Molecular Plant • Volume 6 • Number 2 • Pages 369–385 • March 2013 RESEARCH ARTICLE

Metabolomic Profiling in Selaginella lepidophylla at Various Hydration States Provides New Insights into the Mechanistic Basis of Desiccation Tolerance

Abou Yobia, Bernard W.M. Woneb, Wenxin Xuc, Danny C. Alexanderc, Lining Guoc, John A. Ryalsc, Melvin J. Oliverd and John C. Cushmana,1

a Department of Biochemistry and Molecular Biology, University of Nevada, Reno, NV 89557–0330, USA b Department of Biological Sciences, University of Nevada, Reno, NV 89557–0314, USA c Metabolon Inc., 800 Capitola Drive, Suite 1, Durham, NC 27713, USA d US Department of Agriculture—Agricultural Research Service, Plant Genetic Research Unit, University of Missouri, Columbia, MI 65211, USA

ABSTRACT Selaginella lepidophylla is one of only a few species of spike mosses (Selaginellaceae) that have evolved des- iccation tolerance (DT) or the ability to ‘resurrect’ from an air-dried state. In order to understand the metabolic basis of DT, S. lepidophylla was subjected to a five-stage, rehydration/dehydration cycle, then analyzed using non-biased, global metabolomics profiling technology based on GC/MS and UHLC/MS/MS2 platforms. A total of 251 metabolites including 167 10.1093/mp/sss155 named (66.5%) and 84 (33.4%) unnamed compounds were characterized. Only 42 (16.7%) and 74 (29.5%) of compounds showed significantly increased or decreased abundance, respectively, indicating that most compounds were produced con- stitutively, including highly abundant trehalose, sucrose, and glucose. Several glycolysis/gluconeogenesis and tricarboxylic acid (TCA) cycle intermediates showed increased abundance at 100% relative water content (RWC) and 50% RWC. Vanillate, a potent antioxidant, was also more abundant in the hydrated state. Many different sugar alcohols and sugar acids were more abundant in the hydrated state. These polyols likely decelerate the rate of water loss during the drying process as well as slow water absorption during rehydration, stabilize proteins, and scavenge reactive oxygen species (ROS). In contrast, nitrogen-rich and γ-glutamyl amino acids, citrulline, and nucleotide catabolism products (e.g. allantoin) were more abundant in the dry states, suggesting that these compounds might play important roles in nitrogen remobilization during rehydration or in ROS scavenging. UV-protective compounds such as 3-(3-hydroxyphenyl)propionate, apigenin, and naringenin, were more abundant in the dry states. Most lipids were produced constitutively, with the exception of choline phosphate, which was more abundant in dry states and likely plays a role in membrane hydration and stabilization. In contrast, several poly- unsaturated fatty acids were more abundant in the hydrated states, suggesting that these compounds likely help maintain membrane fluidity during dehydration. Lastly, S. lepidophylla contained seven unnamed compounds that displayed twofold or greater abundance in dry or rehydrating states, suggesting that these compounds might play adaptive roles in DT.

Key words: abiotic/environmental stress; mass spectrometry; metabolomics; oxidative/photooxidative stress; desiccation tolerance; Selaginella.

Introduction coal (Beerling, 2002; Gensel, 2008). Today, extant lycophytes comprise a single monophyletic clade of just over 1000 species. Lycophytes represent a key vascular plant lineage that includes the Lycopodiaceae (club mosses), the Isoetaceae (quillworts), and the Selaginellaceae (spike mosses) that arose over 400 1 To whom correspondence should be addressed. E-mail [email protected], million years ago during the Silurian and dominated the tel. +1 775 784 1918, fax +1 775 784 1419. Earth’s flora from the Devonian through the Carboniferous to © The Author 2013. Published by the Molecular Plant Shanghai Editorial the end of the Permian (Friedman, 2011). Dramatic decreases Office in association with Oxford University Press on behalf of CSPB and

in atmospheric CO2 concentrations occurred during these IPPE, SIBS, CAS. Paleozoic epochs and the resulting decomposed flora now doi:10.1093/mp/sss155, Advance Access publication 13 December 2012 provide a major source of energy to mankind in the form of Received 9 August 2012; accepted 6 December 2012 370 Yobi et al. • Rehydration/Dehydration Metabolomics

Within the spike mosses, the single genus Selaginella con- S. tamariscina (Zhang et al., 2011). Antimicrobial alkaloids tains about 700 species that now exploit a diverse array of arc- have also been characterized from S. moellendorfii (Wang tic, temperate, tropical, and semi-arid habitats (Banks, 2009). et al., 2009). Although most spike moss species are susceptible to desicca- Despite their role in traditional medicine, comprehensive tion, a few species have evolved the ability to survive vegeta- metabolite studies on Selaginella species have been limited. tive tissue drying, defined as the near complete loss (80%–95%) Previous studies have characterized only a few, highly abun- of protoplasmic water (Tuba et al., 1998), and referred to as dant compounds such as lignin (White and Towers, 1967) desiccation tolerance (DT) (Oliver et al., 2000a). Such species and trehalose (White and Towers, 1967; Adams et al., 1990; include S. lepidophylla (Iturriaga et al., 2006), S. bryopteris Iturriaga et al., 2000; Vázquez-Ortíz and Valenzuela-Soto, (Deeba et al., 2009), and S. tamariscina (Wang et al., 2010). 2004; Liu et al., 2008). A recent comparative study examined The DT trait is relatively common in algae, lichens, and mosses the major metabolic differences between the desiccation- (Alpert, 2006; Wood, 2007). However, among vascular plants, sensitive (DS) S. moellendorffi and the DT S. lepidophylla. DT is extremely rare; only about ~0.15% of species being are Several sugars (e.g. glucose, sucrose), sugar alcohols (e.g. known as resurrection plants (Oliver et al., 2000a; Porembski inositol-1-phosphate, myo-inositol, mannitol), and betaine, and Barthlott, 2000; Proctor and Pence, 2002; Proctor and which act as osmoprotectants or hydroxyl radical scavengers, Tuba, 2002). were more abundant in S. lepidophylla at 100% and 50% In terms of response to desiccation, resurrection mosses relative water contents (RWC) (Yobi et al., 2012). In addition, (e.g. bryophytes) rely on a combination of constitutive pro- a variety of γ-glutamyl amino acids, which are thought to tection and inducible repair mechanisms that permit dehy- participate in reactive oxygen species (ROS) detoxification, dration to occur within minutes while maintaining their and possible nitrogen remobilization following rehydration, photosynthetic apparatus relatively intact (Tuba et al., 1998; were markedly higher in S. lepidophylla. A suite of second- Oliver et al., 2000b, 2005). In contrast, DT angiosperms ary compounds, such the flavonols apigenin, luteolin, and require a much longer time to dehydrate and still remain via- naringenin, also accumulated to greater concentrations in ble, presumably to allow sufficient time for mobilizing adap- S. lepidophylla, where they are thought to mitigate ultravi- tive responses and accumulation of key metabolites, such as olet light-induced free radical damage as well as aid in the sucrose, to survive in the desiccated state (Bernacchia et al., stabilization of membrane structures (Hoekstra et al., 2001). 1996; Oliver et al., 2000a). In contrast, lycophytes represent Lastly, polyunsaturated fatty acids were significantly more a plant lineage that lies between mosses and angiosperms abundant in S. lepidophylla, possibly reflecting a need for (Oliver et al., 2000a) and thus might be expected to exhibit greater membrane fluidity in the dry state (Yobi et al., 2012). both constitutive and inducible adaptive mechanisms of DT. The exact mechanisms used to withstand DT have not been Many monocot DT species lose their photosynthetic appara- well investigated in lycophytes. To better understand the met- tus, at least partially, during the dehydration process (Farrant, abolic basis of DT at very low RWCs within the Selaginellaceae, 2000; Proctor and Tuba, 2002), whereas many eudicot species the metabolome of S. lepidophylla was compared over a do not (Georgieva et al., 2009). DT Selaginella species retain five-stage, rehydration/dehydration cycle using an unbiased, their chlorophyll (and presumably their photosynthetic struc- high-throughput metabolomics platform. This temporal study tures) during desiccation (Pandey et al., 2010). revealed that S. lepidophylla employs a combination of diverse Members of the Selaginellales are well known for their constitutive and inducible metabolic strategies to survive and use in traditional folk medicine. For example, S. lepidophylla recover from desiccation of vegetative tissues. has been used as a diuretic, and as a treatment for urinary and kidney infections, chronic gastritis, and gastric carci- RESULTS AND DISCUSSION noma (Ruiz-Bustos et al., 2009; Robles-Zepeda et al., 2011). Extracts of S. tamariscina (Ahn et al., 2006; Lee et al., 2011; Metabolomics approaches can provide a detailed Ha et al., 2012), S. doederleinii (Liu et al., 2011), and S. bry- understanding of an organism’s phenotype (Schauer and opteris (Mishra et al., 2011) have been used as anti-cancer Fernie, 2006; Cascante and Marin, 2008), and disposition therapeutics. Total flavonoids from S. tamariscina have also towards and response to environmental stresses (Sanchez been employed as an anti-diabetic treatment (Zheng et al., et al., 2008; Shulaev et al., 2008; Urano et al., 2010). Recent 2011b). Various Selaginella species have been recognized for studies have compared two closely related species of their potential therapeutic value as anitviral, antimicrobial, Sporobolus (Oliver et al., 2011) and Selaginella (Yobi et al., or anticancer bioactivities. For example, several novel fla- 2012) in order to discern key metabolic differences that vones and biflavones have been characterized from S. lepido- might contribute to DT. In the current study, a total of 251 phylla (Aguilar et al., 2008), S. chrysocaulos and S. bryopteris metabolites were identified to provide insights into both (Swamy et al., 2006), S. labordei (Tan et al., 2009), S. moellen- constitutive and inducible metabolite abundance patterns dorffii (Cao et al., 2010; Liu et al., 2010; Wang et al., 2011; essential for the acquisition of DT in the resurrection Wu and Wang, 2011), S. uncinata (Zheng et al., 2011a), and lycophyte, S. lepidophylla. Yobi et al. • Rehydration/Dehydration Metabolomics 371

