Transcriptomic and metabolomic analysis of Yukon Thellungiella plants grown in cabinets and their natural habitat show phenotypic plasticity
© Guevara et al.; licensee BioMed Central Ltd. 2012
Received: 4 June 2012
Accepted: 20 September 2012
Published: 1 October 2012
Thellungiella salsuginea is an important model plant due to its natural tolerance to abiotic stresses including salt, cold, and water deficits. Microarray and metabolite profiling have shown that Thellungiella undergoes stress-responsive changes in transcript and organic solute abundance when grown under controlled environmental conditions. However, few reports assess the capacity of plants to display stress-responsive traits in natural habitats where concurrent stresses are the norm.
To determine whether stress-responsive changes observed in cabinet-grown plants are recapitulated in the field, we analyzed leaf transcript and metabolic profiles of Thellungiella growing in its native Yukon habitat during two years of contrasting meteorological conditions. We found 673 genes showing differential expression between field and unstressed, chamber-grown plants. There were comparatively few overlaps between genes expressed under field and cabinet treatment-specific conditions. Only 20 of 99 drought-responsive genes were expressed both in the field during a year of low precipitation and in plants subjected to drought treatments in cabinets. There was also a general pattern of lower abundance among metabolites found in field plants relative to control or stress-treated plants in growth cabinets. Nutrient availability may explain some of the observed differences. For example, proline accumulated to high levels in cold and salt-stressed cabinet-grown plants but proline content was, by comparison, negligible in plants at a saline Yukon field site. We show that proline accumulated in a stress-responsive manner in Thellungiella plants salinized in growth cabinets and in salt-stressed seedlings when nitrogen was provided at 1.0 mM. In seedlings grown on 0.1 mM nitrogen medium, the proline content was low while carbohydrates increased. The relatively higher content of sugar-like compounds in field plants and seedlings on low nitrogen media suggests that Thellungiella shows metabolic plasticity in response to environmental stress and that resource availability can influence the expression of stress tolerance traits under field conditions.
Comparisons between Thellungiella plants responding to stress in cabinets and in their natural habitats showed differences but also overlap between transcript and metabolite profiles. The traits in common offer potential targets for improving crops that must respond appropriately to multiple, concurrent stresses.
Abiotic stresses such as drought, sub-optimal or extreme temperatures and high salinity decrease plant growth and productivity and are responsible for major losses in crop yield . Incentives for improving abiotic stress tolerance of crops has taken on renewed significance in the face of looming global food shortages that are anticipated consequences of climate change . Plants that grow in high stress habitats have adaptations that can guide the development of more stress tolerant crops.
Plants can exploit different life strategies to cope with their environment . Specialists may show strong local adaptation to very specific environments while generalists may acquire a more “all purpose” phenotype that allows for plant growth under a broader range of conditions. A third strategy has been described as the adaptive plastic response where a plant with a given genotype can display variable phenotypes that are environment dependent but ultimately contribute to improved fitness . Plasticity is adaptive when the phenotype induced in each environment results in greater fitness than would the alternative phenotype . One difficulty in evaluating the extent to which plasticity plays a role in stress tolerance is that stresses imposed in the laboratory fall short of replicating the complex natural environments that induce adaptive gene by environment interactions .
The application of carefully prescribed treatments under controlled laboratory conditions in growth cabinets has successfully identified stress responses and genes whose products are involved in stress tolerance (reviewed in ). However, to date few benefits predicted from tolerance traits identified using plants subjected to stress in laboratories have translated directly to improved crop performance under actual field conditions . This gap reflects the difficulty of faithfully reproducing the complexities of a natural environment in a growth cabinet . For example, the majority of QTLs for flowering time identified in growth cabinets are not the same as the ones identified in the field under natural conditions [8, 9].
The importance of phenotyping complex adaptive traits in an ecologically relevant context is highlighted by research on the genetic regulation of the transition to flowering in plants. These “flowering time” genes have largely been discovered using Arabidopsis growing in climate-controlled environments and are responsive to cues such as variation in photoperiod or light quality, ambient temperature, plant growth, and vernalization (reviewed in ). QTL mapping for date of bolting in chamber conditions as well as at field sites in Rhode Island and North Carolina showed that only a single QTL was common to all environments, several QTL affected flowering time only in controlled environments and a number of novel QTL determined bolting date in the field . More recently, genome-wide association analyses with multiple Arabidopsis ecotypes concluded similarly that most genes associated with flowering time under climate-controlled conditions were not significant for plants in a field plot in the north of France . This outcome underscores the value of validating the role of stress-responsive genes and their products through comparisons involving field and cabinet conditions. With respect to gene expression, studies that have used microarray profiling to compare transcript profiles in plants under controlled conditions to those in field plants show that there are similarities but also major differences among the products found between these sources, underscoring the challenge of selecting suitable genes for crop improvement [12, 13].
The crucifer Thellungiella salsuginea (also described as Thellungiella halophila and Eutrema salsugineum), herein referred to as Thellungiella, is found on highly saline soil in the Yukon Territory of Canada, a semi-arid, subarctic region. Thellungiella exhibits a significant innate capacity to tolerate water deficits, freezing temperatures, and high salinity [14, 15]. Research on Thellungiella has benefited from its close phylogenetic relationship to Arabidopsis thaliana and it shares many properties of this genetic model plant including small stature, prolific seed habit, capacity for genetic transformation, a relatively short life cycle of six to eight weeks and a compact, fully sequenced, diploid genome (http://www.phytozome.net/thellungiella.php, ). The combination of extreme native environment, high stress tolerance and genetic attributes makes Thellungiella an excellent genetic and physiological model for comparative studies involving field and cabinet grown plants.
