The regulatory gamma subunit SNF4b of the sucrose non-fermenting-related kinase complex is involved in longevity and stachyose accumulation during maturation of Medicago truncatula seeds


  • Claire Rosnoblet,

    1. Unité Mixte de Recherche 1191 Physiologie Moléculaire des Semences (Université d’Angers, Institut National d’Horticulture, Institut National de la Recherche Agronomique), 16 boulevard Lavoisier, 49045 Angers, France, and
    Search for more papers by this author
    • Present address: Institut de Biologie de Lille, Interdisciplinary Research Institute, CNRS, Lille, France.

  • Catherine Aubry,

    1. Unité Mixte de Recherche 1191 Physiologie Moléculaire des Semences (Université d’Angers, Institut National d’Horticulture, Institut National de la Recherche Agronomique), 16 boulevard Lavoisier, 49045 Angers, France, and
    Search for more papers by this author
  • Olivier Leprince,

    1. Unité Mixte de Recherche 1191 Physiologie Moléculaire des Semences (Université d’Angers, Institut National d’Horticulture, Institut National de la Recherche Agronomique), 16 boulevard Lavoisier, 49045 Angers, France, and
    Search for more papers by this author
  • Benoit Ly Vu,

    1. Unité Mixte de Recherche 1191 Physiologie Moléculaire des Semences (Université d’Angers, Institut National d’Horticulture, Institut National de la Recherche Agronomique), 16 boulevard Lavoisier, 49045 Angers, France, and
    Search for more papers by this author
  • Hélène Rogniaux,

    1. Unité de Recherche, Biopolymeres, Interactions, Allergie, Institut National de la Recherche Agronomique, 44316 Nantes, France
    Search for more papers by this author
  • Julia Buitink

    Corresponding author
    1. Unité Mixte de Recherche 1191 Physiologie Moléculaire des Semences (Université d’Angers, Institut National d’Horticulture, Institut National de la Recherche Agronomique), 16 boulevard Lavoisier, 49045 Angers, France, and
    Search for more papers by this author

(fax +33 241 22 5549; e-mail


The sucrose non-fermenting-related kinase complex (SnRK1) is a heterotrimeric complex that plays a central role in metabolic adaptation to nutritional or environmental stresses. Here we investigate the role of a regulatory γ-subunit of the complex, MtSNF4b, in Medicago truncatula seeds. Western blot indicated that MtSNF4b accumulated during seed filling, whereas it disappeared during imbibition of mature seeds. Gel filtration chromatography suggested that MtSNF4b assembled into a complex (450–600 kDa) at the onset of maturation drying, and dissociated during subsequent imbibition. Drying of desiccation-tolerant radicles led to a reassembly of the complex, in contrast to sensitive tissues. Silencing of MtSNF4b using a RNA interference (RNAi) approach resulted in a phenotype with reduced seed longevity, evident from the reduction in both germination percentage and seedling vigour in aged RNAi MtSNF4b seeds compared with the wild-type seeds. In parallel to the assembly of the complex, seeds of the RNAi MtSNF4b lines showed impaired accumulation of raffinose family oligosaccharides compared with control seeds. In mature seeds, the amount of stachyose was reduced by 50–80%, whereas the sucrose content was 60% higher. During imbibition, the differences in non-reducing sugar compared with the control disappeared in parallel to the disassembly of the complex. No difference was observed in dry weight or reserve accumulation such as proteins, lipids and starch. These data suggest that the regulatory γ-subunit MtSNF4b confers a specific and temporal function to SnRK1 complexes in seeds, improving seed longevity and affecting the non-reducing sugar content at later stages of seed maturation.


During maturation, seeds of most species acquire the capacity to withstand drying, i.e. they become desiccation tolerant. In the dry state, they are capable of remaining alive for long periods of time, over decades and even centuries (Walters et al., 2005). This characteristic renders them particularly useful for preserving plant genetic resources. Seed longevity is determined not only by external factors, such as water content and temperature, but also by mostly unidentified intrinsic factors that are acquired at the later stages of seed maturation (Bailly et al., 2001; Sanhewe and Ellis, 1996). The identification of these factors and their regulatory pathways is of practical importance for designing breeding strategies to enhance seedling vigour and germination capacity after storage.

The mechanisms that underlie desiccation tolerance involve the accumulation of non-reducing sugars together with stress proteins such as heat shock proteins and late embryogenesis abundant (LEA) proteins (Hoekstra et al., 2001). These molecules are involved in stabilizing macromolecular structures in the dry state, thereby protecting macromolecules and retaining the functional integrity of membranes upon desiccation and rehydration (Hoekstra et al., 2001). In addition, these non-reducing sugars might serve as a readily available energy source during early imbibition (Downie and Bewley, 2000). During the final stages of maturation of seeds of Medicago truncatula, non-reducing sugar levels, and especially the raffinose family oligosaccharides (RFO), accumulate strongly (Djemel et al., 2005). In dry mature seeds, stachyose represents over 90% of the total soluble sugars, and 12% of the dry weight of a seed. In addition to the synthesis of protective compounds, a concerted downregulation of energy metabolism prior to or during desiccation appears to be required to ensure prolonged survival in the dried state (Buitink et al., 2006; Leprince and Hoekstra, 1998; Leprince et al., 2000). As such, developing seeds must have evolved mechanisms to regulate their metabolism accordingly.