S. lepidophylla and Dehydration samples were collected from plants sampled at full dehydra- In order to discern relative metabolite abundance changes in tion 1 (7% RWC, DRY-1), partial rehydration (50% RWC, REH- S. lepidophylla during a rehydration/dehydration cycle, plants 50), full rehydration (100% RWC, HYD), partial dehydration were allowed to hydrate, then dehydrate, and their RWC was (50% RWC, DEH-50), and full dehydration 2 (7% RWC, DRY- tracked over a 24-h period. Upon hydration, water uptake was 2), and analyzed using non-biased, global metabolome tech- 2 relatively rapid; 70% RWC was achieved within 1 h and 100% nology based on GC/MS and UHLC/MS/MS platforms (Evans RWC after 24 h (Figure 1A). When hydrated, S. lepidophylla et al., 2009). The combined platforms detected a total of plants displayed fully expanded green microphylls (Figure 1B). 251 metabolites, of which 167 (66.5%) were named and 84 In contrast, water loss during dehydration was markedly (33.4%) were unnamed (Figure 2 and Supplemental Table 1). slower, with 70% water loss taking more than 4 h and the After the named metabolites were mapped onto general plants retaining 7% RWC after 24 h of drying. The rate of water biochemical pathways, they were categorized into classes of loss in S. lepidophylla was similar to S. bryopteris, a related DT which amino acids were the most prevalent (19%), followed species, which required about 6 h to lose 80%–90% of its RWC by carbohydrates (16%), lipids (13%), cofactors (6%), nucleo- (Deeba et al., 2009; Pandey et al., 2010). During dehydration, tides (5%), peptides (4%), and secondary metabolites (3%). the microphylls curl to form a tight ball (Figure 1B), which is Lastly, unnamed compounds in S. lepidophylla accounted for an adaptive response to protect the plant against damage due one-third of all compounds surveyed (Figure 2). to high irradiance or temperatures, or both (Lebkuecher and Eickmeier, 1991, 1992, 1993). Chlorophyll is maintained at least Metabolomic Differences during a Rehydration/ partially on the inner surface of the microphylls, indicating Dehydration Cycle that S. lepidophylla is homiochlorophyllous (Tuba et al., 1998). Partial Least Squares-Discriminant Analysis (PLS-DA) was used to build a comparative model of the five different Metabolome Composition of S. lepidophylla hydration states in S. lepidophylla. Metabolites missing three To assess the metabolomic response of S. lepidophylla to a or more out of six (50%) biological replicate values were rehydration/dehydration cycle, six biological replicate tissue excluded, resulting in a total of 204 metabolites for compari- son. The first three PLS-DA components explained 52.2% of the variation (R2 = 0.50; Q2 = 0.40) and showed distinct clus- tering among each of the five states, suggesting that various metabolites likely accounted for the differences observed in the model (Figure 3). The metabolomes of the two dry states

Figure 1. Rehydration/Dehydration Rates in S. lepidophylla. (A) Rehydration time course (h) measured by tracking the percent gain in relative water content (RWC) (blue line and squares) over time and dehydration time course (h) measured by tracking the percent loss in RWC over time (red line and diamonds). Data represent the mean of three independent replicate experiments (n = 9). Error bars represent standard error of the mean. (B) Representative images of S. lepidophylla plants at initial dry state (DRY-1), 50% rehydrated (REH-50), fully rehydrated (HYD), 50% dehy- Figure 2. Functional Categorization of 251 Named and Unnamed drated (DEH-50), and second dry state (DRY-2). Size bars = 2 cm. Metabolites Detected in S. lepidophylla over Five Different RWCs. 372 Yobi et al. • Rehydration/Dehydration Metabolomics

identical conditions were used for each dehydration episode, these differences between the two dry states might be due to variation arising from repeated hydration and dehydration cycles associated with the life history of different sets of plants. Amino acids, peptides, and nucleotide metabolites were notably overrepresented in the dry states (Supplemental Table 1). In contrast, various carbohydrates, particularly 4C, 5C, and 6C sugars and sugar alcohols, lipids, or lipid metabo- lites (with the exception of choline phosphate), and cofac- tors were clearly overrepresented in the hydrated state (Supplemental Table 1).

Carbohydrates and Markers of Energy Metabolism Many differences were observed in the abundance of metabo- lites of glycolysis/gluconeogenesis and the tricarboxylic acid (TCA) cycle. For example, glucose-6-phosphate, fructose-6-phosphate, Figure 3. Metabolites at Different Hydration States in S. lepidophylla maltose-6-phosphate, glycerate, and pyruvate showed a signifi- as Defined by Partial Least-Squares-Discriminant Analysis (PLS-DA) Constructed from 204 Metabolites. cantly (p < 0.05) increase and peak abundance during dehydra- The plot shows the three components that explains 52.2% of the tion (DEH-50) (Figure 4). Several TCA cycle intermediates, such metabolite signatures (R2 = 0.50; Q2 = 0.40). Upside-down triangles as oxaloacetate, fumarate, succinate, and alpha-ketoglutarate, (red) represent initial dry state (DRY-1), squares (brown) represent also showed peak abundance during dehydration, whereas suc- 50% rehydration (REH-50), asterisks (green) represent fully hydrated cinate was more abundant in the hydrated state (Figure 4 and (HYD), crosses (cyan) represent 50% dehydrated (DEH-50), and dia- monds (moss green) represent the second dry state (DRY-2). Supplemental Table 1). Similarly, glycerate, lactate, and pyruvate were significantly (p < 0.05) more abundant in the hydrated at 7% RWC were similar to one another. The metabolic pro- state, with the exception of the DEH-50 state (Supplemental file of the fully hydrated state at 100% RWC was distinct from Table 1). The peak accumulation of many glycolysis/gluconeo- these dry states, and that of the 50% RWC rehydration state genesis metabolites suggests that metabolic flux through these was intermediate between the dry and fully hydrated plants. pathways is essential for the production of downstream prod- Interestingly, the 50% RWC dehydration state was also clearly ucts important for the acquisition of DT. distinct from those of the 50% RWC rehydration and the ini- Various oligosaccharides accumulate in response to drying tial dry states, suggesting that distinct metabolic pathways in the vegetative tissues of all DT tracheophytes studied to operate during rehydration compared with dehydration. date (Alpert and Oliver, 2002). Based on the total relative ion In comparing the S. lepidophylla metabolite profiles counts, trehalose, sucrose, and glucose were the most abun- across all five hydration states, a total of 114 (45.4%) dant compounds within S. lepidophylla tissues and accounted compounds were detected that showed significantly altered for an estimated 50% of the total metabolites. These com- abundance (p < 0.05), including 42 (16.7%, 27 named and pounds tended to be more abundant in the hydrated state 15 unnamed) that were more abundant in one or more of relative to the dry state(s) (Supplemental Table 1) and their the dry or partially dry states, and 74 (29.5%, 57 named and relative abundances were trehalose > sucrose > glucose. 17 unnamed) that showed a greater abundance in the fully Sucrose alone, or in combination with other sugars, such as hydrated state. Two metabolites (i.e. pyruvate, galactose) raffinose-series oligosaccharides and cyclitols, or with macro- showed variable abundance patterns (Supplemental Table 1). molecules, such as late embryogenesis abundant (LEA) pro- A total of 11 metabolites were significantly (p < 0.05) more teins (Wolkers et al., 2001; Shimizu et al., 2010), likely protects abundant in both dry states compared with the fully hydrated vegetative tissues of resurrection species upon drying through state. In addition, 11 metabolites were more abundant only the formation of anhydrous glass or by replacement of water in the initial dry state, whereas the abundance of only one to prevent damaging interactions, such as membrane fusion compound was higher in the second dry state. The precise (Hoekstra et al., 2001). S. lepidophylla synthesizes high consti- origin of these differences remains unclear, but might be due tutive concentrations of sucrose (and trehalose and glucose), to differential metabolic activity changes that occur between and thus resembles DT mosses that maintain constitutively the initial and subsequent rehydration/dehydration cycles. high sucrose levels that do not further increase during drying In contrast, comparing the hydrated state with the two fully (Smirnoff, 1992). In contrast, in a DT fern (Farrant et al., 2009) dehydrated states, 26 metabolites showed significantly (p < and all angiosperm species examined to date, sucrose accu- 0.05) greater abundance than in both dry states. An additional mulation is induced following the imposition of dehydration 46 compounds were significantly (p < 0.05) more abundant stress (Ghasempour et al., 1998; Whittaker et al., 2001; Peters when compared with the second dry state only. Although et al., 2007; Toldi et al., 2009; Oliver et al., 2011). Yobi et al. • Rehydration/Dehydration Metabolomics 373

Figure 4. Metabolites Associated with Energy Metabolism (Glycolysis and TCA Cycle) in S. lepidophylla Profiled during a Rehydration/ Dehydration Cycle. Compounds in red indicate greater relative abundance in one or more dry states. Compounds in green indicate greater relative abundance in the hydrated state. Dotted arrows indicate a gap in the pathway and double-headed arrows indicate a reversible reaction. The x-axis represents the percentage of relative water content (% RWC): initial dry state (DRY-1); 50% rehydrated (REH-50); fully rehydrated (HYD); 50% dehydrated (DEH- 50); and second dry state (DRY-2). The y-axis box plots indicate the scaled intensity mean (closed triangles) and median (––––) values, top/bottom ranges of boxes indicate upper/lower quartiles, respectively, and top/bottom whiskers indicate the maximum/minimum distribution of the data, respectively. Open circles indicate extreme data points. The p- and q-values for the comparisons are presented in Supplemental Table 1. G-6-P, glucose-6-phosphate; F-6-P, fructose-6-phosphate; M-6-P, mannose-6-phosphate; 3-PG, 3-phosphoglycerate; PEP, phosphoenolpyruvate.

