Polyamines are involved in the gynogenesis process in onion


  • Emmanuel Geoffriau,

    Corresponding author
    1. INRA, Station de Génétique et Amélioration des Plantes, Laboratoire de Physiologie et Culture in Vitro, B.P. 86510, 21065 Dijon Cedex, France
    2. Institut National d'Horticulture (INH), UMR GenHort, 2 rue le Nôtre, 49045 Angers, France.
    Search for more papers by this author
  • Rémi Kahane,

    1. INRA, Station de Génétique et Amélioration des Plantes, Laboratoire de Physiologie et Culture in Vitro, B.P. 86510, 21065 Dijon Cedex, France
    2. CIRAD-FLHOR, Boulevard de la Lironde, 34398 Montpellier Cedex 5, France.
    Search for more papers by this author
  • Josette Martin-Tanguy

    1. INRA, Station de Génétique et Amélioration des Plantes, Laboratoire de Physiologie et Culture in Vitro, B.P. 86510, 21065 Dijon Cedex, France
    Search for more papers by this author

  • Edited by J. Almeida

*e-mail: emmanuel.geoffriau@inh.fr


Haploidization in onion (Allium cepa L.) using immature flower buds simulates zygotic embryogenesis with no fecundation. In order to know the involvement of polyamines (PAs) in this process, we determined the concentration of endogenous PAs in flower buds and experimented the addition of various combinations of PA molecules in the medium. At the inoculation stage, high levels of free and conjugated spermidine and low putrescine + hydroxyputrescine/spermidine + spermine ratio characterized the highest responsive varieties. During in vitro culture, high levels of putrescine and its derivatives characterized the lowest responsive varieties, whereas high levels of spermidine and spermine characterized responsive varieties. The putrescine + hydroxyputrescine + homospermidine/spermidine + spermine ratio remained low in responsive varieties. The addition of spermidine or spermine (2 × 10−3 M) to the culture medium improved significantly the embryo production. Our results suggest that the arginine decarboxylase pathway is involved in PA biosynthesis during the in vitro culture of flower buds. Our study showed that specific ratios of PAs are required for successful gynogenesis in onion.

Abbreviations – 

arginine decarboxylase










ornithine decarboxylase














2,4-dichlorophenoxyacetic acid


Obtaining haploid plants is a valuable way of creating real homozygous lines in onion. Attempts through anther culture have failed (Campion et al. 1985, Keller 1990), and onion has been classified as an androgenesis recalcitrant species. For a decade, research has focused on gynogenesis (Bohanec 2002), through the use of irradiated pollen (Dore and Marie 1993), and mainly through immature flowers cultured in vitro. Progress has been made in controlling the induction conditions (Muren 1989, Campion et al. 1992, Bohanec et al. 1995), the growing conditions of the mother plants (Puddephat et al. 1999) and chromosome doubling (Campion et al. 1995, Geoffriau et al. 1997a, Jakse et al. 2003). However, the rate of gynogenic embryo induction per flower explant still remained low and strongly genotype dependent (Bohanec et al. 2003, Geoffriau et al. 1997b, Michalik et al. 2000).

