Abstract
Metabolome analysis of flavored vegetables, green spring onion (Allium fistulosum), Chinese chive (A. tuberosum), and their interspecies hybrid Negi-Nira chive, was conducted using liquid chromatography-Fourier transform ion cyclotron resonance-mass spectrometry, with ca. 2 ppm mass accuracy. Ion peaks in the chromatograms of four biological replicates of the vegetable leaves were processed using the alignment software PowerGet for metabolite comparison, from which we obtained the potential chemical formulae. In total, 860 ion peaks were reproducibly detected; of these, 506, 525, and 336 peaks were found in the hybrid, A. tuberosum, and A. fistulosum, respectively. There were 130 peaks specific to the hybrid; from these, 31 metabolites were annotated by searching compound databases. The sulfur-containing compounds and flavonoids were further analyzed using bioinformatics, to examine the sulfur metabolism of Allium volatiles and the flavonoid pathways in these species. In conclusion, our metabolome analysis of this interspecies hybrid and its parents provides a unique opportunity to elucidate their metabolic background.
Keywords: Fourier transform ion cyclotron resonance-mass spectrometry, interspecies hybrid, metabolome
The genus Allium includes economically important vegetables such as onions, garlic, leeks, and chives, which possess characteristic aromas, and which are cultivated extensively worldwide. Green spring onion (A. fistulosum) and Chinese chive (A. tuberosum), known as “Negi” and “Nira” in Japanese, have a long history of cultivation as flavored vegetables in Japan. Various cultivars of these vegetables are used in Japanese cuisine. Their aromatic compounds, especially sulfides, have been well studied (Yoshimoto and Saito 2019).
Interspecies crossing of A. fistulosum and A. tuberosum was achieved in 1991 at Tochigi Agricultural Experimental Station (Tochigi, Japan) (Amagai et al. 1995, in Japanese): when A. tuberosum was pollinated using A. fistulosum pollen, the resulting seeds contained genetic material only from the A. tuberosum parent, due to apomixis (Kojima et al. 1991; Nakazawa et al. 2006). However, A. fistulosum pollinated by A. tuberosum produced no seeds (Amagai et al. 1995): the embryos from 1,276 ovules of these A. fistulosum specimens were cultivated under sterile conditions on callus induction medium, and the calluses were placed on regeneration medium: two calluses obtained from pollination of A. fistulosum ‘Nissato’ using pollen from A. tuberosum ‘Kinumidori’ (an F1 hybrid cultivar) regenerated whole plants, one of which grew normally. This plant had 24 chromosomes, while the A. tuberosum ‘Kinumidori’ cultivar has 32 chromosomes (tetraploid), and the A. fistulosum cultivar has 16 (diploid); further, its esterase isozymes exhibited a mixed pattern of the isozymes of its parents (Amagai et al. 1995), thereby confirming that this plant, named “Negi-Nira chive” as A. fistulosum×A. tuberosum, was an interspecies hybrid between its parent cultivars. The overall morphological phenotype of the hybrid was more similar to that of the paternal parent A. tuberosum, although with the hollow leaf cross section that is characteristic of A. fistulosum (Figure 1). The hybrid was registered as a new type of “Nira” cultivar, and was named ‘Nakamidori,’ in accordance with the plant variety protection system of Japan. The cultivar is sterile, and has been propagated vegetatively for commercial purposes, mainly by farmers of Tochigi Prefecture (Japan). This hybrid has a garlic-like flavor, unlike its parental cultivars.
Figure 1. Three Allium species were grown in the same greenhouse for three months after seeding. Left: Chinese chive, Allium tuberosum ‘Kinumidori’; middle: Negi-Nira chive, A. fistulosum×A. tuberosum ‘Nakamidori’; right: Green spring onion, A. fistulosum ‘Nissato.’ Each inset shows the cross section of a leaf of the cultivar.

To compare the aromatic composition of the interspecies hybrid and its parental cultivars, Kobayashi et al. (1997) analyzed the leave volatiles using gas chromatography-mass spectrometry (GC-MS): the chromatographic intensities of diallyl disulfide and other sulfur-containing compounds were much higher for the hybrid than for its parents, which may explain the garlic-like aroma in the hybrid. There were 41 volatiles, including five unknown chemical peaks; 11 volatiles were common to the hybrid and parental cultivars; 25 were specific to the hybrid and A. tuberosum; and four were specific to the hybrid and A. fistulosum. Methyl (Z)-1-propenyl disulfide was specific to the hybrid; this suggests that the mixing of the genomes of the parents produced sulfur metabolism in the hybrid that was different from that in the parents.
