Rapid, reversible induction of Crassulacean acid metabolism (CAM) in conjunction with pronounced reconfiguration of carbon metabolism enables Talinum to survive cycles of severe drought.
Abstract
Drought tolerance is a key factor for agriculture in the 21st century as it is a major determinant of plant survival in natural ecosystems as well as crop productivity. Plants have evolved a range of mechanisms to cope with drought, including a specialized type of photosynthesis termed Crassulacean acid metabolism (CAM). CAM is associated with stomatal closure during the day as atmospheric CO2 is assimilated primarily during the night, thus reducing transpirational water loss. The tropical herbaceous perennial species Talinum triangulare is capable of transitioning, in a facultative, reversible manner, from C3 photosynthesis to weakly expressed CAM in response to drought stress. The transcriptional regulation of this transition has been studied. Combining mRNA-Seq with targeted metabolite measurements, we found highly elevated levels of CAM-cycle enzyme transcripts and their metabolic products in T. triangulare leaves upon water deprivation. The carbohydrate metabolism is rewired to reduce the use of reserves for growth to support the CAM-cycle and the synthesis of compatible solutes. This large-scale expression dataset of drought-induced CAM demonstrates transcriptional regulation of the C3–CAM transition. We identified candidate transcription factors to mediate this photosynthetic plasticity, which may contribute in the future to the design of more drought-tolerant crops via engineered CAM.
Drought is a major determinant of both plant survival in natural ecosystems and plant productivity in agriculture (Lobell and Gourdji, 2012). Plants have evolved a range of physiological and nonphysiological traits to cope with water deficit stress (Bartels and Sunkar, 2005). Drought adaptation includes leaf shedding in perennials or completing the life cycle while enough water is present in annuals. Few species (e.g. so-called resurrection plants) tolerate extreme dehydration (Ingram and Bartels, 1996) and resume life upon water resupply. Other plants have evolved strategies to cope with drought, such as specialized biochemical pathways, cells, tissues, and organs to survive water scarcity by mitigating the reduction in tissue water loss (Chaves et al., 2003).
Water limitation sensed in the roots or leaves triggers stress signals, which include but are not limited to abscisic acid (ABA; Bray, 1997; Zhu, 2002; Chaves et al., 2003; Bartels and Sunkar, 2005). The immediate consequence of an ABA signal is stomatal closure (Kollist et al., 2014). This can lead to short-term carbon dioxide (CO2) limitation of photosynthesis, potentially causing oxidative stress, which is mitigated by protective systems (Bartels and Sunkar, 2005; Flexas et al., 2006). The accumulation of compatible solutes is induced to protect the cellular machinery from consequences of leaf water loss and to lower the water potential of the leaf (Hare et al., 1998). These molecules include sugars, such as raffinose, trehalose, and Suc; sugar alcohols like mannitol and inositol; and amino acids and their derivatives, such as Pro or Gly betaine (Hare et al., 1998; Elbein et al., 2003). The signaling cascade has been largely elucidated. ABA is bound by the PYRABACTIN RESISTANCE1 AND PYR1-LIKE REGULATORY COMPONENTS OF ABA RECEPTOR family proteins, which upon binding inhibit the TYPE 2C AND TYPE 2A PROTEIN PHOSPHATASES (PP2C; Cutler et al., 2010). This inhibition, in turn, relieves the inhibition of the protein kinase OPEN STOMATA1 and other kinases, which phosphorylate their targets including the ABA RESPONSIVE ELEMENT BINDING FACTORS. They finally trigger ABA-responsive gene expression (Park et al., 2009). The capability to reduce leaf water loss is variable between species, and leaf water loss exceeding a threshold irreversibly damages a leaf (Lawlor and Cornic, 2002).