The Snf1/AMP-activated protein kinase (AMPK) family is essential for metabolic regulation in eukaryotes (for review see Hardie et al., 1998; Kemp et al., 1999). In mammals, AMPK controls metabolic enzymes in response to stresses that affect cellular energy supply, including nutrient limitation, hypoxia, heat shock and exercise. In yeast, the Snf1 kinase plays an essential role in metabolic adaptation to different carbon sources, in response to environmental stress such as salt stress and heat shock and in developmental processes such as sporulation, life span and aging (Ashrafi et al., 1998, 2000; Honigberg and Lee, 1998). Like its yeast counterparts, several lines of evidence suggest that the SNF1 homologue in plants, SnRK1, has the potential to regulate carbon metabolism through both gene expression and direct control of enzyme activity. For instance, antisense expression of the catalytic α-subunit SnRK1 resulted in the loss of sucrose-inducibility of sucrose synthase gene expression in leaves and in the reduction of sucrose synthase gene expression in tubers (Purcell et al., 1998). The role of SnRK1 in starch metabolism was demonstrated by antisensing the α-subunit in various plant organs. This resulted in a loss of ADP-Glu pyrophosphorylase (AGPase) activity in potato tubers (Tiessen et al., 2003), and a decrease in α-amylase expression in wheat embryos (Laurie et al., 2003) and led to abnormal pollen development in barley and a lack of starch accumulation (Zhang et al., 2001). In addition to affecting sugar metabolism, antisensing the catalytic α-subunit in pea produced transgenic seeds with an array of maturation defects, such as reduced conversion of sucrose into storage products, lower globulin content, altered cotyledon morphology, as well as occasional precocious germination (Radchuk et al., 2006).

The SnRK1 complex is a heterotrimeric complex composed of a catalytic α-subunit that interacts with two other subunits: the γ-subunit and the β-subunit (Hardie et al., 1998). The complex is activated when SNF4 binds to the SNF1 regulatory domain and stabilizes the kinase in an open, active conformation (Jiang and Carlson, 1997). Several regulatory β- and γ-subunits have been identified in M. truncatula and Arabidopsis thaliana that are differentially regulated at the transcriptional level according to stress and developmental stage (Bouly et al., 1999; Buitink et al., 2003a). In seeds, several genes encoding different subunits are expressed, suggesting that several complexes might coexist. For instance, two γ-subunits, MtAKINβγ and MtSNF4b, are expressed during the re-establishment of desiccation tolerance (DT) (Buitink et al., 2003a), and homologues of both these genes were found to complement a snf4-deficient yeast mutant, indicating that they are both true regulatory subunits of the SnRK1 complex (Bradford et al., 2003; Kleinow et al., 2000; Lumbreras et al., 2001). The coexistence of several SnRK1 complexes formed by different regulatory subunits might indeed explain the strong pleiotropic effects observed when the catalytic subunit was downregulated in pea seeds (Radchuk et al., 2006). Investigating the role of a regulatory γ-subunit might not only provide insight into the role of these subunits, but might also help in dissecting the different pathways that are regulated by the SnRK1 complexes in seeds.

In this study we addressed the question of when the regulatory γ-subunit MtSNF4b is assembled in a complex and attempted to elucidate the function of this particular complex by RNA interference (RNAi)-mediated silencing of MtSNF4b in M. truncatula. We show that SNF4b plays a role at the later stages of seed maturation by enhancing seed longevity and modifying sucrose and oligosaccharide content.


SNF4b accumulates during seed maturation and disappears during imbibition

To evaluate the possible time frame in which MtSNF4b could function, its abundance was determined by Western analysis. For this purpose, an antibody was raised against a synthetic N-terminal peptide (KEKKVKDLMVNKKC-N2) and purified further by affinity using the synthetic peptide coupled to CNBr-sepharose. A single band was detected around 47 kDa, slightly higher that the theoretical molecular mass of MtSNF4b (42 kDa) in dry mature seeds, whereas this band could not be detected in leaves, either fresh or after drying (Figure 1a). To confirm that this band indeed corresponded to SNF4b, proteins of an extract from mature seeds of cv. Paraggio were immunoprecipitated using the antiserum against MtSNF4b or the pre-immune antiserum (Figure 1b, lanes 1 and 2). A polypeptide around 47 KDa was specifically immunoprecipitated by the anti-MtSNF4b antibody, and was recognized by Western blot using the same antibody (Figure 1b, lanes 2 and 3). This band was analysed using liquid chromatography-tandem mass spectrometry (LC-MS/MS). After tryptic digestion, one peptide was sequenced (AAMLNALPIVR) and was used to screen the public databases. The only hit found was that of MtSNF4b of M. truncatula (Q7XYX9).

Figure 1.

 Identification and Western blot analysis of MtSNF4b during maturation, imbibition and re-establishment of DT in seeds of M. truncatula.
(a) Immunoprecipitation and Western blot analysis of MtSNF4b using a rabbit polyclonal antibody raised against the N-terminal synthetic peptide of MtSnf4b or the pre-immune serum. The arrow indicates the band that was analysed by LC-MS and identified as MtSNF4b.
(b–f) Western blot analysis of MtSNF4b. Fifteen micrograms of soluble proteins was separated on 12% SDS-PAGE and MtSNF4b was detected by chemiluminescence using the anti-MtSnf4b antibody. (b) Mature, dry seeds and fresh and dried leaves of 6-week-old plants. Leaves were dried for 3 days. (c) Seed embryos at different stages of maturation. Fresh weight accumulation and percentage of DT are indicated. (d) Radicles during imbibition at 20°C. The percentage of germination and DT is indicated. (e) Cotyledons during imbibition at 20°C. (f) Radicles during re-establishment of DT by polyethylene glycol (PEG) treatment. DT, desiccation tolerance; DAP, days after pollination; imb, imbibition; IP, immunoprecipitation.

The amount of MtSNF4b increased during maturation, first becoming detectable around 14 days after pollination (DAP) and increasing until 33 DAP (Figure 1c). The accumulation of fresh weight (Figure 1c) shows that seed filling occurred between 12 and 26 DAP, whereas maturation drying commenced around 26–28 DAP. Desiccation tolerance was acquired between 14 and 20 DAP (Figure 1c). During imbibition, changes in MtSNF4b levels were monitored both in the radicles and in the cotyledons of the embryos (Figure 1d,e). The amount of protein in the radicles decreased during imbibition, and was no longer detectable at 24 h, when the protruded radicle length was approximately 1 cm (Figure 1d). In older roots, the protein could not be detected (data not shown). In the cotyledons, the protein was detected until 24–48 h post-imbibition, in parallel to the loss of DT in these tissues (Figure 1e). Considering the parallel disappearance of DT and MtSNF4b abundance, we also determined whether synthesis of the protein was re-induced upon the re-establishment of DT in sensitive, germinated radicles. It has been shown previously that an incubation of 17-h imbibed germinated seeds in an osmotic polyethylene glycol (PEG) solution of −1.7 MPa at 10°C leads to the re-establishment of DT in the radicles, which also resulted in the re-appearance of MtSNF4b mRNA (Buitink et al., 2003a). However, no change in protein level could be detected during the 72 h of PEG incubation (Figure 1f).