The endogenous concentrations of polyamines (PAs) in plants are high in flowers (Applewhite et al. 2000, Kakkar and Rai 1993, Rastogi and Sawhney 1990, Slocum and Galston 1985). In strawberry and chrysanthemum, PA conjugates accumulate upon the transformation of vegetative bud into a floral bud (Aribaud and Martin-Tanguy 1994b, Tarenghi and Martin-Tanguy 1995), suggesting a regulatory function in plant development. The application of exogenous hormones has shown relationships between these two classes of compounds (Evans and Malmberg 1989, Tiburcio et al. 1993). High levels of growth regulators (2,4-D and BAP at 2 mg l−1 each) are required to obtain gynogenic embryos in onion (Campion et al. 1992), although Martinez et al. (2000) have produced haploids without any, or with exogenous PAs. Several auxin-induced organogenesis processes were reported to require the promotion of PA biosynthesis, e.g. root formation in mung bean hypocotyl cuttings (Friedman et al. 1985) or in leaf explants of tobacco (Burtin et al. 1990), growth of immature cotyledons of Prunus avium (Garin et al. 1995) and adventitious organogenesis (Scoccianti et al. 2000) or somatic embryogenesis (Yadav and Rajam 1997) from eggplant explants. The effect of auxin was explained as an inhibition of arginine decarboxylase (ADC) activity in carrot cell culture (Fienberg et al. 1984) or as a stimulation in pea ovaries (Pérez-Amador and Carbonell 1995) or in BAP-treated Cucumis sativus cotyledons (Suresh et al. 1978). Comparisons between embryogenic and non-embryogenic carrot cultures revealed that the former displayed a significant increase of PA content (Fienberg et al. 1984). When treated with DFMA, an irreversible inhibitor of ADC, embryogenic cultures of carrot presented a 50% reduction in embryogenesis along with a decrease in PA content, demonstrating the involvement of the ADC pathway (Feirer et al. 1984). Only a few studies focused on gametic embryogenesis and solely through androgenesis. An increase of Spd content was noticed during maize (Santos et al. 1995) and tobacco (Garrido et al. 1995) embryogenesis by pollen culture. PAs have been reported to play a major role in both zygotic and somatic embryogenesis, but to our knowledge, they have not been studied during the development of gynogenic embryos. Only Martinez et al. (2000) studied the effect of additions of PA to culture media as an alternative to hormones, showing an effect on embryo induction in onion. The main goal of the present work was to further investigate the role of PAs in the gametic embryogenesis process in onion. Several approaches were used including the quantification of endogenous PAs at different stages of the process in vitro and the effect of exogenous PAs and PA biosynthesis inhibitors in the induction medium.

Materials and methods

Plant material

Six accessions of long-day onion (Allium cepa L.) represented three genetic types: open-pollinated populations (PG49, VSUZ Olomouc; PG52 and PG53, ENSH Versailles France), inbred (IG34, Rio Colorado Seeds, Bakersfield, CA) and clones (CG11 and CG12, COOPD'OR, Auxonne, France). The clones were derived from single plants maintained by micropropagation (Kahane et al. 1992). Bulbs were planted in March in greenhouse conditions for flower production, and in vitro initiation in June. Flowers used as explants were collected 3 days before anthesis. Varieties IG34, CG12, CG11 and PG49 expressing different gynogenic abilities (respectively 7.0, 1.6, 0.7 and 0.5% embryo-production rate) (Geoffriau et al. 1997b) were used for PA analysis (only at the inoculation stage for IG34). Two distinct experiments were set up to assess the effect of the addition in the induction medium of PAs (varieties CG11, CG12 and PG49) or inhibitors of PA biosynthesis (varieties CG11, CG12, PG49, PG52 and PG53).

Gynogenesis procedure and culture media

The procedure described by Geoffriau et al. (1997b) was used for explant preparation, flower bud inoculation, standard culture medium preparation and in vitro growth conditions. Tyramine (Tyr), Spermidine (Spd) or Spermine (Spm) (Sigma, St. Louis, MO) water solutions were filter-sterilized (0.22 µm-mesh filter) and added to the autoclaved culture medium at 1 × 10−5, 5 × 10−4 or 2 × 10−3M. Inhibitors of PA biosynthesis, difluoromethylornithine (DFMO, Merrell Dow Research, Strasbourg, France) or cyclohexylamine (CHA, Sigma) were similarly added at a final concentration of 10−3 and 5 × 10−4 M, respectively. Five hundred flower buds were inoculated per variety and per treatment. Embryo-production rate represents the number of structures emerging from the ovaries for 100 inoculated buds, the regeneration rate is the number of 2-month-old gynogenic plantlets for 100 inoculated buds, and the mean sprouting delay in days corresponds to the time course between inoculation and embryo emergence. The embryos have been shown to be of ovule and gametic origin (Bohanec et al. 1995, Geoffriau et al. 1997b). Hyperhydricity and callus rates are, respectively, the number of vitreous ovaries (Kevers et al. 1984) and the number of ovaries producing external callus for 100 inoculated buds.

