This Article Abstract Full Text (PDF) Supplemental Data All Versions of this Article: 140/4/1418    most recent pp.105.074039v1 Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Boudet, J. Articles by Leprince, O. Search for Related Content PubMed PubMed Citation Articles by Boudet, J. Articles by Leprince, O. Agricola Articles by Boudet, J. Articles by Leprince, O.

ENVIRONMENTAL STRESS AND ADAPTATION TO STRESS

Comparative Analysis of the Heat Stable Proteome of Radicles of Medicago truncatula Seeds during Germination Identifies Late Embryogenesis Abundant Proteins Associated with Desiccation Tolerance^1,[W]

Unité Mixte de Recherche 1191, Physiologie Moléculaire des Semences (Université d'Angers, Institut National d'Horticulture, Institut National de la Recherche Agronomique), Anjou Recherche Semences, 49045 Angers, France (J. Boudet, J. Buitink, P.S., O.L.); Laboratory of Plant Physiology, Department of Plant Sciences, Wageningen University, 6703 BD Wageningen, The Netherlands (F.A.H.); and Unité de Recherche, Biopolymères, Interactions, Allergie, Institut National de la Recherche Agronomique, 44316 Nantes, France (H.R., C.L.)

	ABSTRACT

TOP ABSTRACT RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

 
A proteomic analysis was performed on the heat stable protein fraction of imbibed radicles of Medicago truncatula seeds to investigate whether proteins can be identified that are specifically linked to desiccation tolerance (DT). Radicles were compared before and after emergence (2.8 mm long) in association with the loss of DT, and after reinduction of DT by an osmotic treatment. To separate proteins induced by the osmotic treatment from those linked with DT, the comparison was extended to 5 mm long emerged radicles for which DT could no longer be reinduced, albeit that drought tolerance was increased. The abundance of 15 polypeptides was linked with DT, out of which 11 were identified as late embryogenesis abundant proteins from different groups: MtEm6 (group 1), one isoform of DHN3 (dehydrins), MtPM25 (group 5), and three members of group 3 (MP2, an isoform of PM18, and all the isoforms of SBP65). In silico analysis revealed that their expression is likely seed specific, except for DHN3. Other isoforms of DNH3 and PM18 as well as three isoforms of the dehydrin Budcar5 were associated with drought tolerance. Changes in the abundance of MtEm6 and MtPM25 in imbibed cotyledons during the loss of DT and in developing embryos during the acquisition of DT confirmed the link of these two proteins with DT. Fourier transform infrared spectroscopy revealed that the recombinant MtPM25 and MtEm6 exhibited a certain degree of order in the hydrated state, but that they became more structured by adopting helices and sheets during drying. A model is presented in which DT-linked late embryogenesis abundant proteins might exert different protective functions at high and low hydration levels.

LEA proteins are classified in at least five groups by virtue of similarity in their amino acid sequences (Cuming, 1999; Wise, 2003). They are low complexity, highly hydrophilic, and mostly unordered proteins in the hydrated state, and heat stable (HS) after boiling (Cuming, 1999; Wise, 2003). Generally, the presence of LEA proteins correlates well with DT. LEA proteins accumulate to high levels in developing seeds during late maturation (Blackman et al., 1995; Cuming, 1999; Buitink et al., 2002) and in dehydrating vegetative tissues of resurrection plants (Ramanjulu and Bartels, 2002). Correlations between the disappearance of various members of groups 1, 2, and 3 LEA proteins and loss of DT during germination have also been reported (Ried and Walker-Simmons, 1993; Whitsitt et al., 1997; Capron et al., 2000; Gallardo et al., 2001). Seeds of the double mutant aba,abi3 of Arabidopsis (Arabidopsis thaliana) that are deficient in several HS polypeptides are also desiccation sensitive, although they exhibit an array of pleiotropic defects ranging from decreased accumulation of storage proteins to vivipary (Meurs et al., 1992). A role of LEA proteins in DT has been demonstrated in the bacterium Deinococcus radiodurans. Inactivation of a group 3 LEA protein, homolog of the plant LEA76, leads to 75% reduction in viability of desiccated cultures (Battista et al., 2001). In contrast, in seeds, direct in vivo evidence for a role of LEA proteins in tolerance to complete water loss has not yet been secured. There exist several Arabidopsis mutants that produce seeds devoid of one or two LEA proteins belonging to group 1, but remaining desiccation tolerant (Carles et al., 2002), either arguing against a role for these proteins in DT or revealing a possible functional redundancy between the different group members as suggested by Ditzer et al. (2001). Conversely, dehydrins have been detected in seeds that remain desiccation sensitive at shedding (Kermode, 1997). Nonetheless, in vitro experiments do point to a protective role of LEA proteins against the deleterious effects of drying. For instance, several LEA proteins from groups 2, 3, and 4 were found to protect enzymes against nearly complete loss of water brought about by rapid evaporation or vacuum drying (Goyal et al., 2005; Grelet et al., 2005).