Changes in gene expression and metabolite content have been reported in Thellungiella subjected to specific stresses under controlled environment conditions. Using a genotype of Thellungiella originating from China (Shandong Province), salt stress treatments in growth cabinets led to the up-regulation of only six genes compared to 40 genes in Arabidopsis plants . Several salt-responsive genes were notable due to their higher constitutive expression in unstressed Thellungiella compared to Arabidopsis. This observation suggests that a number of genes associated with salt exposure are constitutively expressed in Thellungiella, resulting in a plant with heightened natural resilience toward salt stress. Gene expression patterns have also been reported for the Yukon Thellungiella ecotype subjected to cold, drought or salinity treatments in controlled environment cabinets . Exposure to drought stress resulted in the differential expression of 101 genes, compared to cold stress (76 genes) or salt stress (22 genes). Interestingly, only three genes showed increased expression with exposure to each of the three stresses studied leading to the conclusion that Thellungiella has rather specific responses to these stresses . Metabolite profiling of Shandong Thellungiella responding to osmotic stresses including desiccation and salinity [20, 21] offers a means to link the expression of stress-responsive genes to metabolic phenotypes. For example, Shandong Thellungiella undergoing salinity stress under controlled environment conditions was shown to accumulate higher levels of proline and sugar alcohols than unsalinized controls and this outcome was associated with the enhanced expression of transcripts encoding enzymes involved in their synthesis .
In this study we focused upon Thellungiella growing under native Yukon field conditions where multiple, concurrent stresses are the norm. We compared gene expression and metabolite profiles of field plants with those subjected to prescribed stress treatments in cabinets . In addition to differences in morphology between plants found in the field and those grown in cabinets, we discovered that gene expression and metabolites associated with stress tolerance in growth cabinets are only partially recapitulated under field conditions in the Canadian Yukon. In particular, while Yukon Thellungiella plants salt-stressed in growth cabinets produced and accumulated proline in a stress-responsive manner, the proline content in leaves of field plants found growing on highly saline soil was, by comparison, low. Using seedlings grown on defined media we show that nitrogen availability can influence the array of organic solutes accumulated by Yukon Thellungiella under osmotic stress. The natural plasticity displayed by Thellungiella from morphology to chemical composition makes this plant a valuable model to test the functional significance of adaptive traits under stress including that associated with proline accumulation.
Results and discussion
In this study we used the crucifer Thellungiella salsuginea, which is native to the semi-arid, subarctic region of Canada, to study gene products expressed under natural conditions and compared them with those identified in plants subjected to regimes of simulated stress in growth cabinets. Given the extreme environmental conditions present at our Yukon field site compared to tightly controlled conditions in our growth cabinets, we fully expected to find differences that were location-specific. However, we also anticipated that stress-responsive gene products essential for survival under salt, cold, and/or drought conditions should be expressed in plants regardless of whether the plants were grown and stressed in the field or in cabinets.
Yukon field site conditions
The season for plant growth in the Yukon Territory of Canada is typically short and for Thellungiella this extends from early May to late July (Bruce Bennett, personal communication). In 2002 we developed sampling and shipping protocols suitable for RNA and metabolite extractions after collecting specimens from remote field locations. We then collected plants at a field site located approximately 40 km NW of Whitehorse returning to the same site over 1- to 2-week periods in early July 2003 and late June 2005.
At the field site Thellungiella plants grow on soil encrusted with salt deposits. Soil at the site is composed of an upper, mineral-organic layer approximately 18 cm deep that overlays a clay horizon. Chemical analysis of the upper layer (Additional file 1: Table S1) shows the soil to be highly alkaline with a pH of 8.3 and saline (the average soil electrical conductivity for the upper 20 cm is 15.7 dS m-1 or roughly equivalent to 160 mM NaCl). This soil is high in magnesium, sodium and sulfates, a feature shared by other saline soils in the area . With respect to other nutrients, the soil has adequate levels of potassium and phosphorus (236 and 26 mg kg-1, respectively), whereas the total nitrogen content is low (0.26%; ). The Ca/Mg ratio is less than 1.0, characteristic for a serpentine soil while a typical soil usually has a Ca/Mg ratio exceeding 1.0 (reviewed in ).
For the Whitehorse, Yukon area, the average temperature for the growing season in 2003 and 2005 was approximately 20°C (Additional file 1: Table S1). In 2003, average high and low temperatures for the ten-day period prior to harvest were 22°C and 15°C, respectively. At the field site on the day of harvest, soil temperature was 17°C, the air temperature was 24°C, and the light intensity was 1500 μmoles m–2 s–1. In 2005, the average high and low temperatures for the ten-day period before harvest were 18°C and 7°C, respectively. At the field where plants were collected in 2005, the soil temperature was 18°C, the daytime air temperature was 24°C, and the light intensity was 1433 μmoles m–2 s–1.
The average cumulative precipitation for this region is 86.9 mm from May through July, the growing season for Thellungiella, classifying this area as semi-arid (Environment Canada climate normals 1971 – 2000; http://www.climate.weatheroffice.gc.ca/climate_normals/index_e.html). In 2003, however, total precipitation prior to harvest was 50% lower than the long-term norms and, as such, can be considered a drought year even for a semi-arid region (Additional file 1: Figure S1; ). In contrast, precipitation prior to the 2005 harvest was 2.4-fold greater than the long-term norms. Thus prevailing levels of light and temperature during harvest for the two years were comparable but precipitation was very low in 2003 relative to 2005.
Thellungiellaplants in the field and in growth chambers have different phenotypes
Plants collected in 2003 experienced below-average precipitation in a 60 d period before harvest. In the same timeframe, plants collected in 2005 experienced more frequent episodes of rainfall leading to above-average cumulative precipitation. The contrasting precipitation patterns allowed us to compare field plants exposed to water deficits in 2003 with those experiencing above-normal cumulative precipitation in 2005. Figure 1C shows the grass surrounding Thellungiella to be dead or dying in 2003. Thellungiella plants were small (8 to 18 cm tall) but green and either flowering or setting seed. In contrast, at the same site in 2005 the above-average precipitation patterns contributed to an abundance of green vegetation around the Thellungiella plants that were themselves almost 2- to 3-fold larger than the plants found in 2003 (Figure 1B and D).