The tropical herbaceous dicot Talinum triangulare (Jacq). Willd. [according to The Plant List Kew now considered a synonym of the accepted name Talinum fruticosum (L.) Juss., but referred to within this work as T. triangulare to relate to extensive ecophysiological work on “T. triangulare” by Herrera et al. (1991), Taisma and Herrera (1998), Herrera (1999), Taisma and Herrera (2003), and Herrera et al. (2015)], in the family Talinaceae (formerly Portulacaceae), responds to intermittent drought (Harris and Martin, 1991; Herrera et al., 1991) in two ways. Leaves change from a horizontal to a vertical orientation and exhibit leaf-rolling (Herrera et al., 1991; Taisma and Herrera, 1998; Herrera, 1999). Leaves induce Crassulacean acid metabolism (CAM) in a reversible, facultative manner (Taisma and Herrera, 1998; Herrera, 2009; Winter and Holtum, 2014; Herrera et al., 2015).
CAM is a carbon concentrating mechanism allowing stomatal closure during the day as atmospheric CO2 is primarily assimilated during the night by PHOSPHOENOLPYRUVATE CARBOXYLASE (PEPC; Osmond, 1978). The produced organic acids (mainly malic acid) are decarboxylated during the following day to provide CO2 for the secondary, light-driven carboxylation via Rubisco. The magnitude of CAM varies between and within plant species (Borland et al., 2011). While constitutive CAM plants are ontogenetically determined to engage in this photosynthetic mode at some point in their life cycle, in facultative CAM plants, a transition to CAM occurs in response to water-deficit stress (Winter et al., 2008; Winter and Holtum, 2014). The ability to experimentally control the timing of CAM induction and CAM-to-C3 reversal makes facultative CAM plants excellent systems to identify key components of CAM.
Extensive physiological, molecular, and mutant-based studies of facultative CAM were focused on, but not restricted to, the halophyte Mesembryanthemum crystallinum, and aided to understand central concepts of the CAM-cycle (Winter and von Willert, 1972; Holtum and Winter, 1982; Cushman et al., 2008a). These included a microarray-based large scale gene expression data set (Cushman et al., 2008b), which described major changes in mRNA steady-state levels during the salinity-induced transition from C3 to CAM. Reversibility of CAM back to C3 photosynthesis after irrigation was reinitiated has been documented for several species including T. triangulare (for review, see Winter and Holtum, 2014). To date, it is not fully understood how this metabolic plasticity is transcriptionally accomplished (Yang et al., 2015). We studied the T. triangulare transition from well-watered to water-limited and back to well-watered conditions (1) to test how CAM induction and its reversion are controlled at the level of mRNA abundance and (2) to evaluate CAM- and drought-specific changes in metabolite levels.
RESULTS
Nocturnal Acidification
Following germination, T. triangulare was grown well-watered for 28 d. Subsequently, water was withheld for 12 d before rewatering. Fresh weight to dry weight ratio after 4, 9, and 12 d of water deprivation showed slight, continuous (1.1%, 4.8%, and 15.0%, respectively) decreases compared to day 0, which were significant on day 12 (Fig. 1A). Measurements of titratable acidity were used as an indicator of the presence or absence of CAM activity in leaves (Fig. 1B). In well-watered plants, acidity levels were low (10 µmol H+ g−1 fresh weight) and did not change in the course of the night. Withholding water for 4 d did not significantly alter titratable acidity compared to well-watered plants. Significant nocturnal acidification, indicative of CAM, was observed in plants from which water was withheld for 9 and 12 d, consistent with previous findings of Herrera et al. (1991) and Winter and Holtum (2014). Drought stress also resulted in pronounced leaf rolling (Fig. 1C). Following rewatering for 2 d, plants returned to a full C3 photosynthetic pattern without nocturnal increases in tissue acidity and unrolled leaves (Fig. 1, B and C). However, fresh weight to dry weight ratio was significantly lower (22.7%) compared to day 0 (Fig. 1A).
Figure 1.
Time course in response to 0 (darkgreen), 4 (light green), 9 (orange), and 12 (red) d of drought and 2 d after rewatering (re2, blue) in T. triangulare. A, Fresh weight to dry weight ratio (mean ± se of leaves harvested at the middle of the day and the middle of the night, n = 14–16). B, Levels of titratable acidity of leaves (mean ± se, n = 8–16). C, Representative pictures of plants during the course of the experiment. Asterisks indicate Student’s t test significance in comparison to day 0 at ***P < 0.001, **P < 0.01, and *P < 0.05.
