1016/j.foodchem.2021.131534

    CrossRef   Google Scholar

    [42]

    Chopade AR, Somade PM, Somade PP, Mali SN. 2021. Identification of Anxiolytic Potential of Niranthin: In-vivo and Computational Investigations. Natual Products and Bioprospecting 11(2):223−33

    doi: 10.1007/s13659-020-00284-8

    CrossRef   Google Scholar

    [43]

    Hattori K, Dupuis B, Fu BX, Edwards NM. 2015. Effects of monoglycerides of varying fatty acid chain length and mixtures thereof on sponge-and-dough breadmaking quality. Cereal Chemistry 92:481−86

    doi: 10.1094/cchem-12-14-0267-r

    CrossRef   Google Scholar

    [44]

    Nartea A, Fanesi B, Pacetti D, Lenti L, Fiorini D, et al. 2023. Cauliflower by-products as functional ingredient in bakery foods: fortification of pizza with glucosinolates, carotenoids and phytosterols. Current Research in Food Science 6:100437

    doi: 10.1016/j.crfs.2023.100437

    CrossRef   Google Scholar

    [45]

    Montesano D, Rocchetti G, Putnik P, Lucini L. 2018. Bioactive profile of pumpkin: an overview on terpenoids and their health-promoting properties. Current Opinion in Food Science 22:81−87

    doi: 10.1016/j.cofs.2018.02.003

    CrossRef   Google Scholar

    [46]

    Sánchez-Hernández L, Puchalska P, García-Ruiz C, Crego AL, Marina ML. 2010. Determination of trigonelline in seeds and vegetable oils by capillary electrophoresis as a novel marker for the detection of adulterations in olive oils. Journal of Agricultural and Food Chemistry 58:7489−96

    doi: 10.1021/jf100550b

    CrossRef   Google Scholar

    [47]

    Lee HG, Kim HS, Je JG, Hwang J, Sanjeewa KKA, et al. 2021. Lipid Inhibitory Effect of (−)-loliolide Isolated from Sargassum horneri in 3T3-L1 Adipocytes: Inhibitory Mechanism of Adipose-Specific Proteins. Marine Drugs 19(2):96

    doi: 10.3390/MD19020096

    CrossRef   Google Scholar

    [48]

    Van Puyvelde H, Dimou N, Katsikari A, Indave Ruiz BI, Godderis L, et al. 2023. The association between dietary intakes of methionine, choline and betaine and breast cancer risk: a systematic review and meta-analysis. Cancer Epidemiology 83:102322

    doi: 10.1016/j.canep.2023.102322

    CrossRef   Google Scholar

    [49]

    Dede B, Avci D, Varkal D, Bahçeli S. 2018. Molecular, spectroscopic, NBO and NLO properties of 4-methyl-5-thiazoleethanol: a comparative theoretical study. Acta Physica Polonica A 134:1083−92

    doi: 10.12693/aphyspola.134.1083

    CrossRef   Google Scholar

    [50]

    Hanh TTH, My NTT, Cham PT, Quang TH, Cuong NX, et al. 2020. Diterpenoids and flavonoids from Andrographis paniculata. Chemical and Pharmaceutical Bulletin 68:96−99

    doi: 10.1248/cpb.c19-00662

    CrossRef   Google Scholar

  • Cite this article

    Zhou Y, Tian Y, Ollennu-Chuasam P, Kortesniemi M, Selander K, et al. 2024. Compositional characteristics of red clover (Trifolium pratense) seeds and supercritical CO2 extracted seed oil as potential sources of bioactive compounds. Food Innovation and Advances 3(1): 11−19 doi: 10.48130/fia-0024-0002
    Zhou Y, Tian Y, Ollennu-Chuasam P, Kortesniemi M, Selander K, et al. 2024. Compositional characteristics of red clover (Trifolium pratense) seeds and supercritical CO2 extracted seed oil as potential sources of bioactive compounds. Food Innovation and Advances 3(1): 11−19 doi: 10.48130/fia-0024-0002

Figures(1)  /  Tables(4)

Article Metrics

Article views(4174) PDF downloads(497)

ARTICLE   Open Access    

Compositional characteristics of red clover (Trifolium pratense) seeds and supercritical CO2 extracted seed oil as potential sources of bioactive compounds

Food Innovation and Advances  3 2024, 3(1): 11−19  |  Cite this article

Abstract: Plant seeds from the Fabaceae (Leguminosae) family are commonly edible. However, little has been done to study the phytochemicals of red clover (Trifolium pratense) seeds. Our study aims to obtain comprehensive and novel findings on red clover seeds and supercritical fluid extraction (SFE)-extracted oil, with the purpose of exploring their potential as a new source of functional ingredients for food and health care products. In our study, red clover seed oil was extracted by supercritical CO2. Forty-four phytochemical compounds were preliminarily identified in red clover seeds and the extracted oil by UPLC-ESI-MS/MS metabolomics method. These compounds mainly belong to lipids, phenolic compounds, terpenoids and phytosterols. Red clover seeds contain fatty acids (4,676.1 mg/100 g dried seeds) and bioactive components such as phenolic compounds (228.4 mg/100 g) and tocopherols (94.9 mg/100 g). In red clover seed oil, unsaturated fatty acids are over 83% and are rich in linoleic acid (54.7 g/100 g oil) and oleic acid (14.0 g/100 g oil). These findings provide important guidance for introducing red clover seed oil into pharmaceutical products or as functional foods.

