The Signature of Seeds in Resurrection Plants: a Molecular and Physiological Comparison of Desiccation Tolerance in Seeds and Vegetative Tissues1

INTEGR.COMP.BIOL., 45:771±787 (2005)

The Signature of Seeds in Resurrection Plants: A Molecular and Physiological Comparison of Desiccation Tolerance in Seeds and Vegetative Tissues1

NICOLA ILLING,KATHERINE J. DENBY,HELEN COLLETT,ARTHUR SHEN, AND JILL M. FARRANT2 Department of Molecular and Cell Biology, University of Cape Town, Private Bag, Rondebosch, 7701, South Africa.

SYNOPSIS. Desiccation-tolerance in vegetative tissues of angiosperms has a polyphyletic origin and could be due to 1) appropriation of the seed-speci®c program of gene expression that protects orthodox seeds against desiccation, and/or 2) a sustainable version of the abiotic stress response. We tested these hypotheses by comparing molecular and physiological data from the development of orthodox seeds, the response of desiccation-sensitive plants to abiotic stress, and the response of desiccation-tolerant plants to extreme water loss. Analysis of publicly-available gene expression data of 35 LEA proteins and 68 anti-oxidant enzymes in the desiccation-sensitive Arabidopsis thaliana identi®ed 13 LEAs and 4 anti-oxidants exclusively expressed in seeds. Two (a LEA6 and 1-cys-peroxiredoxin) are not expressed in vegetative tissues in A. thaliana, but have orthologues that are speci®cally activated in desiccating leaves of Xerophyta humilis. A comparison of antioxidant enzyme activity in two desiccation-sensitive species of Eragrostis with the desiccation-tolerant E. nindensis showed equivalent responses upon initial dehydration, but activity was retained at low water content in E. nindensis only. We propose that these antioxidants are housekeeping enzymes and that they are protected from damage in the desiccation-tolerant species. Sucrose is considered an important protectant against desiccation in orthodox seeds, and we show that sucrose accumulates in drying leaves of E. nindensis, but not in the desiccation-sensitive Eragrostis species. The activation of ``seed-speci®c'' desiccation protection mechanisms (sucrose accumulation and expression of LEA6 and 1-cys-peroxiredoxin genes) in the vegetative tissues of desiccation-tolerant plants points towards acquisition of desiccation tolerance from seeds.

INTRODUCTION former, it is based on induction of several relatively Desiccation tolerance (DT) is the ability of an or- complex protection mechanisms during drying, with ganism to survive the loss of most (Ͼ95%) of its cel- minimal reliance on repair of desiccation-induced lular water. It is considered to be a complex trait that damage during rehydration (Gaff, 1989; Vertucci and is present in reproductive structures (pollen and seeds) Farrant, 1995; Oliver, 1996; Oliver et al., 2000). In the of vascular plants. Desiccation tolerance in vegetative lower order plants, DT is constitutive, less reliant on tissues is also relatively common in less complex complex cellular protection and more on repair during plants such as bryophytes (Proctor, 1990) and lichens rehydration (Bewley and Oliver, 1992; Oliver and (Kappen and Valadares, 1999) but rare in pterido- Bewley, 1997; Oliver et al., 2000). Thus, although DT phytes and angiosperms (Gaff, 1977, 1987; Alpert and in seeds appears to include some mechanisms present Oliver, 2002; Porembski and Barthlott, 2000) and ab- in DT vegetative tissues of less-complex plants (re- sent in gymnosperms (Gaff, 1980). Oliver et al. (2000) viewed in Oliver, 1996; Dickie and Pritchard, 2002), have speculated on the evolution of DT in vegetative additional, more sophisticated mechanisms of protec- tissue and argue that DT is the ancestral state for early tion must have evolved. We examine the possibility land plants (e.g., mosses), which was lost early in the that DT in the vegetative tissues of angiosperms is a evolution of tracheophytes. The subsequent successful subsequent adaptation of these seed developmentally- radiation of vascular plants on land was probably a regulated mechanisms. consequence of the evolution of DT in seeds, in par- Whereas DT is developmentally regulated in seeds allel to the evolution of structural and morphological with the putative mechanisms of tolerance being ac- modi®cations in vegetative tissue which allowed great- cumulated only at precise times after fertilisation, er control of water status. Oliver et al. (2000) speculate drought is stochastic and DT vegetative tissues must that the emergence of DT in seeds was a modi®cation respond to environmental signals to activate protective of vegetative DT in early ancestors. They suggest fur- mechanisms for the whole plant. Conceivably, envi- thermore that vegetative DT in angiosperms subse- ronmental signals for extreme water loss activate an quently re-evolved independently at least eight times existing repertoire of protective seed-speci®c genes in as an adaptation of seed DT. vegetative tissue. DT in seeds and vegetative tissues of angiosperms An alternative hypothesis is that DT evolved from is different from that in extant lower orders. In the the response of desiccation-sensitive plants to abiotic stresses such as cold, salt and drought. Desiccation- 1 From the Symposium Drying Without Dying: The Comparative sensitive plants use an interconnected signaling net- Mechanisms and Evolution of Desiccation Tolerance in Animals, work to activate a common repertoire of responses to Microbes, and Plants presented at the Annual Meeting of the Society abiotic stress (Knight and Knight, 2001). These re- for Integrative and Comparative Biology, 4±8 January 2005, at San Diego, California. sponses appear to overlap with those described for DT 2 E-mail: [email protected] plants during extreme water loss as they include the

771 772 N. ILLING ET AL.