Trehalose, a non-reducing disaccharide, is known to accumu- Interestingly, the osmoprotectant glucosylglycerol late in high amounts in lycophytes (Adams et al., 1990; Iturriaga (2-O-alpha-D-glucopyranosyl-sn-glycerol) was also detected et al., 2000; Liu et al., 2008) as well as in many lower-order DT in S. lepidophylla and was more abundant in the hydrated organisms (Elbein et al., 2003). However, a critical role for this state relative to the dry state(s) (Supplemental Table 1). disaccharide in DT is questionable for several reasons. First, sev- Glucosylglycerol is typically found in cyanobacteria (Klähn eral resurrection ferns and angiosperms accumulate trehalose in and Hagemann, 2011). This compatible solute can stabilize only very small amounts (Ghasempour et al., 1998; Farrant et al., various enzymes against thermal denaturation as well as 2009). Second, a Saccharomyces cerevisiae mutant defective in improve enzyme stability eightfold following lyophilization trehalose biosynthesis can maintain DT, indicating that treha- and rehydration (Sawangwan et al., 2010). The constitutive lose is neither necessary nor sufficient for DT (Ratnakumar and and concerted production of sucrose, trehalose, and glucosyl- Tunnacliffe, 2006). However, trehalose has been shown to act glycerol are very likely critical to the preservation of diverse synergistically with LEA proteins to stabilize client enzymes by enzyme activities during and following desiccation. reducing their aggregation (Goyal et al., 2005). While trehalose alone might not be sufficient for DT, it might act cooperatively, as Sugar Alcohol Metabolism do other sugars, as a chemical chaperone to enhance the molec- Sugar alcohols or polyols are known to play important roles ular shield function of LEA proteins in reducing desiccation- in water-deficit and salinity-stress adaptation and tolerance induced aggregation of proteins (Chakrabortee et al., 2012). (Nuccio et al., 1999; Chen and Murata, 2002; Rontein et al., 374 Yobi et al. • Rehydration/Dehydration Metabolomics

2002). A variety of polyols showed significant (p < 0.05) hydrated tissue with osmolytes (Oliver et al., 2011). Slowing changes in abundance in S. lepidophylla undergoing the rate of water loss might permit more time for the biosyn- rehydration/dehydration, with peak accumulation occurring thesis of desiccation-adaptive proteins and metabolites that in the fully hydrated state (Figure 5 and Supplemental are required for the establishment of DT (Alpert and Oliver, Table 1). For example, xylitol abundance was 10–33-fold 2002). Conversely, biosynthesis of polyols might also slow the greater at 100% RWC compared with the DRY-1 and DRY-2 rate of water absorption during rehydration. Notably, several states, respectively (Figure 5). Relative amounts of other polyols such as sorbitol and erythritol exhibited increased polyols were also significantly higher in the fully hydrated accumulation at the 50% RWC rehydration stage (REH-50), state: ribitol was 16.5–20.0-fold higher, sorbitol was 5.8–20.0- consistent with this hypothesis. A direct comparison of DT fold higher, erythritol was 3.3–9.0-fold higher, compared S. lepidophylla and DS S. moellendorffii indicated that the with the DRY-1 and DRY-2 states, respectively, and glycerol DT species accumulated significantly greater amounts of no was 1.4–2.6-fold higher compared with the DRY-1 and DRY-2 fewer than seven different polyols in the fully hydrated state states, respectively, and 5.0-fold higher in the REH-50 state. (Yobi et al., 2012). Furthermore, a recent comparison of DT Threitol was 3.0–5.0-fold higher compared with the DRY-1 Sporobolus stapfianus and DS S. pyramidalis showed that and DRY-2 states, respectively; however, these differences several polyols, including arabitol, erythritol, and mannitol, were not significant. In addition, related oxidation products were several-fold more abundant in fully hydrated leaves of of these sugars included gluconate and erythronate, which the DT species (Oliver et al., 2011). These observations sug- were 2.8–16.5-fold and 2.0–2.5-fold more abundant, gest that constitutive expression of polyols in the hydrated respectively, in HYD compared with the DRY-1 and DRY-2 state might be critical for many resurrection species to slow states. Lastly, N-acetyglucosamine, a monosaccharide derived the rate of water loss during dehydration as well as water from glucose and an important structural sugar, was 2.0- uptake during rehydration. Indeed, S. lepidophylla shows a fold more abundant in HYD compared with the dried states slower rate of rehydration than does S. moellendorffii (Yobi (Figure 5). et al., 2012). In contrast, mannitol and galactose were significantly In addition to a role in water equilibrium, several poly- (p < 0.05) more abundant in the DRY-1 state (2.0-fold and ols can also act as osmoprotectants by stabilizing protein 1.9-fold, respectively) compared with the fully hydrated state (yeast hexokinase A) structure against thermal (Tiwari and (Supplemental Table 1). In addition to its probable role as Bhat, 2006) or acidic (Devaraneni et al., 2012) denaturation. an osmolyte, mannitol has been shown to act as a hydroxyl Glycerol, sorbitol, and xylitol exhibit better stabilization radical scavenger (Shen et al., 1999). Inositol 1-phosphate than glucose or trehalose through preferential hydration of was 1.2–1.8-fold more abundant in all dehydration states proteins in their denatured state (Xie and Timasheff, 1997) compared with the fully hydrated state, but these differ- by increasing surface tension (Kaushik and Bhat, 1998). ences were not significant (Supplemental Table 1). In contrast Both sorbitol and xylitol showed the greatest relative abun- to other DT species (Farrant et al., 2009; Oliver et al., 2011), dance in S. lepidophylla compared with S. moellendorffii raffinose-series sugars (e.g. galactinol, raffinose, stachyose) (Yobi et al., 2012) and in this study (Figure 5), which is were not detected and probably do not participate in DT in consistent with such a protein-stabilizing function. Several S. lepidophylla. Investigation into the metabolomes of other sugar alcohols, such as myo-inositol, mannitol, and sorbitol, lycophytes is needed to determine whether this observation also possess the ability to scavenge hydroxyl radicals as a holds true for all members of this ancient order within the way of reducing oxidative damage (Sanchez et al., 2008). plant kingdom. Raffinose-series sugars have been shown Thus, polyol accumulation might limit the oxidative stress to act as ROS scavengers in Arabidopsis (Nishizawa et al., burden of DT plants during the early phases of dehydration 2008; Nishizawa-Yokoi et al., 2008). The lack of detectable and rehydration. In addition to osmotic adjustment, the raffinose-series sugars in S. lepidophylla indicates that this accumulation of polyols might also facilitate other meta- function might be fulfilled by other sugars or other diverse bolic functions such as consumption of NADH, which might compounds with ROS-scavenging activities. enhance redox control and reduce ROS production in the The accumulation of polyols in the fully hydrated state in cells (Shen et al., 1999). Lastly, sugar alcohols might stimu- S. lepidophylla indicates several important roles for these late the expression of stress-adaptive genes, as has been sugars. Because polyols, especially glycerol, reduce the sur- observed in Arabidopsis plants engineered to express man- face tension of water (Tiwari and Bhat, 2006) and display nitol (Chan et al., 2011). strong water-binding activity via hydrogen bonding (Kataoka et al., 2011; Weng et al., 2011), they might decelerate the Amino Acid and Peptide Metabolism rate of water loss during the drying process (Figure 1A). Of the 48 amino acids and derivatives detected, relative Indeed, the ability of S. stapfianus to reduce the rate of water amounts of 10 of them occurred in significantly (p < 0.05) loss during dehydration, compared to the DS S. pyramidilis, greater amounts in one or more of the dry states than in the appears to be a result of a similar metabolic ‘priming’ of the fully hydrated state, whereas only eight amino acids were Yobi et al. • Rehydration/Dehydration Metabolomics 375

Figure 5. Differences in Sugar Alcohols and Other Sugars in S. lepidophylla Profiled during a Rehydration/Dehydration Cycle. Compounds in green indicate greater relative abundance the hydrated state as shown in box plots. Dotted arrows indicate a gap in the pathway. Box plots are as described in the Figure 4 legend. The p- and q-values for the comparisons are presented in Supplemental Table 1. G-6-P, glucose-6-phosphate; F-6-P, fructose-6-phosphate; Ru5P, ribulose 5-phosphate; Xu5P, xylulose 5-phosphate; E4P, erythrose 4-phosphate; T4P, threose 4-phosphate. significantly (p < 0.05) more abundant in the hydrated state nitrogen-limiting conditions typically encountered by S. stap- compared with one or more of the dry states (Supplemental fianus and other DT species in their natural habitat of rocky Table 1). Those more abundant in a dry state included outcrops with nitrogen-poor soils (Burke, 2002). Alternatively, nitrogen-rich amino acids such as glutamine, glutamate, these amino acids might provide a nitrogen reservoir useful arginine, aspartate, citrulline (Figure 6), asparagine, during the early stages of rehydration before the recovery of 3-(3-hydroxyphenyl)propionate, N-6-trimethyllysine, trans- photosynthetic activity (Martinelli et al., 2007). In addition, 4-hydroxyproline, and ophthalmate (Supplemental Table 1). these accumulated amino acids could provide precursors for In contrast, glycine, quinate, and alanine were significantly the production of protective nitrogenous compounds, such as (p < 0.05) more abundant in fully hydrated S. lepidophylla the antioxidant glutathione (see Figure 7). compared with both dry and REH-50 states (Supplemental Citrulline, a structural analog of arginine, was the only Table 1). In comparison, quinate accumulation was induced amino acid that was significantly (p < 0.05) more abundant by dehydration (at 60% RWC) in DT S. stapfianus, but not in in all dry and partially dry states compared with the hydrated DS S. pyramidalis (Oliver et al., 2011). state (Figure 6). This non-protein amino acid is thought to The accumulation of nitrogen-rich amino acids (e.g. function either as a nitrogen reserve or as a hydroxyl radi- asparagine, aspartate, arginine, glutamate, and glutamine) cal scavenger, or both, and accumulates along with arginine, partially resembles the desiccation response of the DT grass glutamate, and glutamine in response to drought stress in S. stapfianus, in which several nitrogen-rich amino acids (e.g. a drought-tolerant wild watermelon (Kawasaki et al., 2000; asparagine, arginine, glutamine) accumulated to significantly Akashi et al., 2001). Thus, citrulline might play similar roles in greater concentrations in nearly dry or dry states (Martinelli S. lepidophylla. et al., 2007; Oliver et al., 2011). This increased accumulation of The amino acid derivative 3-(3-hydroxyphenyl)propionate nitrogen-rich amino acids might reflect an adaptation to the was also more abundant in all dry and partially dry states 376 Yobi et al. • Rehydration/Dehydration Metabolomics