    • Red clover (Trifolium pratense) is a perennial species of flowering plant in the family Fabaceae or Leguminosae. Red clover is native to Europe, Western Asia, and northwest Africa, but it is also grown widely in the temperate regions such as Pacific Northwest and the US[1]. It is a quick growing crop adapted to varied environmental conditions, which can be used as an important winter annual in the Deep South[2]. Its finer roots have nodules containing Rhizobia bacteria, with the capacity of fixing nitrogen. Red clover can capture nitrogen (150−250 kg N/ha/year) from the air and feed it into the soil, enriching the soil fertility and providing nourishment to surrounding plants in between seasons[3]. Red clover is also used as a cover crop. The aims include preventing nutrient leaching, reducing soil erosion as well as replacing pesticides to control pests and phytopathogenic diseases in addition to producing pasture mixture, hay, and silage[4]. These advantages of red clover show great value in sustainable agriculture.

      According to the EU Novel Food Regulation, red clover is allowed to be used as a food supplement, but not as food or food ingredients[5]. Extracts of red clover flowers have been used extensively for a variety of health protective purposes, such as promoting wound healing and maintenance of bone health[6]. Additionally, red clover flower showed potential value in alternative therapies for menopausal problems, cardiovascular disease, and a range of hormone-dependent diseases such as breast cancer. Several studies indicated that these health-beneficial effects of red clover are derived from secondary plant metabolites belonging to polyphenols with emphasis on the isoflavonoid compounds[69]. Bioactive compounds, such as flavonoids, from plant seeds raised a lot of interest[10]. These compounds play active roles reducing the risk of health problems associated with oxidative stress, such as coronary heart disease, and cancer initiation and progression[11]. Although the beneficial compounds have been reported in the seeds of several Trifolium species, the study of Trifolium pratense is little[12]. Currently, wild clover seeds are commonly used as animal feed though they are sources of many nutrients, especially with high protein content (> 16.0% DM)[13].

      Oils extracted from Trifolium seeds have shown various health benefits due to high concentrations of bioactive lipids and tocopherols[14]. Clinical studies demonstrated the therapeutic effects of polyunsaturated fatty acids on cardiovascular diseases, Alzheimer’s disease, depression, and various other degenerative neurological disorders[15]. In addition to fatty acids, tocopherols are also important bioactive compounds in plant seed oils. As lipophilic antioxidants, tocopherols show positive effects in protecting polyunsaturated fatty acids from peroxidation[14]. Although red clover leaves were used to produce essential oils, red clover seeds (RCS) oil has been rarely studied[16]. Compared to Soxhlet extraction, supercritical fluid extraction (SFE) has many advantages. SFE provides solvent-free extracts with high selectivity of the targeted components[17]. Extraction using supercritical CO2 is simple and eco-friendly since carbon dioxide is a non-toxic, inert gas with low price. These superiorities make supercritical CO2 extraction a promising method of recovering valuable compounds from plant materials for food and pharmaceutical application[18].

      Therefore, it is of great importance to gain an in-depth understanding of the phytochemical profile of RCS and SFE-extracted oil due to their richness of health-benefiting compounds. Recent studies of RCS were limited to measuring the total content of specific group of compounds using colorimetric method[14, 19]. Yet, to explore the potential of RCS for application in food and health care products, it is important to obtain thorough knowledge on the composition of different groups of nutrients and bioactive compounds of RCS at molecular level. This study aims to comprehensively identify the metabolites of RCS and its SFE-extracted oil. The novelties of our study lie in: (1) new findings on comprehensive phytochemical profiles of the study materials using a non-targeted metabolomics approach and; (2) identification and quantification of the compound groups with nutritional, sensorial, and health-promoting properties using targeted mass spectrum (MS) approach. To the best of our knowledge, this is the first report on the thorough investigation of the chemical compositions of RCS and SFE-extracted RCS oil.

    • Organic RCS were purchased from the local store Hankkija in Turku Finland. Trimethylsilyl was purchased from Thermoscientific (Bellefonte, US). Sodium hydroxide, boric acid, potassium chloride, acetyl chloride and potassium carbonate were provided by Sigma-Aldrich (Darmstadt, Germany). The reference standards used in this study included sugars, organic acids, tocopherols, isoflavonoids and flavonoids (Sigma-Aldrich, St. Louis, MO, US). The internal standard of fatty acids was bought from Larodan AB (Solna, Sweden). Other chemicals of liquid chromatography (LC) and MS grade were purchased from Honeywell Riedel-de Haën Co. (Seelze, Germany).

    • Oil was extracted from RCS by a pilot-scale SFE facility (Chematur Engineering, Karlskoga, Sweden) at Aromtech Ltd (Tornio, Finland). RCS (470 g) were milled and loaded into the extraction vessel. The extraction was carried out for 120 min with CO2 at a flow rate of 0.4 L/min using ethanol (7 g/min) as co-solvent for extraction. The extraction temperature and pressure were 40 °C and 200 bar, respectively. The separation pressure was 50 bars, and the separation temperature was 30 °C. After the extraction, the oil was separated from ethanol by centrifugation and in storage at −80 °C until further analysis.

    • The RCS were ground using a mortar and pestle with addition of liquid nitrogen. Approximately 1 g RCS powder was used for drying in the oven at 105 °C until it reached constant weight.

    • The total protein content of RCS was determined by a Kjeldahl autoanalyzer (Foss Tecator Ab, Höganäs, Sweden), which was calculated using a conversion factor of 6.25 (Protein content = nitrogen content × 6.25)[20].