accumulation of compatible osmolytes, and the up-reg- monly acknowledged to play an important and varied ulation of anti-oxidants and anti-oxidant enzymes role in DT, although the exact nature of this in vivo (Knight and Knight, 2001). Under this scenario, we has yet to be demonstrated. The general protective would predict that there should be signi®cant overlap roles ascribed to sucrose are as water replacement mol- between genes that are induced in response to abiotic ecules and a facilitator, together with proteins, of glass stresses and DT. formation (Leopold et al., 1994; Vertucci and Farrant, The stresses associated with extreme water loss in- 1995; Crowe et al., 1998). It has also been suggested clude the mechanical stress associated with turgor loss, that the formation of sucrose reduces the monosaccha- oxidative stress from free radical-mediated processes ride pool, which in turn reduces the chances of dam- and the destabilization or loss of macromolecular in- aging Maillard-type reactions occurring and puts a sta- tegrity (reviewed by Vertucci and Farrant, 1995; Oli- sis on respiratory metabolism, both of which reduce ver and Bewley, 1997; Walters et al., 2002). Protection free radical formation (Vertucci and Farrant, 1995). mechanisms associated with prevention of damage Free radical scavenging systems are ubiquitous in from these stresses have been extensively reviewed plants and include well known antioxidants such as and the maintenance of subcellular integrity has been ascorbate, glutathione and tocopherol, and enzymes widely attributed to the accumulation of stress-asso- such as the peroxidases (ascorbate peroxidase [AP], ciated proteins, non-reducing sugars and free radical glutathione peroxidase, thioredoxin peroxidase, cata- scavenging systems (for example Farrant, 2000; Scott, lase), glutathione reductase (GR) and superoxide dis- 2000; Oliver et al., 2000; Vicre et al., 2004). mutase (SOD) inter alia. Other anti-oxidant enzymes Late Embryo Abundant (LEA) proteins accumulate such as 1-cys peroxiredoxin have been identi®ed in the during the onset of DT in orthodox seeds and have seeds of desiccation-sensitive angiosperms (Aalen, been shown to occur in response to the drying of DT 1999) and recently in the vegetative tissues of the res- vegetative tissues (Bray, 1997; Ingram and Bartels, urrection plant X. viscosa (Mowla et al., 2002). 1996). Expression of LEAs has also been reported to We have focused on the above-mentioned mecha- be up-regulated during abiotic stress such as cold, nisms of DT to test the hypothesis that DT in vege- drought and osmotic stress (Wise and Tunnacliffe, tative tissue in resurrection plants is an adaptation of 2004). Several LEA mRNA transcripts and proteins DT in seeds. We have compared some of the putative have been identi®ed as being up-regulated in leaves of mechanisms of tolerance in the seeds of desiccation- the resurrection plants Craterostigma plantagineum sensitive angiosperms with those found to occur dur- (4) and Xerophyta viscosa (2) during a cycle of des- ing desiccation of resurrection plants (viz. the accu- iccation (Piatkowski et al., 1990; Mundree and Far- mulation of LEA proteins, anti-oxidants and sucrose) rant, 2000; Ndima et al., 2001). The largest set of using a combination of molecular and physiological LEAs (16) from a single resurrection plant was re- studies. In this study we aim to distinguish between cently identi®ed in a mini-microarray screen of 400 seed-speci®c and abiotic stress-speci®c LEA and anti- cDNAs from Xerophyta humilis (Collett et al., 2004). oxidant enzyme gene expression in Arabidopsis with LEAs are low complexity proteins which have been a view to showing that some of the ``seed-speci®c'' classi®ed into several unrelated groups on the basis of genes are expressed in the vegetative tissue of resur- conservation of peptide motifs (or Pfam domains) rection plants such as X. humilis during desiccation. (Close, 1997; Wise, 2003). It has not been possible to We also compare antioxidant activity and sucrose ac- assign structures to LEAs, or experimentally determine cumulation in the seeds and vegetative tissue of mis- their exact cellular role because they are unfolded in cellaneous desiccation-sensitive and desiccation-toler- the hydrated state. Based on evidence of their abun- ant plants to test our hypotheses. dant expression, and rich hydrophilic amino acid content, it has been proposed that LEAs maintain subcel- MATERIALS AND METHODS lular integrity by protecting cellular structures from the Plant material effects of water loss by either acting as a hydration buffer, by sequestering ions, by direct protection of Resurrection plants were collected and maintained other proteins or membranes, or by renaturing unfold- in a glasshouse at the University of Cape Town as ed proteins (Bray, 1991). Recently, it was shown that previously reported (Sherwin and Farrant, 1996; Dace two LEAs can prevent protein aggregation during wa- et al., 1998; Vander Willigen et al., 2001). Plants of ter stress (Goyal et al., 2005) and the ability of plant Eragrostis curvula and E. teff were grown from seeds LEAs to confer increased tolerance to water de®cit sown in seedling ¯ats (1:1 potting soil:river sand mix) stress on yeast and other plants (for example, Swire- and maintained in the glasshouse for three months. Pri- Clark and Marcotte [1999], Xu et al. [1996] and Siv- or to experimentation, fully hydrated plants were trans- amani et al. [2000]) also suggests LEAs play an im- ferred to a controlled environment chamber where they portant role in protecting tissues from the effects of were acclimated to conditions of 50±65% relative hu- water loss. midity, 14 hr light (500 ␮mol photons´mϪ2´sϪ1;25ЊC): Sucrose is the only sugar commonly accumulated in 10 hr dark (18ЊC). After two weeks plants were used DT tissues of seeds and resurrection plants (Scott, for dehydration/rehydration experiments. Plants were 2000; Ghasempour et al., 1998). Like LEAs, it is com- naturally dried by withholding water until the plants SIGNATURE OF SEEDS IN RESURRECTION PLANTS 773 had reached an air-dry state (Ͻ5% relative water con- TABLE 1. Assignment of A. thaliana locus IDs for LEAs, LEA-like, tent [RWC]). Leaves were sampled at regular intervals putative LEAs and dehydrins to Interpro Superfamilies (IPR) according to conserved Pfam domains. during dehydration for procedures described below. They were held in the dry state for a minimum of two LEA Interpro PFAM Expression weeks prior to rehydration, which was by soil water- group Locus ID superfamily domain cluster ing. LEA-1 At2g40170 IPR000389 PF00477 A Seeds of E. nindensis were obtained from the Ag- At3g51810 A ricultural Research Council (Pretoria, South Africa) LEA-2 At2g21490 IPR000167 PF00257 A and E. curvula seeds were purchased from Silverhill At3g50980 A At4g39130 A Seeds (Cape Town, South Africa). E. teff seeds were At5g66400 B a kind donation from Mr Z. Ginbot. At1g76180 C At3g50970 C Molecular studies and bio-informatics At4g38410 C At1g20440 D Sequencing and classi®cation of X. humilis LEAs At1g20450 D At1g54410 D The 5Ј ends of 16 cDNAs described as LEAs, de- LEA-3 At2g36640 IPR004238 PF02987 A hydrins, putative LEAs or LEA-like proteins on the At3g15670 A basis of blastx analysis (Collett et al., 2004) were se- At4g36600 A quenced on a MegaBACE 500 sequencer (Amersham At5g38760 A*s At1g52690 B Biosciences, Little Chalfont, UK) using the DYEnamic At3g02480 B ET Dye terminator Cycle sequencing Kit for Mega- At3g17520 B BACE (Applied Biosystems). Sequences were ana- At4g13230 B lysed by InterProScan (EBI) to identity conserved At2g03850 C At2g42530 D Pfam domains and to assign the LEAS, putative LEAs, At5g53820 E LEA-like proteins, and dehydrins to InterPro (IPR) Su- At1g72100 not rep perfamilies (Zdobnov and Apweiler, 2001). At2g42530 not rep At3g53040 not rep Classi®cation of LEAs from other resurrection plants At5g44310 not rep LEA-4 At2g35300 IPR005513 PF03760 A Sequences of LEAs that have been reported from At5g06760 B other resurrection plants were obtained from GenBank, At1g32560 not rep analysed by Interproscan, and assigned to IPR Super- LEA-6 At3g22490 IPR007011 PF04927 A families. At3g22500 A At5g27980 A*s At1g03120 not rep Northern blot analysis At5g53260 not rep The relative abundance of LEA mRNA transcripts At5g53270 not rep during a cycle of desiccation and rehydration was ex- LEA-7 At3g53770 IPR004926 PF03242 C At1g02820 D amined by northern blot analysis of total RNA samples At4g02380 E extracted from X. humilis leaves at early and late stag- At4g15910 not rep es of desiccation and rehydration, as described by Col- LEA-8 At1g01470 IPR004864 PF03168 C lett et al. (2004). At2g44060 not rep At2g46140 not rep Microarray analysis LEA-9 At2g41280 PD68804 A At2g41260 not rep The AGI gene locus number for all 49 Arabidopsis LEA-10 At1g04560 IPR008390 PF05512 A LEA and LEA-like proteins grouped under the differ- At5g18970 E ent LEA superfamilies (IPRs) were downloaded from At1g29520 not rep At5g46530 not rep the Interproscan website (Table 1). These superfamilies included the LEA groups de®ned by Wise (2003) A ϭ seed-speci®c genes i.e., genes that are expressed at maximum value during seed development, with less than 30% expression in and IPR00839, since several of the genes in this Su- leaves or roots in response to abiotic stress treatment; B ϭ genes perfamily are listed as encoding LEA-like proteins expressed at maximum value during seed development, and ex- (e.g., AAD10377 hydrophobic LEA-like protein, Ory- pressed at greater than 30% of maximum value in leaves or roots in za sativa). response to abiotic stress treatments; C ϭ genes expressed at max- Scaled microarray expression data during seed de- imum value in leaves or roots in response to abiotic stress treatments and at greater than 30% of this value during seed development; D velopment and abiotic stress treatments corresponding ϭ stress-speci®c genes: genes expressed at maximum value in leaves to the Arabidopsis locus IDs listed in Table 1, and for or roots in response to abiotic stress treatments and at less than 30% 71 anti-oxidants listed by Mittler et al. (2004) was of this value during seed development; E ϭ housekeeping genes: downloaded from the Genevestigator database genes which are expressed at greater than 40% of maximum value in control leaves or roots for stress array experiments; not rep ϭ (www.genevestigator.ethz.ch) (Zimmerman et al., genes not represented on the Affymetrix 25K genechip. *s expres- 2004) (Experiments 90, 120, 121, 122, 124 and 124). sion restricted to silique development and not mature seed. These experiments are part of AtGenExpress, a mul- tinational effort to pro®le the transcriptome of Arabi- 774 N. ILLING ET AL.