Analysis of PA content

The onion gynogenic process was reported to last from 80 to 100 days (Geoffriau et al. 1997b). PA analyses were performed on inoculated flower buds after 0, 4, 8, 15, 40, 80 and 100 days of culture on the induction medium (days post-inoculation). The organs were rinsed in deionized water and stored in 1N HCl at 4°C until grinding. Samples were prepared according to Flores and Galston (1982). Tissues were ground in 1N HCl, and paper filtered (Whatman N°3). The filtrate was concentrated under vacuum at 30°C by rotary evaporation, dissolved in 2 ml 1N HCl and stored at −20°C until HPLC analysis. Free PAs were directly analysed from this extract, whereas conjugated PAs were derived from extract hydrolysis in 6N HCl at 110°C for 10 h. PAs were separated by HPLC (Bondapak C18 column, methanol : water solvent gradient) and detected by fluorescence after dansylation (Flores and Galston 1982, Smith and Davies 1985, Tarenghi et al. 1995). PA standard solutions were prepared and analysed the same way, and standard curves were established for each batch of samples in order to quantify PA contents. Three samples of 24 flower buds were systematically analysed per treatment. All the results are expressed in µmol/g fresh weight (FW).

Statistical analysis

The PA content data were subjected to Analysis of Variance using the SAS (Cary, NC) GLM procedure, and homogenous groups were identified by the Duncan test. Gynogenesis results expressed in frequencies were analysed by the G-test (Sokal and Rohlf 1981) based on Chi-square tables.


PA content at the inoculation stage

Spermidine in both free and conjugated forms was the most abundant PA quantitatively present in flower buds (Table 1). It represented 71–81% of the total free PAs and 86–98% of the total conjugated PAs. The conjugated form was prevalent with up to 6.36 µmol/g FW for the accession IG34. Differences in Spd between varieties were found to be significant: IG34 and CG12 had the highest values, while CG11 had the lowest and PG49 was intermediate. In contrast to Spd, free Putrescine (Put) was detected equally in all varieties and was lower in concentration (maximum at 0.36 µmol g−1 FW). For conjugated Put, CG11 had the highest concentration. Hydroxyputrescine (OHPut) was only detected in PG49 and CG11. Spm was found at significantly lower level in CG11 than in the other varieties. This accession was characterized by the highest Put-OHPut/Spd-Spm ratio (0.31 compared to 0.11 and 0.14 for CG12 and IG34, respectively). The total free PA content was similar for PG49, CG12 and IG34, but PG49 contained OHPut. Conjugated OHPut and Spm were not detected in any accessions.

Table 1.  Composition of free and conjugated polyamines (µmol g−1 FW) in onion flower buds at the inoculation stage. The ratio (Put+OHPut/Spd+Spm) was computed. All data are means of three replicates of 24 flower buds each. For each amine, values in columns with the same letter are not significantly different (P < 0.05). OHPut, hydroxyputrescine; 3OH4MP, 3-hydroxy,4-methoxyphenylethylamine; Put, putrescine; Spd, spermidine; Spm, spermine; Tyr, tyramine.
 Free polyaminesConjugated polyaminesConjugated aromatic amines

Among the aromatic PAs, only Tyr and 3-hydroxy,4-methoxyphenylethylamine (3OH4MP) were detected (Table 1). Interestingly, free Tyr was present only in IG34 (0.260 µmol g−1 FW), the inbred that expressed the highest gynogenic ability. The same accession displayed the highest levels of conjugated Tyr and 3OH4MP, followed in a decreasing order by CG12, PG49 and CG11.

Endogenous PAs following in vitro induction

An increase of free Put was observed in all varieties in the first week of culture, with a peak at 8 days post-inoculation (Fig. 1A). After 15 days, the accumulation depended on accession: both CG11 and PG49 presented a second peak of free Put at 40 days post-inoculation, suggesting a biphasic accumulation, whereas CG12 showed a decreasing content. CG11 and PG49 had Put derivatives [OHPut and homospermidine (HSpd)], which were absent in CG12 during the entire culture period. In the first two accessions, OHPut contents decreased drastically between 0 and 8 days post-inoculation, and OHPut disappeared at 40 days post-inoculation. Only in PG49 was an increase observed at 15 days post-inoculation (Fig. 1B). The curve of HSpd had a different pattern from the other PAs: absent at the inoculation time, HSpd was synthesized only in CG11 and PG49, with a peak at 15 days post-inoculation (Fig. 1C), and disappeared between 15 and 40 days for CG11. In PG49, its levels decreased between 40 and 100 days post-inoculation.