    • RCS powder (4 g) was extracted with 70% ethanol by 20 min ultrasonication for three times. Each time, the supernatants were separated by centrifuging (15 min, 1,500× g at 4 °C) and collected. The volume of combined supernatants was set to 50 mL with extraction solvent. One milliliter of the supernatant was filtered (0.22 μm PTFE syringe filter, Phenomenex) for gas chromatography (GC) analysis.

      The identification and quantification of sugars and organic acids were conducted by an internal standard method in our previous research[21]. The chromatographic system Shimadzu GC-2010 consisted of flame ionization detector (Shimadzu corp., Kyoto, Japan) using a SPB-1 column (30 m × 0.25 mm i.d., 0.25 µm, Supelco, Bellefonte, PA, US). Briefly, RCS extract (0.4 mL) was mixed with 0.2 mL of sorbitol and tartaric acid. The sample was dried by nitrogen flow and mixed with trimethylsilyl (600 µL). The correction factors of the study sugars and organic acids were calculated using the corresponding commercial standards. The identification was based on the comparison of the compound retention times of our samples with those of external standards.

    • The aqueous ethanolic RCS extract described was used for the analysis of phenolic compounds using our previous method[21]. An ultra-high performance liquid chromatography (UPLC) system equipped with a quadrupole time-of-flight tandem mass spectrometer (Q-TOF) (Bruker Corp., Billerica, MA, US) was applied for the identification using an Aeris Peptide XB-C18 column (150 mm × 4.60 mm, 3.6 μm, Torrance, CA, US). The injection volume was 10 μL. The total flow rate was 1 mL/min. The chromatogram was recorded under 260 nm (for isoflavonoids) and 280 nm, 360 nm (for other flavonoids). MS flow rate was 0.3−0.4 mL/min. Mass full-scan was operated in a range of 20–2,000 m/z under both positive and negative ionization modes. The compounds were identified based on the retention times, UV-absorption spectra and MS spectra by comparing with those of reference compounds and referring to the library and the literatures.

      The quantification of the identified phenolic compounds was performed on Shimadzu LC-30AD liquid chromatograph equipped with an SPD-M20A photodiode array detector (Shimadzu Corp., Kyoto, Japan). The same column and chromatographic conditions were used for the UPLC-QTOF analysis. The identified phenolic compounds were quantified using the standard calibration curves (Supplemental Table S1).

    • Fatty acids in RCS and SFE-extracted oil were determined following the procedure of Christie's research[22]. The total lipid content of RCS was studied using a modified Folch method. Briefly, all the glass tubes were rinsed with chloroform. RCS powder (approximately 100 mg) was weighed to the tube with addition of 900 µL chloroform and 1.5 mL methanol. After 1 h soaking at room temperature, the mixture was homogenized with Ultra Turrax (IKA®-Werke, Germany). Chloroform (2 mL) was added and the homogenization was continued for 2 min followed by 30 min sonication. KCl solution (0.88%, 1.2 mL) was added and vortexed for 5 s, followed by 3 min centrifuging at 3500 rpm. After collecting the lower liquid fraction, 1.5 mL chloroform was added to the tube. The sample was vortexed for 30 s before centrifuging (3 min at 3,500 rpm). The combination of lower liquid fraction with the previous fraction was dried by nitrogen flow at 50 °C. Lipid content was calculated by weighing the glass tube with lipid fraction. Lipid fraction was re-dissolved in 3 mL chloroform. For preparation of fatty acid methyl esters, the lipid solution (containing approximately 10 mg lipid) was transferred to a Pyrex tube, adding the internal standard triheptadecanoin (0.34 mg). After evaporating the solvent by nitrogen flow, acetyl chloride/methanol (1/10, v/v, 2 mL) was added. The tube was incubated at 50 °C overnight. Potassium carbonate (1 M, 2 mL) was carefully added to the tube after cooling down to room temperature. N-hexane (2 ml) was added to the tube and the sample was centrifuged (3 min, 1,000 rcf) after vortexing (10 s). The upper liquid fraction was transferred to a GC vial for analysis.

      A Shimadzu GC-2030 equipped with an FID (Shimadzu Corporation, Kyoto, Japan) and DB-23 column (60 m × 0.25 mm × 0.25 μm; Agilent Technologies, J.W. Scientific, Santa Clara, CA, US) was used for the determination and quantification of fatty acids. GC conditions were set as helium flow 1.7 mL/min; 130 °C held 1 min, 6.5 °C/min to 170 °C with no hold, 3.0 °C/min to 205 °C, held for 18 min, 30 °C/min to 230 °C and held for 2 min. The peaks were identified by using external standards GLC-490 (Nu-Check-Prep, Elysian, MN, US) and 37 Component FAME mix (Supelco, St. Louis, MO, US) by comparing the retention times. The quantification was conducted by the internal standard (fatty acid 17:0). The correction factors obtained by analysis of standard mixtures were applied in the quantification of different fatty acids.

    • Tocopherols were analyzed using the lipid extraction method described previously. After evaporation of chloroform by nitrogen flow at 50 °C, the lipid extract of RCS and SFE oil were dissolved in heptane for HPLC analysis. Tocopherols were analyzed by NP-HPLC-FLD with Shimadzu Nexera XR LC-30 HPLC instrument and RF-20A prominence fluorescence detector (Shimadzu, Kyoto, Japan) using an Phenomenex OOG-4162-EO Luna 3 µm silica column (250 mm × 4.6 mm, pore size 100 Å; Torrance, CA, US). The injection volume was 2 μL. Mobile phase (isocratic, 0.4 mL/min) consisted of 2% 1,4-dioxane and 98% heptane. For quantification, standard curves were prepared using α-, β- and γ-tocopherol solutions at different concentrations. The excitation wavelength was 292 nm and emission wavelength was 325 nm.