Figure 6. Differences in Amino Acid and Nitrogen Metabolism in S. lepidophylla Profiled during a Rehydration/Dehydration Cycle. Compounds in red indicate greater relative abundance in one or more dry states as shown in box plots. Box plots are as described in the Figure 4 legend. The p- and q-values for the comparisons are presented in Supplemental Table 1. compared with the hydrated state, and significantly so in partially dry states compared with the fully hydrated state, the DRY-1 and DRY-2 states (Supplemental Table 1). Notably, and significantly so in DRY-1 (Supplemental Table 1). This the structure of 3-(3-hydroxyphenyl)propionate suggests compound was only recently reported in plants (Oliver et al., that it might play a protective role against UV-light dam- 2011). Its accumulation pattern resembles that observed in age because closely related phenolic compounds, such as the DT grass S. stapfianus, in which ophthalmate abundance 3,4′-dihydroxypropiophenone-3-glucoside and 4-(2-amino- increased when RWC reached 50% or less; however, the func- 3-hydroxyphenyl)-4-oxobutanoic acid O-β-d-glucoside, serve tional significance of its accumulation remains unclear (Oliver as UV-B absorbing sunscreens or filters in plants (Tegelberg et al., 2011). et al., 2002) and human eye lenses (Bova et al., 1999), respec- Betaine (glycine betaine), a well-known osmoprotectant in tively. Additional research into the potential UV-protective many organisms including E. coli (Miller and Ingram, 2007) effects of this compound is warranted. and plants (Chen and Murata, 2011), failed to display signifi- N-6-trimethyllysine is a biosynthetic precursor of carnitine cant changes in abundance with changes in hydration states (Hoppel et al., 1980), a known osmoprotectant in Escherichia (Supplemental Table 1). However, this compound was more coli (Cánovas et al., 2007; Verheul et al., 1998). In a recent abundant in DT S. lepidophylla relative to DS S. moellen- sister species comparison, both carnitine and acetyl carni- dorffii (Yobi et al., 2012). tine were more abundant in DT S. lepidophylla than in DS S. moellendorffii (Yobi et al., 2012). However, in the cur- Nucleotide Metabolism rent study, carnitine was not detected and acetyl carnitine Aside from amino acids, the relative abundances of other was more abundant in the hydrated state, except in DRY-1 nitrogen-rich metabolites within the purine and pyrimidine (Supplemental Table 1). nucleotide pathways were altered during the rehydration/ Ophthalmate (glu-2-aminobutyrate-gly), a glutathione dehydration cycle in S. lepidophylla. Of the 14 nucleotides pathway metabolite and analog of glutathione that lacks and derivatives detected, the relative amounts of three antioxidant properties, was more abundant in all dry or of them (e.g. allantoin, 1-methyladenosine, and uridine Yobi et al. • Rehydration/Dehydration Metabolomics 377

5’-monophosphate (UMP)) were significantly (p < 0.05) more also similar to those observed in the DT grass S. stapfianus, in abundant in all dry or partially dry states compared with which allantoin abundance increased when RWC fell to 40% the fully hydrated state, and significantly so in DRY-1 or less (Oliver et al., 2011). (Supplemental Table 1). Inosine was also more abundant in all of the dry or partially dry states, but this difference was not Glutathione and γ-glutamyl Peptide Metabolism significant. In contrast, 2’-deoxyadenosine was significantly Within glutathione metabolism and the associated (p < 0.05) more abundant in fully hydrated S. lepidophylla γ-glutamyl amino acid biosynthetic pathway, significant compared with both dry and REH-50 states (Supplemental differences in the relative abundance of many metabolites Table 1). were apparent over the course of the rehydration/dehydra- The ureides allantoin and allantoate serve as nitrogen- tion cycle. For example, of the nine γ-glutamyl amino acid rich intermediates of purine catabolism that are involved dipeptides detected, seven were more abundant in a dry in abiotic stress protection, apparently by quenching ROS state or partially dry state relative to the hydrated state, and (Werner and Witte, 2011). Both allantoin and allantoate can four of these (e.g. γ-glutamylisoleucine, γ-glutamylleucine, attenuate cell death and ROS accumulation in an Arabidopsis γ-glutamylmethionine, and γ-glutamylthreonine) were signif- mutant defective in xanthine dehydrogenase, the key icantly (p < 0.05) more abundant (Figure 7 and Supplemental enzyme in the catabolism of purine leading to the production Table 1). In contrast, γ-glutamyltryptophan was more abun- of hypoxanthine and xanthine and then to uric acid, which dant in the hydrated state, except in the DEH-50 state is subsequently degraded to allantoin and allantoate (Supplemental Table 1). Other pathway intermediates, includ- (Brychkova et al., 2008). Similarly, RNA interference-mediated ing 5-oxoproline, were up to fivefold more abundant in the suppression of xanthine dehydrogenase in drought-stressed hydrated state compared with the dry or partially dry states Arabidopsis led to a marked reduction in plant growth and (Figure 7 and Supplemental Table 1). Glycine and oxidized significant increases in cell death and H2O2 accumulation glutathione (GSSG) were also significantly (p < 0.05) more (Watanabe et al., 2010). However, this stress-hypersensitive abundant in the hydrated state, with the exception of the phenotype can be reversed if the plants are pretreated DEH-50 state (Figure 7 and Supplemental Table 1). As men- with exogenous uric acid. These results demonstrate the tioned above, ophthalmate, a glutathione pathway metabo- importance of these ureides in nutrient recovery and lite, was more abundant in all dry or partially dry states than remobilization during stress. The greater relative abundance in the fully hydrated state (Supplemental Table 1). of allantoin in all dehydrated or partially dehydrated states In Arabidopsis, and presumably in other plants, γ-glutamyl examined here points to similar roles for this compound in amino acid cycle γ-glutamyltranspeptidase (also known DT survival and recovery in S. lepidophylla. These results are as γ-glutamyltransferase) catalyzes the interconversion of