Changes in Transcript and Metabolite Levels
Samples for mRNA-Seq were taken at midday and midnight prior to the drought treatment (day 0, well-watered), on day 4, day 9, and day 12 of drought, and again 2 d after rewatering (Fig. 2, A and B). Biological triplicates yielded, on average, 40.7 million reads over the time course (Supplemental Table S1). Of those, an average of 52% could be mapped for quantification to the reference genome of Arabidopsis (Arabidopsis thaliana; Supplemental Table S1; “Materials and Methods”). The mapped reads matched 16,766 Arabidopsis genes with at least one read per gene. This is comparable to the yield of cross species mapping with equidistant species for both number of genes matched and percentage of mapped reads (Gowik et al., 2011). Within-species mapping on the assembled contigs resulted in 81% mapped reads but indicated that the assembly suffers from the known limitations of transcriptome assembly such as contig fragmentation (Franssen et al., 2011; Supplemental Dataset S2).
Figure 2.
Changes in leaf transcriptomes and metabolomes under varying levels of water availability. A, Histograms of log2-fold changes in gene expression compared to day 0 (log10-scaled). Colored bars indicate significant changes (DESeq2, q < 0.01). B and C, Venn diagrams representing overlapping changes (↑: increased, ↓: depleted) in gene expression (B, DESeq2 q < 0.01, n = 3, 16,766 genes analyzed in total) or metabolite levels (C, Student’s t test, P < 0.05, n = 3–4, 39 metabolites measured in total) at the middle of the day between water-limited stages compared to day 0. See Supplemental Figure S1 for analogous Venn diagrams of changes at the middle of the night.
Statistical evaluation including multiple hypothesis testing correction (by DESeq2; for details, see “Materials and Methods”) identified significantly differential gene expression for 4,628 genes (28% of the whole transcriptome) at midday and 5,191 genes (31% of the transcriptome) at midnight on day 9 as well as 6,143 genes (37% of the transcriptome) at midday and 6,565 genes (39% of the transcriptome) at midnight on day 12 compared to well-watered plants (day 0), i.e. during the 2 d during which pronounced CAM activity was observed (Fig. 2B).
On day 9, 2,117 (12.6%) and 2,213 (13.2%) genes were upregulated, while 2,511 (15%) and 2,978 genes (17.8%) were downregulated at midday and midnight, respectively. On day 12, 2,713 (16.2%) and 2,897 (17.3%) genes were upregulated, while 3,430 (20.5%) and 3,668 genes (21.9%) were downregulated at midday and midnight, respectively. A Venn analysis indicated that 1,634 genes and 2,180 genes were shared among the genes up- and downregulated at midday, respectively (Fig. 2B), while at midnight, 1,545 upregulated and 1,865 downregulated genes were shared (Supplemental Fig. S1.). A small percentage (1.9% and 2.1% of upregulated and 1.2% and 1.9% of downregulated genes at midday and midnight, respectively) was exclusively differentially regulated on day 9. On day 4 of water deprivation, no differentially expressed genes (DEGs) were detectable compared to well-watered conditions at midday, while at midnight 16 genes were up- and 35 were downregulated (Supplemental Fig. S1). Upon rewatering, the number of significantly DEGs was reduced to 485 at midday (3% of the transcriptome) and 1,632 (10% of the transcriptome) at midnight. To compare the changes of steady-state levels of mRNA and metabolites, 39 metabolites were quantified at midday and midnight on days 0, 4, 9, 12, and after rewatering for 2 d (Fig. 2C; Supplemental Fig. S1; Supplemental Table S2). On day 4, nine metabolites were significantly different from well-watered plants at either midday or midnight. In leaves harvested at day 9 and 12, 13 and 14 of the 39 metabolites differed significantly at midday, while 6 and 11 metabolites differed significantly at midnight, respectively. The highest total number of significant changes was observed in rewatered plants, 10 at midday and 16 at midnight compared to day 0. Both the transcriptome analysis and the targeted metabolite profiling showed that transcriptome and metabolome were markedly altered in T. triangulare plants, which experience drought. While changes in mRNA abundance were largely reversed upon rewatering (Fig. 2A), the metabolic state remained altered.