To gain an insight into the metabolism of the hybrid in comparison with those of its parents, we conducted metabolome analysis using liquid chromatography coupled with FT/ICR mass spectrometry (LC-MS) and bioinformatics. The cultivars A. fistulosum ‘Nissato,’ A. tuberosum ‘Kinumidori’ and A. fistulosum×A. tuberosum ‘Nakamidori’ were grown in a greenhouse of Tochigi Agricultural Experimental Station (Figure 1). Mature leaves were harvested (four biological replicates per cultivar), immediately frozen in liquid nitrogen, and then kept at −80°C until analysis. The frozen leaves were homogenized into powder using a mortar and pestle in liquid nitrogen. We then mixed 200 mg of the powder with 600 µl of 80% methanol containing 25 μM of 7-hydroxy-5-methylflavone as an internal standard. A Mixer Mill MM300 (QIAGEN, Valencia, CA, USA) was used to homogenize the mixture at 25 Hz for 2 min twice. After centrifugation (20,000×g, 5 min, 4°C), the supernatant was filtered through a 0.2 µm pore-membrane filter (Millipore, Bedford, MA, USA).
A liquid chromatography system (Agilent 1100, Agilent, Palo Alto, CA, USA) coupled with a Fourier transform ion cyclotron resonance (FT/ICR) mass spectrometer (LTQ-FT, Thermo Fisher Scientific, Waltham, MA, USA) was used for the LC-MS analysis of the methanol extracts as described by Iijima et al. (2008), but with modified LC elution: briefly, the extracts (20 µl each) were analyzed using LC-MS and a C18 reverse-phase column (Tosoh TSKgel ODS-100 V 5 µm, 4.6×250 mm, Tosoh Co., Ltd., Tokyo, Japan), under linear gradient elution of acetonitrile (3–97% for 45 min, 97% for 5 min, 97 to 3% for 0.1 min and then 3% for 7 min), in 0.1% formic acid. Mass detection was performed in the electrospray ionization (ESI) positive mode for a mass range of 100–1,500 m/z with a resolution of 100,000. The detailed analytical conditions are available at the metabolome metadata database Metabolonote (http://metabolonote.kazusa.or.jp/SE43:/) (Ara et al. 2015).
Total ion chromatogram of the hybrid was not simply summation of those of its parents, either qualitatively or quantitatively (Figure 2). Although many of the ion peaks of the parents occurred in the chromatogram of the hybrid, the hybrid had several unique peaks. It seems unlikely that the relative intensities of the ions occurring in the hybrid reflect those of its parents when they were comparable ions. However, the chromatogram of the hybrid generally resembles that of the paternal A. tuberosum, rather than that of A. fistulosum. Consistent with this finding, the gas chromatograms of this hybrid and A. tuberosum were similar (Kobayashi et al. 1997). These results reflect that the metabolism of the hybrid is influenced by complex interactions between the parents’ genomes.
Figure 2. Total ion chromatograms of leaf metabolites of Allium tuberosum ‘Kinumidori,’ A. fistulosum×A. tuberosum ‘Nakamidori,’ and A. fistulosum ‘Nissato.’ Methanol leaf extracts were separated using liquid chromatography, and the relative intensity of each ion peak was recorded at each time point.

Twelve chromatograms were generated (from the four biological replicates for three cultivars). The ion peaks were extracted and aligned using ion information with the m/z value and retention time of each ion by the software PowerGet (Sakurai et al. 2014). A possible chemical formula (or formulae) for each ion was calculated from its m/z value. Ions that occurred reproducibly in all four replicates in any of the cultivars were analyzed further. In total, 860 reliable ion peaks were selected, and are depicted in a Venn diagram: 506, 525, and 336 ions occurred in the hybrid, A. tuberosum, and A. fistulosum, respectively; 130 ions were specific to the hybrid (Figure 3). For the hybrid-specific ion peaks, we then used the calculated mass values to search the compound databases, KEGG (https://www.genome.jp/kegg/), KNApSAcK (http://www.knapsackfamily.com/KNApSAcK/), LIPIDMAPS (https://www.lipidmaps.org/), and HMDB (http://www.hmdb.ca/) yielded 31 matching metabolites, which belonged to metabolic categories including aminocarboxylic acids, fatty acid derivatives, flavonoids, iridoids, phenolics, and steroids (Table 1). Although none of the predicted molecules was identified by using authentic chemicals on the same chromatographic conditions, this metabolomic profile information helps to elucidate the metabolism of Allium species. The datasets analyzed here are publicly available at Metabolonote (http://metabolonote.kazusa.or.jp/SE43:/).