    • Sample preparation and compound identification for non-targeted metabolomics were based on previous methods with minor modifications[23]. Briefly, RCS powder (approximately 100 mg) or oil (100 µL) were added to a tube containing 400 µL of cold acetonitrile, respectively. The tubes were kept in an ice box during the sample preparation. The mixture was vertexed for 10 s and centrifuged for 5 min (7,000× g at 4 °C). The supernatants (1 mL) were filtered (0.22 μm PTFE syringe filters, Phenomenex) for further analysis.

      The compound identification was performed on the same Q-TOF mass spectrometer as described previously. A Zorbax Eclipse XDB-C18 column (2.1 mm × 100 mm,1.8 µm, Santa Clara, CA, US) was used. The gradient elution were water (solvent A) and methanol (solvent B) both containing 0.1% (v/v) of formic acid. The injection volume was 2 μL. The flow rate was 0.4 mL/min. LC gradient program was followed by 0−10 min to 2%−100% solvent B, 10−14.5 min to 100% B, 14.5−14.51 min to 100%−2% B, 14.51−16.5 min to 2% B. Mass full scan was operated under both positive and negative ionization modes. MS parameters were reported in Supplemental Table S2.

    • All data were expressed as mean ± standard deviation based on dry matter content. MS DIAL (version 4.8) was applied to process the raw data from Q-TOF mass spectrometer. The parameters of MS DIAL were set as MS1 tolerance 0.01 Da, MS2 tolerance 0.025 Da, m/z range 50-1500, minimum peak amplitude 3,000 signal counts, and mass slice width 0.1 Da for peak selections. The smoothing level and minimum peak width were 3 scans and 5 scans, respectively. The selected adduct ions include: [M + H]+, [M + NH4]+, [M + Na]+, [M + K]+, [M + CH3OH + H]+, [M + ACN + H]+, [M + H - H2O]+, [2 M + H]+ for the positive mode and [M − H], [M + Cl], [M − H2O − H], [M + FA − H], [2 M − H], [3 M − H] for the negative mode. The m/z tolerance was set to 0.015 Da and retention time tolerance was 0.05 min. The database MassBank was utilized in MS-DIAL for additional annotations and mass spectral comparison.

      Heatmap was performed using MetaboAnalyst 5.0 (www.metaboanalyst.ca). Heatmap was used to analyze the semi-quantitative results of component distribution in RCS and the oil extracted by the equal weight of the seeds.

    • As shown in Table 1, the total content of proteins in RCS was 6.9 g/100 g DM. Higher protein contents were reported by a previous research, in which seed protein contents of wild red clover were 14.2−17.3 g/100 g DM using Kjeldahl method[13]. Our study RCS contained 50.0 mg/100 g DM of organic acids, including citric acid and malic acid. Sugars were identified from RCS, including disaccharides (maltose and sucrose), monosaccharides (glucose, fructose, xylose, arabinose and mannose), sugar alcohols (inositol, mannitol and xylitol) and sugar acid (galacturonic acid). In our study, sucrose (2,856.2 mg/100 g DM) and fructose (534.1 mg/100 g DM) were found to be dominant in the sugars, which accounted for 77% and 14% of the total sugars (3,702.8 mg/100 g DM), respectively. To date, no report has been published on the contents and profiles of sugars in RCS despite their postive effect on gut microbe modulation and immunological properties. For instance, arabinosyl substitutional position and ratio in the xylan backbone contributes to the health-beneficial properties of arabinoxylans[24]. The arabinose to xylose (A/X) ratios of the study RCS showed that RCS can be a potential source of arabinoxylans. Arabinoxylan oligosaccharides have prebiotic functions, and they can positively modulate human gut health by promoting the growth of beneficial gut microbes and suppressing pathogenic gut microbes. The mechanism is that oligosaccharides are not digested in the upper gastrointestinal tract but are only fermented in the colon. Short chain fatty acids produced by oligosaccharides during gut fermentation decrease the pH in the colon, which inhibits the growth of pathogenic gut microbes[24,25]. Additionally, the presence of mannose showed the potential of RCS in nutraceutical and pharmaceutical applications due to its health-promoting functions of relieving constipation, upregulating blood lipid metabolism and reducing blood cholesterol and triglyceride levels[26].

      Table 1.  Content of proteins, lipids, sugars, and organic acids in RCS.