Profiling of Thellungiellaplants grown under field conditions
To prepare transcript profiles of field plants we used a cDNA microarray containing 3628 unique sequences derived from libraries enriched in stress-responsive cDNAs as described in Wong et al. . Field expression data was quantified relative to the signal intensity of control cDNA prepared from cauline leaves of plants grown in controlled environment chambers. Of the gene products represented on the microarray, 673 (19%) were differentially expressed within cauline leaves of Thellungiella growing in the field relative to cabinet-grown controls (Additional file 2). In 2003, transcripts associated with 216 cDNAs were differentially expressed with 132 and 84 being present at higher or lower levels, respectively, relative to cabinet-grown plants. For plants sampled in 2005, transcripts associated with 548 genes were differentially expressed with 301 and 247 being present at higher or lower levels, respectively, relative to plants grown in chambers. Of the 673 genes differentially expressed in field samples, only 7% were expressed at higher levels over chamber-grown controls in both years and 6% were expressed to a lesser extent in both field years.
Patterns among gene products of field-grown Thellungiellaand cabinet-grown plants subjected to environmental stress
Metabolite profiling of Thellungiellaplants grown under field conditions
The GO annotation analysis of Figure 2 shows that transcripts associated with metabolism comprise one of the largest categories of differentially regulated genes in Thellungiella field plants. In order to gain greater insight into the biochemical activities of these plants we extracted and profiled polar metabolites by gas chromatography/mass spectrometry (GC/MS). Using this approach, it was possible to simultaneously monitor over 300 mass spectral tags (MSTs) corresponding to chemically diverse compounds.
Plant metabolites can undergo diurnal changes in abundance . With this consideration and the restricted access to the field site, it was important to identify which metabolites undergo diurnal changes in abundance and the time point that would provide the most comprehensive “snapshot” of chemical components. We sampled leaf tissue harvested at the field site at three time-points corresponding to 2, 7 and 12 h after sunrise. Measurements of MST abundance at any given time point were expressed as an average fold-change of the three time points selected for tissue analysis and the 29 MSTs found to undergo significant changes in abundance were analyzed by HCA (Additional file 1: Figure S2). The majority of the compounds that underwent changes in content showed maximum abundance in the samples obtained 7 h after sunrise. This group includes myo-inositol, raffinose, galactinol, quinic acid, fructose, sucrose, and 16 unidentified MSTs. Members of the raffinose family oligosaccharides have been implicated in desiccation and freezing tolerance of a variety of plants. While one might expect the content of protective compounds to remain stable in the face of a persistent osmotic stress, such as that found at a highly saline field site, diurnal changes have been observed for myo-inositol, fructose, and raffinose for unwatered coleus  and so diurnal changes in these components neither supports nor precludes a role for raffinose family oligosaccharides in osmotic adjustment of Thellungiella. A smaller group whose abundance was highest in early morning (2 h) samples included phosphate, citrate, glycine, and unknown 1320 while one MST, unknown 966, was most abundant in the late time-point sample (12 h after sunrise). Given the difficulty of sampling at a remote field site and the pattern of highest abundance tending to be near midday, we selected a single time-point corresponding to 7 h after sunrise for routine sampling and synchronized this sampling time for comparisons with plants in growth cabinet material. Our findings agree with those of Gibon et al.  who reported that metabolite abundance in Arabidopsis is relatively low early in the day and increases over the course of the day.
Physiological adjustments to salt exposure include plasticity in organic solutes accumulated by Yukon Thellungiellaplants
A general mechanism for tolerance to osmotic stress among plants involves the accumulation of organic solutes to serve various roles including scavenging free radicals, detoxification and osmotic adjustment . There is ample experimental evidence that the Shandong ecotype of Thellungiella accumulates proline in response to NaCl treatments, a trait observed in numerous plant species, [17, 20–22, 43]. In contrast, we found little evidence that Thellungiella plants harvested at the Yukon field site contained proline at levels exceeding those found in rosette leaves of well-watered chamber-grown plants (Figure 5). This outcome was unexpected because our soil analysis and visual observations of the field location show that the plants are growing under saline conditions (Figure 1 and Additional file 1: Table S1). Tissue sampled from the field was enriched in Na+ compared to cabinet controls and leaf ψS values from field plants were lower than non-salinized plants in growth chambers. Furthermore, our HCA of differentially expressed genes between field and chamber-grown plants (Figure 3) does not implicate gene products associated with proline synthesis under stress such as the Thellungiella ortholog of Δ 1-pyrroline-5-carboxylate synthetase 1[19, 44].
Nitrogen availability influences the array of solutes accumulated by Yukon Thellungielladuring salt exposure
GC/MS was used to identify qualitative and quantitative changes in organic solutes extracted from harvested shoots. We identified 49 MSTs as showing significantly different relative abundance between the four treatments; 12 were identified as probable sugars based upon mass spectral information but 23 could not be associated with any chemical class (Additional file 4).
On this PCA, observations obtained from individual plants exposed to the same test media group together (Figure 8A), suggesting that variance between treatments was greater than variance within treatments. Observations derived from seedlings grown in the absence of salt (LN and HN) and those from plants grown in the presence of salt (LN + S, HN + S) are discriminated along the PC1 axis, indicating that PC1 is primarily associated with variance in metabolites due to seedling salt exposure. Observations derived from the LN and LN + S are separated from HN and HN + S treatments along the PC2 axis, indicating that PC2 is associated with variance due to seedling exposure to nitrogen (Figure 8A).
The PCA data was analyzed further to identify metabolite variables that influence the position of the observations along PC1 and PC2. Projections of metabolite variables onto the two principal components yielded values near the origin for the majority of metabolites (Figure 8B). However, ten of the 49 variables were clearly separable from the rest and these included fructose, myo-inositol, sugar 2100 and raffinose. These particular variables lie below zero on the PC2 axis and contribute towards the position of the LN + S observations. Points corresponding to the variables pyroglutamic acid and serine are found on the left side of the PC1 axis and above zero on the PC2 axis where they contribute to the position of the HN observations. In contrast, proline has a strong positive loading for both PC1 and PC2 axes and so contributes to the position of the HN + S observations. These results show that a change in abundance among comparatively few MSTs strongly influenced the position and separation of observations produced following PCA of MST data from the Thellungiella seedlings used in this study. Furthermore, much of the variance in this dataset is due to seedling proline content. Additional file 4 shows that the 45-fold increase in proline abundance between HN and HN + S plants can be contrasted with the fold-increases for several sugars between LN and LN + S plants (including raffinose, myo-inositol, fructose, sugars 2100 and 1579). This observation is consistent with our hypothesis that carbohydrates replace proline when plants are salt-stressed under nitrogen limiting conditions.