Figure 3. Venn diagram of the 860 ion peaks found in the interspecies hybrid, A. fistulosum×A. tuberosum, its maternal parent A. fistulosum, and paternal parent A. tuberosum. The numbers of putative sulfides are shown in parentheses.
Table 1. Metabolite annotation found specific in the interspecies hybrid, Negi-Nira chive (A. fistulosum×A. tuberosum).
| No. | Retention time (min) | Detected m/z | Adducts | Chemical category | Annotation |
|---|---|---|---|---|---|
| 1 | 6.5 | 231.134 | [M+H]+ | Aminocarboxylic acids | Peptide C10H18N2O4 |
| 2 | 17.9 | 233.15 | [M+H]+ | Aminocarboxylic acids | Peptide C10H20N2O4 |
| 3 | 33.7 | 330.264 | [M+H]+ | Fatty acid derivatives | Fatty acid derivative C18H35NO4 |
| 4 | 34.4 | 348.274 | [M+H]+ | Fatty acid derivatives | Fatty acid derivative C18H37NO5 |
| 5 | 39.1 | 227.128 | [M+H]+ | Fatty acid derivatives | Fatty acid derivative C12H18O4 |
| 6 | 15.8 | 873.204 | [M+H]+ | Flavonoids | Flavonoid (trimer) C47H36O17 |
| 7 | 17.3 | 1,111.312 | [M+H]+ | Flavonoids | Flavonoid (+4Hex) C49H58O29 |
| 8 | 17.8 | 771.198 | [M+H]+ | Flavonoids | Flavonoid (+2 or 3Hex) C33H38O21 |
| 9 | 17.9 | 757.218 | [M+H]+ | Flavonoids | Flavonoid (+3Hex) C33H40O20 |
| 10 | 18.8 | 697.161 | [M+H]+ | Flavonoids | Flavonoid (+2Hex) C30H32O19 |
| 11 | 19.4 | 757.219 | [M+H]+ | Flavonoids | Flavonoid (+3Hex) C33H40O20 |
| 12 | 22.3 | 595.166 | [M+H]+ | Flavonoids | Flavonoid (+2Hex) C27H30O15 |
| 13 | 23.2 | 625.176 | [M+H]+ | Flavonoids | Flavonoid (+2Hex) C28H32O16 |
| 14 | 24.6 | 757.198 | [M+H]+ | Flavonoids | Flavonoid (+2Hex) C36H36O18 |
| 15 | 24.9 | 787.208 | [M+H]+ | Flavonoids | Flavonoid (+2Hex) C37H38O19 |
| 16 | 25.6 | 609.181 | [M+H]+ | Flavonoids | Flavonoid (+2Hex) C28H32O15 |
| 17 | 27.9 | 595.145 | [M+H]+ | Flavonoids | Flavonoid (+0 or 1Hex or dimer) C30H26O13 |
| 18 | 28.2 | 625.155 | [M+H]+ | Flavonoids | Flavonoid (+1Hex) C31H28O14 |
| 19 | 17.8 | 421.17 | [M+H]+ | Iridoids | Iridoid (+1Hex) C18H28O11 |
| 20 | 20.4 | 391.16 | [M+H]+ | Iridoids | Iridoid (+1Hex) C17H26O10 |
| 21 | 19.1 | 558.255 | [M+H]+ | Phenolics | Phenolic C27H35N5O8 |
| 22 | 22.2 | 269.102 | [M+H]+ | Phenolics | Phenolic C13H16O6 |
| 23 | 22.2 | 431.155 | [M+H]+ | Phenolics | Phenolic (+1 or 2Hex) C19H26O11 |
| 24 | 22.6 | 443.155 | [M+H]+ | Phenolics | Phenolic (+1Hex) C20H26O11 |
| 25 | 24.2 | 627.244 | [M+H]+ | Phenolics | Phenolic C33H38O12 |
| 26 | 26.6 | 503.212 | [M+H]+ | Phenolics | Phenolic (+2Hex) C23H34O12 |
| 27 | 33.6 | 269.102 | [M+H]+ | Phenolics | Phenolic C13H16O6 |
| 28 | 35.8 | 303.123 | [M+H]+ | Phenolics | Phenolic C17H18O5 |
| 29 | 36.9 | 253.107 | [M+H]+ | Phenolics | Phenolic C13H16O5 |
| 30 | 41.3 | 273.112 | [M+H]+ | Phenolics | Phenolic C16H16O4 |
| 31 | 33.7 | 769.4 | [M+H]+ | Steroids | Steroid (+3Hex) C39H60O15 |
Abbreviations; Hex: hexose.