      CompositionContent
      Dry matter (%)88.6 ± 0.1
      Proteins (g/100 g DM)6.9 ± 0.1
      Lipids (g/100 g DM)7.0 ± 1.1
      Sugars (mg/100 g DM)3,702.8 ± 632.7
      Arabinose40.9 ± 2.8
      Xylose13.1 ± 2.1
      Xylitol26.1 ± 4.0
      Fructose534.1 ± 96.3
      Glucose114.9 ± 19.3
      Mannose11.4 ± 1.5
      Mannitol29.4 ± 4.2
      Inositol38.6 ± 5.5
      Sucrose2,856.2 ±496.5
      Maltose31.4 ± 4.3
      Galacturonic acid6.7 ± 0.4
      Arabinose to xylose ratio3.2 ± 0.6
      Organic acids (mg/100 g DM)50.0 ± 4.5
      Malic acid14.0 ± 2.7
      Citric acid36.1 ± 1.9
    • The phenolic compounds were characterized by MS and MS2 (Table 2). The identification was based on the fragmentation pattern, the retention time and UV absorption spectra, and the comparison of those with the commercial standards and relevant literatures. Two flavanonols (taxifolin, taxifolin hexoside), eight flavonols (quercetin hexoside 1, quercetin hexoside 2, quercetin hexoside 3, quercetin 3-O-galactoside, quercetin 3-O-glucoside, quercetin, kaempferol and isorhamnetin), one isoflavone (formononetin coumaroyl hexoside) were identified or preliminarily identified in RCS. In addition, two unknown phenolic compounds were detected.

      Table 2.  Identification of phenolic compounds in the RCS by UPLC-DAD-ESI-QTOF*.

      No.IdentificationContent
      (mg/100 g DM)
      UV λmax (nm)[M+Na]+/[M+H]+/[M-H] (m/z)MS2 (m/z)
      Total phenolic compounds228.4 ± 51.0
      Flavanonols102.2 ± 22.0
      1Taxifolin hexoside60.0 ± 13.3288489.1321/467.1505/465.1069467.1505 → 305.0989, 260.0501, 139.0380
      465.1069 → 303.0518, 285.0411
      2Taxifolin42.2 ± 8.7289327.0443/305.0623/303.0533305.0623 → 287.0522, 259.0577, 231.0629, 195.0270, 153.0171, 149.0221
      303.0533 → 285.0416, 275.0568, 259.0618, 241.0515, 217.0515, 178.9993, 153.0199, 125.0257
      Flavanols101.6 ± 24.2
      4Quercetin hexoside 12.7 ± 0.6255, 371487.0793/465.0975/463.0895465.0975→ 303.0467
      463.0895→ 301.0365, 151.0048
      5Quercetin 3-O-galactoside1.1 ± 0.2255, 353−/465.0978/463.0893465.0978 → 303.0465
      463.0893 → 300.0287, 271.0258, 255.0309, 243.0309
      6Quercetin 3-O-glucoside11.7 ± 3.0255, 353487.0801/465.0974/463.0903465.0974 → 303.0469
      463.0903 → 300.0286, 271.0257, 255.0307, 243.0309
      7Quercetin hexoside 219.2 ± 4.5252, 365487.0799/465.0977/463.0887465.0977 → 303.0471
      463.0887 → 301.0355
      9Quercetin hexoside 31.3 ± 0.2270, 369487.0792/465.0978/463.0888465.0978 → 303.0465
      463.0888 → 301.0367
      10Quercetin63.1 ± 15.2255, 370325.0286/303.0471/301.0377303.0471 → 229.0473, 153.0169
      301.0377 → 273.0402, 178.9992, 151.0044
      12Kaempferol1.2 ± 0.2270, 365−/287.0523/285.0418
      13Isorhamnetin1.4 ± 0.3260, 371−/317.0624/315.0521
      Isoflavones1.6 ± 0.4
      11Formononetin malonyl- hexoside
      1.6 ± 0.4265, 310−/517.1283/515.1208517.1283 → 269.0786
      1,031.2513 ([2M-H]) → 515.1230, 267.0677
      Unknown compounds23.0 ± 4.5
      3Unknown compound 15.2 ± 0.9270427.1527/405.1711/403.1626427.1527 → 265.1018, 203.0510
      403.1626 → 223.0989, 179.1088
      8Unknown compound 217.8 ± 4.2275, 320643.2288/−/619.2421643.2288 → 449.1734
      619.2421 → 193.0513, 178.0277, 149.0612
      * Quantitative results are shown as means ± standard deviation of triplicate analyses. '−' means the MSMS spectrum was not provided.

      Table 2shows the total content of preliminarily identified phenolic compounds (228.4 mg/100 g DM) in the study seeds. The most abundant groups of phenolic compounds were flavanonols (102.2 mg/100 g DM) and flavonols (101.6 mg/100 g DM), the sum of both accounted for 89% of the total content. Taxifolin hexoside (60.0 mg/100 g DM) and quercetin (63.1 mg/100 g DM) presented the dominant roles in the groups of flavanonols and flavonols, respectively. The isoflavone (formononetin coumaroyl hexoside, 1.6 mg/100 g DM) was found at lower content compared to the other phenolic compounds. Our results suggested lower levels of quercetin, kaempferol, isorhamnetin in comparison than the previous data (671.4, 7.9 and 2.9 mg/100 g DM, respectively) of RCS in the research of Chiriac et al.[19]. This variation was probably caused by different extraction procedures. In their study, phenolic compounds were extracted from RCS with assistance of microwave for 15 min. In addition, Prati et al. reported the presence of taxifolin and taxifolin hexoside in seed extract of red clover, though the content of individual compounds were not determined[10]. The anti-cancer properties of flavanonols and flavanols, and their role in reducing the risk of cardiovascular disease have been well documented[8]. They also showed strong antioxidant and antimicrobial effects[8, 14]. Biochanin A, genistein, daidzein and formononetin, along with their respective glucosides were the common isoflavonoids found in red clover leaves[27]. Interestingly, in our study, only one isoflavonoid glycoside (formononetin coumaroyl hexoside) was observed in RCS, and the content was low. Therefore, the distribution and quantity of isoflavonoids varied in different parts of red clover. Considered as potent phytoestrogens, isoflavonoids have been used as the bioactive compounds for treating hormone-related ailments such as post-menopausal symptoms, pre-menstrual syndrome and dysmenorrhea[28]. Additionally, decoctions of red clover flower and leaves have been orally consumed as ethnomedicines to reduce inflammation as well as to treat chest pain and chronic rheumatism[28].