dopsis (http://web.uni-frankfurt.de/fb15/botanik/mcb/ with 0.1 M phosphate buffer (pH 7.8). Samples were AFGN/atgenex.htm). 35 out of the 49 Arabidopsis eluted with phosphate buffer (0.1 M, pH 7.8) and so- LEAs were present on the Affymetrix 25K chip used lutions collected for analysis of enzyme activities. AP in these experiments. We checked that the distribution was measured as described in Wang, et al. (1991), GR of signal intensities for each of the full datasets was by the method of Esterbauer and Grill (1978) and SOD similar to validate our cross-comparison between the was measured using the method of Giannopolitis and seed development and abiotic stress series. Ries (1971) and modi®ed by Bailly et al. (1996). Average signal values were calculated as the mean from the repeats for each data point in the silique/seed Sucrose determination and the abiotic stress series. A general signal value for Three 100 mg replicates of leaf tissue and seeds each gene in control plants was determined by taking were ground in liquid nitrogen and samples extracted the mean of the control data points for green shoot and in cold 100 mM NaOH (50% v/v ethanol/water). Chlo- root respectively from the abiotic stress series. For roform (15% v/v) was added and the samples incu- each gene, each average signal intensity in the silique bated on ice for 10 min. Thereafter the pH was ad- and seed development series and in the abiotic stress justed to 7.5 with 100 mM HEPES in 100 mM glacial experiments, as well as the general control values were acetic acid. After centrifugation for 20 min at 4ЊCat expressed as a percentage of the maximum value 28,000 g, the supernatant was removed. A repeat ex- across all the experiments. Genes were manually traction was performed on the pellets. The superna- grouped into the following clusters based on their pro- tants were pooled and then centrifuged. Sucrose in the ®les of expression; A) Seed speci®c genes: maximum supernatant was calculated from the spectrophotomet- expression in seed, less than 30% of the maximum ric measurement of NADPH production using a D- expression in any of the abiotic stress treatments, B) glucose/D-fructose sugar assay kit (Boehringer Mann- maximum expression in seed and more than 30% of heim, Germany). the maximum expression in any of the abiotic stress treatments, C) maximum expression in one of the abi- Measurement of photosynthesis otic stress treatments as well as more than 30% of the Light-saturated net photosynthesis (A) of leaves of maximum expression during seed development, D) E. curvula and E. nindensis during desiccation was Stress speci®c genes: maximum expression in one of measured using a Ciras-1 infrared gas analyzer with a the abiotic stress treatments and less than 30% of the Parkinson's Leaf Cuvette and in built illumination unit maximum expression during seed development and E) (PP Systems, Hertfordshire, UK) operated in differ- Housekeeping genes: expression in the control plants ential mode at an ambient CO2 concentration of 350 is more than 40% of the maximum expression value ppm and 22ЊC (50% RH). The parameter A was cal- of either the seed development or abiotic stress treat- culated according to the equations of von Caemmerer ment series. The average expression in controls was and Farquhar (1981) and the data were expressed as a less than 30% of the maximum expression value in percentage of A measured in control hydrated tissues. groups A±D. Measurements were taken on 5 individual plants and repeated during at least two cycles of drying. Physiological studies RESULTS Relative water content (RWC) determination Clustering expression pro®les of Arabidopsis LEA The RWC was calculated as the water content di- mRNAs during seed development and in response to vided by the water content estimated at full turgor. The abiotic stress means of the water content of leaves at full turgor were LEAs were identi®ed as being up-regulated in recorded for each species using more than 20 repre- leaves in the resurrection plant X. humilis following sentative leaf samples from plants that had been fully desiccation (Collett et al., 2004). We posed the ques- hydrated overnight in plastic bags. Water contents tion: Are any of the LEA groups restricted to a pro®le were gravimetrically determined by oven drying at of seed-speci®c or abiotic-stress speci®c gene expres- 70ЊC for 48 hr. sion in a desiccation-sensitive plant? To this end, we ®rst analysed the expression patterns of LEAs from a Antioxidant enzyme activity desiccation-sensitive plant during seed development Leaf tissues and seeds were extracted and analysed and abiotic stress by mining publicly-available Ara- for AP, GR and SOD analysis as described previously bidopsis genome and microarray data. (Farrant et al., 2003, 2004). Three 250 mg replicates All the AGI gene locus IDs that fall into 10 different of leaf tissue were ground in liquid nitrogen and ex- LEA superfamilies, de®ned on InterProScan, were tracted in 3 ml of cold buffer (0.1 M phosphate buffer identi®ed (Table 1). Out of 49 Arabidopsis LEAs, 35 pH 7.8; 2 mM dithiothreitol [DTT]; 0.1 mM EDTA; had speci®c probes on the 25K Affymetrix gene chip 1.25 mM PEG 4000; 0.1 g insoluble polyvinylpyrrol- and the expression of these 35 was compared between idone [PVP]). The extract was centrifuged at 16,000 g silique and seed development and during abiotic stress for 15 min and the supernatant was desalted on a Se- treatments (cold, osmotic, salt and drought stress). phadex G-25 PD10 column (Pharmacia) equilibrated We aimed to distinguish between seed-speci®c and SIGNATURE OF SEEDS IN RESURRECTION PLANTS 775 stress-speci®c LEAs on the basis of expression pro®les When InterProScan failed to identify PFAM domains in the selected AtGenExpress dataset. LEA expression in the X. humilis sequence information, LEAs were pro®les grouped into different clusters illustrated in grouped on the basis of the InterProScan of their clos- Figure 1 and summarized in Table 1. According to this est homologues identi®ed by BlastX analysis (Table analysis, we were able to identify LEA groups that fell 2). only into Cluster A (i.e., seed speci®c), but no LEA With the exception of the Group 1 LEAs groups that were restricted to Cluster D (i.e., abiotic (IPR000389) and Group 9 LEAs (PD688044), genes stress speci®c). The Group 1 LEAs (IPR000389) and belonging to all the other LEA groups were identi®ed a Group 9 LEA (PD68804) were identi®ed as being in the 16 LEAs isolated from 400 sequenced X. humilis seed-speci®c (i.e., cluster A) as there was very low cDNAs (Table 2). In Arabidopsis, 10 genes are clas- expression (Ͻ10% of seed expression) in roots or si®ed in the LEA-2 (IPR000167) and 14 genes are leaves during abiotic stress treatments. Although only classi®ed in the LEA-3 (IPR004238) superfamilies 3 out of the 6 Group 6 LEAs (Table 1) were repre- (Table 1). These two superfamilies were also well-rep- sented on the Affymetrix 25K genechip, these were resented in the X. humilis cDNAs (Table 2). We also also only expressed during silique development and analysed LEAs reported as up-regulated during des- for the latter stages of seed maturation. The maximum iccation in other resurrection plants, namely Crater- absolute level of expression of the silique-speci®c ostigma plantigineum, X. viscosa and the moss, Tor- At5g27980 was very low in comparison to the other tula ruralis These LEAs can be classi®ed into groups expression pro®les reported in this study (Ͻ0.5% of 2, 3, 8 and 10 (Table 3). Again, Groups 2 and 3 are most abundant LEA, i.e., At5g664400/LEA2, cluster predominant. Only a small number of Arabidopsis B). Group 7 LEAs (IPR004926), were expressed under genes are classi®ed in the Group 1 and Group 9 su- stress conditions (cluster C and D) or as housekeeping perfamilies, and the lack of representation of genes in genes in roots (cluster E). The one LEA from Group these groups from resurrection plants could be due to 8 (IPR004864) that was represented in these experi- the fact that the EST datasets analysed to date do not ments was expressed at high levels in both maturing constitute the full LEA complement. seed, and in response to abiotic stress (Table 1, Fig. The LEA6 cDNA (HC248) was the only LEA in the 1). Some members within LEA groups 2, 3, 4, 10 were X. humilis geneset that clearly corresponded to a LEA seed-speci®c, whilst others were activated in response that was expressed only in maturing seed in Arabi- to abiotic stress treatment and seed maturation, or were dopsis. No conclusions could be drawn on the roles housekeeping genes. that the other X. humilis LEAs play in the evolution What is remarkable is the diversity of LEAs ex- of desiccation tolerance in vegetative tissues, since pressed during seed maturation; 74% of the LEAs are they belong to superfamily groups which include both expressed during seed development with 19/35 LEAs seed-speci®c members as well as members which are being expressed at their maximum in seed develop- expressed in both seed and abiotic stress in Arabidop- ment and 7/35 LEAs being expressed at Ͼ30% max- sis (Fig. 1). imum levels in seeds. In contrast, 46% of the LEAs were expressed in response to abiotic stress, with 10/ Northern blot analysis of X. humilis LEAs 35 being expressed at maximum level in either leaves Many of the Group 2 and Group 3 LEAs are up- or roots and 6/35 LEAs being expressed at Ͼ30% lev- regulated in response to abiotic stress treatments in el. Expression of homologues of Arabidopsis LEAs Arabidopsis (Fig. 1). We tested the hypothesis that X. from Groups 1, 6 and 9 in desiccating vegetative tissue humilis Group 2 and Group 3 LEAs corresponding to of resurrection plants, could be considered an indicator the stress LEAs from desiccation-sensitive plants, that vegetative DT is due to appropriation of seed DT. might be up-regulated during the early stages of de- This could arise from a change in regulation of genes hydration (i.e., Ͼ65% RWC), and that a second class that in desiccation-sensitive plants, such as Arabidop- of LEAs, which correspond to those expressed during sis, are only expressed during the desiccation phase of the desiccation phase of seed development, are ex- seed maturation. pressed during the later stages of desiccation (Ͻ65% RWC). Northern blot analysis of 13 of the 16 LEAs Sequence analysis of sixteen X. humilis LEAs (Fig. 2) clearly showed that the X. humilis LEAs in- To see whether any of the sixteen desiccation up- vestigated all have similar pro®les of expression. They regulated LEA ESTs identi®ed in the leaves of X. hu- are all signi®cantly up-regulated only in the later stag- milis (Collett et al., 2004) belonged to the unique seed- es of desiccation (i.e., Ͻ65% RWC) in leaves, and speci®c LEAs we identi®ed in Arabidopsis (i.e., most LEA mRNA transcripts are stably stored in dry groups 1, 6 or 9), we sequenced the 5Ј end of each X. leaves (Ͻ6% RWC). humilis cDNA. The closest homologue for each X. humilis putative LEA was identi®ed by a BlastX search Clustering Arabidopsis anti-oxidant mRNA and the nucleotide sequences for each LEA were an- expression pro®les during seed development and in alysed by InterProScan, and grouped into LEA super- response to abiotic stress families on the basis of their IPR Superfamily classi- We extended our analysis of the selected Arabidop- ®cation (Table 2) (Zdobnov and Apweiler, 2001). sis AtGenExpress microarray dataset, posing the ques- 776 N. ILLING ET AL.

FIG. 1. Manual cluster analysis of expression pro®les of 35 Arabidopsis LEA genes during silique and seed development and in response to abiotic stress treatments (cold, osmotic, salt and drought) in green shoots (A) and roots (B). 100% represents the highest signal intensity for each particular gene across these experiments. *1: The maximum intensity signal for these genes was less than 1% of the maximum signal in the LEA subset of genes. *2: The maximum intensity signal for these genes was less than 2.5% of the maximum signal in the LEA subset of genes. SIGNATURE OF SEEDS IN RESURRECTION PLANTS 777

FIG. 1. Continued. tion: can we identify any anti-oxidant genes that are tissue of desiccation tolerant plants? Microarray data expressed exclusively during the desiccation phase of from the seed development and abiotic stress series seed maturation in this desiccation-sensitive plant, and corresponding to AGI gene locus IDs for 68/71 dif- are up-regulated during desiccation in the vegetative ferent anti-oxidants listed by Mittler et al. (2004) 778 N. ILLING ET AL.

TABLE 2. Classi®cation via InterPro Scan of sixteen X. humilis LEAs that are up-regulated during desiccation in leaves.