In addition to being present in anhydrobiotes, LEA proteins are also expressed in desiccation-sensitive vegetative tissues as a response to stress involving changes in cellular water potential (Cuming, 1999). Most of the experimental evidence shows that LEA proteins that are overexpressed in vegetative tissues can improve tolerance to various degrees of hyperosmotic stress (–1 to –6 MPa), induced by a partial loss of water, salt, or freezing (Imai et al., 1996; Swire-Clark and Marcotte, 1999; Cheng et al., 2002; Houde et al., 2004; Riera et al., 2004). Whether LEA proteins play a similar role in seeds as they do in drought-tolerant systems is unclear. LEA proteins are extremely diversified in terms of genotypic variability, regulation, and localization at the tissue and cellular level (Dure, 1993; Cuming, 1999; Wise, 2003). Several LEA genes appear to be specifically expressed in seeds, such as the Em1 and Em6 in Arabidopsis (group 1; Bies et al., 1998) and rab28 (group 5) in Arabidopsis (Arenas-Mena et al., 1999) and maize (Zea mays; Niogret et al., 1996). When these genes are overexpressed in desiccation-sensitive systems such as yeast (Saccharomyces cerevisiae; Swire-Clark and Marcotte, 1999), leaves of rice (Oryza sativa; Cheng et al., 2002), and Arabidopsis seedlings (Borrell et al., 2002) an improved tolerance to salt or drought is observed. During drying, seed tissues pass through hydration ranges that also necessitate protection against drought. Thus, it can be argued that in seeds, the role of LEA proteins might be similar as in drought-tolerant vegetative tissues, their action being confined to relatively high water contents (approximately –3.5 MPa; Hoekstra et al., 2001; Ramanjulu and Bartels, 2002). In this case, no LEA proteins would be found to be specifically correlated with tolerance to low water contents. Alternatively, they might exert several functions that differ according to the hydration level reached by the seed tissues during drying, as is the case for nonreducing sugars (Hoekstra et al., 2001; Ramanjulu and Bartels, 2002). Sugars are thought to act as compatible solutes during the initial water loss and when the bulk water is removed, they protect macromolecules by replacing water with OH groups and by forming a glass, which stabilizes the macromolecular structures for long periods of time (Hoekstra et al., 2001). Similarly, apart from their protective role in drought conditions, LEA proteins have been shown in vitro to prevent conformational changes of hydrophilic polypeptides when the last hydration layer is removed (Wolkers et al., 2001) and to participate in the formation of a glassy state, occurring at a water content below 0.10 g/g (g water/g dry weight; Wolkers et al., 2001; Shih et al., 2004). Therefore, this study investigates if there exist specific LEA proteins whose abundance is associated with DT rather than drought tolerance in Medicago truncatula seeds.

To comprehend the changes in LEA proteins simultaneously, comparative proteomic analysis was carried out in desiccation-tolerant and -sensitive radicles during germination. In addition, this approach allows to assess whether putative posttranslational modifications are also associated with DT, considering that some LEA proteins from groups 1, 2, and 3 are submitted to posttranscriptional and posttranslational modifications during seed maturation and germination (Bies et al., 1998; Campalans et al., 2000). Furthermore, the phosphorylation status of the acidic dehydrins was found to determine the protective activity (Riera et al., 2004; Alsheikh et al., 2005). To facilitate the detection of LEA proteins, we focused on the HS proteome that resists coagulation upon heating at 95°C. By this method, the soluble protein extract containing hydrophilic proteins should be enriched with LEA proteins and devoid of storage proteins, which can represent up to 60% of the total proteome of M. truncatula seeds (Gallardo et al., 2003; Watson et al., 2003). Apart from LEA proteins, other unidentified proteins with protective functions might be present in the HS fraction. This is based on a recent argumentation that LEA proteins are members of a larger family of osmotic stress proteins called hydrophilins, which are defined as proteins that have a Gly content >6% and a hydrophylicity index >1 (Garay-Arroyo et al., 2000).

Several transcriptomic and proteomic analyses both during seed development (Gallardo et al., 2003; Hajduch et al., 2005) and germination (Gallardo et al., 2001; Soeda et al., 2005) have given some insights into several seed-specific events such as synthesis of storage reserves, desiccation, radicle protrusion, and germination performance. Although several stress proteins were identified, their abundance was not studied in relation to DT. In our work, profiles of HS proteins were compared between desiccation-tolerant radicles of nongerminated (NG) seeds and sensitive radicles after emergence out of the seed coat. To confirm the link with DT, profiles were also studied in emerged radicles in which DT was reestablished. This can be brought about by exposing germinated seeds to an osmotic treatment for several days (Leprince et al., 2000; Buitink et al., 2003). Those proteins that were expressed in treated radicles upon reestablishment of DT were further analyzed in older emerged radicles, for which DT can no longer be reinduced by the osmotic treatment. This strategy allowed for the discrimination of putative proteins linked to DT and osmotic tolerance. Eleven polypeptides representing several forms of LEA proteins were found to be associated with DT. Among them, two proteins belonging to group 1 and group 5 were further characterized during maturation and germination by western blotting. To gain further insights into their function, their secondary structure was compared in the hydrated and dry state after fast and slow drying using Fourier transform infrared (FTIR) spectroscopy.