    • The total lipid content in our study RCS, as determined by a modified Folch method was 7.0 g/100 g DM (Table 3), which was consist with the result (9.8 g/100 g DM) of the research by Ahmed et al.[14]. Nineteen fatty acids were identified in our seeds with a total content of 46,760.5 µg/g DM, in which polyunsaturated fatty acids were 79%. Polyunsaturated linoleic acid took the dominant role, which accounted for 72% of the total fatty acids. Consistent to our findings, linoleic acid presented as the dominant fatty acid in the seeds of Trifolium pratense (44%), Trifolium repens (65%) and Trifolium resupinatum (52%) in previous researches[14, 29]. Comparatively, linoleic acid was observed with low content (5%−11%) in five other Trifolium species (T. balansae, T. stellatum, T. nigrescens subsp. petrisavi, T. constantinopolitanum, and T. resupinatum var. resupinatum)[12]. As essential fatty acids, high content of linoleic acid in our study samples confers the seeds significant nutritional value. For instance, linoleic acid is the precursor of a number of potent pro-inflammatory mediators, including prostaglandins and leukotrienes, which has led to the development of anti-inflammatory pharmaceuticals[30]. The total content of unsaturated fatty acids in the study RCS was 3.9 folds higher than saturated fatty acids. Saturated fatty acids such as palmitic acid (14%) and steric acid (6%) in RCS have the potential of enhancing the oxidative stability in food products. As reported in previous studies, palmitic acid and stearic acid can be used in solid fat applications, such as margarine, shortening and confectionary industries[31].

      Table 3.  Composition of and content of lipids, fatty acids and tocopherols in RCS*.

      CompositionContent
      Lipids (g/100 g DM)7.0 ± 1.1
      Fatty acids (µg/g DM)46,760.5 ± 2,686.1
      Myristic acid (C14:0)58.3 ± 6.7
      Palmitic acid (C16:0)6,369.6 ± 338.9
      Stearic acid (C18:0)2,993.1 ± 149.8
      Arachidic acid (C20:0)65.7 ± 3.2
      Behenic acid (C22:0)55.7 ± 3.6
      Lignoceric acid (C24:0)48.1 ± 21.6
      Total SFA9,590.5 ± 507.2
      Palmitoleic acid (16:1 n-7)45.8 ± 2.2
      Oleic acid (C18:1 n-9)178.9 ± 4.6
      Vaccenic acid (18:1 n-7)7.6 ± 2.7
      Eicosenoic acid (20:1 n-9)23.3 ± 4.4
      Erucic acid (22:1 n-9)43.0 ± 8.5
      Nervonic acid (24:1 n-9)8.0 ± 4.6
      Total MUFA310.5 ± 10.4
      Linoleic acid (18:2 n-6)33,769.6 ± 1,971.1
      α-Linolenic acid (18:3 n-3)2,863.8 ± 191.1
      γ-Linolenic acid (18:3 n-6)24.2 ± 5.5
      Eicosadienoic acid (20:2 n-6)2.9 ± 0.4
      Dihomo-α-linolenic acid (20:3 n-3)5.7 ± 1.4
      Arachidonic acid (20:4 n-6)5.7 ± 1.5
      Eicosapentaenoic acid (20:5 n-3)187.6 ± 37.0
      Total PUFA36,859.5 ± 2,188.1
      Total n-33,057.1 ± 219.0
      Total n-633,802.4 ± 1,969.7
      Tocopherols, mg/100 g DM94.9 ± 4.4
      α-Tocopherol91.9 ± 4.2
      β-Tocopherol1.3 ± 0.1
      γ-Tocopherol1.7 ± 0.1
      δ-Tocopherol
      * Results are shown as means ± standard deviation of triplicate analyses. '−' means the compound was not detected in the sample.
    • Three tocopherols (α-, β- and γ-tocopherol) were identified in RCS, with a total content of 94.9 mg/100 g DM. Alfa-tocopherol accounted for 97% (91.9 mg/100 g DM), γ-tocopherol for 2% and β-tocopherol for 1%. Our results in Table 3 suggested the same level of tocopherols as that in previous RCS research (101.7 mg/100 g DM)[14]. Alfa-tocopherol is an important antioxidant for humans with positive effects on the immune system. Alfa-tocopherol-rich RCS showed potential in being an antioxidant and anti-inflammatory agent in disease management[32].

    • Using SFE, the oil yield from RCS was 3.7% on a dry weight basis, which was significantly lower than the yield (7%) obtained with the Folch method from the same batch of RCS in the laboratory. This difference could have been ascribed to the lower efficiency of CO2 for extracting polar lipids. Increasing the pressure for SFE could potentially result in a higher yield of oil.

      Higher oil yields were reported in Soxhlet-extracted underutilized oilseed crop species of the Fabaceae (Leguminosae) family (19%−36%, dry wt. basis), which included Bauhinia purpurea, Phanera vahlii, Butea monosperma, Caesalpinia crista, Gliricidia sepium, Mimosa pudica and Millettia pinnata[33]. Environmental and genetic factors as well as the extract methods can significantly affect the oil yield, fatty acid composition and physiochemical properties of the oil[34].