Log2 Genbank Interpro Interpro FH/D accession superfamily of Pfam superfamily of LEA Lab ref ratios² SD² no.* X. humilis protein domain Best BlastX hit Description E-value best BlastX hit Pfam group** HC189 3.17 Ϯ0.35 CK906385 IPR000167 PF00257 None *** LEA-2 HC265 4.55 Ϯ1.78 CK988413 **** AAL83428 44kDa dehy- 1E-05 IPR000167 PF00257 LEA-2 drin-like protein [C. sericea] HC330 2.49 Ϯ0.7 CK906432 IPR000167 PF00257 S12095 embryonic 1E-18 IPR000167 PF00257 LEA-2 abundant protein-rad- ish HC421 1.72 Ϯ0.31 CK906386 **** AAL15651 dehydrin-like 2Eϩ00 IPR000167 PF00257 LEA-2 protein [M. sativa] D5 4.12 Ϯ0.29 CK906406 IPR004238 PF02987 BAB97392 LEA-like pro- 6E-10 IPR004238 PF02987 LEA-3 tein [L. lon- gi¯orum] D10 1.96 Ϯ1.36 CK906427 IPR004238 PF02987 BAD44201 late embryo- 1E-22 IPR004238 PF02987 LEA-3 genesis abundant protein-like [A. thaliana] At5g44310 HC85 2.18 Ϯ0.33 CK906404 IPR004238 PF02987 P13934 LEA protein 5E-18 IPR004238 PF02987 LEA-3 76-rape HC339 4.51 Ϯ0.64 CK906402 **** AAN74637 LEA1 protein 2E-05 IPR004238 PF02987 LEA-3 [T. aestivum] HC429 2.48 Ϯ0.29 CK906398 **** BAD22766 LEA protein 3E-17 IPR004238 PF02987 LEA-3 [B. inermis] HC338 1.58 Ϯ0.52 CK906399 **** XP࿞481643 putative seed 5E-08 IPR005513 PF03760 LEA-4 maturation protein [O. sativa] HC248 1.80 Ϯ0.96 CK906408 IPR007011 PF04927 AAF21312 seed matura- 3E-26 IPR007011 PF04927 LEA-6 tion protein PM26 [G. max] HC96 1.40 Ϯ0.52 CK906401 IPR004926 PF03242 T07161 LEA homo- 1E-15 IPR004926 PF03242 LEA-7 log-tomato HC332 4.78 Ϯ1.08 CK906400 IPR004864 PF03168 T09875 LEA protein 5E-50 IPR004864 PF03168 LEA-8 Lea14-A- upland cot- ton HC22 2.17 Ϯ0.80 CK906403 AAC37469 LEA protein 7E-06 IPR008390 PF05512 LEA-10 with hydrophobic domain [G. max] HC63 2.70 Ϯ0.47 CK906405 IPR008390 PF05512 AAD10377 hydrophobic 3E-44 IPR008390 PF05512 LEA-10 LEA-like protein [O. sativa] HC216 2.83 Ϯ0.72 CK906407 IPR008390 PF05512 XP࿞477728 putative plas- 1E-36 IPR008390 PF05512 LEA-10 ma membrane associated protein [O. sativa]

² Average log2 expression ratios (fully-hydrated (FH) versus dehydrating (D) leaf tissue) and standard deviations (SD) from microarrays and reverse northern blots described by Collett et al. (2004). * Genbank Accession Numbers from Collett et al. (2004) which are based on single pass sequence from 3Ј end. Analysis based on contigs of 5Ј and 3Ј sequence information. ** LEA groups correspond to LEA groups 2±8, de®ned by Wise 2003, and LEA-10 de®ned by IPR008390. *** No Blastx homologue identi®ed. **** No Pfam domain identi®ed. SIGNATURE OF SEEDS IN RESURRECTION PLANTS 779

TABLE 3. Classi®cation of LEAs from other resurrection plants.

Genbank InterproScan Species Lab ref. accession number superfamily Pfam domain LEA group Description Craterostigma plantagineum Pcc3-06 P23283 IPR004238 PF02987 LEA-3 Desiccation -related protein [C. plantagineum] pcC27-45 P22241 IPR004864 PF03168 LEA-8 Desiccation -related protein [C. plantagineum] pcC27-04 P22238 IPR000167 PF00257 LEA-2 Desiccation -related protein [C. plantagineum] pcC6-19 S43775 IPR000167 PF00257 LEA-2 Desiccation -related protein [C. plantagineum] Xerophyta viscosa AAP22171 AAP22171 IPR000167 PF00257 LEA-2 Xerophyta viscosa XVT8 IPR000167 PF00257 LEA-2 Ndima et al., 2001 (no Genbank submission) Tortula ruralis T. ruralis EST TrEMBL ID for Cluster closest homologue*1 37 O16527 IPR004238 PF02987 LEA-3 C. elegans CE-LEA 28 P13934 IPR004238 PF02987 LEA-3 Late embryogenesis abundant (LEA) protein 76 [B. napus] 22 O16527 IPR004238 PF02987 LEA-3 C. elegans CE-LEA 21 Q9RV58 IPR004238 PF02987 LEA-3 Protein DR1172 LEA type 1 family [D. ra- diodurans] 40 Q9XFD0 IPR004238 PF02987 LEA-3 ABA-inducible protein WRAB1 (cold-responsive LEA) [T. aestivum] 18 Q9LF88 IPR004238 PF02987 LEA-3 Putative late embryogenesis abundant protein [A. thaliana] 32 Q9ZRF8 IPR008390 PF05512 LEA-10 Hydrophobic LEA-like protein [O. sativa] * 1 TrEMBL homologues for LEAs represented in the 30 most abundant transcripts present in T. ruralis rehydration cDNA libraries taken from Oliver et al. (2004), Table 1. which were present on the Affymetrix 25K genechip, Anti-oxidants was analysed. Genes were classi®ed into the same seed The percent change in AP, GR and SOD enzyme development and abiotic stress clusters that were used activity, taken from published data on desiccation-tol- for the LEA analysis. In contrast to the LEA genes, erant (resurrection plants and orthodox seeds) and 72% of the Arabidopsis antioxidant genes were ex- -sensitive tissues upon dehydration from full turgor is pressed at high levels in control plants (Ͼ40% maximum seed or stress expression) and were therefore shown in Figure 3. Discrepancies in reporting (for ex- classi®ed as housekeeping genes (cluster E) (Table 4). ample many studies, particularly those on vegetative Very few (4/71) of the anti-oxidants were seed-specif- tissues, do not report tissue water content but only dry- ic. However, At1g48130 (1-cys peroxiredoxin) gave ing time) has limited the number of species that could the highest intensity reading of all the antioxidants as- be compared. The data in Figure 3 show that there is sayed in this study, with less than 0.1% expression in no consistent trend in AP activity among tolerant tis- vegetative tissue. Homologues of 1-cys peroxiredoxins sues, nor between tolerant and sensitive types. Among have been identi®ed in both X. viscosa (Mowla et al., the resurrection plants only the Xerophyta spp., and 2002) and X. humilis (Collett et al., 2004). This gene among orthodox seeds only Acer platanoides, show an is thus another example of a seed-speci®c gene from overall increase in AP activity on drying. In other tol- desiccation-sensitive plants that is expressed at high erant tissue, with the exception of wheat seeds, AP levels in vegetative tissues during desiccation in des- activity is down regulated. Amongst sensitive tissues iccation tolerant plants. a small increase occurred on drying of wheat seedlings, the remainder also showing down-regulation of Physiological studies AP activity. A decrease in AP activity during seed des- To further address the main question of whether DT iccation appears common and Bailly (2004) has sug- mechanisms in vegetative tissue of desiccation-tolerant gested that the ascorbate system is probably not in- plants is an adaptation of DT mechanisms in seeds, we volved in DT. GR activity, on the other hand, is ele- examined two physiological responses of DT in seeds, vated in dry relative to hydrated tissues in all tolerant namely, anti-oxidant activity and sucrose accumula- species, but also in all but one (Quercus robur) sen- tion. sitive type (Fig. 3). This suggests that GR is a general 780 N. ILLING ET AL.

FIG. 2. Northern blot analysis of mRNA transcript abundance of 13 LEAs in X. humilis leaves during a cycle of desiccation and rehydration. 18s rRNA signal intensity was used to indicate equal loading of RNA in each lane. A galactinol synthase probe (Genbank Acc No CN517268) was used to con®rm integrity of mRNA in samples lacking LEA expression. The assigned LEA group (Table 2) for each gene is given on the right hand side of the ®gure. stress responsive enzyme (and may be considered a also play a role in response to various degrees of stress ``housekeeping'' protectant) rather than being speci®c (Pammenter and Berjak, 1999; Bailly, 2004). to DT. Studies aimed at understanding the putative The data assembled in Figure 3 is a comparison be- role(s) of GR and the glutathione system give no clear tween hydrated and dry states and gives no informa- trend in relation to DT. For example GR activity in- tion on changes in enzyme activity during the process creased during maturation drying of French bean seeds of drying (the data are not reported in most studies). (Bailly et al., 2001) but declined in the case of wheat It is possible that there was a change in enzyme activ- seeds (De Gara et al., 2003) and remained unchanged ity in response to initial drying. Our study on dehy- in sun¯ower seeds (Bailly et al., 2003). Reviews on dration of vegetative tissue of three Eragrostis spp. vegetative tissues and recalcitrant seeds similarly in- with differing degrees of tolerance to water de®cit dicate differences among species in response to water shows that anti-oxidant enzymes were active in hy- de®cit (Kermode and Finch-Savage, 2002; Farrant, drated tissues of all species (Fig. 4), suggesting a 2000; Kranner and Grill, 1997; Navari-Izzo et al., housekeeping role (as predicted by the analysis above). 1997; Kranner, 2002). SOD activity (Fig. 3) is elevated There was an initial increase in activity of AP, GR and in all desiccation-tolerant tissues surveyed but only in SOD in all the species at 70% RWC. In the desicca- one sensitive species (Digitaria sanguinalis leaves). tion-sensitive E. teff and E. curvula enzyme activity SOD activity was down-regulated in all other desic- ceased when the plants were dried below their critical cation-sensitive types. While this might suggest a role water contents (Balsamo et al., 2005) of 50 and 40%, unique to DT, reviews of the literature suggest that the RWC respectively. In contrast, the activities of all SOD enzymes are probably also housekeeping and three enzymes remained elevated at lower RWC in the SIGNATURE OF SEEDS IN RESURRECTION PLANTS 781

TABLE 4. Analysis of expression of anti-oxidant genes during seed development and in response to abiotic stress in A. thaliana.