MtSNF4b assembles in a high-molecular-weight complex at the final stages of seed maturation and dissociates after germination

Having established that MtSNF4b accumulated during maturation, whereas it disappeared during germination, we investigated whether the subunit was present in a complex and in which time frame this assembly would occur. Previous studies using yeast and plant native extracts demonstrated, using chromatography techniques, that the catalytic subunits, together with their regulatory counterparts were present in high-molecular-weight (Mw) fractions, and that this assembly corresponded to the activity of the SNF1 complex (Estruch et al., 1992; Sugden et al., 1999). In our study, native protein extracts extracted from M. truncatula cv. Paraggio seeds were separated using gel filtration chromatography, followed by Western blotting, to determine at which Mw fraction MtSNF4b was present. During the seed filling phase (16–20 DAP), MtSNF4b protein was found in the low-Mw fractions (40–60 kDa), corresponding to its own Mw, indicating that MtSNF4b was not assembled in any complex (Figure 2a). At 24 DAP, MtSNF4 was also detected in the fractions between 60 and 200 kDa. From 28 DAP onwards, when maturation drying commenced (see Figure 1c), MtSNF4b was also present in high-Mw fractions, corresponding to 450–600 kDa. To visualize the amount of SNF4b protein in the different fractions after separation of an extract of 24 DAP and 40 DAP, the intensity of the bands of each fraction was quantified and expressed as a percentage of the total intensity (Figure 2b). At 24 DAP, all the SNF4b was found in the low-Mw fractions (1–8), whereas in dry seeds about 40% of the amount of the subunit was present in fractions >400 kDa. It was verified that the presence of the protein in these high-Mw fractions was not an artefact related to the concentration of the protein extracts. Indeed, it could be dissociated by guanidine chloride (data not shown). Complex formation was also studied during maturation of the genotype R108 (the genotype used for the RNAi approach) and was found to be comparable to that of cv. Paraggio (Figure 2). Thus, it appears that the MtSNF4b protein assembles into a high-Mw complex at the final stages of maturation, during the early phase of maturation drying.

Figure 2.

 Western blot analysis of MtSNF4b at different developmental stages of M. truncatula seeds after separation of the protein extracts by their molecular weight using gel filtration chromatography.
(a) Seeds (40–50) were harvested at different days after pollination (DAP). Native protein extracts (1.5–2 mg) were separated by gel filtration chromatography on Superdex 200. Western blot analysis was performed as described in Figure 1. The arrows indicate the band corresponding to MtSNF4b.
(b) Estimation of the amount of MtSNF4b in each fraction at 24 and 40 DAP calculated as a percentage of the total intensity.

The MtSNF4b was shown to disappear during imbibition, after radicle protrusion, whereas its abundance remained high in cotyledons upon further imbibition until 24–48 h (Figure 1c). Gel filtration chromatography was used to investigate when the MtSNF4b complex dissociated in both tissues during imbibition (Figure 3a,b). The protein was still present in a complex in the radicles during early imbibition (6 h). In radicles of 17-h imbibed seeds that had become sensitive to desiccation, MtSNF4b was only detected in low-Mw fractions, indicating complete disassembly of the complex in these tissues. In imbibing cotyledons, SNF4b remained present in a complex for longer than in imbibing radicles (Figure 3b). After 30 h of imbibition, SNF4b could still be detected in the high-Mw fractions, although most of it had dissociated compared to 20 h of imbibition (Figure 3b). After 49 h of imbibition, when desiccation tolerance was lost, no SNF4b was detected in any of the gel filtration fractions. Native extracts were also prepared from germinated radicles (17 h of imbibition at 20°C) after 3 days of PEG treatment at −1.7 MPa and analysed by gel filtration chromatography. This treatment did not lead to the re-association of MtSNF4b in a high-Mw complex (Figure 3c). However, when these desiccation-tolerant radicles were dried, SNF4b reappeared in the high-Mw fractions (Figure 3c). This observation was not due to an artefact, because drying of desiccation-sensitive radicles prior to the PEG treatment and re-establishment of DT did not lead to the reappearance of SNF4b at high Mw [Figure 3c; compare desiccation sensitive (DS) with DT]. Thus, as was found during maturation drying, the SNF4b complex appears to be induced during the loss of water in the tolerant, germinated seed radicles.

Figure 3.

 Western blot analysis of MtSNF4b in cotyledons and radicles of M. truncatula seeds during imbibition and re-establishment of DT after separation of the protein extracts by their molecular weight using gel filtration chromatography.
Native protein extracts (1.5–2 mg) were separated by gel filtration chromatography on Superdex 200. Western blot analysis was performed as described in Figure 1. The arrows indicate the band corresponding to MtSNF4b.
(a) Radicles at 6 h and 17 h of imbibition.
(b) Cotyledons after 20, 30 and 49 h of imbibition at 20°C.
(c) Fresh or dried radicles of germinated seeds at 17 h of imbibition with or without a 3-day incubation in polyethylene glycol solution (−1.7 MPa, 10°C).
DS, desiccation sensitive. DT, desiccation tolerant. DAP, days after pollination.

In an attempt to identify the other subunits that participate in the formation of the complex, MtSNF4b was immunoprecipitated using the antibody raised against the N-terminal synthetic peptide. MtSNF4b could indeed be immunoprecipitated from a native protein extract of dry seeds (Figure 1a), but this was not possible with a 600–500 kDa protein fraction from the gel filtration experiment of mature seeds containing MtSNF4b (data not shown). This suggests that it is not possible to immunoprecipitate the protein when it is assembled in a complex, possibly because the N-terminal region of the protein was not accessible to the antibody due to its involvement with the other subunits. The MtSNF4b immunoprecipitated from the total protein extract therefore represents only ‘free’ MtSNF4b, and explains why no other subunits could be identified by this technique.