Figure 1.

Free (A) putrescine, (B) hydroxyputrescine and (C) homospermidine in onion flower buds during in vitro gynogenesis. For each stage (in days post-inoculation), a same letter indicates that values are not significantly different (P < 0.05).

Spd remained quantitatively the most abundant PA during the culture period. All accessions exhibited a similar pattern: a sharp decline just after inoculation followed by an increase to a peak at 8 days post-inoculation and a slow decline till the end of the culture (Fig. 2A). CG12 had the highest concentration: at 8 days post-inoculation, Spd level in CG12 was twice the other accessions. The concentration of Spm decreased sharply in all accessions in the first 8 days of culture (Fig. 2B). This decrease was less pronounced in CG12, where Spm contents remained high until 40 days after inoculation. Noticeably, Spm disappeared in CG11 and PG49 from day 8 onwards.

Figure 2.

Free (A) spermidine and (B) spermine in onion flower buds during in vitro gynogenesis. For each stage (in days post-inoculation), a same letter indicates that values are not significantly different (P < 0.05).

The ratio Put+OHPut+HSpd/Spd+Spm was calculated (Fig. 3). It remained low for CG12, whereas it increased gradually for CG11 and PG49 until 40 days post-inoculation. It further increased in PG49 while it remained stable in CG11, so that at 80 days post-inoculation, CG11 and PG49 ratios were, respectively, 6 and 25 times higher than the CG12 ratio.

Figure 3.

Calculated ratio (Put+OHPut+HSpd)/(Spd+Spm) in onion flower buds during in vitro gynogenesis.

We also detected Agmatine (Agm) and aromatic amines. Agm was only detected at day 8 post-inoculation in all three accessions, with significantly higher content in CG12 (0.346 µmol g−1 FW) than in CG11 or PG49 (0.047 and 0.050 µmol g−1 FW, respectively). Among the two aromatic amines detected only in the first days post-inoculation, Tyr was the most abundant compared to 3OH4MP, present only in CG11 8 and 15 days post-inoculation. Total PAs and aromatic amine concentration patterns presented a peak 8 days post-inoculation followed by a sharp (aromatic) or slight (PAs) decline (Fig. 4). Aromatic amines, at a lower level than PAs, were hardly detectable from 40 days post-inoculation onwards, while PAs were still present in the flower tissues at the end of the culture.

Figure 4.

Total (A) free polyamine and (B) aromatic amine in onion flower buds during in vitro gynogenesis. For each stage (in days post-inoculation), a same letter indicates that values are not significantly different (P < 0.05).

Effect of exogenous PAs

A positive effect of exogenous Tyr, Spd or Spm was observed on the embryo-production rate. The embryo production increased with increasing concentrations of Spd or Spm (Table 2). PG49 exhibited a positive response with the addition of Tyr at 5 × 10−4 M, while no effect was detected with this amine for the two other accessions. CG11 was most responsive to the highest values of Spd or Spm, whereas CG12 was also responsive to the lowest values of Spd or Spm. Additions of Spm at all concentrations enhanced the regeneration rate. The effect of exogenous PAs on the sprouting delay was significant but inconsistent: the addition of 1 × 10−5 M Spd or 5 × 10−4 M Spm reduced it by 16 and 11 days for CG12, while the addition of 2 × 10−3 M Spm increased it by 13 days in CG11. No significant effect was obtained with Tyr or for PG49. The addition of PAs in the medium influenced embryo maturation, but no trend could be identified. The addition of amines also increased the rate of callus developed on flowers buds, especially at the highest concentrations, and had no effect on hyperhydricity of ovaries (data not shown).