Cluster A Cluster E Max seed, Ͻ30% stress Control Ͼ50% seed or stress At1g20630 Catalase 1 (Cat1) At1g07890 Ascorbate peroxidase (APX1) At1g48130 1-Cys peroxiredoxin (1-cys-PrxR) At1g08830 Cu/Zn superoxide dismutase (CSD1) At2g28190 Cu/Zn superoxide dismutase (CSD2) At1g09090 NADPH oxidase (RbohB) At3g11050 Ferritin 4* At1g19230 NADPH oxidase (RbohE) At3g56350 Mn superoxide dismutase like At1g20620 Catalase 3 (Cat3) At1g33660 Ascorbate peroxidase (APX7) Cluster B At1g60740 Peroxiredoxin Type 2 (Type 2 PrxR D) Max seed, Ͼ30% stress At1g63460 Glutathione peroxidase 8 (GPX8) At1g65990 Peroxiredoxin Type 2 (Type 2 PrxR A) At1g63940 Monohydroascorbate reductase (MDAR1) At4g11600 Phospholipid glutathione peroxidase (GPX6) At1g64060 NADPH oxidase (RbohF) At5g64210 Alternative oxidase (AOX2) At1g65980 Peroxiredoxin Type 2 (Type 2 prxR B) At1g75270 Dehydroascorbate reductase (DHAR3) Cluster C At1g77490 Thylakoid ascorbate peroxidase (Thylakoid-APX) Max stress, Ͼ30% seed At2g25080 Glutathione peroxidase 1 (GPX1) At1g19570 Dehydroascorbate reductase 5 (DHAR5) At2g40300 Ferritin 3 At2g31570 Glutathione peroxidase 2 (GPX2) At2g43350 Glutathione peroxidase 3 (GPX3) At3g24170 Glutathione reductase 1 (GR1) At2g48150 Glutathione peroxidase (GPX4) At3g56090 Ferritin 2 At3g06050 Peroxiredoxin 2-cys (2-cys-PrxR F) At5g01600 Ferritin 1 At3g10920 Mn superoxide dismutase (MSD1) At3g11630 Peroxiredoxin 2-cys (2-cys-PrxR A) Cluster D At3g26060 Peroxiredoxin (PrxR O) Max stress, Ͻ30% seed At3g27620 Alternative oxidase (AOX1C) At1g32350 Putative alternative oxidase (AOX putative) At3g27820 monodehydroascorbate reductase (MDAR3) At3g09640 Ascorbate peroxidase 2 (APX2) At3g45810 NADPH oxidase J (RbohJ) At3g09940 Monohydroascorbate reductase 2 (MDAR2) At3g52880 Monohydroascorbate reductase (MDAR4) At3g22360 Alternative oxidase 1B (AOX1B) At3g52960 Peroxiredoxin Type 2 (Type 2 PrxR E) At3g22370 Alternative oxidase 1A (AOX1A) At3g54660 Glutathione reductase (GR2) At5g47910 NADPH oxidase (RbohD) At3g63080 Glutathione peroxidase 5 (GPX5) At4g08390 Stromal ascorbate peroxidase (Stromal-APX) At4g09010 Ascorbate peroxidase (APX4) At4g11230 NADPH oxidase (RbohH0) At4g22260 Alternative oxidase (AOX1B) At4g25090 NADPH oxidase (RbohG) At4g25100 Fe superoxide dismutase (FSD1) At4g31870 Glutathione peroxidase 7 (GPX7) At4g32320 Ascorbate peroxidase (APX6) At4g35000 Ascorbate peroxidase (APX3) At4g35090 Catalase 2 (Cat2) At4g35970 Ascorbate peroxidase 5 (APX5) At5g03630 Monodehydroascorbate reductase 5 (MDAR5) At5g06290 Peroxiredoxin 2-cys (PrxR B) At5g07390 NADPH oxidase (RbohA) At5g16710 Dehydroascorbate reductase (DHAR1) At5g18100 Cu/Zn superoxide dismutase (CSD3) At5g23310 Fe superoxide dismutase (FSD3) At5g36270 Dehydroascorbate reductase (DHAR2) At5g51060 NADPH oxidase (RbohC) At5g51100 Fe superoxide dismutase (FSD3) At5g60010 NADPH oxidase (RbohH) * Silique speci®c.

DT E. nindensis compared to the desiccation-sensitive enzyme activity therefore corroborates the conclusions species. Mature air-dry orthodox seeds of both E. nin- from the Arabidopsis antioxidant expression pro®les, densis and E. teff retained signi®cant anti-oxidant en- in that anti-oxidant activity does not appear to be a zyme activity (Fig. 5). Since the enzyme assays are mechanism speci®c to DT. done in vitro this re¯ects only their potential for activity and it is unlikely that they are active in situ at Sucrose 5% RWC. Nonetheless, the ability to remain active Our survey of changes in sucrose content in re- when extracted from dry tissue is evidence that the sponse to drying in tissues of desiccation-tolerant and enzymes had been suf®ciently protected in the dry -sensitive species (Fig. 6) shows that this sugar does state. The lack of enzyme activity in dry tissues of E. indeed increase, to varying extents, on drying in teff and E. curvula is likely to be due to their desic- angiosperm resurrection plants and orthodox seeds. cation-induced denaturation (rather than down-regu- There was also a general increased level in response lation) since the plants did not survive drying below to drying in those sensitive tissues from which we 50% RWC (data not shown). Analysis of anti-oxidant could obtain drying course data. However, all but one 782 N. ILLING ET AL.

FIG. 3. Percent change in A) ascorbate peroxidase, B) glutathione reductase and C) superoxide dismutase activity from full turgor to air dry state in vegetative tissues of resurrection plants (RP), orthodox seeds (OS) and desiccation-sensitive tissues (SENS). SENS data are a combination of vegetative tissues and recalcitrant seeds. Res- urrection Plants (RP): CW, Craterostigma wilmsii; XH, Xerophyta FIG. 4. Changes in activity of the antioxidant enzymes ascorbate humilis; EN, Eragrostis nindensis;MF,Myrothamnus ¯abellifolius; peroxidase (A), glutathione reductase (B) and superoxide dismutase XV, Xerophyta viscosa. Orthodox seeds (OS): PV, Phaseolus vul- (C) during dehydration of the desiccation-tolerant grass Eragrostis garis;TA,Triticum aestivum;AP,Acer platanoides; HA, Helianthus nindensis (ࡗ) and desiccation-sensitive grasses E. teff (Ⅲ) and E. annuus; BC, Brassica campestris. Sensitive tissues (SENS): TA, curvula (᭡). E. nindensis survives drying to 5% RWC; E. teff and Triticum aestivum (seedlings); AH, Aesculus hippocastanum; QR, E. curvula die below 50% and 40% RWC respectively. Values Quercus robur; APs, Acer pseudoplatanus; AS, Acer saccharinum; shown are means of nine replicates (three separate extracts, three TC, Theobroma cacao; SR, Shorea robusta; DS, Digitaria sanguin- internal replicates). Vertical bars denote standard deviation. alis. Data taken from: Leprince et al. (1990); Hendry et al. (1992); Chaitanya and Naithani, 1994; Li and Sun (1999); Sherwin and Far- rant (1998); Greggains et al. (2000); Bailly et al. (2001, 2003); De Gara et al. (2003); Farrant et al. (2003, 2004); Ekmekci et al. (2004). SIGNATURE OF SEEDS IN RESURRECTION PLANTS 783

FIG. 5. Antioxidant enzyme activity in mature dry seeds of E. nindensis (white bars) and E. teff (black bars). AP, ascorbate peroxidase (units, nmol´minϪ1´mg proteinϪ1); GR, glutathione reductase (units, nmol NADP.mg proteinϪ1); SOD, superoxide dismutase (units, Units.mg proteinϪ1). Values shown are the means of nine replicates FIG. 6. Percent change in sucrose content from full turgor to air (three separate extracts, three internal replicates). Vertical bars de- dry state in vegetative tissues of resurrection plants (RP), orthodox note standard deviation. seeds (OS) and desiccation-sensitive tissues (SENS). SENS data are a combination of vegetative tissues and recalcitrant seeds. Resurrec- tion Plants (RP): CW, Craterostigma wilmsii; XH, Xerophyta hu- (wheat seedling) of the reports on sensitive tissues deal milis; EN, Eragrostis nindensis;MF,Myrothamnus ¯abellifolius; with recalcitrant seeds (not desiccation-tolerant) and BC, Borya constricta; CS, Coleochloa setifera; MK, Microchloa kunthii;XV,Xerophyta viscosa; SS, Sporobolus stap®anus. Ortho- the accumulation of sucrose may well be as storage dox seeds (OS): PS1, PS2 & PS3, Pisum sativum, genotypes SD1, reserve rather than pertaining to water de®cit during SD5 and SD7 respectively from Karner et al. (2004); HV, Hordeum maturation. At least one species (Avicennia marina) vulgaris;PV,Phaseolus vulgaris;TA,Triticum aestivum; BC, Bras- does not dry during the terminal stages of development sica campestris;AP,Acer platanoides. Sensitive tissues (SENS): TA, and yet considerable levels of sucrose accumulate Triticum aestivum (seedlings); QR, Quercus robur; AM, Avicennia marina; APs, Acer pseudoplatanus; AS, Acer saccharinum; AA, Ar- (Farrant et al., 1992). The work on wheat coleoptiles aucaria angustifolia; CS, Camellia sinensis. Data taken from: Le- indicated that these tissues were newly germinated and prince et al. (1990); Farrant et al. (1992); Finch-Savage and Blake still relatively plastic with respect to re-induction of (1994); Steadman et al. (1996); Farrant and Walters, (1997); Ghas- DT (Farrant et al., 2004) and thus sucrose accumula- empour et al. (1998); Li and Sun, (1999); Farrant et al. (2003, 2004); Karner et al. (2004); Vander Willigen et al. (2001); Whittaker et al. tion in that system might well be related to DT. (2004). Our study on Eragrostis spp. provides additional evidence that sucrose accumulation is a mechanism linked to DT. In vegetative tissues, sucrose accumulates only in the desiccation-tolerant species E. nindensis in response to drying (Fig. 7). In contrast, we ®nd that the mature orthodox seeds of both E. nindensis and E. teff accumulate high sucrose levels of 88 (Ϯ12) and 145 (Ϯ16) ␮mol´mg dwϪ1 respectively. The accumulation of sucrose in E. nindensis (Fig. 7) is not likely to be due to photosynthesis, as this is shut down early in the drying time course before max- imal sucrose accrual. This is true too of other resurrection plants studied to date (Farrant, 2000; Mundree et al., 2002; Whittaker et al., 2001, 2004) and sucrose accumulation is proposed to be from mobilization of other storage oligo- and/or polysaccharides such as oc- tulose, stachyose and starch (Schwall et al., 1995; Nor- wood et al., 2000, 2003). In seeds, sucrose is imported FIG. 7. Changes in photosynthetic assimilation rate (A) as a per- from the parent plant. Although photosynthesis contin- centage of that in control hydrated tissue (lefthand Y axis, closed ues to lower water contents in E. curvula, sucrose is symbols) and sucrose content (␮mol. mg dwϪ1) (righthand Y axis, not accumulated in this species (Fig. 7) and this spe- open symbols) during dehydration of the desiccation-tolerant grass cies apparently does not have the ability to up-regulate Eragrostis nindensis (᭡) and desiccation-sensitive grasses E. curvula (ࡗ) and E. teff (Ⅺ) (sucrose only). E. nindensis survives drying alternative means of sucrose accumulation upon water to 5% RWC; E. teff and E. curvula die below 50% and 40% RWC de®cit stress. Sucrose accumulation is therefore com- respectively. The values shown are means of three separate extrac- mon to both seed development and DT in vegetative tions. Vertical bars denote standard deviation. 784 N. ILLING ET AL.