One β-subunit, MtAKINβ2, whose transcript levels are comparable to those of MtSNF4b (Buitink et al., 2003a), was further investigated as a potential partner of the SnRK1/MtSNF4b complex. In tomato, the homologue (92% similarity) of this β-subunit, LeSIP1, has been shown to interact with LeSNF4, the homologue of MtSNF4b, using a two-hybrid assay (Bradford et al., 2003). Using an antibody raised against an N-terminal synthetic peptide of MtAKINβ2, Western analysis demonstrated that this subunit accumulates at the final stages of maturation, although the signal was very faint (data not shown). Likewise, MtAKINβ2 could also be detected in the high-Mw fractions in which SNF4b was found during maturation and early imbibition (data not shown). However, considering the likelihood that several SnRK1 complexes will be present in these high-Mw fractions, the presence of other subunits is no direct evidence that they actually interact with MtSNF4b, so further analysis was abandoned.

Ribonucleic acid interference silencing of MtSNF4b leads to reduced seed longevity

To investigate the role of SNF4b in seeds of M. truncatula, plants were constructed using RNAi technology to disrupt the MtSNF4b gene, using the pFGC5941/Gateway vector, in which a 609-bp sense and antisense fragment of MtSNF4b was inserted in two Gateway cassettes separated by a CHSA intron and driven by a 35S promoter. The strategy for selection is indicated in Data S1. Briefly, after subsequent steps of transformation and in vitro culture using the R108 genotype, seeds of 12 T0 lines that were previously selected based on RT-PCR using the primers of the 609-bp MtSNF4b fragment were analysed using half of the cotyledons for Western blot analysis to screen for the presence or absence of the SNF4b protein, and the other remaining half of the seed with the radicle was used to grow the T1 generation. Those plants for which the seeds showed 75%:25% segregation for SNF4b absence and presence were retained and their seeds were analysed to select homozygous plants (absence of SNF4b in 60 individual seeds). Three independent transformants were retained for further analysis as confirmed by Southern analysis (data not shown). All studies were performed on the seeds of the T2 or T3 generation.

Reverse transcriptase-PCR shows that in mature seeds of both the wild-type and transgenic seeds containing the empty plasmid, MtSNF4b was highly expressed (Figure 4a). In contrast, the mRNA levels in the three independent transformants were barely detectable, indicating that the RNAi-mediated silencing was efficient (Figure 4a). Western blot analyses of the protein extracts of mature control and transgenic lines using anti-SNF4 antibody demonstrated that the protein was highly abundant in the control lines (wild type and the line transformed with the empty plasmid), whereas the protein was undetectable in the three transgenic lines (Figure 4b). In addition, for the developmental stage for which MtSNF4b abundance in the wild-type seeds was highest (40 DAP), the absence of the protein in the high-Mw fractions of the gel filtration chromatography analysis was verified on two transgenic lines (Figure 4c). No band corresponding to MtSNF4b could be detected in the fractions corresponding to 600–250 kDa for both RNAi MtSNF4b lines. A faint signal was still apparent for line 2 in fractions 9–11, corresponding to 150–80 kDa.

Figure 4.

 Characterization of RNAi-mediated silencing of MtSNF4b lines.
(a) Expression of MtSNF4b and 18S in control and RNAi MtSNF4b transgenic lines of M. truncatula determined by RT-PCR. WT, wild type; empty plasmid, transgenic control transformed with empty plasmid.
(b) Western blot on protein extracts from mature seeds of wild-type and RNAi MtSNF4b lines, as described in Figure 1.
(c) Western blot on protein extracts of mature seeds of WT and RNAi MtSNF4b lines, separated in different fractions by gel filtration chromatography.

To detect putative pleiotropic effects of the regulatory γ-subunit of MtSNF4b during seed development, as was the case for antisense lines of the catalytic α-subunit in pea (Radchuk et al., 2006), plants of transgenic lines together with the wild type were grown at the same time in a growth chamber. Seed fresh weight, water content and dry weight accumulated in a similar manner in both RNAi lines and the wild type, indicating that seed filling processes were not affected by the strong reduction of SNF4b in the transgenic seeds and that seed development was not delayed (Data S3). At 28 DAP, seed filling had terminated and maturation drying commenced (see dashed line in Data S3). Seeds were harvested from the pods at different time intervals during maturation and dried back using a 43% relative humidity (RH) airflow for 3 days, after which they were rehydrated on a moist filter paper to determine the acquisition of DT. Desiccation tolerance was acquired between 14 and 20 DAP. Both the wild-type seeds and those of the RNAi lines acquired their tolerance to desiccation within the same time frame during development, and mature seeds of the RNAi lines did not show any difference in the final percentage of DT (Data S3c). After germination, radicle growth was indistinguishable from that of the wild-type seeds. These data concur with the observation that the complex formation of SNF4b takes place after DT has been acquired. In view of the reappearance of the complex after drying of germinated seeds in which DT was re-established, the percentage of re-establishment of DT was also determined in radicles of the wild type and the two RNAi lines (Data S3d). Desiccation tolerance could be re-induced in the transgenic lines, indicating that the absence of the SnRK1/SNF4b complex did not result in a loss of DT.