Even for a plant that grows in a high stress environment, there was considerable plasticity in variance to abiotic stressors. Predictably, in our comparisons between field- and cabinet-grown Thellungiella plants we found many differences in gene expression and metabolite content. However, there were consistent patterns among stress-responsive transcripts and metabolites that strongly associate a subset of gene products with a role in the adaptive capacity of this plant to survive on saline soils and under drought conditions. The converse is also true. One trait that has a long association with stress tolerance, including salinity tolerance of Thellungiella[20, 21, 43], is that of proline accumulation. Surprisingly, we found no evidence for elevated proline content in leaves of plants that grew in salt-laden Yukon soils despite the fact that plants descended from seeds of the same population grown in cabinets accumulate this amino acid in a stress-responsive manner. In this report we show evidence that for at least this trait, the availability of nutrients influences the array of organic solutes accumulated by Thellungiella growing under saline conditions. We propose that this plasticity is not an exclusive property of proline metabolism but a trait that reflects a more fundamental flexibility required for Thellungiella’s extremophile lifestyle. As such, we hypothesize that stress-responsive traits in common between field and growth cabinet locations is not a serendipitous overlap but likely reflects their essential requirement for stress tolerance. Comparisons between field and laboratory grown plants offer a promising approach to revealing candidate gene leads for improved crop performance under adverse environmental conditions.
Thellungiellaat Yukon field sites
In 2002, 2003, and 2005 cauline leaf tissue was harvested from mature, flowering Yukon Thellungiella at a field site near Whitehorse, Yukon (location: 60° 55.928′N, 135° 10.249′ W; elevation 647 m). Leaf tissue from individual plants was transferred quickly to 2 mL Nalgene cryovials, immediately flash frozen in liquid nitrogen, and then transferred to a charged MVE XC20/3V vapour shipper (Jencons Scientific Inc, Bridgeville, PA) where samples were kept frozen at -150°C for transport. Vials were transferred to -80°C for storage pending analysis.
Controlled environment plant growth conditions and stress treatments
Seeds of the Yukon genotype that had been sourced from the field site and bulked at McMaster University were sterilized using a vapour-phase gas technique  and then mixed with 0.1% (w/v) Phytagel (Sigma) and applied by pipette onto a moistened soil mixture containing six parts Promix BX (Premier Horticulture, Rivière-du-Loup, PQ) and one part Turface (Profile Products LLC, Buffalo, NY) in individual 5 × 5 × 7 cm pots. Seeds in pots were stratified for 2 d at 4°C before transfer to growth chambers (AC 60 Enconair, Winnipeg, MB) set with a 21 h day and irradiance of 250 μmol m-2s-1 and 22°C/10°C day/night temperature regime. Plants were watered daily as needed and fertilized one time per week with 1 g L-1 20-20-20 (N-P-K) fertilizer. For salinity treatments, four-week-old plants were watered with 50 mM NaCl for 3 d and the salinity level of the irrigating solution was increased in increments of 50 mM NaCl every 3 d to final concentrations of 100 mM, 300 mM, and 500 mM NaCl. Plants were watered for 3 d at their final salinity level prior to physiological analysis and harvest. For comparative determinations, treated plants were paired with age-matched unsalinized controls.
Yukon Thellungiella plants grown in controlled environment chambers as described above produce cauline leaves at approximately 8 weeks. Thus, cauline leaves collected from 12 week old cabinet-grown plants served as the source of unstressed tissue in microarray and metabolite profiling experiments involving cauline leaves from field-grown plants. Rosette leaves of chamber-grown plants served as the source of tissue for metabolite profiles comparing healthy, cold, drought, drought/re-watered, and salinized plant samples. Tissue from individual plants (un-pooled) was used for all analyses.
For seedling nutrient stress experiments, Thellungiella or Arabidopsis (Col-0) seeds were sterilized and stratified as above on MS agar plates. Four days after germination, seedlings were transferred to solid media plates composed of Hoagland's nutrient solution  in which the amount of nitrogen (nitrate) was 0.1 or 1 mM and salt (NaCl) was 0 or 100 mM . Treatment plates containing seedlings were incubated vertically in a growth chamber with a 12 h photoperiod set at 22°C and 150-200 μmol m-2s-1 photon flux density. For metabolite analysis, shoots were harvested at the same root length stage of development across treatments. Tissue was pooled in 50 mg aliquots, flash frozen in liquid nitrogen and stored at -80°C until further use.
Soil and physiological analyses
Soil samples were collected to a depth of 18 cm at the site where Yukon Thellungiella tissue was harvested. Soil pH, electrical conductivity, and chemical content were determined and the elemental composition of soil samples, field and chamber leaf tissues were analyzed .
Leaf water (ψW) and solute (ψS) potential measurements were performed using a HR33T psychrometer fitted with a C52 chamber (Wescor Inc., Logan, UT) using a 6 mm diameter disc excised from a mature fully expanded leaf . Leaf turgor (ψp) pressure was estimated as the difference between the water and solute potential measurements .
Differential gene expression
Due to the scarcity of field-sourced material, total RNA extraction and microarray hybridization conditions were as reported by Wong et al.  except the mRNA was first amplified using the Ambion MessageAmp aRNA Amplification kit (Applied Biosystems) and the cDNA was primed with a random primer mix (Invitrogen). A single round of RNA amplification was carried out according to the manufacturer’s protocol using 1 μg of total RNA. To generate the probe for microarray hybridization, 40 μg of amplified RNA was reverse-transcribed to synthesize aminoallyl-labeled cDNA followed by coupling of the aminoallyl groups to either Cyanine 3 or 5 (Cy3/Cy5). The cDNA microarrays were hybridized with Cy3- and Cy5-labeled probe pairs prepared using RNA extracted from untreated cauline leaves of chamber grown Thellungiella and compared with similarly labeled products from cauline leaves of Thellungiella harvested at the Yukon field site.