	RESULTS

TOP ABSTRACT RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

 

Changes in HS Protein Patterns in Relation to the Loss and Reestablishment of DT

In seeds of M. truncatula, DT of the radicle is maintained during the early phase of imbibition and is lost when the radicle protrudes the seed coat (Table I ). Germinated seeds with 2.8 mm long protruded radicles are not able to survive a 3 d drying at 42% relative humidity (RH) at 20°C (Table I). Previously, Buitink et al. (2003) showed that DT can be reestablished in these sensitive radicles by incubating the germinated seeds in a solution of polyethylene glycol (PEG) having a water potential of –1.7 MPa for 2 d. Table I shows that in these conditions, DT was restored to 91% in 2.8 mm long emerged radicles. However, when germinated seeds are selected at a later stage during postgerminative growth, corresponding to seeds with protruded radicles of 5 mm in length, DT can no longer be reestablished after the same PEG treatment (Table I).

View larger version (17K): [in this window] [in a new window]   Figure 2. The relation between DT and the water content of the radicles of M. truncatula at different intervals during fast drying. Germinated seeds exhibiting a protruded radicle of 2.8 and 5 mm were first incubated or not (control) in a PEG solution for 2 d then dried for different intervals. The data from two independent experiments are pooled together.

View larger version (86K): [in this window] [in a new window]   Figure 1. Two-dimensional electrophoresis of the HS proteome of NG radicles excised from M. truncatula seeds that were imbibed for 16 h using 24 cm nonlinear immobilized pH gradient strips (3–10). pI and molecular mass (in kilodaltons) are noted. Numbers indicate the spots that were identified.

A total of 23 spots showed a higher abundance in both desiccation-tolerant stages compared to the sensitive stage (profile 9, Table II). These spots were further analyzed using two additional stages: 5 mm long, desiccation-sensitive emerged radicles before and after a PEG treatment (Table I). As a result, the spots could be separated in two subgroups: those that are only induced in the 2.8 mm long radicles after PEG treatment and are thus linked specifically to DT (subgroup A) and those that are also induced by the PEG incubation in 5 mm long radicles that remain desiccation sensitive (subgroup B). Out of the 23 spots, 11 were found to be associated specifically with the induction of DT (subgroup A) and seven were induced both in 2.8 and 5 mm long PEG-treated radicles (subgroup B), albeit not always to similar levels in both tissues. Among these seven spots, five exhibited a significantly higher intensity in 2.8 mm PEG-treated radicles than in the 5 mm ones. They were therefore also linked to DT. The remaining five spots could not be categorized in either of these two groups and are not further considered in this study.

It is noteworthy that, although the PEG incubation of the 5 mm long radicles did not lead to the reestablishment of DT, this treatment did result in an improvement of the tolerance to drying. This is demonstrated by the assessment of the water content to which 50% of the population of germinated seeds can be dried and rehydrated without loss of viability of their radicle (threshold water content, Fig. 2 ; Table I). Germinated seeds with an emerged radicle of 2.8 mm long were able to survive a desiccation treatment down to 1 g/g, but died at lower water content. Fifty percent of survival was obtained at 0.3 g/g (Table I). After the PEG treatment, 2.8 mm long emerged radicles were able to survive nearly complete removal of water and were thus considered desiccation tolerant (Fig. 2). In contrast, 5 mm long radicles were very sensitive to drying. Fifty percent of death was obtained when the radicles were dried to 3.6 g/g. However, after a 2 d incubation in the osmoticum, they had become more tolerant to desiccation since the threshold water content decreased to 0.8 g/g (Table I). Whitsitt et al. (1997) observed a similar effect for soybean (Glycine max) seedlings: an incipient water deficit decreased the sensitivity of seedlings to further dehydration. The division of profile 9 into the two subgroups A and B showed that certain spots could only be induced in those tissues that become desiccation tolerant, whereas others could also be reinduced by the PEG incubation in 5 mm long radicles that remained desiccation sensitive (data not shown).

Identification of HS Proteins Associated with DT

The 16 spots belonging to profile 9A and 9B and linked with DT were analyzed using matrix-assisted laser desorption ionization time of flight (MALDI-TOF) mass spectrometry and liquid chromatography-tandem mass spectrometry (LC-MS/MS). Out of these 16 spots, 11 were identified as six different LEA proteins, some of them being present as different isoforms (Table III ). Another polypeptide was identified as a homolog of a pea (Pisum sativum) legumin precursor (Table III). Since the M_r of this spot was much lower than expected, we suspected that the onset of the digestion of storage proteins, which is known to occur during radicle growth (Capron et al., 2000; Gallardo et al., 2001), yielded small, hydrophilic peptides. This legumin fragment represented less than 0.05% of the HS proteome. It was not therefore taken into account for the remainder of this study. Finally, four polypeptides were not identified.