To assess whether downregulation of MtSNF4b had an effect on long-term survival in the dry state (e.g. longevity), storage experiments were carried out under moderate accelerated ageing conditions (75% RH, 35°C). The seed population of the wild type took approximately 28 days to lose 50% of germination capacity (Figure 5a). The seeds of two transgenic RNAi MtSNF4b lines aged noticeably faster, exhibiting 50% loss of germination after 18 and 21 days of ageing for lines 1 and 2, respectively. For example, at 21 days of ageing, 84% of the wild-type seed population germinated, whereas for the two transgenic lines, the germination percentage was reduced to 29% and 46%. In addition to a faster decrease in germination percentage, a strong reduction in growth was observed in the seedlings that had survived 2–4 weeks of ageing of the RNAi MtSNF4b lines compared to the wild type. This is illustrated in Figure 5(b) and (c), where seedling length was measured after 3 days of imbibition for those seeds that still germinated after 21 days of ageing. Wild type seedlings were approximately 27 mm long, whereas the length reduced to 18 mm in the transgenic lines (Figure 5b). Whereas hypocotyls appeared to retain their capacity to elongate, the radicles in particular showed a strong reduction in growth (Figure 5c, arrow). The length of the radicle in the transgenic lines was reduced by 80% compared with the wild-type seedlings. Thus, the lack of SNF4b clearly appears to affect seed longevity, expressed as germination capacity and seed vigour upon subsequent imbibition.

Figure 5.

 Effect of accelerated ageing on germination percentage and seedling vigour of wild-type and RNAi MtSNF4b seeds.
Accelerated ageing was carried out at 35°C and 75% RH.
(a) Germination percentage at different time intervals of ageing. Germination was scored when radicles protruded the seed envelope. Data are the average of two independent experiments of 100 seeds obtained from two different cultures. The LSD at the 5% level of probability for comparison between means is shown.
(b) Seedling length of 21-day aged, germinated seeds after 3-day imbibition at 20°C in the dark. Statistical analysis was performed using the Dunnet multiple comparisons test, the control corresponding to the wild-type (WT) values. An asterisk indicates a significant difference at the 99% level.
(c) Representative picture of wild-type and RNAi MtSNF4b seedlings of line 1 that were imbibed for 3 days after an ageing treatment of 21 days [see part (b)]. Similar results were obtained for line 2.

Silencing of MtSNF4b affects non-reducing sugar metabolism at the final stages of seed maturation and early imbibition

To investigate the possible involvement of the MtSNF4b/SnRK1 complex in sugar metabolism, soluble sugar and starch contents were analysed in mature seeds of wild-type and transgenic lines containing the empty plasmid or the MtSNF4b RNAi constructs. In mature, dry seeds of the control lines, the most abundant sugars present are sucrose and stachyose, representing 13% and 79%, respectively, of the total soluble sugar content (Figure 6). Silencing MtSNF4b resulted in large differences in the amount of sucrose and stachyose between the transgenic lines and the wild-type and transgenic control seeds (Figure 6). Sucrose levels remained over 60–70% higher in the transgenic lines than in wild-type seeds (Figure 6a). In contrast, stachyose, the most abundant sugar in the mature wild-type seeds (approximately 8% of the total dry weight), accumulated to only half the amount in the three transgenic lines (Figure 6b). In seeds of line 3, with the strongest phenotype, only 20% of the amount of oligosaccharides accumulated compared with the control seeds. Other soluble sugars that could be detected in the seeds (myo-inositol, galactose, glucose, fructose, raffinose and verbascose) showed no significant differences between control and silenced lines (Data S2).

Figure 6.

 Effect of RNAi-mediated silencing of MtSNF4b on sucrose and stachyose content in mature seeds.
Sugar contents were measured in mature seeds of control [wild type (WT) and transgenic control transformed with an empty plasmid, closed bars] and three transgenic RNAi MtSNF4b lines (open bars): (a) sucrose; (b) stachyose. Sugar contents (μg mg−1) were determined on five replicates of five to ten seeds harvested from five to seven plants per line (average ± SE). Statistical analysis was performed using the Dunnet multiple comparisons test, the control corresponding to the wild-type or transgenic control values. An asterisk indicates a significant difference at the 99% level.

To further investigate the differences in sugar content in the RNAi MtSNF4b lines, the sugar content in two MtSNF4b silenced lines and the wild type was analysed in a time-dependent manner during seed development and imbibition. Figure 7 depicts the stachyose and sucrose levels in embryos during development (a, c) and radicles (b, d) during imbibition. In wild-type embryos, glucose and fructose levels started to decrease around 12–14 DAP (Data S2). At this stage, sucrose levels were high [100 μg mg−1 dry weight (dw)] (Figure 7c). During subsequent maturation, levels of RFO increased in wild-type seeds, commencing with the trisaccharide raffinose and followed by the tetra- and pentasaccharides stachyose and verbascose, respectively (Figure 7a, closed symbols and Data S2). At the final stages of maturation, 25–28 DAP, raffinose content decreased progressively (Data S2) and the most abundant sugars present in dry wild-type seeds were sucrose and stachyose. Silencing of MtSNF4b resulted in an impaired accumulation of stachyose and a higher content of sucrose compared to the control seeds from 28 DAP onwards (Figure 7a,c; compare closed with open symbols). No significant differences were observed for the accumulation of raffinose and verbascose, or for the decrease in the monosaccharides glucose and fructose between wild type and the transgenic lines (see Data S2). Altogether these sugars represented only 8% of the total soluble sugar content in dry seeds.

Figure 7.

 Changes in stachyose (a, b) and sucrose (c, d) content during maturation (a, c) and in radicles during imbibition (b, d) of seeds of control (solid symbols) and RNAi MtSNF4b lines (open symbols).
Sugar contents (μg mg−1) were determined on five replicates of 5–10 embryos or 15–20 radicles harvested from five to seven plants per line (average ± SE). Statistical analysis was performed using the Dunnet multiple comparisons test, the control corresponding to the wild-type values. An asterisk indicates a significant difference at the 99% level. The arrow indicates the time frame in which MtSNF4b is found in high-Mw fractions, indicating its presence in a complex (derived from Figures 2 and 3).