Figure 7. Differences in γ-Glutamyl Amino Acid and Glutathione Metabolism in S. lepidophylla Profiled during a Rehydration/Dehydration Cycle. Compounds in red or green indicate greater relative abundance in one or more dry states or hydrated state, respectively, as shown in box plots. Dotted arrows indicate a gap in the pathway. Box plots are as described in the Figure 4 legend. The p- and q-values for the comparisons are presented in Supplemental Table 1. Cys-Gly, cysteinylglycine; AA, amino acid; GSH, glutathione (reduced); GSSG, glutathione disulfide (oxidized). 378 Yobi et al. • Rehydration/Dehydration Metabolomics glutathione (GSH, L–Glu–Cys–Gly) and free amino acids to consistent with the constitutive production of several other cysteinylglycine and γ-glutamyl amino acids (Ohkama-Ohtsu classes of metabolites including sugars and sugar alcohols dis- et al., 2008). In animal systems, γ-glutamyl cyclo-transferase cussed earlier. converts γ-glutamyl amino acids into 5-oxoproline, releas- ing the amino acid previously captured in the extracellular Secondary Metabolites space into the cytoplasm (Ohkama-Ohtsu et al., 2008). In There were no significant differences in the relative abun- Arabidopsis, GSH degradation occurs within the cytoplasm dance of secondary metabolites over the course of the rehy- and transmembrane recycling of amino acids is thought to dration/dehydration cycle with the exception of vanillate, be a minor event (Ohkama-Ohtsu et al., 2008). In S. lepido- which was more abundant in the hydrated states and sig- phylla, the significant increases in γ-glutamyl amino acids, as nificantly so compared with the DRY-2 state (Supplemental well as nitrogen-rich amino acids such as glutamate, during Table 1). Vanillate (and vanillic acid esters) is a well-known all stages of dehydration suggest that these peptides might and potent antioxidant that exhibits protective effects against have an important function in the acquisition of DT. For free radical-induced biomembrane damage (Tai et al., 2012). example, transmembrane amino acid recycling might occur Vanillate appears to be important in DT, as this compound as it does in animal systems, and thus provide an important was found to be significantly more abundant in both sister- mechanism for nitrogen storage during desiccation, as sug- group comparisons between DS and DT species of Sporobolus gested to occur in S. stapfianus (Oliver et al., 2011). Moreover, (Oliver et al., 2011) and Selaginella (Yobi et al., 2012). the accumulation of γ-glutamyl amino acids in response to In contrast, other phenolics (e.g. caffeate), flavonols desiccation appears to be an evolutionarily well-conserved (e.g. apigenin and naringenin), and phenylpropanoids event. For example, γ-glutamylisoleucine, γ-glutamylleucine, (e.g. coniferyl alcohol) were more abundant in all dry and γ-glutamyl-phenyalanine accumulate in S. lepidophylla or partially dry states, but these differences were not (Figure 7), as well as in T. ruralis and S. stapfianus, follow- significant (Supplemental Table 1). A flavonoid closely ing desiccation (Oliver et al., 2011). In addition, both S. lepi- related to apigenin, called saponarin, is known to be dophylla and S. stapfianus accumulate γ-glutamylglutamine induced in barley seedlings upon 6–8 d of UV-B exposure and γ-glutamylmethionine, whereas both S. lepidophylla and (Kaspar et al., 2010). Thus, apigenin is very likely to play a T. ruralis accumulate γ-glutamylvaline (Oliver et al., 2011). In protective role against irradiation with UV light. In addition, addition, S. lepidophylla accumulates γ-glutamylthreonine, apigenin likely ameliorates oxidative stress, as shown but not in a significant manner (Supplemental Table 1). Sister- in mammalian cells (Chan et al., 2012; Suh et al., 2012). group comparisons between DS and DT species of Sporobolus Naringenin, which is closely related in structure to apigenin, (Oliver et al., 2011) and Selaginella (Yobi et al., 2012) dem- is also likely to play protective roles against both exposure onstrated that DS species fail to accumulate γ-glutamyl to UV light and oxidative stress damage, as demonstrated amino acids, providing further support for an important role in rat kidney (Renugadevi and Prabu, 2009) and hepatic for these compounds in DT. Although additional research is (Prabu et al., 2011) cells. Apigenin, naringenin, and vanillate needed on these compounds, storage of γ-glutamyl amino showed significantly (p < 0.05) greater abundance in DT acids in the dried state could provide a readily mobilized form S. lepidophylla compared with DS S. moellendorffii at two of nitrogen for protein synthesis following rehydration. different hydration states (Yobi et al., 2012), reinforcing the GSH exists in either an oxidized or a reduced form. The importance of these compounds in DT. oxidized form of glutathione (GSSG) showed significantly (p < 0.05) greater abundance in S. lepidophylla in the Lipids, Phospholipids, and Fatty Acids hydrated or partially hydrated states (Figure 7). Upon becom- Of the 32 lipid metabolism compounds surveyed, only one ing oxidized, it donates a reducing equivalent to unstable (i.e. choline phosphate) showed greater relative abundance reactive oxygen intermediates such as hydrogen peroxide in response to dehydration at all stages of dehydration and or dehydroascorbate, and then reacts instantly with a simi- significantly (p < 0.05) so in both dry states, whereas 12 com- lar molecule to form glutathione disulfide (GSSG) (Mittler, pounds were significantly more abundant in the fully hydrated 2002). Therefore, the greater abundance of GSSG observed state compared with one or more re- or dehydration states in hydrated S. lepidophylla suggests active protection against (Supplemental Table 1). The zwitterionic choline phosphate oxidative stress at early stages of dehydration. In contrast, head group is known to possess a strong water-association S. stapfianus did not show significant accumulation of GSSG capacity and appears to promote water-lipid association during DT until RWC reached <40% (Oliver et al., 2011). between the onium head group and O-C=O groups in fatty In terms of other antioxidant systems, S. lepidophylla failed acid chains (Foglia et al., 2010). Water molecules bound at this to display the desiccation-induced accumulation of alpha- interphase might promote interaction with biomolecules such and beta-tocopherols as was observed in S. stapfianus (Oliver as sugars, amino acids, and membrane proteins (Disalvo et al., et al., 2011). This observation suggests that ROS scavenging 2008). Membrane interaction with simple sugars or oligosac- occurs in a constitutive manner in S. lepidophylla, which is charides has been postulated to stabilize or protect the native Yobi et al. • Rehydration/Dehydration Metabolomics 379 structure of lipid bilayers through inhibition of the fluid-to-gel state (Supplemental Table 1). This observation stands in stark membrane phase transition at low hydration (Ohtake et al., contrast to the 14 unnamed metabolites that showed 5–120- 2006; Valluru and Van den Ende, 2008). Thus, the increased fold greater abundance in S. lepidophylla from a sister-group relative abundance of choline phosphate might promote contrast between S. lepidophylla and S. moellendorffii at 50% membrane hydration as well as aid in membrane stabilization and 100% RWC hydration states (Yobi et al., 2012). Nonetheless, during the various stages of the rehydration/dehydration cycle. the seven unnamed metabolites that showed >2.0-fold Several lipoxygenase (LOX) activity markers (e.g. 13-hydroxy greater abundance under dry or partially dry conditions octadecadienoic acid (13-HODE) and 2-hydroxypalmitate) in S. lepidophylla represent important targets for future showed greater relative abundance under a majority of dry research, as these compounds might play important roles in or partially dry states, but these differences were not signifi- the acquisition of DT and could also serve as useful targets for cant (Supplemental Table 1). These lipid peroxidation prod- metabolic engineering to improve drought tolerance in crops. ucts might arise during dehydration from either ROS attack Lastly, 11 unnamed compounds showed significantly increased or the action of LOXs (Marin et al., 1998). In addition, com- abundance in the REH50 state (Figure 8), which suggests that pounds such as 13-HODE, 2-hydroxypalmitate, and 2-hydroxy- these compounds might play roles during the recovery from stearate exhibited significantly (p < 0.05) greater abundance desiccation. Because S. lepidophylla is positioned evolutionarily in DT S. lepidophylla compared with DS S. moellendorffii at between bryophytes, which rely mostly upon repair-based two different hydration states, suggesting these compounds mechanisms during rehydration, and angiosperms, which rely participate in DT mechanisms (Yobi et al., 2012). Several unsat- mainly upon dehydration-induced protection (Oliver et al., urated fatty acids (e.g. arachidate, dihomo-linoleate, dihomo- 2000a), S. lepidophylla provides a useful model to bridge linolenate, linoleate, and linolenate) were significantly (p < our understanding of the evolutionary progression of DT 0.05) more abundant in the hydrated state compared with the mechanisms among distantly related resurrection species. DRY-1 or DRY-2 states (Supplemental Table 1). A decrease in the relative amounts of various unsaturated lipids upon dehy- Conclusions dration has been documented in the DT angiosperms S. stap- This temporal evaluation of a five-stage, rehydration/dehy- fianus (Quartacci et al., 1997) and Ramonda serbica (Quartacci dration cycle in S. lepidophylla afforded a global, non-biased, et al., 2002). These alterations in unsaturated fatty acid concen- high-resolution account of the metabolite abundance changes trations might contribute to maintaining or increasing mem- that occur as part of the ability of this spike moss to ‘resurrect’ brane fluidity to allow for recovery following dehydration from an air-dried state. Like DT mosses, S. lepidophylla exhibits (Upchurch, 2008). Thus, the observed trend towards greater an obvious predisposition for DT by expressing a majority of unsaturated fatty acid abundance in the hydrated state and metabolites, such as highly abundant trehalose, sucrose, and during dehydration suggests a predisposition towards the glucose in a constitutive manner, whereas only 16.7% and maintenance of membrane fluidity in S. lepidophylla. 29.5% of all compounds show significant increases or decreases in abundance upon change in hydration state, respectively. Unnamed Metabolites Many glycolysis/gluconeogenesis and TCA cycle intermediates Of the 84 unnamed compounds identified in S. lepidophylla, and many different sugar alcohols and sugar acids were more 15 (17.8%) and 17 (20.2%) were significantly (p < 0.05) more abundant in the hydrated state. Polyols appear to play vari- abundant in one or more dry or partially dry states than the ous roles, including the slowing of water loss during drying or fully hydrated state, respectively. Unnamed compounds sig- water absorption during rehydration. Polyols might also serve nificantly more abundant in dry or partially dry states showed as osmoprotectants in stabilizing or preventing denaturation 2.1–7.3-fold greater abundance, respectively. The unnamed of proteins during dehydration, in hydroxyl radical scaveng- compounds significantly more abundant in the fully hydrated ing, in redox control, in reducing ROS production, and in trig- state displayed 2.0–4.0-fold greater abundance compared gering stress-adaptive gene expression. Vanillate was more to dry or partially dry states, respectively. In addition, nine abundant in the hydrated states and is a well-known and and five unnamed compounds were always more abundant potent antioxidant that is likely to protect against free radi- in all dehydrated states and in the hydrated state, respec- cal-induced biomembrane damage. Protection against photo- tively. However, these hydration-state-specific (dry/par- and oxidative damage during dehydration is also known to tially dry or fully hydrated) differences were not significant be critical for DT. Citrulline, a non-protein amino acid analog (Supplemental Table 1). of arginine, and allantoin, a nitrogen-rich product of purine The relatively small magnitude of significant differences catabolism, which likely serve as nitrogen reserves or attenu- among the five different hydration states in S. lepidophylla ate ROS accumulation and cell death, were more abundant suggests that metabolites that might play key roles in DT in all dry and partially dry states. Several amino acid-derived are mainly expressed constitutively. This idea is reinforced by secondary metabolites also appeared to play protective roles the observation that no unnamed compound showed more against UV-light damage including 3-(3-hydroxyphenyl) pro- than a 1.7-fold increase in relative abundance in the DEH-50 pionate and the flavonols apigenin and naringenin in all dry 380 Yobi et al. • Rehydration/Dehydration Metabolomics and partially dry states. In addition, several nitrogen-rich and Hirt’s Gardens (Medina, OH) and kept dry until the plants γ-glutamyl amino acids that were markedly more abundant in were subjected to an initial rehydration/dehydration cycle in dry or partially dry S. lepidophylla, likely play critical roles in order to identify and remove nonviable tissues. The rate of the acquisition of DT through the scavenging of ROS via glu- RWC change during the rehydration/dehydration cycle was tathione metabolism, or the remobilization of amino acids fol- established by submerging nine approximately 1-year-old lowing rehydration, or both. Lastly, S. lepidophylla contained plants in distilled water in a growth chamber under constant seven unnamed compounds that displayed twofold or greater (175 μmol m–2 s–1) cool-white fluorescent and incandescent abundance in dry or rehydrating states, suggesting that these light at 26°C and 37% relative humidity (RH). For hydration, compounds might play adaptive roles in the acquisition of DT the plants were removed from the water, excess water was or in the recovery from the desiccated state. removed by gentle blotting, and the plants were weighed at regular intervals over a 24-h period. For dehydration, the plants were then placed in dry trays and weighed at regular METHODS intervals over a 24-h period. The plants were incubated at 65°C for 2 d, then reweighed to obtain dry weights. The RWC Plant Material and Water-Deficit Stress Treatments was calculated using the following formula: RWC (%) = (Fwt- Selaginella lepidophylla (Hooker and Greville) Spring (flower Dwt)/(FTwt-Dwt) × 100, where Fwt is the fresh weight at spe- of stone) plants were purchased in the dried state from cific time points during the rehydration/dehydration cycle,

Figure 8. Differences in Nine Unnamed Compounds in S. lepidophylla Profiled during a Rehydration/Dehydration Cycle. Compounds with greater relative abundance in one or more dry states as shown in box plots are presented. Box plots are as described in the Figure 4 legend. The p- and q-values for the comparisons are presented in Supplemental Table 1. Yobi et al. • Rehydration/Dehydration Metabolomics 381