Table 2.  Effect of various concentrations of polyamines in the induction medium on embryo production rate, regeneration rate and delay in embryo sprouting. Five hundred flower buds per treatment and per accession (1500 buds for the total). *, **indicates that the values are significantly different from the control at 5 or 1% level, respectively (G-test and Student's t-test). For each treatment, values in columns with the same letter are not significantly different (P < 0.05, G-test).
  Tyramine (M)Spermidine (M)Spermine (M)
AccessionControl1 × 10−55 × 10−42 × 10−31 × 10−55 × 10−42 × 10−31 × 10−55 × 10−42 × 10−3
Embryo-production rate (%)
Regeneration rate (%)
Sprouting delay (days)

Effect of PA inhibitors

The embryo and regeneration rates decreased significantly by the addition of CHA in the induction medium (Table 3). Moreover, callus production on the ovary tissues was significantly reduced, and the frequency of hyperhydricity of inoculated flower buds almost increased three-fold. Exogenous DFMO had no noticeable effect on the measured variables (Table 3).

Table 3.  Effect of inhibitors of polyamine biosynthesis in the induction medium (legend as in Table 2; nd, not determined).
Addition of cyclohexylamine (CHA) (5 × 10−4 M)
 Embryo rate (%)Regeneration rate (%)Sprouting delay (days)Hyperhydricity (%)Callus production (%)
Total2.7a0.9**a1.4b0.7*a  6.2a17.1**b4.3b1.5**b
Addition of difluoromethylornithine (DFMO) (2 × 10−3 M)
 Embryo rate (%)Regeneration rate (%)Sprouting delay (days)Hyperhydricity (%)Callus production (%)
Total1.7ab1.2ab0.4a0.4a  6.8b6.9b0.2a0.2a


The influence of PAs during gynogenesis, a type of gametic embryogenesis, was shown through PA analysis of onion floral tissues and the addition of PAs in the media, this being the first report of PA analysis of onion floral tissues. The genotypic variability in gynogenesis in onion has been demonstrated (Geoffriau et al. 1997b, Michalik et al. 2000), and this study suggests that PA composition of flower buds can be used to characterize the responsiveness in culture. A high Spd content, the absence of OHPut, a low Put-OHPut/Spd-Spm ratio, low conjugated Put and high aromatic amine contents characterized the most responsive accessions. This pattern has even been associated with good development of gametic embryos through our observations. In various species, high concentration of endogenous Spd has been associated with floral induction and development (Aribaud and Martin-Tanguy 1994a, Kaur-Sawhney et al. 1988, Tiburcio et al. 1990) and mentioned in studies on embryogenic tissues (Cvikrova et al. 1999, Minocha et al. 1999). Levels of OHPut in embryogenic tissues have not been reported in the literature, although its absence was a clear marker for gynogenesis capacity in onion (Table 1). Interestingly, IG34 flower buds were the only ones that contained free Tyr. This compound could be related to gametic embryogenesis as it has been positively correlated with pro-embryogenic cell mass formation in alfalfa explants (Cvikrova et al. 1999). Since the ratio Put+OHPut/Spd+Spm had low values during flowering in high responsive accessions (data not shown), it could be considered as an early marker of competency to undergo in vitro gynogenesis and therefore discriminate between accessions, a trend also observed during the culture (Fig. 3). Similarly, a low Put/Spd ratio was reported to be critical for the development of grape embryos into plants (Faure et al. 1991) and for callus regeneration in rice (Bajaj and Rajam 1995). We also observed significant differences in conjugated PAs, but their role in embryogenesis is not clear. Some measured differences might be due to variation of the reproduction potential, as these compounds have important functions in the reproductive development (Aribaud and Martin-Tanguy 1994a, Martin-Tanguy 1997).

Although the early addition of PAs in the medium could lead to significant improvement of embryo induction or maturation, no trend was identified. Exogenous PAs increased callus development rate on floral tissues, especially at the highest concentrations, and had no effect on hyperhydricity rate (data not shown). Exogenous Spd or Spm promoted embryo production, as if these PAs were the limiting factor for onion gynogenesis. The positive effect of Spd was confirmed by the use of CHA, an inhibitor of Spd and Spm synthases, that reduced embryo and regeneration rates and increased organ hyperhydricity. Exogenous CHA led to some inhibition of gynogenesis, perhaps due to the promotion of ethylene and Put accumulation. Martinez et al. (2000) showed that the addition of combined Put and Spd in a hormone-free medium improved gynogenesis in onion, while the addition of Put alone had no significant effect. These results are in agreement with the influence of Spd in embryogenesis reported for carrot cell cultures (Mengoli et al. 1989) or in maize androgenesis (Santos et al. 1995). The distinct PA profiles that we observed between CG12 on one hand and PG49 and CG11 on another hand, could be due to distinct PA regulations. A study of ODC, ADC and amine oxidase activities would reveal such differences.