Three biological replicates with dye swap as a technical replicate were used for leaf tissue harvested from the field in 2003 and 2005 and statistical analysis was therefore based on a total of six replicates per field year. ScanArray confocal scanning system and QuantArray data acquisition software (Perkin-Elmer) were used to obtain data, and after normalization, differentially expressed genes between growth chamber and field-grown Yukon Thellungiella were detected using a Bayesian model as previously described .
Polar metabolites for profiling by GC/MS were obtained from 200 mg of frozen Thellungiella tissue taken from an individual leaf or pooled seedling shoot tissue to which 50 μL of ribitol (2 mg mL-1) was added. Tissue was ground together with the ribitol standard to a fine powder in liquid nitrogen. Metabolite extraction using methanol and chloroform was as described by Fiehn et al.  and the aqueous phase containing polar metabolites was retained for this study. The retention times were converted to retention indices by adding a mixture of fatty acid standards . Methoxymation and derivatization were as reported by Roessner et al. .
Samples were analyzed using a Trace DSQ GC-MS system (Thermo Finnigan) operated in the positive ion electron impact (EI+) full scan mode. Samples were diluted 25-fold with hexane and 1 μL was injected using a MPS 2 autosampler (Gerstel GmbH & Co., Mϋlheim, Germany). Chromatography was performed on a 30 m long, 0.25 mm ID and 0.25 μm film thickness Restek Rtx-5MS integra column (crossbond 5% diphenyl – 95% dimethyl polysiloxane; Chromatographic Specialties Inc., Brockville, ON) fused with a 10 m guard column of the same composition. The injection temperature was 230°C, and the ion source was kept at 200°C. The carrier gas was helium (>99.999% purity; VitalAire, Hamilton, ON) delivered at a constant flow rate of 1 mL min-1. The temperature program was set at 50°C for 2.5 min then temperature was first increased at a rate of 7.5°C min-1 to 70°C followed by 5°C min-1 ramp to a final temperature of 310°C where it remained for 6 min. Mass spectra were recorded at three scans per second with a scanning range of 50 to 650 m/z.
Automated mass spectral deconvolution and identification system (AMDIS) software was used to extract peak abundance and mass spectra information for each trimethylsilyl derivative resolved in GC/MS chromatograms . This information was imported in tab-delimited format to the GC/MS Analysis Software Package (GASP) (available at: http://www.flintbox.com, ) where the retention index (RI) was calculated as described by Roessner et al.  and the relative abundance for each component was adjusted to reflect the estimated recovery of the internal standard ribitol. The term MST refers to an individual trimethylsilyl derivative identified by a characteristic RI and MS . Identical MSTs from multiple GC/MS analyses were aligned with GASP and subjected to statistical tests to identify those associated with specific samples and/or treatments. An arbitrary threshold detection limit of 0.00005 for relative response factors associated with MSTs was used in determining abundance and fold-differences among chemical components from different samples.
The chemical identity of an MST was determined by comparing its RI and MS parameters to those obtained for authentic standards analyzed using the same instrument and experimental conditions. MSTs that shared m/z ratios with authentic standards that had different RIs were classified according to predicted chemical classes such as sugars or sugar phosphates. When necessary for confirmation of MST identities, chemical standards were co-injected with samples prepared from leaves or seedlings. Leaf proline content was determined by comparing sample relative response factor values for proline to those calculated from running eight known amounts of proline on the GC-MS.
Transcripts found to be differentially expressed in the field relative to the cauline leaves of cabinet-grown plants were compared to those showing stress-related changes in abundance reported previously by Wong et al. . The Wong et al.  dataset is comprised of 147 transcripts showing statistically significant stress-responsive changes in leaves of Yukon Thellungiella plants grown in cabinets and subjected to cold, drought, salinity, and drought followed by re-watering treatments. Transcripts showing a 1.5-fold greater or lesser difference in expression in Thellungiella harvested from the field compared to cabinet-grown controls or stress treated versus control cabinet-grown plants (P < 0.01) were converted to a ratio of the control, log10 transformed, then subjected to HCA. MSTs identified in metabolite profiling whose abundance was significantly different by ANOVA (P < 0.05) between treated and control or field and control plants were expressed as a ratio of the control, and were then log10 transformed before being subjected to HCA. A minimum of five plants were used from each source of tissue.
Euclidean distance was used to calculate the distance matrix and a complete linkage method was used for hierarchical clustering of MSTs and transcripts using the program Cluster . Heat maps were constructed using the Java Treeview program (http://jtreeview.sourceforge.net, ).
For seedling analyses, GC-MS profiles were generated from three biological replicate experiments testing the four combinations of nitrogen and salt. Extracts for analysis by GC-MS were prepared from duplicate, seedling shoot samples pooled from each treatment. Handling of data with respect to peak identification was as described above. The aligned data of all three trials was subjected to ANOVA and only those MSTs that were statistically significant (P < 0.05) were related to the mean value for all four treatments. This relative value was then log 10 transformed before being subjected to PCA using a covariance matrix.
Hierarchical cluster analysis
Mass spectral tags
Principal component analysis
Standard error of the mean.
We thank Bruce Bennett, Biologist, Environment Yukon Biodiversity Program, Randy Mulder, and Sandra Little for their assistance at the Yukon field site. Field work was conducted with permits issued under the Yukon Scientists and Explorers Act. We thank Dr. Susan A. Dudley for her input on multivariate statistics. This work was supported by funding from Natural Sciences and Engineering Research Council Discovery grant (EAW), NSERC/AAFC/Performance Plants Inc. CRD, ANAC Northern Scientific Training Program grants (DRG, JD) and the Ontario Research Fund-Research Excellence program.
The microarray experiments reported by Wong et al.  investigating gene expression in plants stressed in growth cabinets is accessible at NCBI’s Gene Expression Omnibus (GEO) database , accession GSE2981. The microarray data involving field plant and cabinet control plant comparisons discussed in this publication have been deposited in the NCBI GEO database and are accessible through GEO Series accession number GSE41182 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE41182).