As sulfuric volatiles characterize Allium flavors (Kusano et al. 2016), we searched for sulfur-containing compounds among the predicted formulae, and annotated 18 sulfides, 19 disulfides, and 6 trisulfides; of these, 24 ion peaks were confirmed as 34S isotope-containing compounds, using the software MassChroViewer (Table 2) (Sakurai and Shibata 2017). Most of the predicted sulfuric compounds had m/z values of less than 240; some of these compounds might relate to flavor production. There were also two large disulfides (C26H45NO8S2 and C56H84O23S2).
Table 2. The list of predicted sulfur-containing compounds in the three vegetables.
| Detected vegitables | Formula of predicted sulfur-containing compounds |
|---|---|
| A. tuberosum | C9H16O2S2, C56H84O23S2* |
| A. fistulosum×A. tuberosum | C7H11NO4S, C7H12N2O3S*, C7H12N2O3S*, C10H13NO7S, C15H20N4O8S |
| A. fistulosum | C7H2N2O7S, C9H14OS2, C9H14OS3*, C9H14OS3*, C9H16O2S2*, C9H16O2S2*, C10H17N7O7S*, C24H23NO10S* |
| A. fistulosum and A. fistulosum×A. tuberosum | C6H11NO3S* |
| A. tuberosum and A. fistulosum×A. tuberosum | C4H8OS2, C4H9NOS2, C6H10OS2*, C6H10OS2*, C6H10OS2*, C6H11NOS2*, C6H11NOS2*, C6H11NOS2, C6H11NO2S2*, C7H14OS3*, C7H14OS3*, C7H14OS3*, C8H13NOS2*, C8H13NOS2*, C9H16OS3*, C19H25N3O7S, C21H24O13S, C21H24O13S, C26H45NO8S2, C33H44N2O17S* |
| All three vegitables | C4H11N7S, C6H11NO3S*, C9H16O2S2, C11H13NOS2, C14H23N3O8S, C19H45O4PS, C21H27N5O5S |
The same formula in the table mean the isomers that have distinct retention times on the chromatography. * 34S peak was confirmed by the software MassChroViewer.
There were three isomers of C6H10OS2 with distinct retention times on the chromatogram; one of these might be the main constituent of the flavor of garlic, namely allicin. C6H11NO3S, which occurred in all three cultivars, was annotated as alliin, a precursor to allicin, although the possibility that it indicated the presence of S-acetonylcysteine or other compounds cannot be excluded without further targeted chemical analysis. C14H23N3O8S occurred in all three cultivars; it was annotated as S-(2-Carboxypropyl)glutathione, a precursor peptide of the sulfur-containing flavor molecules. C6H11NO2S2, which was specific to the hybrid and A. fistulosum, was annotated as S-(Allylthio)-L-cysteine, a sulfur-containing amino acid. Future analysis targeting these compounds will clarify their chemical structures.
We were able to annotate five sulfides, including two isomers of C7H12N2O3S, and C7H11NO4S, C10H13NO7S, and C15H20N4O8S, which were specific to the hybrid. Methyl (Z)-1-propenyl disulfide (C4H8S2, M.W. 120.0067), which was found only in the hybrid when analyzed by GC-MS (Kobayashi et al. 1997), was not found in this study, probably due to the low sensitivity of FT/ICR mass spectrometry at the mass range of ca. 100 to 150 m/z. Interestingly, we annotated two molecules with the formula C9H14OS3; this is also the formula of ajoene, an antioxidant found in garlic (Naznin et al. 2010). These annotations of sulfides provide potential avenues for further study of the flavor of this hybrid.