View larger version (15K): [in this window] [in a new window]   Figure 3. Phylogenic tree of LEA proteins of M. truncatula. ClustalX was used to create an alignment of the following translated tentative consensus that was found in the Medicago database (http://www.tigr.org/tdb/tgi/mtgi). The alignment was bootstrapped (n = 1,000 replicates) to create the final tree with the bootstrap values indicated. Underlined LEA proteins indicate those that have been found associated with DT (see Table III). BudCar5 (TC 100264 homolog of Q9M603 of Medicago sativa); CapLEA-1a (TC94389 similar to 049816 of Cicer arietinum); CapLEA-1b (TC112317 similar to 049816 of C. arietinum); DHN (dehydrin; TC101013 similar to 023957 of G. max); DHN3 (or dehydrin like, TC 100921, similar to Q945Q7 pea); DHN-Cog (dehydrin cognate, TC106659 similar to Q43430 of pea); DIP (drought induced protein, TC95389 similar to Q941N0 of R. raetam); ECP31 (TC 96862 similar Q96245 of Arabidopsis); Em6 (TC96799 homolog to P93510 of R. pseudoacacia); Lea14 (TC 101891, similar to P46519 of G. max); LEA5 (TC105834 similar to O24422 of G. max), LEA D34a (TC102411 weakly similar to P09444 LEA D34 of G. hirsutum); LEA D34b (combination of TC107159 and –158, both similar to P09444 LEA D34 of G. hirsutum); LEA likea (TC108292 weakly similar to Q6Z4J9 of rice); LEA likeb (TC 95012 similar to Q8LCW6 of Arabidopsis); PM32 (mitochondrial LEA, TC101811 and TC78559 similar to Q5NJL5 of pea); MP2 (maturation polypeptide; TC95538 and TC87025 similar to Q39871 of G. max); MtPM25 (this study); PM1 (TC96465 similar to Q01417 of G. max); PM10 (TC100258, TC85220 similar to Q39801 of G. max); PM18 (or 35 kD seed maturation protein, TC96265 similar to Q9ZTY1 of G. max); PM22 (translated BQ124186 similar to Q9XER5 of G. max); PM24 (TC96521 weakly similar to Q9SEL0 of G. max); RAB21 (TC 107383 similar to P12253 of rice); SBP65 (seed biotinylated protein, TC102224 similar to Q41060 of pea); SMP LEA-4 (TC94509 similar to Q9FNW8 of G. tabacina). Whenever possible, the classification according to Dure (1993) and Cuming (1999) together with the PFam assignments are indicated. The scale bar (number of substitutions/site) corresponds to the relative branch length.

View larger version (60K): [in this window] [in a new window]   Figure 4. Changes in three HS proteins identified as PM25 (A and B), Em6 (C and D), and MP2 (E and F) associated with DT in germinating radicles of M. truncatula. Representative 2D gels (A, C, and E) and relative spot quantities (B, D, and F) during germination (white bars) and after a PEG treatment (black bars) in 2.8 and 5 mm long emerged radicles. Spot numbers are taken from Figure 1. pI and molecular mass (in kilodaltons) are indicated on the left gel. For MP2, the histogram corresponds to the most preponderant isoform (spot 19). Different letters shown above the bars represent significant differences after multiple comparison of the means (P < 0.05).

View larger version (64K): [in this window] [in a new window]   Figure 5. Changes in spots identified as PM18 (A and B) and SBP65 (C and D) that are associated with DT in germinating radicles of M. truncatula. Representative 2D gels (A–C) and relative spot intensities (B and D) corresponding to PM18 (B, spots 15–18) and SBP65 (D, spots 1–5) during germination (white bars) and after a PEG treatment (black bars) in 2.8 and 5 mm long radicles. Spot number assignment is taken from Figure 1. pI and molecular mass (in kilodaltons) are indicated on the left gel. Different letters shown above the bars represent significant differences after multiple comparison of the means (P < 0.05).

View larger version (65K): [in this window] [in a new window]   Figure 6. Changes in several spots identified as DHN3 (A and B) and Budcar5 (C and D). Representative 2D gels (A) and relative spot intensities (B) during germination (white bars) and after a PEG treatment (black bars) in 2.8 and 5 mm long emerged radicles. Spot number assignment is taken from Figure 1. pI and molecular mass (in kilodaltons) are indicated on the left gel. Different letters shown above the bars represent significant differences after multiple comparison of the means (P < 0.05).

View larger version (53K): [in this window] [in a new window]   Figure 7. In silico expression analysis of LEA genes in M. truncatula identified in this study (see Table III). The number of ESTs present in various libraries that were found at http://www.tigr.org/tdb/tgi/mtgi and exhibited similar characteristics (organs, developing seeds, stress, etc.) were averaged and expressed as percent of detected EST x 1,000. The absence of value means that the EST was absent in the libraries. The corresponding libraries are: leaf (T10110, 4046, 4049), flower (#9D5, #ARC), nodulated root (T1617, T10109, 4047), seed (#ARD, T10493, T10494, T11127), drought (5413), biotic stress (#A8V, T11031, T1581, T10014), N deficit (#8GI, 5518, #8GF), P starvation (5415), and plant microbe interaction (T1748, 7263, T10173, T1815, T1510, T1707, T1682, #ARE, #CDE, 5520).