- Boyer JS: Plant productivity and environment. Science. 1982, 218: 443-448. 10.1126/science.218.4571.443.PubMedView ArticleGoogle Scholar
- Lobell DB, Schlenker W, Costa-Roberts J: Climate trends and global crop production since 1980. Science. 2011, 333: 616-620. 10.1126/science.1204531.PubMedView ArticleGoogle Scholar
- Des Marais DL, Juenger TE: Pleiotropy, plasticity, and evolution of plant abiotic stress tolerance. Ann NY Acad Sci. 2010, 1206: 56-79. 10.1111/j.1749-6632.2010.05703.x.PubMedView ArticleGoogle Scholar
- Nicotra AB, Atkin OK, Bonser SP, Davidson AM, Finnegan EJ, Mathesius U, Poot P, Purugganan MD, Richards CL, Valladares F, van Kleunen M: Plant phenotypic plasticity in a changing climate. Trends Plant Sci. 2010, 15: 684-692. 10.1016/j.tplants.2010.09.008.PubMedView ArticleGoogle Scholar
- Dudley SA, Schmitt J: Testing the adaptive plasticity hypothesis: Density-dependent selection on manipulated stem length in Impatiens capensis. Amer Nat. 1996, 147: 445-465. 10.1086/285860.View ArticleGoogle Scholar
- Bartels D, Sunkar R: Drought and salt tolerance in plants. Crit Rev Plant Sci. 2005, 24: 23-58. 10.1080/07352680590910410.View ArticleGoogle Scholar
- Yang S, Vanderbeld B, Wan J, Huang Y: Narrowing down the targets: Towards successful genetic engineering of drought-tolerant crops. Molec Plant. 2010, 3: 469-490. 10.1093/mp/ssq016.View ArticleGoogle Scholar
- Weinig C, Ungerer MC, Dorn LA, Kane NC, Toyonaga Y, Halldorsdottir SS, Mackay TFC, Purugganan MD, Schmitt J: Novel loci control variation in reproductive timing in Arabidopsis thaliana in natural environments. Genetics. 2002, 162: 1875-1884.PubMedPubMed CentralGoogle Scholar
- Anderson JT, Lee C-R, Mitchell-Olds T: Life-history QTLs and natural selection on flowering time in Boechera stricta, a perennial relative of Arabidopsis. Evolution. 2010, 65: 771-787.PubMedPubMed CentralView ArticleGoogle Scholar
- Amasino R: Seasonal and developmental timing of flowering. Plant J. 2010, 61: 1001-1013. 10.1111/j.1365-313X.2010.04148.x.PubMedView ArticleGoogle Scholar
- Brachi B, Faure N, Horton M, Flahauw E, Vazquez A, Nordborg M, Bergelson J, Cuguen J, Roux F: Linkage and association mapping of Arabidopsis thaliana flowering time in nature. PLoS Genet. 2010, 6: e1000940. 10.1371/journal.pgen.1000940.PubMedPubMed CentralView ArticleGoogle Scholar
- Dhanaraj AL, Alkharouf NW, Beard HS, Chouikha IB, Matthews BF, Wei H, Arora R, Rowland LJ: Major differences observed in transcript profiles of blueberry during cold acclimation under field and cold room conditions. Planta. 2007, 225: 735-751. 10.1007/s00425-006-0382-1.PubMedView ArticleGoogle Scholar
- Miyazaki S, Fredricksen M, Hollis KC, Poroyko V, Shepley D, Galbraith DW, Long SP, Bohnert HJ: Transcript expression profiles of Arabidopsis thaliana grown under controlled conditions and open-air elevated concentrations of CO2 and of O3. Field Crops Res. 2004, 90: 47-59. 10.1016/j.fcr.2004.07.010.View ArticleGoogle Scholar
- Wong CE, Li Y, Whitty BR, Diaz-Camino C, Akhter SR, Brandle JE, Golding GB, Weretilnyk EA, Moffatt BA, Griffith M: Expressed sequence tags from the Yukon ecotype of Thellungiella reveal that gene expression in response to cold, drought and salinity shows little overlap. Plant Molec Biol. 2005, 58: 561-574. 10.1007/s11103-005-6163-6.View ArticleGoogle Scholar
- Griffith M, Timonin M, Wong AC, Gray GR, Akhter SR, Saldanha M, Rogers MA, Weretilnyk EA, Moffatt B: Thellungiella: an Arabidopsis-related model plant adapted to cold temperatures. Plant Cell Environ. 2007, 30: 529-538. 10.1111/j.1365-3040.2007.01653.x.PubMedView ArticleGoogle Scholar
- Bailey CD, Koch MA, Mayer M, Mummenhoff K, O'Kane SL, Warwick SI, Windham MD, Al-Shehbaz IA: Toward a global phylogeny of the Brassicaceae. Molec Biol Evol. 2006, 23: 2142-2160. 10.1093/molbev/msl087.PubMedView ArticleGoogle Scholar
- Inan G, Zhang Q, Li PH, Wang ZL, Cao ZY, Zhang H, Zhang CQ, Quist TM, Goodwin SM, Zhu JH, Shi HH, Damsz B, Charbaji T, Gong QQ, Ma SS, Fredricksen M, Galbraith DW, Jenks MA, Rhodes D, Hasegawa PM, Bohnert HJ, Joly RJ, Bressan RA, Zhu JK: Salt cress. A halophyte and cryophyte Arabidopsis relative model system and its applicability to molecular genetic analyses of growth and development of extremophiles. Plant Physiol. 2004, 135: 1718-1737. 10.1104/pp.104.041723.PubMedPubMed CentralView ArticleGoogle Scholar
- Taji T, Seki M, Satou M, Sakurai T, Kobayashi M, Ishiyama K, Narusaka Y, Narusaka M, Zhu JK, Shinozaki K: Comparative genomics in salt tolerance between Arabidopsis and Arabidopsis-related halophyte salt cress using Arabidopsis microarray. Plant Physiol. 2004, 135: 1697-1709. 10.1104/pp.104.039909.PubMedPubMed CentralView ArticleGoogle Scholar
- Wong CE, Li Y, Labbe A, Guevara D, Nuin P, Whitty B, Diaz C, Golding GB, Gray GR, Weretilnyk EA, Griffith M, Moffatt BA: Transcriptional profiling implicates novel interactions between abiotic stress and hormonal responses in Thellungiella, a close relative of Arabidopsis. Plant Physiol. 2006, 140: 1437-1450. 10.1104/pp.105.070508.PubMedPubMed CentralView ArticleGoogle Scholar