We compared the distribution of the compounds annotated as flavonoids in the three cultivars (Supplementary Table S1). Interestingly, only C25H26O6 was common among the three cultivars. Three were common between the hybrid and A. fistulosum, and four between the hybrid and A. tuberosum. Thirteen were specific to the hybrid, suggesting that the genomic mixing results in significant alternation of flavonoid synthesis. Targeted analyses of flavonoids, using MS/MS/MS analysis, for example, will provide greater clarity about the effects of genomic mixing on flavonoid pathways in Allium.
This study provided a comparative metabolome profile of the interspecies hybrid and its parents. Although the chemical structure of each molecule cannot be described from the metabolome profiles that we present here, they nonetheless provide a basis for hypotheses about the potential metabolomic differences that can occur in hybrid genomes. In particular, knowing which molecules are found in the parents but not in the hybrid makes it possible to hypothesize about the origins of the molecules that are specific to the hybrid. Genomic and transcriptomic information about interspecies hybrids and their parents will stimulate future research in this regard.
Acknowledgments
This work was supported by a grant from the Kazusa DNA Research Institute Foundation.
Abbreviations
- GC-MS
gas chromatography-mass spectrometry
- LC-FT/ICR-MS
liquid chromatography-Fourier transform ion cyclotron resonance-mass spectrometry
Accession numbers
The raw data were deposited in the MassBase metabolome database (http://webs2.kazusa.or.jp/massbase/; accession numbers MDLC1_32601-32607 and 32611-32618), and are publicly available.
Supplementary Data
References
- Amagai M, Ohashi K, Kimura S, Oguri N, Kojima A (1995) Breeding of interspecies hybrid between Allium fitsulosum and A. tuberosum by embryo culture. Bull Tochigi Agr Exp Stn 43: 87–94 (in Japanese) [Google Scholar]
- Ara T, Enomoto M, Arita M, Ikeda C, Kera K, Yamada M, Nishioka T, Ikeda T, Nihei Y, Shibata D, et al. (2015) Metabolonote: A wiki-based database for managing hierarchical metadata of metabolome analyses. Front Bioeng Biotechnol 3: 38. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Iijima Y, Nakamura Y, Ogata Y, Tanaka K, Sakurai N, Suda K, Suzuki T, Suzuki H, Okazaki K, Kitayama M, et al. (2008) Metabolite annotations based on the integration of mass spectral information. Plant J 54: 949–962 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kobayashi A, Amagai M, Kubota K, Morisawa C (1997) Aroma composition of an interspecies hybrid cultivar named “Nakamidori” between Allium fistulosum and A. tuberosum by Embryo Culture. Nippon Nogeikagaku Kaishi 71: 1273–1279 (in Japanese) [Google Scholar]
- Kojima A, Nagato Y, Hinata K (1991) Degree of apomixis in Chinese chive (Allium tuberosum) estimated by esterase isozyme analysis. Japan J Breed 41: 73–83 [Google Scholar]
- Kusano M, Kobayashi M, Iizuka Y, Fukushima A, Saito K (2016) Unbiased profiling of volatile organic compounds in the headspace of Allium plants using an in-tube extraction device. BMC Res Notes 9: 133. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Nakazawa Y, Namai K, Kojima A, Kobayashi T, Tazaki K, Amagai M (2006) Estimating the inheritance for parthenogenesis in tetraploid Chinese Leek (Allium ramosum, syn. A. tuberosum). Ikushugaku Kenkyu 8: 89–98 (in Japanese) [Google Scholar]
- Naznin MT, Maeda T, Morita N (2010) Antioxidant functions of E- AND Z-Ajoene derived from Japanese garlic. Int J Food Prop 13: 821–829 [Google Scholar]
- Sakurai N, Ara T, Enomoto M, Motegi T, Morishita Y, Kurabayashi A, Iijima Y, Ogata Y, Nakajima D, Suzuki H, et al. (2014) Tools and databases of the KOMICS web portal for preprocessing, mining, and dissemination of metabolomics data. BioMed Res Int 2014: 1–11 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sakurai N, Shibata D (2017) Tools and databases for an integrated metabolite annotation environment for liquid chromatography-mass spectrometry-based untargeted metabolomics. Carotenoid Science 22: 16–22 [Google Scholar]
- Yoshimoto N, Saito K (2019) S-Alk(en)ylcysteine sulfoxides in the genus Allium: Proposed biosynthesis, chemical conversion, and bioactivities. J Exp Bot 70: 4124–4138 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.