Changes in MtPM25 and MtEm6 in Relation to DT during Seed Maturation and Germination

The expression profiles of two of the six LEA proteins that were linked to DT were further characterized to confirm the data obtained from the 2D proteomic analysis. The analysis was extended to cotyledons during germination and embryos during seed development to ascertain the abundance of these LEA proteins with DT. MtPM25 and MtEm6 were chosen because they were represented by a single spot in the gels, thereby alleviating any complication with the interpretation of western blots that were performed in one dimension. Full-length cDNAs corresponding to the MtPM25 (DQ206870) and MtEm6 (DQ206712) were obtained by RACE. Sequences corresponding to an N-terminal poly-HIS tag and cleavage site for enterokinase were added to the full-length encoding sequence and the recombinant proteins were expressed in Escherichia coli. Rabbit polyclonal antibodies were raised against the purified recombinant MtPM25 and MtEm6. For each antibody, a signal at the expected molecular size was detected both with protein extracts from radicles and the recombinant protein. The signals were absent when the preimmune serums were used (data not shown). During seed imbibition, contents of both MtPM25 and MtEm6 in the radicles remained high for up to 15 h (Fig. 8, A and E ). In 20 h imbibed radicles (approximately 2.8 mm in length), MtPM25 was barely detectable, whereas MtEm6 had already disappeared. In accordance with the proteomic analysis, the osmotic treatment was found to reinduce the expression of both proteins in 2.8 mm long emerged radicles, albeit to lower levels than those found in NG radicles (Fig. 8, C and G). In 5 mm long radicles, the PEG treatment only resulted in the appearance of a very faint signal. The relationship between DT and the presence of both proteins was also confirmed for the cotyledons during germination (Fig. 8, B and F). In contrast to radicles, DT in cotyledons was maintained for up to 24 h of imbibition and lost at 48 h. In parallel, MtEm6 and MtPM25 amounts decreased to barely detectable levels and disappeared. During seed development, tolerance to rapid enforced drying was acquired between 14 and 22 d after pollination (DAP; Fig. 8). Contents of MtPM25 increased at 14 DAP in parallel with the acquisition of DT (Fig. 8D), whereas those of MtEm6 started to accumulate later at 18 DAP.

View larger version (41K): [in this window] [in a new window]   Figure 8. Western-blot analysis of MtPM25 (A–D) and MtEm6 (E–H) in relation to DT: in radicles (A and E) and cotyledons (B and F) during seed imbibition; in emerged radicles (C and G) having a length of 2.8 and 5 mm, before and after the PEG treatment; and in embryos (D and H) during the acquisition of DT during seed maturation (10–40 DAP). Twenty (embryos and radicles) and 50 µg (cotyledons) of soluble proteins were hybridized with rabbit antibodies against recombinant MtPM25 and MtEm6. Independent protein extractions and immunoblots were performed in triplicates and yielded identical results. Percentage of DT and molecular marker mass (kilodaltons) are indicated. DS, Dry seeds.

It has been established that LEA proteins of groups 3 and 4 undergo an unordered-to-ordered structure transition during the loss of water (Wolkers et al., 2001; Goyal et al., 2003; Shih et al., 2004). Considering their divergence in the Kyte and Doolittle hydrophilicity profile (http://ca.expasy.org/tools/protscale.html), possible differences in secondary structure of the recombinant form of MtPM25 and MtEm6 were investigated. Whether changes in protein conformation were induced upon drying was also assessed. FTIR spectra of recombinant proteins in the hydrated and dried state were recorded after removal of the His_6x tag (Fig. 9 ). To avoid interference of the H-O-H scissoring vibration of water around 1,646 cm^–1 with the amide-I band between 1,700 and 1,600 cm^–1, D₂O instead of water was used for the proteins in solution. Wolkers et al. (1998) demonstrated that intermolecular -sheet formation can effectively be prevented by fast drying, probably because the time required for such nonintramolecular rearrangements is too short. For this reason, we studied the recombinant proteins after fast and slow drying in air of approximately 3% and 67% RH, respectively.

View larger version (11K): [in this window] [in a new window]   Figure 9. FTIR absorption spectra in the amide region of detagged, recombinant MtPM25 and MtEm6. Conditions of the proteins were: A, hydrated in D2O; B, after fast drying (FD) in an air stream of 3% RH; and C, after slow drying (SD) in circulating air of 67% RH.