Dwt is the weight after incubation at 65°C for 2 d, and FTwt in the Kyoto Encyclopedia of Genes and Genomes (KEGG) is the weight after 24 h of rehydration. Samples represent- (www.genome.jp/kegg/) and Plant Metabolic Network (PMN) ing five RWCs—the initial dry state (DRY-1), 50% rehydrated (www.plantcyc.org/). (REH-50), 100% rehydrated (HYD), 50% dehydrated (DEH-50), and completely dehydrated (DRY-2)—were then selected to Data Imputation and Statistical Analysis determine differences in metabolite composition over the course of the rehydration/dehydration cycle. The samples were analyzed over the course of 2 d. This required data correction for minor variations resulting from instrument inter-day tuning differences (Evans et al., 2009). If Metabolomic Profiling values for a given metabolite were missing, then these were Six biological replicates from each species were collected at assigned the observed minimum detection value, based on each time point, lyophilized, and kept at –80°C under her- the assumption that the missing values were below the limits metic conditions prior to extraction. Then, 20 mg of each leaf of detection. For the convenience of data visualization, the sample was extracted in 400 μl of methanol containing recov- raw area counts for each biochemical were rescaled by divid- ery standards using an automated MicroLab STAR® system ing each sample value by the median value for the specific (Hamilton Company, Salt Lake City, UT). biochemical. Global unbiased metabolic profiling in S. lepidophylla was Statistical analysis of the data was performed using JMP performed using an integrated platform consisting of a com- (SAS, www.jmp.com), a commercial software package, and bination of three independent approaches: (1) ultrahigh per- ‘R’ (http://cran.r-project.org/), a freely available open-source formance liquid chromatography/tandem mass spectrometry software package. A log transformation was applied to the (UHLC/MS/MS2) optimized for basic species, (2) UHLC/MS/MS2 observed relative abundances for each biochemical because optimized for acidic species, and (3) gas chromatography/mass the variance generally increased as a function of each bio- spectrometry (GC/MS) as described in detail in Evans et al. chemical’s average response. (2009). Out of the 251 metabolites that were detected, a total UHLC/MS/MS2 analyses were performed using a Waters of 204 metabolites with no missing values were imported Acquity UPLC (Waters Corporation, Milford, MA) and a into the chemometrics software Solo (Eigenvector Research, Thermo-Finnigan LTQ mass spectrometer (Thermo Fisher Inc., Wenatchee, WA) for Partial Least Squares-Discriminant Scientific, Inc., Waltham, MA) equipped with an electrospray Analysis (PLS-DA) to determine classification of the differ- ionization (ESI) source and linear ion-trap (LIT) mass analyzer. ent treatments (Figure 3). PLS-DA is a regression extension Each sample was subjected to two independent UHPLC/ of Principal Component Analysis (PCA) used to model group MS runs using separate dedicated columns optimized for classification. R2 and Q2 are used as measures for the robust- either positive ions or for negative ions. For GC/MS, samples ness of a PLS-DA model, where R2 is the cumulative variance were re-dried under vacuum desiccation for a minimum of explained by the components. Cross-validation of R2 esti- 24 h, derivatized under dried nitrogen using bistrimethyl- mates Q2, which is the cumulative variance predicted by the silyl-triflouroacetamide (BSTFA), and then analyzed on a model. Thus, both of these values indicate how well the over- Thermo-Finnigan Trace DSQ fast-scanning single-quadrupole all model classifies and predicts group membership in a data MS using electron impact ionization operated at unit mass set. Both of these values range from 0 to 1 and the higher the resolving power. Chromatographic separation, followed by number, the more robust the model. full scan mass spectra, was carried out to record retention In order to visualize the entire data set, a heat map was time, molecular weight (m/z), and MS/MS2 of all detectable generated to show fold change for each compound identified ions present in the samples (see Supplemental Table 2) in from the LC–MS/MS2 or GC–MS analyses of the tissue samples compliance with recommendations for reporting metabolite (see Supplemental Table 1). Fold change was calculated for data (Fernie et al., 2011). each compound defined as the means ratio of each treatment Automated comparison of the ion features in the experi- in S. lepidophylla. Welch’s two-sample t-tests were then used mental samples was used to identify metabolites using a to determine whether or not each metabolite had significantly custom reference library of chemical standards. Reference increased or decreased in abundance. The False Discovery Rate library standard entries included the retention time, molecu- (FDR) was then calculated to correct for multiple Welch’s two- lar weight (m/z), preferred adducts, and in-source fragments sample t-test comparisons for the hundreds of compounds of compounds, as well as their associated MS/MS2 spectra. detected. Metabolite expression studies are very similar to Comparison of experimental samples to process blanks (water gene array studies with a very large number of statistical com- only) and solvent blanks allowed for the removal of artifac- parisons. As with microarray studies, metabolomic profiling tual peaks, whereas the reference library allowed the rapid generates large numbers of metabolites, and thus FDR is typi- identification of metabolites in the experimental samples cally calculated with multiple comparison adjustment instead with high confidence. Mapping of named metabolites was of family-wise error rate adjustments, such as the Bonferroni performed onto general biochemical pathways, as provided or Tukey corrections. The FDR for a given set of metabolites 382 Yobi et al. • Rehydration/Dehydration Metabolomics was estimated by the q-value (Storey, 2002). Box plots were plant Craterostigma plantagineum. Plant Physiol. 111, generated for those compounds that showed a significant 1043–1050. increase or decrease using both the Welch two-sample t-test Bova, L.M., Wood, A.M., Jamie, J.F., and Truscott, R.J.W. (1999). and FDR (i.e. p < 0.05 and q < 0.10) significance values. UV filter compounds in human lenses: the origin of 4-(2-amino- 3-hydroxyphenol)-4-oxobutanoic acid O-beta-D-glucoside. SUPPLEMENTARY DATA Invest. Opthamol. Vis. Sci. 40, 3237–3244. Brychkova, G., Alikulov, Z., Fluhr, R., and Sagi, M. (2008). A critical Supplementary Data are available at Molecular Plant Online. role for ureides in dark and senescence-induced purine remo- bilization is unmasked in the Atxdh1 Arabidopsis mutant. Plant J. 54, 496–509. FUNDING Burke, A. (2002). Properties of soil pockets on arid nama Karoo This work was supported, in part, by a grant from the USDA inselbergs: the effect of geology and derived landforms. J. Arid National Institute of Food and Agriculture (2007–55100– Environ. 50, 219–234. 18374 to M.J.O. and J.C.C.) and by Hatch funding (NAES-0341) Cánovas, M., Bernal, V., Sevilla, A., and Iborra, J. (2007). Salt stress from the Nevada Agricultural Experiment Station. effects on the central and carnitine metabolisms of Escherichia coli. Biotechnol. Bioeng. 96, 722–737. Cao, Y., Tan, N., Chen, J., Zeng, G., Ma, Y., Wu, Y., Yan, H., Yang, Acknowledgments J., Lu, L., and Wang, Q. (2010). Bioactive flavones and bifla- vones from Selaginella moellendorffii Hieron. Fitoterapia. 81, The authors would also like to thank Mary Ann Cushman 253–258. for her helpful and clarifying comments on the manuscript. Cascante, M., and Marin, S. (2008). Metabolomics and fluxomics Mention of a trademark or proprietary product does not approaches. Essays Biochem. 45, 67–81. constitute a guarantee or warranty of the product by the Chakrabortee, S., Tripathi, R., Watson, M., Schierle, G.S., United States Department of Agriculture and does not imply Kurniawan, D.P., Kaminski, C.F., Wise, M.J., and Tunnacliffe, A. its approval to the exclusion of other products that may also (2012). Intrinsically disordered proteins as molecular shields. be suitable. No conflict of interest declared. Mol. Biosyst. 8, 210–219. Chan, L.P., Chou, T.H., Ding, H.Y., Chen, P.R., Chiang, F.Y., Kuo, P.L., References and Liang, C.H. (2012). Apigenin induces apoptosis via tumor necrosis factor receptor- and Bcl-2-mediated pathway and Adams, R.P., Kendall, E., and Kartha, K.K. (1990). Comparison of enhances susceptibility of head and neck squamous cell carci- free sugars in growing and desiccated plants of Selaginella noma to 5-fluorouracil and cisplatin. Biochim. Biophys. Acta. lepidophylla. Biochem. System. Ecol. 18, 107–110. 1820, 1081–1091. Aguilar, M.I., Romero, M.G., Chávez, M.I., King-Díaz, B., and Lotina- Chan, Z., Grumet, R., and Loescher, W. (2011). Global gene expres- Hennsen, B. (2008). Biflavonoids isolated from Selaginella sion analysis of transgenic, mannitol-producing, and salt-toler- lepidophylla inhibit photosynthesis in spinach chloroplasts. J. ant Arabidopsis thaliana indicates widespread changes in abiotic Agric. Food Chem. 56, 6994–7000. and biotic stress-related genes. J. Exp. Bot. 62, 4787–4803. Ahn, S.H., Mun, Y.J., Lee, S.W., Kwak, S., Choi, M.K., Baik, S.K., Kim, Chen, T.H.H., and Murata, N. (2002). Enhancement of tolerance of Y.M., and Woo, W.H. (2006). Selaginella tamariscina induces abiotic stress by metabolic engineering of betaines and other apoptosis via a caspase-3-mediated mechanism in human pro- compatible solutes. Curr. Opin. Plant. Biol. 5, 250–257. myelocytic leukemia cells. J. Med. Food. 9, 138–144. Chen, T.H.H., and Murata, N. (2011). Glycinebetaine protects Akashi, K., Miyake, C., and Yokota, A. (2001). Citrulline, a novel plants against abiotic stress: mechanisms and biotechnological compatible solute in drought-tolerant wild watermelon leaves, applications. Plant Cell Environ. 34, 1–20. is an efficient hydroxyl radical scavenger. FEBS Lett. 508, 438–442. Deeba, F., Pandey, V., Pathre, U., and Kanojiya, S. (2009). Alpert, P. (2006). Contraints of tolerance: why are desiccation-tol- Proteome analysis of detached fronds from a resurrection plant erant organisms so small or rare? J. Exp. Biol. 209, 1575–1584. Selaginella bryopteris: response to dehydration and rehydra- Alpert, P., and Oliver, M.J. (2002). Drying without dying. In tion. J. Proteomics Bioinformatics. 2, 108–116. Desiccation and Survival in Plants: Drying without Dying, Black, Devaraneni, P.K., Mishra, N., and Bhat, R. (2012). Polyol osmolytes M., and Pritchard, H.W., eds (New York, USA: CABI Publishing), stabilize native-like cooperative intermediate state of yeast pp. 1–43. hexokinase A at low pH. Biochimie. 94, 947–952. Banks, J.A. (2009). Selaginella and 400 million years of separation. Disalvo, E.A., Lairion, F., Martini, F., Tymczyszyn, E., Frías, M., Annu. Rev. Plant Biol. 60, 223–238. Almaleck, H., and Gordillo, G.J. (2008). Structural and func- Beerling, D.J. (2002). Low atmospheric CO(2) levels during the tional properties of hydration and confined water in mem- Permo-Carboniferous glaciation inferred from fossil lycopsids. brane interfaces. Biochim. Biophys. Acta. 1778, 2655–2670. Proc. Natl Acad. Sci. U S A. 99, 12567–12571. Elbein, A.D., Pan, Y.T., Pastuszak, I., and Carroll, D. (2003). New Bernacchia, G., Salamini, F., and Bartels, D. (1996). Molecular Insights on trehalose: a multifunctional molecule. Glycobiology. characterization of the rehydration process in the resurrection 13, 17R–27R. Yobi et al. • Rehydration/Dehydration Metabolomics 383