The difference in sugar composition in mature seeds was also found in radicles of the three transgenic lines, isolated 3 h after imbibition (Figure 7b,d). The total soluble sugar content in the radicles was higher that in whole seeds (15.6% vs. 9.4%), but sucrose (17%) and stachyose (73%) were still the most abundant sugars. Again, sucrose levels were over twofold higher in the MtSNF4b RNAi lines, whereas stachyose levels were 40–74% lower (Figure 7c,d). Upon further imbibition, stachyose content decreased in both wild-type and transgenic lines, albeit at a slower rate in the transgenic lines. At 20 h of imbibition, no differences were observed between the lines. Likewise, sucrose content increased in the wild-type seeds and decreased in the transgenic lines to reach a comparable value at 20 h of imbibition. As for maturation, the differences in sugar content were only observed when MtSNF4b could be found in a complex (see solid arrow in Figure 7).

Previously, it has been shown that the SnRK1 catalytic subunit is involved in starch metabolism (Laurie et al., 2003; Tiessen et al., 2003). To investigate whether starch levels are affected by the SNF4b complex, starch content was determined at different stages during maturation and early imbibition in two RNAi lines compared with wild-type seeds (Data S4). Starch levels are high in seeds harvested at 14 DAP (8% of the dry weight) and gradually decrease during further maturation. Around 28 DAP, starch is almost completely degraded and represents 0.1% of the dry weight in mature seeds. No differences were found in starch content between wild-type and RNAi MtSNF4b seeds throughout the development. Only at 24 and 28 DAP starch content in RNAi line 2 was significantly higher in than the wild type, but this difference was not found in line 1 (Data S4). Likewise, during imbibition starch accumulated in imbibing radicles but no difference could be found between the wild type and RNAi lines. Unlike antisensing of the catalytic subunit SnRK1, down-regulation of SNF4b does not affect starch metabolism.

In addition, to investigate whether silencing of MtSNF4b led to a change in total fatty acids (FA), their amount and composition were measured during development (Data S5 and S6). Both FA content and composition were comparable to those found previously by Djemel et al. (2005). Total FA content increased during early seed development to reach 13% of the dry weight and remained constant from 16 DAP onwards. No significant different in the amount or composition was observed between the wild-type or transgenic RNAi MtSNF4b lines. In conclusion, the only differences in metabolites that this study revealed between the seeds in which MtSNF4b was silenced and the wild-type seeds was a difference in di- and oligosaccharides.


This study shows that MtSNF4b accumulates at the half-way stage of seed development (around 12 DAP) and disappears after radicle emergence (Figure 1). Despite its abundance at earlier stages, MtSNF4b appears to assemble in a complex only at the final stages of maturation (28 DAP), when the seed filling process is complete and maturation drying begins. MtSNF4b shares 87% similarity with LeSNF4 of tomato, which is capable of complementing yeast snf1 and snf4 mutants and physically interacts with LeSNF1 and LeSIP1 in a glucose-dependent manner in yeast two-hybrid assays (Bradford et al., 2003). In addition, the homologue of LeSIP1 in M. truncatula, MtAKINβ2, was detected in the same high-Mw fractions as MtSNF4b, making it possible that they are both participating in the same complex. This suggests that the complex in which MtSNF4b participates is likely to correspond to a SnRK1 complex. The findings in our study are comparable to those of Estruch et al. (1992) who analysed the yeast Snf1 complex using comparable gel filtration techniques and detected Snf4 and Snf1 in a wide range of fractions, ranging from 700 to 40 kDa. Complex formation of SnRK1 with its subunits has been shown to concur with the activity of the complex (Crawford et al., 2001; Sugden et al., 1999). Therefore, it seems plausible that MtSNF4b activates a SnRK1 complex in seeds of M. truncatula when maturation drying commences (from 28 DAP onwards). The complex dissociates during imbibition in the radicles, at the point of radicle protrusion.

Accelerated ageing of the transgenic lines led to reduced germination percentage and seedling vigour compared with the wild-type seeds. Longevity of seeds is acquired at the later stages of maturation (Bailly et al., 2001; Sanhewe and Ellis, 1996), i.e. within the same time frame in which SNF4b is present in a complex. Considering the importance of the AMPK family in metabolic regulation in eukaryotes (Hardie et al., 1998; Kemp et al., 1999), it might be possible that SNF4b is important for the correct regulation of metabolism prior to entry into a state of quiescence. The importance of carbon metabolism in the acquisition of DT has become evident from the large number of genes involved in carbon metabolism that were differentially expressed in relation to DT in M. truncatula (Buitink et al., 2006). It has been previously suggested that oligosaccharides, or rather the ratio between sucrose versus oligosaccharides, is correlated to shelf life (Obendorf, 1997). However, the importance of RFOs in desiccation tolerance and seed longevity has been a contentious issue. Recent evidence from genetic, physiological and biophysical studies argues against a direct link between a protective effect of RFOs and seed survival in the dry state (Bentsink et al., 2000; Black et al., 1999; Buitink et al., 2000). It is possible that oligosaccharides serve as a readily available energy source during imbibition (Downie and Bewley, 2000; Peterbauer and Richter, 2001). This might be a plausible explanation for M. truncatula seeds, because stachyose makes up between 9 and 12% of the dry weight of a mature seed. Main et al. (1983) showed that an accelerated ageing treatment in soybean seeds had a significant effect on the differentiation of the embryonic axis into a seedling and that the reduced development may be related to an inability to mobilize or utilize stored carbon reserves.