tissues, and we suggest that sucrose accumulation for protection of proteins. An alternative explanation protection against desiccation damage is not associated could be that different LEAs are speci®cally targeted with photosynthesis but involves up-regulation of al- to different organelles or cellular structures, where ternative pathways, with signals that might be common they play a local role in protecting proteins, nucleic to those present during orthodox seed development. acids and membranes from the effects of water loss. LEAs have been reported to be expressed at high lev- DISCUSSION els in several other desiccation tolerant plants includ- This study offers an approach to exploring the ori- ing the bryophyte T. ruralis, and the angiosperms C. gins of DT in angisosperms and resurrection plants. plantagineum and X. viscosa (Table 3). These LEAs Two possibilities were entertained which are not nec- all represent the LEA-2, LEA-3 LEA-8 and LEA-10 essarily mutually exclusive: DT in resurrection plants superfamilies. Six different LEA-3 ESTs were was acquired via DT of seeds and/or via effective ad- amongst the most abundantly represented mRNA tran- aptation of abiotic stress responses. We have looked at scripts in a T. ruralis rehydration cDNA library (Oli- both molecular and physiological/biochemical re- ver et al., 2004) and a LEA-3 gene has been shown sponses to desiccation and abiotic stress to ask whether to expressed in the anhydrobiotic nematode Aphelen- responses commonly associated with DT in vegetative chus avenae (Browne et al., 2002). The LEA-3 super- tissues are also active in desiccating seeds and/or dur- family may thus represent very ancient proteins that ing abiotic stress. play important roles during desiccation. The nomenclature used in this study to describe the LEAs LEA superfamilies corresponds to the convention de- First we systematically compared the expression scribed by Wise (2003), based on differences in pep- levels of 35 LEAs represented on an Arabidopsis Af- tide composition (see Wise and Tunnacliffe, 2004). We fymetrix 25K gene chip during seed development, and have included a LEA-10 group to accommodate an under a standardized set of conditions for abiotic stress unclassi®ed LEA-like protein in the Pfam database. It treatments. A large diversity of LEAs were expressed remains to be seen whether different functions can be during seed development in comparison to the re- assigned to the different LEA superfamilies. We used sponse of vegetative tissue to abiotic stress. Whereas the terms `LEA' and `LEA-like' to identify superfam- there is no particular class of LEAs that is uniquely ilies which represent this large class of proteins in the expressed during abiotic stress, genes belonging to Pfam database. It is important to bear in mind that the LEA-1, -6 and -9 superfamilies are only signi®cantly annotation of sequences as LEAs tends to be arbitrary, expressed during seed development. These LEAs may originally assigned according to the abundance of a thus be uniquely associated with defense against se- transcript during late embryogenesis. A unifying fea- vere water loss such as would occur in desiccation- ture of this group of proteins is their large number of tolerant angiosperms/resurrection plants or orthodox homologous repetitive hydrophilic peptide motifs and seeds. their high percentage of glycine residues. Recent re- Many of the abiotic stress-responsive LEAs identi- ports on functional analysis of recombinant proteins of ®ed in the expression dataset have also been identi®ed AvLEA1, a Group 3 LEA protein from the anhydro- in other Arabidopsis microarray studies on abiotic biotic nematode A. avenae, and Em, a group I LEA stress. These include At1g01470 (LEA-8), At1g20440 protein from wheat (Goyal et al., 2005) have shown (LEA-2), At1g52690 (LEA-3), At5g06760 (LEA-4) that both these LEAs show anti-aggregation properties and At5g66400 (LEA-2) (Bray, 2004). Our compari- and protected enzyme activity under conditions of wa- son with expression data for seed development shows ter loss. These properties were synergistically en- that only At1g20440 (LEA-2) is speci®c to the abiotic hanced in the presence of sucrose. stress response. Notably, no LEA-1, -6, -7, -9 or -10 genes were identi®ed as stress up-regulated in these Antioxidants studies. In contrast to the LEAs, mRNA transcripts for most We have shown that the expression of at least 16 of the anti-oxidant enzymes were abundant in Arabi- different LEA genes, representing the LEA-2, -3 -4, dopsis under control conditions. Very few anti-oxi- -6 -7, -8 and -10 superfamilies, is activated during des- dants were seed-speci®c but notably, one of these, a iccation in X. humilis leaves (Collett et al., 2004). 1-cys-peroxiredoxin, has been previously shown to be LEA-6 was identi®ed as a `seed-speci®c' group in Ar- abundantly expressed during desiccation in the moss abidopsis. Northern blot analysis has shown that the T. ruralis and in the leaves of X. humilis and X. vis- LEAs investigated are speci®cally activated during the cosa. late stages of desiccation, and not during the early Overall, the physiological data on antioxidant en- stages of water loss, suggesting that they are a unique- zymes such as AP, GR and SOD suggest that these are ly desiccation-speci®c set of LEAs. We speculate that `housekeeping' protectants, responsive in the case of the simultaneous activation of such a large comple- most abiotic stresses. We propose that they while they ment of LEAs under conditions of water loss could are part of the protection systems in desiccation-tol- point towards the formation of an interacting network erant tissues, they are not unique to them and thus are necessary for the stabilization of membranes and the not useful in the evaluation of evolution of DT. Only SIGNATURE OF SEEDS IN RESURRECTION PLANTS 785 in true desiccation-tolerant tissues can the activity re- ACKNOWLEDGMENTS main elevated, but this is likely to be a consequence We thank Cathal Seoighe and Brigitte Hamman for of mechanisms that protect the anti-oxidant enzymes, assistance in compiling the data for the study and rather than a unique DT mechanism. Bronwen Aken and Zek Ginbot for some of the anti- However, there are some antioxidants that appear to oxidant and sucrose data generated in the Eragrostis be novel to DT, expressed only in maturation drying study. This research was supported by funding from of orthodox seeds and desiccation of resurrection the University of Cape Town, the National Research plants. For example, a 1-Cys peroxiredoxin has been Foundation, South Africa and the National Bioinfor- reported to be seed speci®c (Aalen, 1999; Haslekas et matics Network, South Africa. al., 1998) but is induced on drying of the resurrection plant X. viscosa (Mowla et al., 2002). Interestingly, a REFERENCES 1-Cys peroxiredoxin is also expressed during rehydra- Aalen, R. B. 1999. Peroxiredoxin antioxidants in seed physiology. tion of the desiccation-tolerant moss T. ruralis (Oliver, Seed Sci. Res. 9:285±295. 1996) and thus might be indicative of the evolutionary Alpert, P. and M. J. Oliver. 2002. Drying without dying. In M. Black process of these antioxidants in DT. and H. W. Pritchard (eds.), Desiccation and survival in plants: Drying without dying, pp. 3±43. CABI publishing, Wallingford Sucrose and New York. Bailly, C. 2004. Active oxygen species and antioxidants in seed Sucrose appears to be universally present in mature biology. Seed Sci. Res. 14:93±107. orthodox seeds (Chen and Burris, 1990; Vertucci and Bailly, C., A. Audigier, F. Ladonne, M. H. Wanger, F. Ocste, F. Con- Farrant, 1995) and increases in response to desiccation bineau, and D. Come. 2001. Changes in oligosaccharide content in all angiosperm resurrection plants studied to date and antioxidant enzyme activities in developing bean seeds as related to acquisition of drying tolerance and seed quality. J. (Ghasempour et al., 1998; Scott, 2000; Whittaker et Exp. Bot. 52:701±708. al., 2001, 2004). Although the precise role of sucrose Bailly, C., A. Benamar, F. Corbineau, and D. CoÃme. 1996. Changes in DT is still unclear, it is generally agreed that the in malondialdehyde content and in superoxide dismutase, cat- accumulation of this metabolite is important for pro- alase and glutathione reductase activities in sun¯ower seeds as tection against desiccation damage (Koster and Leo- related to deterioration during accelerated aging. Physiol. Plant. 97:104±110. pold, 1988; Scott, 2000; Walters et al., 2002). Our data Bailly, C., J. Leymarie, S. Rousseau, D. Come, A. Feutry, and F. show that sucrose accumulates in vegetative tissues of Corbineau. 2003. Sun¯ower seed development as related to an- only the desiccation-tolerant species of Eragrostis and, tioxidant enzyme activities. In G. Nicolas, K. J. Bradford, D. like other resurrection plants studied to date (Farrant, Come, and H. W. Pritchard (eds.), The biology of seeds: Recent research advances, pp. 69±75. CABI publishing, Wallingford 2000; Mundree et al., 2002; Whittaker et al., 2001, and New York. 2004), appears unrelated to photosynthesis. Thus while Balsamo, R., C. Vander Willigen, and J. M. Farrant. 2005. Relating sucrose itself is a common metabolite, its method of leaf tensile properties to drought tolerance for selected species accumulation under stress conditions might be an im- of Eragrostis. Ann. Bot. (In press) portant factor in DT. Breakdown of oligo- and poly- Bewley, J. D. and M. J. Oliver. 1992. Desiccation-tolerance in vegetative plant tissues and seeds: protein synthesis in relation to saccharides and their mobilization and resynthesis into desiccation and a potential role for protection and repair mech- sucrose within tissues that become protected against anisms. In C. B. Osmond and G. Somero (eds.), Water and life: desiccation damage is a common phenomenon in seeds A comparative analysis of water relationships at the organis- and resurrection plants. It is not known whether the mic, cellular and molecular levels, pp. 141±160. Springer-Ver- lag, Berlin. mechanisms and signals for this accumulation are sim- Bray, E. A. 1991. Molecular responses to water de®cit. Plant Phy- ilar in seeds and resurrection plants but it is clear that siol. 103:1035±1040. they are absent from vegetative tissues of desiccation- Bray, E. A. 1997. Plant responses to water de®cit. Trends Plant Sci. sensitive species E. teff and E. curvula. 2:48±54. Browne, J., A. Tunnacliffe, and A. Burnell. 2002. Plant desiccation IN SUMMARY gene found in nematode. Nature 416:38. Chaitanya, K. S. and S. C. Naithani. 1994. Role of superoxide, lipid Some of the LEAs and the majority of anti-oxidants peroxidation and superoxide dismutase in membrane perturba- analysed in the present study appear to be ``house- tion during loss of viability in seeds of Shorea robusta Gaertnf. keeping'' or are activated under conditions of abiotic New Phytol. 126:623±627. Chen, Y. and J. S. Burris. 1990. Role of carbohydrates in desiccation stress (cold, osmotic, salt or drought) in desiccation- tolerance and membrane behaviour in maturing maize seed. sensitive and -tolerant tissues. We propose that these Crop Sci. 30:971±975. are active and protect only at higher water contents Close, T. J. 1997. Dehydrins: A commonality in the response of (Ͼ65% RWC), but that for true desiccation-tolerance plants to dehydration and low temperature. Physiol. Plant. 100: a further repertoire of protectants are necessary and 291±296. Collett, H., A. Shen, M. Gardner, J. M. Farrant, K. J. Denby, and that these are activated below 65% RWC in desicca- N. Illing. 2004. Towards transcript pro®ling of desiccation tol- tion-tolerant tissues. The activation of a number of erance in Xerophyta humilis: Construction of a normalized 11k `seed-speci®c' desiccation protection mechanisms, X. humilis cDNA set and microarray expression analysis of 424 such as sucrose accumulation and expression of a cDNAs in response to dehyration. Physiol. Plant. 122:39±53. Crowe, J. H., J. F. Carpenter, and L. M. Crowe. 1998. The role of LEA-6 and a 1-cys-peroxiredoxin gene, in the vege- vitri®cation in anhydrobiosis. Ann. Rev. Physiol. 60:73±103. tative tissues of desiccation-tolerant plants points to- Dace, H., H. W. Sherwin, N. Illing, and J. M. Farrant. 1998. Use of wards acquisition of DT from seeds. metabolic inhibitors to elucidate mechanisms of recovery from 786 N. ILLING ET AL.