    • Extracted by SFE, the oil quality remained almost unchanged in terms of the profiles of fatty acids. As shown in Table 4, linoleic acid was identified as the major component in RCS oil. Linoleic acid was reported as a common fatty acid in the seed oil of many Fabaceae (Leguminosae) family plants[12, 35]. Omega-6 fatty acids are acknowledged as bioactive lipids with health-beneficial properties, in which linoleic acid accounts for 85%−90% of the dietary intake amount[36]. RCS oil contained rich unsaturated fatty acids such as linoleic acid (61%), oleic acid (16%) and α-linolenic acid (6%), with 91% of which are omega-6 fatty acids. Although essential oil were extracted from red and white clover flowers, our study is the first one showing RCS can be a source for producing oil which is rich in n-6 polyunsaturated fatty acids (PUFAs)[37].

      Table 4.  Composition and content of fatty acids and tocopherols in RCS oil*.

      CompositionContent (mg/100 g)% in the
      fatty acids
      Fatty acids89,482.4 ± 2,515.0
      Myristic acid (C14:0)84.7 ± 1.00.09
      Palmitic acid (C16:0)9,445.0 ± 394.810.55
      Stearic acid (C18:0)4,955.4 ± 207.35.54
      Arachidic acid (C20:0)137.2 ± 13.00.15
      Behenic acid (C22:0)27.4 ± 4.70.03
      Lignoceric acid (C24:0)63.0 ± 7.60.07
      Total SFA14,712.7 ± 599.016.44
      Palmitoleic acid (16:1 n-7)81.2 ± 3.80.09
      Oleic acid (C18:1 n-9)14,036.8 ± 2,598.015.69
      Eicosenoic acid (20:1 n-9)59.6 ± 9.20.07
      Erucic acid (22:1 n-9)40.4 ± 4.70.05
      Nervonic acid (24:1 n-9)5.0 ± 1.70.01
      Total MUFA14,223.0 ± 2,607.415.89
      Linoleic acid (18:2 n-6)54,665.6 ± 2,159.761.09
      α-Linolenic acid (18:3 n-3)5,432.4 ± 81.86.07
      γ-Linolenic acid (18:3 n-6)31.2 ± 3.60.03
      Eicosadienoic acid (20:2 n-6)74.2 ± 51.40.08
      Dihomo-α-linolenic acid (20:3 n-3)5.4 ± 2.20.01
      Arachidonic acid (20:4 n-6)5.4 ± 0.80.01
      Eicosapentaenoic acid (20:5 n-3)332.3 ± 5.30.37
      Total PUFA60,546.6 ± 2,258.067.66
      Total n-35,770.1 ± 88.36.45
      Total n-654,776.5 ± 2,169.761.21
      Tocopherols (mg/100 g)40.0 ± 5.9
      α-Tocopherol38.9 ± 5.8
      β-Tocopherol0.7 ± 0.1
      γ-Tocopherol0.5 ± 0.1
      δ-Tocopherol
      * Results are shown as means ± standard deviation of triplicate analyses. '−' means the compound was not detected in the sample.
    • The tocopherols identified in RCS and SFE-extracted RCS oil followed the same pattern. The total content of tocopherols investigated in the present study was 40.0 mg/100 g oil, of which α-tocopherol was predominant (97%), followed by β-tocopherol (2%) and γ-tocopherol (1%). The tocopherol content of seed oils from five Fabaceae species (Vigna angularis, Phaseolus vulgaris, Phaseolus lunatus, Phaseolus coccineus and Glycine soja,) were reported in the range of 1.5−26.7 mg/100 g oil[36]. Gamma-type was the most abundant tocopherol of all five Fabaceae species. Compared to these legume species, our study showed that the RCS oil extracted by SFE contained higher levels of total tocopherols and different profiles of tocopherols, being a richer source of α-tocopherol.

    • In our study, 44 phytochemicals were putatively identified in RCS and RCS oil, which belonged to more than 13 different chemical classes. In terms of the number of individual compounds, lipids were the most abundant class with 16 compounds annotated, followed by phenolic compounds (n = 9), terpenoid (n = 6) and phytosterol (n = 4). As shown in Fig. 1, the metabolites in RCS and the SFE-extracted oil were arranged with hierarchical clustering based on their abundance.

      Figure 1. 

      Heatmap of correlation between phytochemical compounds present in RCS and SFE-extracted oil.

      Formononetin, medicarpin and homopterocarpin were identified in the isoflavonoids group in our study. Medicarpin showed significant antiangiogenic and cytotoxic activity, and homopterocarpin has hepatoprotective and antioxidant potential[27, 38, 39]. The presence of these isoflavonoids suggests an interesting potential application of RCS extract and RCS oil in cancer therapy. Unlike the isoflavonoids equally presented in RCS and its oil, other phenolic compounds such as flavonoids (genkwanin, 6,3'-dimethoxyflavone, isoanhydroicaritin), and phenolic acid (erionic acid E) were observed to be more abundant in RCS oil than RCS. Modern pharmacological studies have shown that these compounds were of high importance due to their functions of anti-inflammatory, antioxidant and anti-anxiety[4042].