More detailed information on secondary structures in these proteins was obtained by a curve-fitting procedure on the original amide-I band according to Wolkers et al. (2001). An example of the curve-fitting procedure is given in Figure 10 with fast-dried MtEm6 protein. Peaks representing different secondary structures were selected on account of the second derivative spectrum (Fig. 10A). Coaddition of all the dashed peaks that were mathematically produced should result in a fit (crosses) that resembles the original absorption spectrum (Fig. 10B). Individual contributions by the various protein secondary structures can thus be estimated. Table IV shows the curve-fitting results of the amide-I region of both proteins in D₂O and after fast or slow drying. In solution both proteins had between 30% and 40% -helical structure, with MtPM25 having more extended sheet than MtEm6. If both random and turn structures were to be combined and considered as unordered structures, even though a certain degree of order might exist for some of them, then MtEm6 would have 37% helix, 10% sheet, and 53% unordered structures, whereas MtPM25 would consist of 33% helix, 18% sheet, and 49% unordered structures (Table IV). These figures concur fairly well with PELE predictions (San Diego Supercomputer Center Biology workbench: http://workbench.sdsc.edu). In water solution, MtEm6 is predicted to form 33% helix, 3% sheet, and 64% unordered structures, whereas 38% helix, 14% sheet, and 48% unordered structures are expected for MtPM25. Support for the largely unordered nature of both Lea proteins in water also comes from the behavior of the amide-II band around 1,540 cm^–1 in D₂O (Fig. 9). This band was considerably smaller than that upon fast drying and partly downshifted to 1,450 cm^–1. The effect of D₂O was stronger for MtPM25 than for MtEm6. Apparently, the amide protons (N-H) were, to a considerable extent, open for ²H exchange from D₂O, which is interpreted to mean that both proteins have a fairly unordered structure in water (Haris et al., 1989).

View larger version (17K): [in this window] [in a new window]   Figure 10. Curve-fitting procedure illustrated for the recombinant MtEm6 after fast drying in an air stream of 3% RH. The absorption maxima of the different protein secondary structures in the amide-I band were selected on account of peak positions in the second-derivative spectrum (A). With these different components a least square iterative curve fitting was performed to fit Voigt line shapes to the original spectrum between 1,720 and 1,600 cm–1 (B). Prior to curve fitting a straight base line passing through the ordinates at 1,720 and 1,600 cm–1 was subtracted. Coaddition of all of the dashed peaks that were mathematically produced resulted in a fit (crosses) that resembled the original absorption spectrum (gray line).

	DISCUSSION

TOP ABSTRACT RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

 
To identify proteins involved in DT, a proteomic screening of the HS fraction of soluble proteins from imbibed radicles of M. truncatula was combined with a physiological system that enables the reestablishment of DT in 2.8 mm long emerged radicles by an osmotic treatment. To separate the proteins induced by the osmoticum from those involved in DT, the comparison was extended to emerged radicles of 5 mm long, for which DT could no longer be reinduced by the same osmotic treatment. In total, 15 polypeptides were found, whose abundance was linked to DT. Among them 11 were identified, which represented six LEA proteins from different groups: MtEm6 (group 1), one isoform of DHN3 (dehydrins), MtPM25 (group 5), and three members of group 3 (MP2, the basic isoform of PM18, and all the isoforms of SBP65 ;Table III). Our in silico analysis revealed that the expression of all the DT-linked LEA genes was apparently seed specific, except for one isoform of DHN3.

The abundance of all these proteins was associated with DT (Figs. 4, 5, and 8). Nonetheless, the causal relationship between the six LEA proteins and DT remains difficult to assess. It is possible to obtain dry and viable seeds from Arabidopsis and maize mutants with very low or undetectable levels of Em transcripts (Williams and Tsang, 1991; Carles et al., 2002). It is not known whether in these mutants, other (LEA) proteins could compensate for the absence of Em proteins. In addition LEA proteins might act synergistically with other protective compounds in the dry state. For instance, the combination of LEA proteins and nonreducing sugars offers better protection against protein aggregation after drying than each component alone (Goyal et al., 2005). In vivo, cytoplasmic glasses are thought to be composed of sugars and other compounds (for review, see Buitink and Leprince, 2004). In vitro, a mixture of LEA proteins and sugars forms a glass upon drying that exhibits physicochemical properties resembling those of cytoplasmic glasses, whereas a glass made of sugars alone has different properties (Wolkers et al., 2001; Buitink and Leprince, 2004; Shih et al., 2004).