- Arbona V, Argamasilla R, Gomez-Cadenas A: Common and divergent physiological, hormonal and metabolic responses of Arabidopsis thaliana and Thellungiella halophila to water and salt stress. J Plant Physiol. 2010, 167: 1342-1350. 10.1016/j.jplph.2010.05.012.PubMedView ArticleGoogle Scholar
- Lugan R, Niogret M, Leport L, Guegan J, Larher FR, Savoure A, Kopka J, Bouchereau A: Metabolome and water homeostasis analysis of Thellungiella salsuginea suggests that dehydration tolerance is a key response to osmotic stress in this halophyte. Plant J. 2010, 64: 215-229. 10.1111/j.1365-313X.2010.04323.x.PubMedView ArticleGoogle Scholar
- Gong QQ, Li PH, Ma SS, Rupassara SI, Bohnert HJ: Salinity stress adaptation competence in the extremophile Thellungiella halophila in comparison with its relative Arabidopsis thaliana. Plant J. 2005, 44: 826-839. 10.1111/j.1365-313X.2005.02587.x.PubMedView ArticleGoogle Scholar
- Day JH: Reconnaissance soil survey of the Takhini and Dezadeash valleys in the Yukon territory. Ottawa, ON: Canada Dept. of Agriculture. 1962.Google Scholar
- Martens DC, Lindsay WL: Testing soils for copper, iron, manganese and zinc. Soil testing and plant analysis. 3rd edition. Edited by: Westerman RL. Madison, WI: Soil Science Society of America, 1990:229-264.Google Scholar
- Brady KU, Kruckeberg AR, Bradshaw HD: Evolutionary ecology of plant adaptation to serpentine soils. Annu Rev Ecol Evol Sys. 2005, 36: 243-266. 10.1146/annurev.ecolsys.35.021103.105730.View ArticleGoogle Scholar
- Hogg EH, Wein RW: Impacts of drought on forest growth and regeneration following fire in southwestern Yukon, Canada. Can J For Res. 2005, 35: 2141-2150. 10.1139/x05-120.View ArticleGoogle Scholar
- Morgan JM: Osmoregulation and water stress in higher plants. Annu Rev Plant Physiol. 1984, 35: 299-319. 10.1146/annurev.pp.35.060184.001503.View ArticleGoogle Scholar
- Huang D, Wu W, Abrams SR, Cutler AJ: The relationship of drought-related gene expression in Arabidopsis thaliana to hormonal and environmental factors. J Exp Bot. 2008, 59: 2991-3007. 10.1093/jxb/ern155.PubMedPubMed CentralView ArticleGoogle Scholar
- Saito S, Hirai N, Matsumoto C, Ohigashi H, Ohta D, Sakata K, Mizutani M: Arabidopsis CYP707As encode (+)-abscisic acid 8 ′-hydroxylase, a key enzyme in the oxidative catabolism of abscisic acid. Plant Physiol. 2004, 134: 1439-1449. 10.1104/pp.103.037614.PubMedPubMed CentralView ArticleGoogle Scholar
- Hernandez-Blanco C, Feng DX, Hu J, Sanchez-Vallet A, Deslandes L, Llorente F, Berrocal-Lobo M, Keller H, Barlet X, Sanchez-Rodriguez C, Anderson LK, Somerville S, Marco Y, Molina A: Impairment of cellulose synthases required for Arabidopsis secondary cell wall formation enhances disease resistance. Plant Cell. 2007, 19: 890-903. 10.1105/tpc.106.048058.PubMedPubMed CentralView ArticleGoogle Scholar
- Yang Y, Sulpice R, Himmelbach A, Meinhard M, Christmann A, Grill E: Fibrillin expression is regulated by abscisic acid response regulators and is involved in abscisic acid-mediated photoprotection. Proc Natl Acad Sci USA. 2006, 103: 6061-6066. 10.1073/pnas.0501720103.PubMedPubMed CentralView ArticleGoogle Scholar
- Luhua S, Ciftci-Yilmaz S, Harper J, Cushman J, Mittler R: Enhanced tolerance to oxidative stress in transgenic Arabidopsis plants expressing proteins of unknown function. Plant Physiol. 2008, 148: 280-292. 10.1104/pp.108.124875.PubMedPubMed CentralView ArticleGoogle Scholar
- Wohlbach DJ, Quirino BF, Sussman MR: Analysis of the Arabidopsis histidine kinase ATHK1 reveals a connection between vegetative osmotic stress sensing and seed maturation. Plant Cell. 2008, 20: 1101-1117. 10.1105/tpc.107.055871.PubMedPubMed CentralView ArticleGoogle Scholar
- Cobbett C, Goldsbrough P: Phytochelatins and metallothioneins: Roles in heavy metal detoxification and homeostasis. Annu Rev Plant Biol. 2002, 53: 159-182. 10.1146/annurev.arplant.53.100301.135154.PubMedView ArticleGoogle Scholar
- Shan X, Zhang Y, Peng W, Wang Z, Xie D: Molecular mechanism for jasmonate-induction of anthocyanin accumulation in Arabidopsis. J Exp Bot. 2009, 60: 3849-3860. 10.1093/jxb/erp223.PubMedView ArticleGoogle Scholar
- Du J, Huang Y, Xi J, Cao M, Ni W, Chen X, Zhu J, Oliver DJ, Xiang C: Functional gene-mining for salt-tolerance genes with the power of Arabidopsis. Plant J. 2008, 56: 653-664. 10.1111/j.1365-313X.2008.03602.x.PubMedPubMed CentralView ArticleGoogle Scholar
- Rizhsky L, Liang H, Shuman J, Shulaev V, Davletova S, Mittler R: When defense pathways collide. The response of Arabidopsis to a combination of drought and heat stress. Plant Physiol. 2004, 134: 1683-1696. 10.1104/pp.103.033431.PubMedPubMed CentralView ArticleGoogle Scholar
- Mowla SB, Cuypers A, Driscoll SP, Kiddle G, Thomson J, Foyer CH, Theodoulou FL: Yeast complementation reveals a role for an Arabidopsis thaliana late embryogenesis abundant (LEA)-like protein in oxidative stress tolerance. Plant J. 2006, 48: 743-756. 10.1111/j.1365-313X.2006.02911.x.PubMedView ArticleGoogle Scholar