The results of this study indicate that the activating subunit MtSNF4b is involved in the regulation of carbon metabolism, as has been found for other studies on the SnRK1 complex in plants. Most of these studies in seeds and other sink organs have focused on the catalytic α-subunit, and point to a role in regulating starch metabolism (Laurie et al., 2003; Tiessen et al., 2003; Zhang et al., 2001). For instance, antisensing the α-subunit resulted in a loss of AGPase activity in potato tubers (Tiessen et al., 2003), a decrease in α-amylase expression in wheat embryos (Laurie et al., 1998) and prevented starch accumulation during pollen development in barley (Zhang et al., 2001). In this study, no difference in starch content was detected between the wild-type and RNAi MtSNF4b seeds, either during development or imbibition. The assembly of SNF4b in a complex occurs only at 28 DAP, when the starch content is already low (<1% dw) in the seed embryos. Apparently, the particular SnRK1 complex that is regulated by SNF4b activates other targets than those described previously. In silenced MtSNF4b seeds, stachyose accumulated to only 20–50% of the wild-type content during maturation, whereas sucrose levels were 60% higher than in the wild-type seeds. This suggests that the conversion from sucrose to stachyose is affected. It remains to be investigated whether the reduced stachyose accumulation is the results of different steady-state sucrose levels, for instance via modification of a sugar-sensing pathway that is regulated by the SNF4b complex, or whether stachyose synthase is directly controlled by the complex. It is noteworthy that some stachyose already accumulated prior to the formation of the complex, indicating that additional factors exist that are responsible for controlling RFO metabolism. The different role found for MtSNF4b in carbon metabolism compared with other studies on the SnRK1 complex might indicate that there exist several SnRK1 complexes, each activated by different regulatory subunits. Indeed, downregulation of SnRK1 in pea seeds resulted in a strong pleiotropic phenotype, affecting the accumulation of storage proteins, embryogenesis and inhibition of vivipary (Radchuk et al., 2006). Although RFO content was not reported, sucrose levels were also higher than in the wild-type seeds. In the light of these observations, MtSNF4b appears to serve as a regulatory subunit that confers specificity towards carbon metabolism to the SnRK1 complex within a restricted time frame of seed development.

Experimental procedures

Plant material and treatments

Plants of M. truncatula Gaertn. (cv. Paraggio and R108) were grown in a sterile mix of vermiculite and soil in a growth chamber at 24°C/21°C, 16-h photoperiod at 200 μmol m−2 sec−2. Flowers were tagged and developing seeds were removed from the pods at different time intervals. Fresh weight and dry weight were determined gravimetrically on four replicates of five seeds and water content determined after heating the seeds for 48 h at 96°C. Seeds were imbibed on wet filter paper at 20°C in the dark after scarification by a scalpel, and seeds were considered germinated when the radicle protruded from the surrounding envelope. Desiccation tolerance was re-induced in sensitive radicles by selection of germinated seeds with a protruded radicle size of 2.8 mm and incubation in a PEG solution of −1.7 MPa at 10°C as described in Buitink et al. (2003b). Desiccation tolerance was determined by rapid drying over an airflow of 42% RH and rehydrating isolated seeds at different maturation stages (60–90 seeds) or after re-induction by the PEG treatment (50–70 seeds).

Two independent accelerated ageing experiments (100 seeds for each time point) were performed on seed lots obtained from two different cultures. Seeds were kept over a saturated solution of NaCl (75% RH) at 35°C in hermetically sealed boxes for different time spans, after which they were imbibed at 20°C and the percentage germination determined. Seed vigour was assessed by measuring the seedling length at 3 days of imbibition at 20°C of those seeds (about 30–50 individuals) that still germinated upon imbibition after 21 days of storage.

Plasmid construction and transformation procedure

To construct the MtSNF4b RNAi plasmid a specific 609-bp portion of MtSNF4b was amplified with MtSNF4b-F: 5′-GAGAGTATGAATCTAAAGG-3′ and MtSNF4b-R: 5′-AAATCTTGAACAGAATGTG-3′. The product of the amplification was transferred by recombination using the Gateway Cloning Technology Gateway (Invitrogen, into the binary vector pFGC5941/Gateway (pFRB). This plasmid contains a sense and antisense Gateway cassette, separated by a chalcone synthase A (ChsA) intron and driven by a 35S promoter. Transformation of the plasmid into R108 and in vitro culture was performed according to Trinh et al. (1998) using A. tumefaciens strain EHA105. After a first selection of positive lines by PCR, the seed populations of six transgenic lines were further characterized for a 3:1 segregation of the RNAi MtSNF4b insertion by screening half of the cotyledons of 40 seeds for the presence of MtSNF4b by Western blot (for strategy see Data S1). The remaining halves of the seeds containing the radicle were grown into T1 plants. For three lines with 3:1 segregation, 60 seeds of 12 individual plants were analysed by Western blot for selection of RNAi MtSNF4b homozygotes. Three homozygous lines (T2) were retained for further analysis.

Southern blotting and RT-PCR

Genomic DNA from the wild type and transformed RNAi-MtSnf4b was isolated from young leaves according to Dellaporta et al. (1983). Aliquots (10 μg) were digested with EcoRI, EcoRV and BamHI, fractionated on 0.8% (w/v) agarose gels and transferred to Hybond-N+ membranes (Amersham Biosciences, Hybridization was performed using the 609-bp 32P-radiolabelled (α-dCTP) MtSnf4b cDNA fragment used for MtSNF4b RNAi vector construction (random prime labelling system, Amersham). The efficiency of the silencing of MtSNF4b was determined by RT-PCR. Ribonucleic acid was extracted from dry seeds following Verwoerd et al. (1989) and RT was performed as described in Buitink et al. (2003b). A cDNA fragment (609 bp) of MtSNF4b was amplified using 5′-GAGAGTATGAATCTAAAGG-3′ and 5′-AAATCTTGAACAGAATGTG-3′ as forward and reverse primers, respectively. An 18S rRNA fragment (427 bp) was used as a control and was amplified with 5′-CCAGGTCCAGACATAGTAAG-3′ and 5′-GTACAAAGGGCAGGGACGTA-3′ as forward and reverse primers, respectively. The PCR consisted of a preliminary denaturation step of 4 min at 94°C, followed by 35 cycles for MtSNF4b and 25 cycles for 18S of 1 min at 94°C, 1 min at 60°C and 1 min at 72°C, and a final elongation step of 10 min at 72°C.