This study also identified several LEA proteins that are linked to drought tolerance rather than DT, such as several isoforms of DHN3 (spots 12 and 13) as well as isoforms of BudCar5 (Fig. 6). Indeed, the isoforms of these dehydrins were induced not only in the 2.8 mm long radicles after PEG incubation, but also in the 5 mm long PEG-treated radicles. However, although the treatment on the 5 mm long radicles did not reestablish DT, it did lead nevertheless to an increased tolerance to drying, evident from the reduction in the threshold water content from 3.6 to 0.8 g/g. In silico analysis shows that both dehydrin genes are expressed in drought-stressed plants as well (Fig. 7). This observation concurs with those of Black et al. (1999) who showed that the induction of dehydrins in maturating wheat (Triticum aestivum) embryos is not regulated by the same factors that induce DT. The presence of dehydrins in recalcitrant seeds of temperate climate (Kermode, 1997), the absence of correlation between their amounts and seed longevity (Wechsberg et al., 1994), together with the observation that dehydrins protect enzyme activities only at water potentials above –3 MPa (Reyes et al., 2005) all point to a protective function at high hydration levels. Thus, dehydrins might protect at intermediate hydration levels (>0.8 g/g), whereas the DT-linked LEA proteins might play a role below the hydration level corresponding to the threshold water content of 2.8 mm long radicles (i.e. 0.3 g/g). In this respect, transcript levels of homologs of MtEm6 were correlated with seed longevity of Brassica napus (Soeda et al., 2005) and the wheat homolog of MtEm6 was found to protect citrate synthase from aggregation due to desiccation upon multiple freeze-drying cycles, supporting the hypothesis that Em6 can protect macromolecules in the dry state (Goyal et al., 2005). Conversely, overexpression of the wheat Em in yeast cells (Swire-Clark and Marcotte, 1999) and also of the Arabidopsis homolog of MtPM25 in germinating seeds (Borrell et al., 2002) led to improved growth under high NaCl, KCl, LiCl, and sorbitol conditions. Recently, a similar observation was made for E. coli overexpressing the PM2 (Liu and Zheng, 2005). The cellular water potential resulting from incubation in these osmotic solutions (osmotic potential ranging from –2 to –6 MPa and equivalent to 96%–98% RH) is much higher than those experienced by the dry seeds (in this study: 42% RH equivalent to –180 MPa; Walters et al., 2002). These results argue for a protective role of Em6, PM25, and MP2 during hyperosmotic conditions rather than at low water contents. In the light of these observations, one could envisage that the DT-linked Em6, MP2, MtPM25, and PM18 could exert more than one function upon water loss as hypothesized for nonreducing sugars, which act as osmolytes during hyperosmotic stress and stabilizers of macromolecules in the dry state (Prestrelski et al., 1993; Allison et al., 1999; Hoekstra et al., 2001). For example, at high moisture contents, DT-linked LEA proteins may act as compatible solutes that preferentially exclude chaotropic agents (such as salts) from the surface of macromolecules as suggested by the beneficial effects described above (Swire-Clark and Marcotte, 1999; Borrell et al., 2002; Liu and Zheng, 2005; Reyes et al., 2005). Likewise, when the hydration shell is removed (i.e. water content less than 0.3 g/g), they might exert their protective effects in the dry state, as was found for wheat Em, by replacing water molecules by hydrogen bonding and/or forming a glass which stabilizes the system in the dried state (Hoekstra et al., 2001; Wolkers et al., 2001; Buitink and Leprince, 2004).

Group 5 LEA proteins, to which MtPM25 belongs, have been reported to be a peculiar group, with low hydrophilicity and absence of heat stability (Cuming, 1999; Ramanjulu and Bartels, 2002). Indeed, MtPM25 was the least hydrophilic from the six LEA proteins identified in this study. However, MtPM25 was HS and the contention that group 5 proteins are not HS is questionable. To determine the structure of a member of group 5 LEA protein, FTIR analysis on MtPM25 was carried out and compared to that on Em6. The data on the secondary structure of MtEm6 and MtPM25 in solution complement those obtained for members of groups 3 and 4 using FTIR spectroscopy as well as for dehydrins and a member of the D95 family with other spectroscopy techniques. In the hydrated state, LEA proteins exhibit a wide degree of disorder, ranging from unordered (group 1; Soulages et al., 2002; group 3 LEA proteins from Typha latifolia pollen [Wolkers et al., 2001]; nematodes [Goyal et al., 2003]; and dehydrins [Soulages et al., 2003]), to 60% to 70% unordered (GmPM16, a soybean group 4 LEA protein [Shih et al., 2004] and Em proteins [Table IV; McCubbin et al., 1985]), and finally down to 50% unordered (MtPM25 [Table IV] and an Arabidopsis Lea14, a member of the D 95 family [PF03168; Singh et al., 2005]). Likewise, the nature and contents of ordered structures in solution varies greatly. For example, -sheet amounts range from 10% for MtEm6 to 40% for the Arabidopsis Lea14. Considering that the ectopic expression of members of groups 1 to 5 always improves the tolerance against salt or water stress in bacteria, yeast, and plants, albeit to various degrees, it is tempting to speculate that the unordered domains of LEA proteins are responsible for alleviating the osmotic stress endured by the tissues. It is noteworthy that the presence of sheets appears to be a common feature of LEA proteins and apparently does not affect the heat solubility of the protein.