- Eklof JM, Brumer H: The XTH gene family: An update on enzyme structure, function, and phylogeny in xyloglucan remodeling. Plant Physiol. 2010, 153: 456-466. 10.1104/pp.110.156844.PubMedPubMed CentralView ArticleGoogle Scholar
- Gibon Y, Usadel B, Blaesing OE, Kamlage B, Hoehne M, Trethewey R, Stitt M: Integration of metabolite with transcript and enzyme activity profiling during diurnal cycles in Arabidopsis rosettes. Genome Biol. 2006, 7: 76. 10.1186/gb-2006-7-8-r76.View ArticleGoogle Scholar
- Pattanagul W, Madore MA: Water deficit effects on raffinose family oligosaccharide metabolism in coleus. Plant Physiol. 1999, 121: 987-993. 10.1104/pp.121.3.987.PubMedPubMed CentralView ArticleGoogle Scholar
- Hanson AD, Hitz WD: Metabolic responses of mesophytes to plant water deficits. Annu Rev Plant Physiol. 1982, 33: 163-203. 10.1146/annurev.pp.33.060182.001115.View ArticleGoogle Scholar
- Kant S, Kant P, Raveh E, Barak S: Evidence that differential gene expression between the halophyte, Thellungiella halophila, and Arabidopsis thaliana is responsible for higher levels of the compatible osmolyte proline and tight control of Na+ uptake in T. halophila. Plant Cell Environ. 2006, 29: 1220-1234. 10.1111/j.1365-3040.2006.01502.x.PubMedView ArticleGoogle Scholar
- Szekely G, Abraham E, Cseplo A, Rigo G, Zsigmond L, Csiszar J, Ayaydin F, Strizhov N, Jasik J, Schmelzer E, Koncz C, Szabados L: Duplicated P5CS genes of Arabidopsis play distinct roles in stress regulation and developmental control of proline biosynthesis. Plant J. 2008, 53: 11-28. 10.1111/j.1365-313X.2007.03318.x.PubMedView ArticleGoogle Scholar
- Colmer T, Fan T, Lauchli A, Higashi R: Interactive effects of salinity, nitrogen and sulphur on the organic solutes in Spartina alterniflora leaf blades. J Exp Bot. 1996, 47: 369-375. 10.1093/jxb/47.3.369.View ArticleGoogle Scholar
- Hanson AD, Rivoal J, Paquet L, Gage DA: Biosynthesis of 3-dimethylsulfonioproionate in Wollastonia biflora (L.) DC. Evidence that S-methylmethionine is an intermediate. Plant Physiol. 1994, 105: 103-110. 10.1104/pp.105.1.103.PubMedPubMed CentralView ArticleGoogle Scholar
- Fiehn O, Kopka J, Trethewey RN, Willmitzer L: Identification of uncommon plant metabolites based on calculation of elemental compositions using gas chromatography and quadrupole mass spectrometry. Anal Chem. 2000, 72: 3573-3580. 10.1021/ac991142i.PubMedView ArticleGoogle Scholar
- Clough SJ, Bent AF: Floral dip: a simplified method for Agrobacterium-mediated transformation of Arabidopsis thaliana. Plant J. 1998, 16: 735-743. 10.1046/j.1365-313x.1998.00343.x.PubMedView ArticleGoogle Scholar
- Hoagland DR, Arnon DI: The water-culture method forgrowing plants without soil:. Berkeley: The College of Agriculture, University of California 1950.Google Scholar
- Kant S, Bi Y, Weretilnyk E, Barak S, Rothstein SJ: The Arabidopsis halophytic relative Thellungiella halophila tolerates nitrogen-limiting conditions by maintaining growth, nitrogen uptake, and assimilation. Plant Physiol. 2008, 147: 1168-1180. 10.1104/pp.108.118125.PubMedPubMed CentralView ArticleGoogle Scholar
- Ping CL, Bockheim JG, Kimble JM, Michaelson GJ, Walker DA: Characteristics of cryogenic soils along a latitudinal transect in Arctic Alaska. J Geophys Res-Atmos. 1998, 103 (D22): 28917-28928. 10.1029/98JD02024.View ArticleGoogle Scholar
- Nonami H, Boyer JS: Turgor and growth at low water potentials. Plant Physiol. 1989, 89: 798-804. 10.1104/pp.89.3.798.PubMedPubMed CentralView ArticleGoogle Scholar
- Roessner U, Wagner C, Kopka J, Trethewey RN, Willmitzer L: Simultaneous analysis of metabolites in potato tuber by gas-chromatography-mass spectrometry. Plant J. 2000, 23: 131-142. 10.1046/j.1365-313x.2000.00774.x.PubMedView ArticleGoogle Scholar
- Stein SE: An integrated method for spectrum extraction and compound identification from gas chromatography. J Am Soc Mass Spectrom. 1999, 10: 770-781. 10.1016/S1044-0305(99)00047-1.View ArticleGoogle Scholar
- Nuin P, Weretilnyk EA, Summers PS, Guevara DR, Golding GB, GASP: GC/MS analysis software package. 2005Google Scholar
- Kopka J: Current challenges and developments in GC-MS based metabolite profiling technology. J Biotechnol. 2006, 124: 312-322. 10.1016/j.jbiotec.2005.12.012.PubMedView ArticleGoogle Scholar
- Eisen MB, Spellman PT, Brown PO, Botstein D: Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci USA. 1998, 95: 14863-14868. 10.1073/pnas.95.25.14863.PubMedPubMed CentralView ArticleGoogle Scholar
- Saldanha AJ: Java Treeview–extensible visualization of microarray data. Bioinformatics. 2004, 20: 3246-3248. 10.1093/bioinformatics/bth349.PubMedView ArticleGoogle Scholar
- Edgar R, Domrachev M, Lash AE: Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res. 2002, 30: 207-210. 10.1093/nar/30.1.207.PubMedPubMed CentralView ArticleGoogle Scholar