desiccation stress in the resurrection plant Xerophyta humilis. radical processes and loss of seed viability during desiccation Plant Growth Reg. 24:171±177. in the recalcitrant species Quercus robur. New Phytol. 122:273± De Gara, L., M. C. de Pinto, V. M. C. Moliterni, and M. G. 279. D'Egidio. 2003. Redox regulation and storage processes during Ingram, J. and D. Bartels. 1996. The molecular basis of dehydration maturation in kernels of Triticum aestivum. J. Exp. Bot. 54:249± tolerance in plants. Ann. Rev. Plant Physiol. Plant Mol. Biol. 258. 47:377±403. Dickie, J. B and H. W. Pritchard. 2002. Systematic and evolutionary Kappen, L. and F. Valladares. 1999. Opportunistic growth and des- aspects of desiccation tolerance in seeds. In M. Black and H. iccation tolerance: The ecological success of poikilohydrous au- W. Pritchard (eds.), Desiccation and survival in plants: Drying totrophs. In F. I. Pugnaire and F. Valladares (eds.), Handbook of without dying, pp. 239±260. CABI publishing, Wallingford and functional plant ecology, pp. 10±80. Marcel Dekker, New York. New York. Karner, U., T. Peterbauer, V. Raboy, D. A. Jones, C. L. Hedley, and Ekmekci, Y., J. M. Farrant, J. A. Thomson, and S. M. Mundree. A. Richter. 2004. myo-Inositol and sucrose concentrations affect 2004. Antioxidant response and photosynthetic characteristics the accumulation of raf®nose family oligosaccharides in seeds. of Xerophyta viscosa Baker and Digitaria sanguinalis L. leaves J. Exp. Bot. 55:1981±1987. induced by high light. Israel J. Plant Sci. 52:177±187. Kermode, A. R. and W. E. Finch-Savage. 2002. Desiccation sensi- Esterbauer, H. and D. Grill. 1978. Seasonal variation of glutathione tivity in orthodox and recalcitrant seeds in relation to develop- and glutathione reductase in needles of Picea abies. Plant Phy- ment. In M. Black and H. W. Pritchard (eds.), Desiccation and siol. 61:119±121. survival in plants: Drying without dying, pp.149±184. CABI Farrant, J. M., 2000. A comparison of mechanisms of desiccation- publishing, Wallingford and New York. tolerance among three angiosperm resurrection plant species. Knight, H. and M. R. Knight. 2001. Abiotic stress signaling path- Plant Ecol. 151:29±39. ways: Speci®city and cross-talk. Trends Plant Sci. 6:1360±1385. Farrant, J. M., C. Bailly, J. Leymarie, B. Hamman, D. Come, and Koster, K. L. and A. C. Leopold. 1988. Sugars and desiccation tol- F. Corbineau. 2004. Wheat seedlings as a model to understand erance in seeds. Plant Physiol. 88:829±832. desiccation-tolerance andÐsensitivity. Physiol. Plant. 120:563± Kranner, I. 2002. Glutathione status correlates with different degrees 574. of desiccation tolerance in three lichens. New Phytol. 154:451± Farrant, J. M., S. Bartsch, D. Loffell, C. Vander Willigen, and A. 460. Whittaker. 2003. An investigation into the effects of light on Kranner, I. and D. Grill. 1997. Desiccation and the subsequent re- the desiccation of three resurrection plants species. Plant Cell covery of cryptogamics that are resistant to drought. Phyton 37: Env. 26:1275±1286. 139±150. Farrant, J. M., N. W. Pammenter, and P. Berjak. 1992. Development Leopold, A. C., W. Q. Sun, and I. Bernal-Lugo. 1994. The glassy of the recalcitrant (homoiohydrous) seeds of Avicennia marina state in seeds: Analysis and function. Seed Sci. Res. 4:267±274. (Forsk.) Vierh.: Anatomical, ultrastructural and biochemical Leprince, O., R. Bronchart, and R. Deltour. 1990. Changes in starch events associated with development from histodifferentiation to and soluble sugars in relation to the acquisition of desiccation maturation. Ann. Bot. 7:75±86. tolerance during maturation of Brassica campestris seed. Plant Farrant, J. M. and C. Walters. 1997. Ultrastructural and biophysical Cell Env. 13:539±546. changes in developing embryos of Aesculus hippocastanum in Li, C. and W. Q. Sun. 1999. Desiccation sensitivity and activities of relation to the acquisition of tolerance to drying. Physiol. Plant. free radical-scavenging enzymes in recalcitrant Theobroma ca- 104:513±524. cao seeds. Seed Sci. Res. 9:209±217. Finch-Savage, W. E. and P. S. Blake. 1994. Indeterminate develop- Mittler, R., S. Vanderauwera, M. Gollery, and F. Van Breusegem. ment in desiccation-sensitive seeds of Quercus robur L. Seed 2004. Reactive oxygen gene network of plants. Trends Plant Sci. Res. 4:127±133. Sci. 9:1360±1385. Gaff, D. F. 1977. Desiccation tolerant plants of southern Africa. Oec- Mowla, S. B., J. A. Thomson, J. M. Farrant, and S. G. Mundree. ologia 31:95±109. 2002. A novel stress-inducible antioxidant enzyme identi®ed Gaff, D. F. 1980. Protoplasmic tolerance of extreme water stress. In from the resurrection plant Xerophyta viscosa. Planta 215:716± N. C. Turner and P. J. Kramer (eds.), Adaptions of plants to 726. water and high temperature stress, pp. 207±230. John Wiley Mundree, S. G. and J. M. Farrant. 2000. Some physiological and and Sons, New York. molecular insights into the mechanisms of desiccation tolerance Gaff, D. F. 1987. Desiccation tolerant plants in South America. Oec- in the resurrection plant Xerophyta viscosa Baker. In J. Cherry ologia 74:133±136. (ed.), Plant tolerance to abiotic stresses in agriculture: Role of Gaff, D. F. 1989. Responses of desiccation tolerant ``resurrection'' genetic engineering, pp. 201±222. Kluwer Academic Publish- plants to water stress. In K. H. Kreeb, H. Richter, and T. M. ers, The Netherlands. Hinckley (eds.), Structural and functional responses to environ- Mundree, S. G., B. Baker, S. Mowla, S. Peters, S. Marais, C. Vander mental stress: Water shortage, pp. 255±268. SPB Academic Willigen, K. Govender, A. Maredza, J. M. Farrant, and J. A. Publishers, The Hague. Thomson. 2002. Physiological and molecular insights into Ghasempour, H. R., D. F. Gaff, R. P. W. Williams, and R. D. Gi- drought tolerance. African J. Biotechnol. 1:28±38. anello. 1998. Contents of sugars in leaves of drying desiccation Navari-Izzo, F., S. Meneguzzo, B. Loggini, C. Vazzana, and C. L. tolerant ¯owering plants, particularly grasses. Plant Growth M. Sgherri. 1997. The role of glutathione system during dehy- Reg. 24:185±191. dration of Boea hydroscopica. Physiol. Plant. 99:23±30. Giannopolitis, C. N. and S. K. Ries. 1971. Superoxide dismutases. Ndima, T. B., J. M. Farrant, J. A. Thomson, and S. G. Mundree. I. Occurrence in higher plants. Plant Physiol 59:309±314. 2001. Molecular characterization of XVT8, a stress-responsive Goyal, K., L. J. Walton, and A. Tunnacliffe. 2005. LEA proteins gene from the resurrection plant Xerophyta viscosa Baker. Plant prevent protein aggregation due to water stress. Biochem. J. Growth Reg. 35:137±145. 388:151±157. Norwood, M., M. R. Truesdale, A. Richter, and P. Scott. 2000. Pho- Greggains, V., W. E. Finch-Savage, W. P. Quick, and N. M. Atherton. tosynthetic carbohydrate metabolism in the resurrection plant 2000. Putative desiccation tolerance mechanisms in orthodox Craterostigma plantagineum. J. Exp. Bot. 51:159±165. and recalcitrant seeds of the genus Acer. Seed Sci. Res. 10:317± Norwood, M., O. Toldi, A. Richter, and P. Scott. 2003. Investigation 327. into the ability of roots of the poikilohydric plant Craterostigma Haslekas, C., R. A. Stacy, V. Nygaard, F. A. Culianez-Marcia, and plantagineum to survive dehydration stress. J. Exp. Bot. 54: R. B. Aalen. 1998. The expression of a peroxiredoxin antioxi- 2313±2327. dant gene, AtPerl, in Arabidopsis thaliana is seed-speci®c and Oliver, M. J. 1996. Desiccation tolerance in vegetative plant cells. related to dormancy. Plant Molec. Biol. 36:833±845. Physiol. Plant. 97:779±787. Hendry, G. A. F., W. E. Finch-Savage, P. C. Thorpe, N. M. Atherton, Oliver, M. J. and J. D. Bewley. 1997. Desiccation-tolerance of plant S. M. Buckland, K. A. Nilsson, and W. E. Seel. 1992. Free tissues: A mechanistic overview. Hort. Rev. 18:171±214. SIGNATURE OF SEEDS IN RESURRECTION PLANTS 787