Considering that MtEm6 and MtPM25 are hypothesized to play a role in the dry state, their conformation was also determined after drying. The removal of the water induced a transition from a fairly disordered conformation to the formation of a considerable amount of ordered structures (Table IV). Our study suggests that this behavior is yet another feature that now appears to be common to all LEA proteins. Indeed, originally observed for the groups 3 and 4 LEA proteins mentioned above, this study shows that it is also the case for members of groups 1 and 5. Both proteins show an increase in their -helical and -sheet contents (Table IV). According to Wolkers et al. (1998), the -sheet formation results from the replacement of hydrogen bonding of water by intermolecular hydrogen bonds between peptide backbones. When induced by drying, sheets were fully reversible and could be interconverted by rehydration, in agreement with previous observations on the group 3 LEA protein from pollen (Wolkers et al., 2001) and the group 4 GmPM16 of soybean (Shih et al., 2004). The PELE program did predict the overall contributions of secondary structures of MtEm6 and MtPM25 in the hydrated state, but did not in the dry state. Also for the group 4 GmPM16, the structure predictions cannot be fulfilled in the dried state (Shih et al., 2004). This contrasts with the findings of Goyal et al. (2003) on a group 3 LEA from nematodes. Thus, caution must be taken in extrapolating computer predictions to the dried state to understand the structure-function relationship of LEA proteins at low water contents.

So at which hydration level do these proteins gain structure? Slow drying over saturated salt solutions indicated that the proteins had to be dried below an equilibrium RH of 85% to observe a change in the secondary structure (-sheet formation). The corresponding hydration level is around 0.2 to 0.3 g/g, or to the onset of the removal of the hydration shell (Hoekstra et al., 2001; Walters et al., 2002). This is a significant finding considering the hypothesis that DT-linked LEA proteins could play different protective functions depending on the hydration level. Indeed, the gain of structure at low water contents occurs at a hydration level that is below that experienced during hyperosmotic stress, but before the protein is immobilized in the cytoplasmic glassy state. Furthermore, it was observed that the rate of water loss influences the conformation adopted by MtEm6 and MtPM25 in the dry state, which is in agreement with the data of Wolkers et al. (2001) on the group 3 LEA protein from pollen. The presence of solutes such as Suc also influences the change in conformation during drying (Wolkers et al., 1998, 2001). We do not know whether this is also the case for MtPM25 and MtEm6. However, in light of these observations, it is important to test whether and how the nature and rate of drying influences the protective activity of LEA proteins in vitro.

Both group 2 (DHN3 and BudCar5) and group 3 LEA proteins (MP2, PM18, and SBP65) were detected as several isoforms. This observation extends previous experimental evidence showing that dehydrins are submitted to posttranslational modifications such as phosphorylation during seed development and germination (Campalans et al., 2000; Riera et al., 2004). Phosphorylation of dehydrins has a functional importance during water and cold stress (Riera et al., 2004; Alsheikh et al., 2005). Both PM18 and DHN3 have distinct isoforms that are related to DT, whereas others are clearly not (Figs. 5 and 6). This raises the question as to whether the regulation of posttranslational modifications might be also important for DT.

SBP65 exhibits a peculiar posttranscriptional modification. This DT-linked group 3 LEA protein is known to be biotinylated in seeds of a wide range of species (Dehaye et al., 1997; Capron et al., 2000). This protein has been shown to accumulate during the later stages of seed development and to be degraded during germination (Dehaye et al., 1997; Capron et al., 2000). It has been suggested that SBP65 constitutes a storage form of biotin that can be released during germination and postgerminative growth (Dehaye et al., 1997). Whether its role in DT is simply to store biotin or whether it has an additional, structural function in the protection against drying remains to be ascertained.

Using a computational analysis, it has been argued that LEA proteins are members of a larger family of stress proteins called hydrophilins that could be used as predictors of the responsiveness to osmotic adaptation in prokaryotes and eukaryotes (Garay-Arroyo et al., 2000). Hydrophilins are defined as proteins having a Gly content >6% and a hydrophylicity index >1. The analysis of the HS proteome of M. truncatula radicles does not support this analysis on seeds. On the one hand, MtPM25 and SPB65, two polypeptides linked to DT do not satisfy the hydrophilicity and Gly content criterium, respectively. On the other hand, we identified several hydrophilic proteins, such as several fragments of legumin and vicilin that matched those criteria and yet, they were not induced by the osmotic treatment. Therefore, in seeds the term hydrophilins cannot adequately describe the LEA protein family or water stress proteins.

Altogether, the data reported in this study suggest that the LEA proteins expressed in seeds can be divided in two groups, those that are induced only in tissues that are desiccation tolerant, and those that are also induced in osmotically shocked radicles that remain desiccation sensitive but do increase their tolerance to drying. The first group contains LEA proteins that seem to be seed specific, based on electronic northerns, whereas the second group is represented by proteins that are also expressed in vegetative tissues. Possibly, the proteins that are linked to DT might protect both at high hydration levels and at very low water contents (<0.3 g/g). Further research should be focused on elucidating whether these proteins play a role at water contents where the proteins have gained further order in their secondary structures and whether their functions are regulated by posttranslational modifications.

	MATERIALS AND METHODS

TOP ABSTRACT RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

 

Plant Material and Treatments

Seeds of Medicago truncatula Gaertn. (cv Paraggio; Seedco Australia) were allowed to imbibe on filter paper in distilled water at 20°C in the dark for up to 3 d. For the proteomic analysis, desiccation-tolerant stages were established as described in Buitink et al. (2003). They correspond to 16 h-imbibed seeds prior to radicle emergence and seeds exhibiting a protruded radicle leng