Immunoprecipitation and mass spectrometry

Immunoprecipitation of MtSNF4b using a rabbit polyclonal antibody raised against N-terminal synthetic peptide of MtSnf4b (KEKKVKDLMVNKKC-NH2) or the pre-immune serum on native protein extracts of mature seeds was performed according to Ferrando et al. (2001) using 500 μg of soluble proteins and 2 μg purified anti-SNF4b antibody. The antibody was purified with the peptide coupled to CNBr-activated Sepharose 4b following the manufacturer’s instructions (Amersham). The excised band that was specifically immunoprecipitated using the anti-SNF4b antibody was sequenced using liquid chromatography-tandem mass spectrometry (LC-MS/MS) according to Boudet et al. (2006) using in-gel tryptic digestion. The LC-MS/MS analysis was performed using a nanoscale HPLC (Famos-Switchos-Ultimatesystem, LC Packings, coupled to a hybrid quadrupole orthogonal acceleration time-of-flight mass spectrometer (Q-TOF; Global, Micromass-Waters, Protein identification was performed by comparing the sequence with the UniProt sequence databank and the Institute for Genomic Research Medicago expressed sequence tag (EST) databank (date of release 26 January 2005).

Sodium dodecyl sulphate-PAGE and immunoblot analysis

Material for the transformant selection (two half cotyledons) and analysis of Mtsnf4 abundance (maturation 40–50 seeds, imbibition 50 radicles or 25 cotyledons, PEG incubation 50 radicles, 6-week-old leaves before or after 3 days’ drying on the bench) was frozen into liquid nitrogen directly after excision and stored at −80°C until extraction. Proteins were extracted according to Boudet et al. (2006). Samples were separated by SDS-PAGE on 12% polyacrylamide gel using standard molecular biology techniques. Separated proteins were stained with Coomassie Brilliant Blue R-250 or transferred to nitrocellulose membranes (Schleicher & Schuell Inc., Immunodetection was performed by chemiluminescence as described by Boudet et al. (2006) using a rabbit polyclonal antibody (1:1000) raised against N-terminal synthetic peptide of MtSnf4b.

Gel filtration chromatography

Soluble proteins were extracted at 4°C in 50 mm 2-amino-2-(hydroxymethyl)-1,3-propanediol (TRIS)-HCl pH 7.8, 1 mm EDTA, 1 mm EGTA, 1 mm DTT, 1% (v/v) glycerol, 1 mm NaVO4, 1% (v/v) IGEPAL-CA630 and 150 mm NaCl. After extraction, protease inhibitors (1/50 of final volume) were directly added (Sigma, After two consecutive centrifugations at 13 000 g at 4°C, protein concentrations were assayed according to Bradford (1976). Separation of the freshly extracted proteins by gel filtration was performed at room temperature on a Superdex 200 HR 10/30 column (Amersham Pharmacia Biotech, equilibrated in a native buffer (50 mm TRIS-HCl pH 7.8, 1 mm EDTA, 1 mm EGTA, 1 mm DTT, 1% (v/v) glycerol, 1 mm NaVO4 and 150 mm NaCl) with a flow rate of 0.25 ml min−1. Fractions of 0.5 ml were lyophilized and analysed by Western blotting.

Soluble sugar and starch determination

Five to ten embryos harvested at different stages of maturation (12–40 DAP), five to ten dry mature seeds (10–20 mg of dry matter) or 15–20 radicles (5–10 mg dw) harvested at different times during imbibition were lyophilized and dry weight was determined gravimetrically. Seed parts were ground in a mortar in the presence of 1 ml 80% methanol containing melizitose as the internal sugar standard. After heating at 76°C for 15 min, the liquid was evaporated under vacuum. The residue was dissolved in 1 ml distilled water and centrifuged for 1 min at 13 000 g. Sugars were analysed by HPLC on a Carbopac PA-1 column (Dionex Corp., as described by Hoekstra et al. (1994). Data are the average of five extractions from seed material obtained from five to seven different plants. Starch analysis was performed on the pellets after soluble sugar extraction. Pellets were freeze-dried, washed, weighed and dissolved in 50 μl of 8 n HCl and 200 μl concentrated DMSO. After 1 h agitation at 60°C, 300 μl of distilled water and 80 μl of 5 n NaOH were added before the addition of 360 μl of citrate buffer pH 4.6 (0.1 m citric acid and 0.2 m Na2HPO4.2H2O). After centrifugation (1 min at 13 000 g), the supernatant was used for starch determination using the ‘EnzyPlus Starch’ kit analysis (Diffchamb Group, following the manufacturer’s protocol.

Lipid extraction and analysis

Five seeds (10–20 mg of dry matter) harvested at different stages of maturation (12–40 DAP) were ground in 3 ml of CHCl3/methanol (MeOH; 2/1, v/v) after addition of the internal standard triheptadecanoin (4 mg). The CHCl3/MeOH extract was sonicated for a few minutes to improve extraction and subsequently cleared and dried by passing it over a column of anhydrous Na2SO4. The lipids were transmethylated with 2.5 ml of 0.2 m KOH in methanol for 15 min at 70°C under vigorous shaking. After cooling on ice, 1 ml of a saturated NaCl solution and 1 ml of hexane were added, and the fatty acid methylesters were phase-separated to the hexane (2 min of centrifugation at 1000 g) and passed over an anhydrous Na2SO4 column prior to GC analysis. The fatty acid methylesters were analysed on a Shimadzu GC8A GC (, equipped with a 30-mm J&W DB225 megabore column (J&W Scientific, coupled to a Spectra Physics Chromjet integrator ( Identification was done by comparing retention times with standards. Quantification of the different components was achieved by comparing the total peak surface of methyl esters with the peak area of the methyl ester of the internal standard. Data are the average of three extractions.


We thank Dr Marten Denekamp (Syngenta Seeds, Enkhuizen, The Netherlands) for providing the pFRB RNAi binary vector. D. Jamar and Dr F.A. Hoekstra (Laboratory for Plant Physiology, Wageningen University and Research Centre, The Netherlands) are gratefully acknowledged for their assistance with the sugar and lipid measurements. We thank B. Jettner (Seed-Co Australia Co-Operative Ltd, Hilton Australia) and P. Ratet (Institut des Sciences Végétales, CNRS, Gif sur Yvette, France) for the gift of seeds of M. truncatula cv. Paraggio and R108, respectively. F. Frugier (Institut des Sciences Végétales, CNRS, Gif sur Yvette, France) is gratefully acknowledged for the gift of the transgenic line with the empty pFBR plasmid.