At 8 days post-inoculation, Agm, which is synthesized through the ADC pathway (Tiburcio et al. 1997), was present in all accessions, being at a significantly higher concentration in the most responsive variety. This suggests that the increase in endogenous PA resulted from ADC activity. Furthermore, since the addition of DFMO as ODC inhibitor had no effect on gynogenesis performances (Table 3), ADC pathway might be involved in the biosynthesis of PAs during in vitro gynogenesis in onion. These data are preliminary, and this hypothesis needs to be confirmed by further determination of ADC activity and/or the use ADC inhibitors. Such findings would corroborate, in a distinct system of embryogenesis, the involvement of the ADC pathway, what has been so far demonstrated in somatic embryogenesis (Feirer et al. 1984, Mengoli et al. 1989). In tobacco androgenesis, Garrido et al. (1995) observed that ADC and ODC activities decreased during embryogenesis induction, but ADC activity was required for the embryo formation itself. The onion gynogenesis protocol requires high level of 2,4-D (2 mg l−1) in the induction medium. This growth regulator has been shown to stimulate the ADC activity (Pérez-Amador and Carbonell 1995) but also to reduce or inhibit embryo development (Fienberg et al. 1984, Nissen and Minocha 1993). Further research on the role of 2,4-D in gynogenesis ability and PA regulation is needed.

In parallel to the genetic predisposition to gynogenesis, PA composition could characterize the culture stages of flower buds. We revealed a possible critical stage at 8–12 days post-inoculation when peaks of free Put, Spd, total PAs and aromatic amines were clearly identified. Some of them followed a sharp decline, what would mean new biosyntheses. When subculturing flower buds from induction medium to hormone-free medium 12 days post-inoculation, we observed no significant difference (data not shown), suggesting that embryogenesis was already induced. The present data suggest that embryogenesis took place as early as 8 days post-inoculation, in a process requiring 70–80 days for the earliest developed embryos to come out of the ovary. This hypothesis could be confirmed by measurements of decarboxylase activity and cytological studies. In comparison, the induction of gynogenic embryos from the embryo sac took place as soon as 20 days post-inoculation for sugarbeet ovules (Bossoutrot and Hosemans 1985). A second key-stage around 15 days post-inoculation corresponded to peaks of OHPut and HSpd in low-responsive accessions only, and later (40 days post-inoculation) a new synthesis of Put was observed, whereas Put compounds disappeared in high-responsive accessions. The accumulation of Put and Put derivatives is associated with a low responsiveness, as shown by the Put+OHPut+HSpd/Spd+Spm ratio.

An interesting question relates to the origin of the observed Put accumulation and Put derivative variations. Put accumulation can be due to new biosynthesis but also to other origins. The initial increase (8 days post-inoculation) could be due to the hydrolysis of the conjugated form present at the inoculation stage since the amine conjugates may serve as a storage form of amines (Bonneau et al. 1994, Martin-Tanguy 1997). Conjugated Put was not detected later during in vitro culture. The increase of free Put could also be due to conversion of spermidine to putrescine as shown in maize roots (De Agazio et al 1995). HSpd was shown to occur in many plants and to be synthesized from Put (Ober et al. 2003) which could correspond to the observed variations at 15 days post-inoculation. OHPut was detected in leguminous plants (Hamana and Matsuzaki 1993), but little is known about its metabolic significance.

Both the study of endogenous PAs and the addition of exogenous PAs showed the critical importance of the presence of Spd and a low level of Put and its derivatives for successful gynogenesis in onion. Further research is needed to elucidate PA regulation during this process.

Acknowledgements –  The authors gratefully acknowledge C. Bellamy for in vitro technical support and B. Pasquis for PA analysis.