Oliver, M. J., S. E. Dowd, J. Zaragoza, S. A. Mauget, and P. R. Vander Willigen, C., N. W. Pammenter, S. G. Mundree, and J. M. Payton. 2004. The rehydration transcriptome of the desiccation- Farrant. 2001. Some physiological comparisons between the toleratant bryophyte Tortula ruralis: Transcript classi®cation and resurrection grass, Eragrostis nindensis, and the related desic- analysis. BMC Genomics. 2004 Nov 16; 5(1):89. cation-senstive species, Eragrostis curvula. Plant Growth Reg. Oliver, M. J., Z. Tuba, and B. D. Mishler. 2000. The evolution of 35:121±129. vegetative desiccation tolerance in land plants. Plant Ecol. 151: Vertucci, C. W. and J. M. Farrant. 1995. Acquisition and loss of 85±100. desiccation tolerance. In J. Kigel and G. Galili (eds.), Seed de- Pammenter, N. W. and P. Berjak. 1999. A review of recalcitrant seed velopment and germination, pp. 237±271. Marcel Dekker, New physiology in relation to desiccation-tolerance mechanisms. York. Seed Sci. Res. 9:13±37. Vicre, M., O. Lerouxel, J. M. Farrant, P. Lerouge, and A. Driouich. Piatkowski, D., K. Schneider, F. Salamini, and D. Bartels. 1990. 2004. Composition and desiccation induced alterations of the Characterization of ®ve abscisic acid-responsive cDNA clones cell wall in the resurrection plant Craterostigma wilmsii Phy- isolated from the desiccation-tolerant plant Craterostigma plan- siol. Plant. 120:229±239. tagineum and their relationship to other water stress genes. Plant Von Caemmerer, S. and G. D. Farquhar. 1981. Some relationships Physiol. 94:1682±1688. between the biochemistry of photosynthesis and the gas ex- Porembski, S. and W. Barthlott. 2000. Granitic and gneissic outcrops change of leaves. Planta 153:376±387. (inselfbergs) as centers of diversity for desiccation-tolerant vas- Walters, C., J. M. Farrant, N. W. Pammenter, and P. Berjak. 2002. cular plants. Plant Ecol. 151:19±28. Desiccation and Damage. In M. Black and H. W. Pritchard Proctor, M. C. F. 1990. The physiological basis of bryophyte pro- (eds.), Desiccation and survival in plants. Drying without dying, duction. Bot. J. Linn. Soc. 104:61±77. pp. 263±291. CAB International, Oxon. Schwall, G., R. Elster, J. Ingram, G. Bernacchia, G. Bianchi, L. Wang S. Y., H. J. Jaio, and M. Fuast. 1991 Changes in ascorbate, glutathione, and related enzyme activities during thidiazuron- Gallagher, F. Salamini, and D. Bartels. 1995. Carbohydrate me- induced bud break of apple. Physiol. Plant. 82:231±236. tabolism in the desiccation-tolerant plant Craterostigma plan- Whittaker, A., A. Bochicchio, C. Vazzana, G. Lindsey, and J. M. tagineum Hochst. In H. Pontis, G. Salerno, and E. Echeverria Farrant. 2001. Changes in leaf hexokinase activity and metab- (eds.), Sucrose metabolism, biochemistry, physiology and mo- olite levels in response to drying in the desiccation-tolerant spe- lecular biology. Current Topics in Plant Physiology, Vol 14, cies Sporobolus stap®anus and Xerophyta viscosa. J. Exp. Bot. pp. 245±253. American Society of Plant Physiologists. 352:961±969. Scott, P. 2000. Resurrection plants and the secret of the external leaf. Whittaker, A., T. Martinelli, A. Bochicchio, C. Vazzana, and J. M. Ann. Bot. 85:159±166. Farrant. 2004. Comparison of sucrose metabolism during the Sherwin, H. W. and J. M. Farrant. 1996. Difference in rehydration rehydration of desiccation-tolerant and desiccation-sensitive leaf of three desiccation-tolerant angiosperm species. Ann. Bot. 78: material of Sporobolus stap®anus. Physiol. Plant. 122:11±20. 703±710. Wise, M. J. 2003. LEAping to conclusions: A computational re- Sherwin, H. W. and J. M. Farrant. 1998. Protection mechanisms analysis of late emryogenesis abundant proteins and their pos- against excess light in the resurrection plants Craterostigma sible roles. BMC Bioinformatics 4:52±71. wilmsii and Xerophyta viscosa. Plant Growth Reg. 24:203±210. Wise, M. J. and A. Tunnacliffe. 2004. POPP the question: What do Sivamani, E., A. Bahieldrin, J. M. Wraith, T. Al-Niemi, W. E. Dyer, LEA proteins do. Trends Plant Sci. 9:1360±1385. T-H. D Ho, and R. Qu. 2000. Improved biomass productivity Xu, D., X. Duan, B. Wang, B. Hong, T.-H. D. Ho, and R. Wu. 1996. and water use ef®ciency under water de®cit conditions in trans- Expression of a late embryogenesis abundant protein gene, genic wheat constitutively expressing the barley HVA1 gene. HVA1, from barley confers tolerance to water de®cit and salt Plant Sci. 155:1±9. stress in transgenic rice. Plant Physiol. 110:249±257. Steadman, K., H. W. Pritchard, and P. M. Dey. 1996. Tissue-speci®c Zdobnov, E. M. and R. Apweiler. 2001. InterProScan-an integration soluble sugars in seeds as indicators of storage category. Ann. platform for the signature-recognition methods in InterPro. Bot. 77:667±674. Bioinformatics 17:847±848. Swire-Clark, G. A. and W. R. Marcotte, Jr. 1999. The wheat LEA Zimmermann, P., M. Hirsch-Hoffmann, L. Hennig, and W. Gruis- protein Em functions as an osmoprotective molecule in Saccha- sem. 2004. GENEVESTIGATOR. Arabidopsis microarray da- romyces cerevisiae. Plant Molec. Biol. 39:117±128. tabase and analysis toolbox. Plant Physiol. 136:2621±2632.