Evans, A.M., DeHaven, C.D., Barrett, T., Mitchell, M., and Milgram, Kaushik, J.K., and Bhat, R. (1998). Thermal stability of proteins E. (2009). Integrated, nontargeted ultrahigh performance inaqueous polyol solutions: role of the surface tension of liquid chromatography/electrospray ionization tandem mass water in the stabilizing effect of polyols. J. Phys. Chem. 102, spectrometry platform for the identification and relative quan- 7058–7066. tification of the small-molecule complement of biological sys- Kawasaki, S., Miyake, C., Kohchi, T., Fujii, S., Uchida, M., and tems. Anal. Chem. 81, 6656–6667. Yokota, A. (2000). Responses of wild watermelon to drought Farrant, J. (2000). A comparison of mechanisms of desiccation stress: accumulation of an ArgE homologue and citrulline in tolerance among three angiosperm resurrection plant species. leaves during water deficits. Plant Cell Physiol. 41, 864–873. Plant Ecol. 151, 29–39. Klähn, S., and Hagemann, M. (2011). Compatible solute biosyn- Farrant, J.M., Lehner, A., Cooper, K., and Wiswedel, S. (2009). thesis in cyanobacteria. Environ. Microbiol. 13, 551–562. Desiccation tolerance in the vegetative tissues of the fern Lebkuecher, J.G, and Eickmeier, W.G. (1991). Reduced photoinhi- Mohria caffrorum is seasonally regulated. Plant J. 57, 65–79. bition with stem curling in the resurrection plant Selaginella Fernie, A.R., Aharoni, A., Willmitzer, L., Stitt, M., Tohge, T., Kopka, lepidophylla. Oecologia. 88, 597–604. J., Carroll, A.J., Saito, K., Fraser, P.D., and DeLuca, V. (2011). Lebkuecher, J.G., and Eickmeier, W.G. (1992). Photoinhibition of Recommendations for reporting metabolite data. Plant Cell. photophosphorylation, adendosine triphosphate content, and 23, 2477–2482. glyceraldehyde-3-phosphate dehydrogenase (NADP+) follow- Foglia, F., Lawrence, M.J., Lorenz, C.D., and McLain, S.E. (2010). On ing high-irradiance dessiccation of Selaginella lepidophylla. the hydration of the phosphocholine headgroup in aqueous Can. J. Bot. 70, 205–211. solution. J. Chem. Phys. 133, 145103–145110. Lebkuecher, J.G., and Eickmeier, W.G. (1993). Physiological ben- Friedman, W.E. (2011). Plant genomics: homoplasy heaven in a efits of stem curling for resurrection plants in the field. Ecology. lycophyte genome. Curr. Biol. 21, R554–R556. 74, 1073–1080. Gensel, P.G. (2008). The earliest lant plants. Ann. Rev. Ecol. Evol. Lee, S., Kim, H., Kang, J.W., Kim, J.H., Lee, D.H., Kim, M.S., Yang, Syst. 39, 459–477. Y., Woo, E.R., Kim, Y.M., Hong, J., et al. (2011). The biflavonoid Georgieva, K., Röding, A., and Büchel, C. (2009). Changes in some amentoflavone induces apoptosis via suppressing E7 expres- thylakoid membrane proteins and pigments upon desiccation sion, cell cycle arrest at sub-G1 phase, and mitochondria-ema- of the resurrection plant Haberlea rhodopensis. J. Plant Physiol. nated intrinsic pathways in human cervical cancer cells. J. Med. 166, 1520–1528. Food. 14, 808–816. Ghasempour, H.R., Gaff, D.F., Williams, R.P.W., and Gianello, R.D. Liu, H., Zhao, S., Zhang, Y., Wu, J., Peng, H., Fan, J., and Liao, J. (1998). Contents of sugars in leaves of drying desiccation toler- (2011). Reactive oxygen species-mediated mitochondrial dys- ant flowering plants, particularly grasses. Plant Growth Regul. function is involved in apoptosis in human nasopharyngeal car- 24, 185–191. cinoma CNE cells induced by Selaginella doederleinii extract. J. Goyal, K., Walton, L.J., and Tunnacliffe, A. (2005). LEA proteins Ethnopharmacol. 138, 184–191. prevent protein aggregation due to water stress. Biochem. J. Liu, J.F., Xu, K.P., Li, F.S., Shen, J., Hu, C.P., Zou, H., Yang, F., Liu, 388, 151–157. G.R., Xiang, H.L., Zhou, Y.J., et al. (2010). A new flavonoid from Ha, L.H., Thao, D.T., Huong, H.T., Minh, C.V., and Dat, N.T. (2012). Selaginella tamariscina (Beauv.) Spring. Chem. Pharm. Bull. Toxicity and anticancer effects of an extract from Selaginella (Tokyo). 58, 549–551. tamariscina on a mice model. Nat. Prod. Research. 26, 1130–1134. Liu, M.-S., Chien, C.-T., and Lin, T.-P. (2008). Constitutive compo- Hoekstra, F.A., Golvina, E.A., and Buitink, J. (2001). Mechanisms of nents and induced gene expression are involved in the desic- plant desiccation tolerance. Trends Plant Sci. 6, 431–438. cation tolerance of Selaginella tamariscina. Plant Cell Physiol. Hoppel, C.L., Cox, R.A., and Novak, R.F. (1980). N6-trimethyl-lysine 49, 653–663. metabolism. 3-hydroxy-N6-trimethyl-lysine and carnitine bio- Marin, K., Zuther, E., Kerstan, T., Kunert, A., and Hagemann, M. synthesis. Biochem. J. 188, 509–519. (1998). The ggpS gene from Synechocystis sp. strain PCC 6803 Iturriaga, G., Cushman, M.A.F., and Cushman, J.C. (2006). An EST encoding glucosyl-glycerol-phosphate synthase is involved in catalogue from the resurrection plant Selaginella lepidophylla osmolyte synthesis. J. Bacteriol. 180, 4843–4849. reveals abiotic stress-adaptive genes. Plant Biol. 170, 1173–1184. Martinelli, T., Whittaker, A., Bochicchio, A., Vazzana, C., Suzuki, Iturriaga, G., Gaff, D.F., and Zentella, R. (2000). New desiccation- A., and Masclaux-Daubresse, C. (2007). Amino acid pattern and tolerant plants, including a grass, in the central highlands of glutamate metabolism during dehydration stress in the ‘resur- Mexico, accumulate trehalose. Aust. J. Bot. 48, 153–158. rection’ plant Sporobolus stapfianus: a comparison between desiccation-sensitive and desiccation-tolerant leaves. J. Exp. Kaspar, S., Matros, A., and Mock, H.P. (2010). Proteome and flavo- Bot. 58, 3037–3046. noid analysis reveals distinct responses of epidermal tissue and whole leaves upon UV-B radiation of barley (Hordeum vulgare Miller, E.N., and Ingram, L.O. (2007). Combined effect of betaine L.) seedlings. J. Proteome Res. 9, 2402–2411. and trehalose on osmotic tolerance of Escherichia coli in min- Kataoka, Y., Kitadai, N., Hisatomi, O., and Nakashima, S. (2011). eral salts medium. Biotechnol. Lett. 29, 213–217. Nature of hydrogen bonding of water molecules in aqueous Mishra, P.K., Raghuram, G.V., Bhargava, A., Ahirwar, A., Samarth, solutions of glycerol by attenuated total reflection (ATR) infra- R., Upadhyaya, R., Jain, S.K., and Pathak, N. (2011). In vitro red spectroscopy. Appl. Spectrosc. 65, 436–441. and in vivo evaluation of the anticarcinogenic and cancer 384 Yobi et al. • Rehydration/Dehydration Metabolomics

chemopreventive potential of a flavonoid-rich fraction from a Proctor, M., and Tuba, Z. (2002). Poikilohydry and homoihydry: traditional Indian herb Selaginella bryopteris. Br. J. Nutr. 106, anithesis or spectrum of possibilities? New Phytol. 156, 327–349. 1154–1168. Quartacci, M.F., Forli, M., Rascio, N., Dalla Vecchia, F., Mittler, R. (2002). Oxidative stress, antioxidants and stress toler- Bochicchio, A., and Navari-Izzo, F. (1997). Desiccation- ance. Trends Plant Sci. 7, 405–410. tolerant Sporobolus stapfianus: lipid composition and cel- Nishizawa, A., Yabuta, Y., and Shigeoka, S. (2008). Galactinol and lular ultrastructure during dehydration and rehydration. raffinose contribute a novel function to protect plants from J. Exp. Bot. 48, 1269–1279. oxidative damage. Plant Physiol. 147, 1251–1263. Quartacci, M.F., Glisić, O., Stevanović, B., and Navari-Izzo, F. (2002). Nishizawa-Yokoi, A., Yabuta, Y., and Shigeoka, S. (2008). The Plasma membrane lipids in the resurrection plant Ramonda contribution of carbohydrates including raffinose family oli- serbica following dehydration and rehydration. J. Exp. Bot. 53, gosaccharides and sugar alcohols to protection of plants from 2159–2166. oxidative damage. Plant Signal. Behav. 3, 1016–1018. Ratnakumar, S., and Tunnacliffe, A. (2006). Intracellular trehalose Nuccio, M.L., Rhodes, D., McNeil, S.D., and Hanson, A.D. (1999). is neither necessary nor sufficient for desiccation tolerance in Metabolic engineering of plants for osmotic stress resistance. yeast. FEMS Yeast Res. 6, 903–913. Curr. Opin. Plant Biol. 2, 128–134. Renugadevi, J., and Prabu, S.M. (2009). Naringenin protects Ohkama-Ohtsu, N., Oikawa, A., Zhao, P., Xiang, C., Saito, K., and against cadmium-induced oxidative renal dysfunction in rats. Oliver, D.J. (2008). A gamma-glutamyl transpeptidase-inde- Toxicology. 256, 128–134. pendent pathway of glutathione catabolism to glutamate via Robles-Zepeda, R.E., Velázquez-Contreras, C.A., Garibay-Escobar, 5-oxoproline in Arabidopsis. Plant Physiol. 148, 1603–1613. A., Gálvez-Ruiz, J.C., and Ruiz-Bustos, E. (2011). Antimicrobial Ohtake, S., Schebor, C., and de Pablo, J.J. (2006). Effects of tre- activity of northwestern Mexican plants against Helicobacter halose on the phase behavior of DPPC-cholesterol unilamellar pylori. J. Med. Food. 14, 1280–1283. vesicles. Biochim. Biophys. Acta. 1758, 65–73. Rontein, D., Basset, G., and Hanson, A.D. (2002). Metabolic engi- Oliver, M.J., Guo, L., Alexander, D.C., Ryals, J.A., Wone, B.W.M., neering of osmoprotectant accumulation in plants. Metab. and Cushman, J.C. (2011). A sister group metabolomic contrast Eng. 4, 49–56. using untargeted global metabolomic analysis delineates the Ruiz-Bustos, E., Velazquez, C., Garibay-Escobar, A., García, Z., biochemical regulation underlying desiccation tolerance in Plascencia-Jatomea, M., Cortez-Rocha, M.O., Hernandez- Sporobolus stapfianus. Plant Cell. 23, 1231–1248. Martínez, J., and Robles-Zepeda, R.E. (2009). Antibacterial and antifungal activities of some Mexican medicinal plants. J. Med. Oliver, M.J., Tuba, Z., and Mishler, B.D. (2000a). The evolution of vege- Food. 12, 1398–1402. tative desiccation tolerance in land plants. Plant Ecol. 151, 85–100. Sanchez, D.H., Siahpoosh, M.R., Roessner, U., Udvardi, M., and Oliver, M.J., Velten, J., and Mishler, B.D. (2005). Desiccation toler- Kopka, J. (2008). Plant metabolomics reveals conserved and ance in bryophytes: a reflection of the primitive strategy for plant divergent metabolic responses to salinity. Physiol. Plant. 132, survival in dehydrating habitats? Integr. Comp. Biol. 45, 788–799. 209–219. Oliver, M.J., Velten, J., and Wood, A.J. (2000b). Bryophytes as Sawangwan, T., Goedl, C., and Nidetzky, B. (2010). Glucosylglycerol experimental models for the study of environmental stress and glucosylglycerate as enzyme stabilizers. Biotechnol. J. 5, tolerance: Tortula ruralis and desiccation-tolerance in mosses. 187–191. Plant Ecol. 151, 73–84. Schauer, N., and Fernie, A.R. (2006). Plant metabolomics: towards Pandey, V., Ranjan, S., Deeba, F., Pandey, A.K., Singh, R., and biological function and mechanism. Trends Plant Sci. 11, 508–516. Shirke, P.A. (2010). Desiccation-induced physiological and bio- chemical changes in resurrection plant, Selaginella bryopteris. Shen, B., Hohmann, S., Jensen, R.G., and Bohnert, H.J. (1999). Roles of J. Plant Physiol. 167, 1351–1359. sugar alcohols in osmotic stress adaptation: replacement of glyc- erol by mannitol and sorbitol in yeast. Plant Physiol. 121, 45–52. Peters, S., Mundree, S.D., Thomson, J.A., Farrant, J.M., and Keller, F. (2007). Protection mechanisms in the resurrection plant Shimizu, T., Kanamori, Y., Furuki, T., Kikawada, T., Okuda, T., Xerophyta viscosa (Baker): both sucrose and raffinose family Takahashi, T., Mihara, H., and Sakurai, M. (2010). Desiccation- oligosaccharides (RFOs) accumulate in leaves in response to induced structuralization and glass formation of group 3 late water deficit. J. Exp. Bot. 58, 1947–1956. embryogenesis abundant protein model peptides. Biochemistry. 49, 1093–1104. Porembski, S., and Barthlott, W. (2000). Granitic and gneissic out- crops (inselbergs) as center of diversity for desiccation-tolerant Shulaev, V., Cortez, D., Miller, G., and Mittler, R. (2008). Metabolomics vascular plants. Plant Ecology. 151, 19–28. for plant stress response. Physiol. Plant. 132, 199–208. Prabu, S.M., Shagirtha, K., and Renugadevi, J. (2011). Naringenin Smirnoff, N. (1992). The carbohydrates of bryophytes in relation in combination with vitamins C and E potentially protects oxi- to desiccation-tolerance. J. Bryol. 17, 185–191. dative stress-mediated hepatic injury in cadmium-intoxicated Storey, J.D. (2002). A direct approach to false discovery rates. J. rats. J. Nutr. Sci. Vitaminol. (Tokyo). 57, 177–185. Royal Statist. Soc. Ser. B. 64, 479–498. Proctor, M., and Pence, V. (2002). Vegetative tissues: bryophytes, Suh, K.S., Oh, S., Woo, J.T., Kim, S.W., Kim, J.W., Kim, Y.S., and vascular resurrection plants and vegetative propagules. In Chon, S. (2012). Apigenin attenuates 2-deoxy-D-ribose-induced Desiccation and Survival in Plants: Drying without Dying, Black, oxidative cell damage in HIT-T15 pancreatic β-cells. Biol. Pharm. M., and Pritchard, H., eds (Oxford: CABI Publishing). Bull. 35, 121–126. Yobi et al. • Rehydration/Dehydration Metabolomics 385