      Polyunsaturated fatty acids was observed to be more abundant in the SFE-extracted RCS oil than in the RCS in our study. These dietary omega-6 fatty acids have important health benefits since they can reduce blood cholesterol levels[15]. Additionally, monoacylglycerols (glycerol 1-palmitate, 2-monoolein, glyceryl monostearate) were commonly found in vegetable oils. A previous study reported them to be capable of improving loaf volume and texture, although they were not known for being particularly healthy[43]. In our study, phytosterols (19_norandrosterone, stigmasta-4,22-dien-3-one and ergosterol) were dominant in RCS oil, whereas stigmasta-5,22-dien-3-one was found to be more abundant in the RCS. In previous research, phytosterols exerted beneficial hypolipidemic function, as well as anti-cancer, anti-inflammatory, anti-photoaging, anti-osteoarthritic, immunomodulatory, hepatoprotection and antioxidative activities[44].

      Terpenoid-metabolites are also compounds with remarkable biological properties such as anti-inflammatory, hepatoprotective, and anti-cancer effects[45]. A wide range of terpenoids were found in RCS and RCS oil including diterpenes (6-Oxocativic acid, daniellic acid, 7b,9-Dihydroxy-3-(hydroxymethyl)-1,1,6,8-tetramethyl-5-oxo-1,1a,1b,4,4a,5,7a,7b,8,9-decahydro-9aH-cyclopropa[3,4]benzo[1,2-e]azulen-9a-yl acetate), triterpene derivatives (panaxadiol, glycyrrhetic acid, methyl ester) and tetraterpene (9-epiblumenol B) indicating their potential for pharmaceutical application.

      Furthermore, several compounds with health-promoting properties were determined in our study seeds and oil (Fig. 1). Trigonelline as an alkaloid, was reported to be effective in the hypoglycemic, hypocholesterolemic, antitumor, antimigraine, or antiseptic treatments[46]. Loliolide shows potential in the treatment of patients with obesity as a lipid-lowering agent[47]. Choline is an essential nutrient that is naturally present in some foods and commercially available as a dietary supplement. Choline plays an important role in maintaining the health of the nervous system and in the development of normal brain functions[48]. 4-Methyl-5-thiazoleethanol has pharmacological and biological activities[49]. Previous pharmacological studies showed that andrograpanin exerted anti-inflammatory activity through down-regulating the p38 mitogen-activated protein kinase (MAPKs) signaling pathways[50].

    • To the best of our knowledge, our research is the first study to investigate the phytochemical profiles in RCS and SFE-extracted RCS oil using both targeted and non-targeted analytical methods, which provided a broader image of phytochemical profiles in the study materials. Forty-four phytochemicals were putatively identified in RCS and RCS oil, mainly including lipids, phenolic compounds, terpenoids and phytosterols. Abundant phenolic compounds (including flavanonols, flavonols and isoflavones) were observed in RCS and SFE-extracted oil. Although the oil yield of RCS is not as high as other oilseed crop species in the Fabaceae (Leguminosae) family, the RCS oil has proven to be a rich source of unsaturated fatty acids (mainly linoleic acid and oleic acid) with a high content of α-tocopherol, which has potential for food and pharmaceutical uses.

      • The study is supported by the Finland-China Food and Health Network (funded by the Finnish Ministry of Education and Culture), FOODNUTRI-Climate Smart Food and Nutrition Research Infrastructure (funded by the Research Council of Finland, No. 337980) and the FIRI 2021 call: Non-roadmap research infrastructures as part of the EU Recovery and Resilience Facility (No. 345916).

      • The authors confirm contribution to the paper as follows: conceptualization: Zhou Y, Tian Y, Selander K, Yang B; methodology: Zhou Y, Tian Y; investigation: Zhou Y, Tian Y, Ollennu-Chuasam P; data curation: Zhou Y, Tian Y; formal analysis and visualization: Zhou Y; writing-original draft: Zhou Y; writing-review & editing: Zhou Y, Tian Y, Kortesniemi M, Selander K, Väänänen K, Yang B; resources: Selander K, Väänänen K; funding acquisition and supervision: Yang B. All authors reviewed the results and approved the final version of the manuscript.

      • All data generated or analyzed during this study are included in this published article and its supplementary information files

      • The authors declare that they have no conflict of interest. Baoru Yang is the Editorial Board member of Food Innovation and Advances who was blinded from reviewing or making decisions on the manuscript. The article was subject to the journal's standard procedures, with peer-review handled independently of this Editorial Board member and the research groups.

      • Authors contributed equally: Ying Zhou, Ye Tian

      • Copyright: © 2024 by the author(s). Published by Maximum Academic Press on behalf of China Agricultural University, Zhejiang University and Shenyang Agricultural University. This article is an open access article distributed under Creative Commons Attribution License (CC BY 4.0), visit https://creativecommons.org/licenses/by/4.0/.
    Figure (1)  Table (4) References (50)
  • About this article
    Cite this article
    Zhou Y, Tian Y, Ollennu-Chuasam P, Kortesniemi M, Selander K, et al. 2024. Compositional characteristics of red clover (Trifolium pratense) seeds and supercritical CO2 extracted seed oil as potential sources of bioactive compounds. Food Innovation and Advances 3(1): 11−19 doi: 10.48130/fia-0024-0002
    Zhou Y, Tian Y, Ollennu-Chuasam P, Kortesniemi M, Selander K, et al. 2024. Compositional characteristics of red clover (Trifolium pratense) seeds and supercritical CO2 extracted seed oil as potential sources of bioactive compounds. Food Innovation and Advances 3(1): 11−19 doi: 10.48130/fia-0024-0002

Catalog

  • About this article

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return