Swamy, R.C., Kunert, O., Schühly, W., Bucar, F., Ferreira, D., Rani, Watanabe, S., Nakagawa, A., Izumi, S., Shimada, H., and V.S., Kumar, B.R., and Appa Rao, A.V. (2006). Structurally unique Sakamoto, A. (2010). RNA interference-mediated suppres- biflavonoids from Selaginella chrysocaulos and Selaginella bry- sion of xanthine dehydrogenase reveals the role of purine opteris. Chem Biodivers. 3, 405–413. metabolism in drought tolerance in Arabidopsis. FEBS Lett. Tai, A., Sawano, T., and Ito, H. (2012). Antioxidative properties 584, 1181–1186. of vanillic acid esters in multiple antioxidant assays. Biosci. Weng, L., Chen, C., Zuo, J., and Li, W. (2011). Molecular dynamics Biotechnol. Biochem. 76, 314–318. study of effects of temperature and concentration on hydro- Tan, W.J., Xu, J.C., Li, L., and Chen, K.L. (2009). Bioactive com- gen-bond abilities of ethylene glycol and glycerol: implications pounds of inhibiting xanthine oxidase from Selaginella labor- for cryopreservation. J. Phys. Chem. A. 115, 4729–4737. dei. Nat. Prod. Res. 23, 393–398. Werner, A.K., and Witte, C.P. (2011). The biochemistry of nitro- Tegelberg, R., Aphalo, P.J., and Julkunen-Tiitto, R. (2002). Effects gen mobilization: purine ring catabolism. Trends. Plant. Sci. 16, of long-term, elevated ultraviolet-B radiation on phytochemi- 381–387. cals in the bark of silver birch (Betula pendula). Tree Physiol. White, E., and Towers, G.H.N. (1967). Comparative biochemistry of 22, 1257–1263. the lycopods. Phytochemistry. 6, 663–667. Tiwari, A., and Bhat, R. (2006). Stabilization of yeast hexokinase Whittaker, A., Bochicchio, A., Vazzana, C., Lindsey, G., and Farrant, A by polyol osmolytes: correlation with the physicochemical J. (2001). Changes in leaf hexokinase activity and metabolite properties of aqueous solutions. Biophys. Chem. 124, 90–99. levels in response to drying in the desiccation-tolerant species Toldi, O., Tuba, Z., and Scott, P. (2009). Vegetative desiccation tol- Sporobolus stapfianus and Xerophyta viscosa. J. Exp. Bot. 52, erance: is it a goldmine for bioengineering crops? Plant Sci. 961–969. 176, 187–199. Wolkers, W.F., McCready, S., Brandt, W.F., Lindsey, G.G., and Tuba, Z., Proctor, M.C.F., and Csintalan, Z. (1998). Ecological Hoekstra, F.A. (2001). Isolation and characterization of a D-7 responses of homoiochlorophyllous and poikilochlorophyllous LEA protein from pollen that stabilizes glasses in vitro. Biochim. desiccation tolerance plants: a comparison and an ecological Biophys. Acta. 1544, 196–206. perspective. Plant Growth Regulation. 24, 211–217. Wood, A.J. (2007). New frontiers in bryology and lichenology: the Upchurch, R.G. (2008). Fatty acid unsaturation, mobilization, and nature and distribution of vegetative desiccation-tolerance in regulation in the response of plants to stress. Biotechnol. Lett. hornworts, liverworts and mosses. The Bryologist. 110, 163–177. 30, 967–977. Wu, B., and Wang, J. (2011). Phenolic compounds from Selaginella Urano, K., Kurihara, Y., Seki, M., and Shinozaki, K. (2010). ‘Omics’ moellendorfii. Chem. Biodivers. 8, 1735–1747. analyses of regulatory networks in plant biotic stress responses. Xie, G., and Timasheff, S.N. (1997). Mechanism of the stabilization Curr. Opin. Plant. Biol. 13, 132–138. of ribonuclease A by sorbitol: preferential hydration is greater Valluru, R., and Van den Ende, W. (2008). Plant fructans in stress for the denatured than for the native protein. Protein Sci. 6, environments: emerging concepts and future prospects. J. Exp. 211–221. Bot. 59, 2905–2916. Yobi, A., Wone, B.W.M., Xu, W., Alexander, D.C., Guo, L., Ryals, Vázquez-Ortíz, F.A., and Valenzuela-Soto, E.M. (2004). HPLC deter- J.A., Oliver, M.J., and Cushman, J.C. (2012). Comparative mination of trehalose in Selaginella lepidophylla plants. J. metabolic profiling between desiccation-sensitive and desic- Liquid Chromat. Related Technol. 27, 1937–1946. cation-tolerant species of Selaginella reveals insights into the Verheul, A., Wouters, J.A., Rombouts, F.M., and Abee, T. (1998). resurrection trait. Plant J. 72, 983–999. A possible role of ProP, ProU and CaiT in osmoprotection of Zhang, Y.X., Li, Q.Y., Yan, L.L., and Shi, Y. (2011). Structural char- Escherichia coli by carnitine. J. Appl. Microbiol. 85, 1036–1046. acterization and identification of biflavones in Selaginella Wang, H.S., Sun, L., Wang, Y.H., Shi, Y.N., Tang, G.H., Zhao, F.W., tamariscina by liquid chromatography-diode-array detection/ Niu, H.M., Long, C.L., and Li, L. (2011). Carboxymethyl flavo- electrospray ionization tandem mass spectrometry. Rapid noids and a monoterpene glucoside from Selaginella moellen- Commun. Mass Spectrom. 25, 2173–2186. dorffii. Arch. Pharm. Res. 34, 1283–1288. Zheng, J.X., Zheng, Y., Zhi, H., Dai, Y., Wang, N.L., Fang, Y.X., Wang, X., Chen, S., Zhang, H., Shi, L., Cao, F., Guo, L., Xie, Y., Wang, Du, Z.Y., Zhang, K., Li, M.M., Wu, L.Y., et al.. (2011a). New T., Yan, X., and Dai, S. (2010). Desiccation tolerance mecha- 3’,8’’-linked biflavonoids from Selaginella uncinata displaying nism in resurrection fern-ally Selaginella tamariscina revealed protective effect against anoxia. Molecules. 16, 6206–6214. by physiological and proteomic analysis. J. Proteomics. Res. 9, Zheng, X.K., Zhang, L., Wang, W.W., Wu, Y.Y., Zhang, Q.B., 6561–6577. and Feng, W.S. (2011b). Anti-diabetic activity and potential Wang, Y., Long, C., Yang, F., Wang, X., Sun, Q., Wang, H., Shi, mechanism of total flavonoids of Selaginella tamariscina Y., and Tang, G. (2009). Pyrrolidinoindoline alkaloids from (Beauv.) Spring in rats induced by high fat diet and low dose Selaginella moellendorfii. J. Nat. Prod. 72, 1151–1154. STZ. J. Ethnopharmacol. 137, 662–668.