CAM-Related Transcriptional Changes
Transcript abundance of known enzymes of the CAM-cycle sensu stricto (i.e. carboxylation and decarboxylation) and sensu lato (i.e. encoding auxiliary steps such as starch turnover and glycolysis for phosphoenolpyruvate [PEP] generation) were analyzed during C3-CAM-C3 transitions. Transcripts encoding four CAM-cycle enzymes sensu stricto, namely PEPC, NADP-MALIC ENZYME (NADP-ME), NAD-MALIC ENZYME (NAD-ME), and PYRUVATE, ORTHOPHOSPHATE DIKINASE (PPDK) had higher steady-state levels at both midnight and midday on day 9 and/or day 12 of water-limitation (Fig. 3; Supplemental Dataset S3).
Figure 3.
Abundances of CAM genes sensu stricto and sensu lato. Scheme of carbon assimilation via CAM and gene expression of central enzymes and transporters. Metabolites are represented in gray. Transcript levels were measured at the middle of the day and the middle of the night in leaves of T. triangulare plants under five different stages of water availability. Scaled to largest expression by gene; mean ± sd, n = 3. Asterisks indicate differential gene expression in comparison to day 0 as determined by DESeq2, * q < 0.05, ** q < 0.01, and *** q < 0.001. Blue and yellow arrows represent reactions occurring at night and day, respectively. Abbreviations are explained in the text and in Supplemental Dataset S3. Question mark and dotted arrows indicate putative activity of plastidial NADP-ME, mitochondrial NAD-ME, and cytosolic PPDK as discussed in the text. Separation into phosphorolytic and hydrolytic starch degradation is based on the models presented by Weise et al. (2011) and Streb and Zeeman (2012). 1,3-BPG, 1,3-Bisphosphoglycerate; 2-PGA, 2-phosphoglycerate; 3-PGA, 3-phosphoglycerate; DHAP, dihydroxyacetone phosphate; F1,6BP, Fru 1,6-bisphosphate; F6P, Fru 6-phosphate; G1P, Glc 1-phosphate; G6P, Glc 6-phosphate; GAP, glyceraldehyde 3-phosphate; Mito, mitochondrion; re2, 2 d after rewatering.
CARBONIC ANHYDRASE (CA) catalyzes the hydration of CO2 to HCO3− at physiological pH and thereby is thought to support providing PEPC with its substrate in CAM plants (Tsuzuki et al., 1982). The gene encoding cytosolic BETA-CARBONIC ANHYDRASE3 (BCA3) was unchanged in expression upon drought but already highly expressed in C3 conditions (2,498 rpm on day 0). This is in agreement with an earlier study finding no differences in CA activity in M. crystallinum in C3 and CAM mode (Tsuzuki et al., 1982). Two lowly expressed genes encoding CA isoforms (alpha CA3 between 2.7 and 10 rpm and BCA5 between 23.7 and 80.7 rpm) were upregulated at midday on both days. Upregulation of major carbonic anhydrase isoforms as it was found for M. crystallinum (Cushman et al., 2008b) could not be detected in T. triangulare.
The gene PPC encodes PEPC, which catalyzes the CO2 carboxylation at night. PPC was upregulated 25-fold at midnight (to 15,510 rpm, expression rank 4 on day 12) on day 9 and 12, respectively. The coding sequences of PPC vary between C3, C4, and CAM plants (Bläsing et al., 2000; Paulus et al., 2013). The coding sequences of the four T. triangulare PPC contigs with on average at least 100 reads mapped were extracted from the assembly, translated, and aligned with PEPC sequences from various C3, C4, and CAM plants (http://www.uniprot.org; Supplemental Fig. S2; Supplemental Table S3). The T. triangulare contigs encoding for PEPC showed characteristics of both C3 and C4 PEPCs. PEP saturation kinetics are known to be determined by the amino acid at position 780 (counting based on Zea mays sequence CAA33317), which in C4 plants is typically Ser and in C3 plants Ala (Bläsing et al., 2000). Sensitivity to malate inhibition is determined by the amino acid at position 890, Gly in C4 plants, and Arg in C3 plants (Paulus et al., 2013). While at position 890, all contigs encoded for the C3-typical Arg in T. triangulare; at position 780, contig Tt63271 (7,638 rpm at midnight on day 9) and Tt9871_8 (1,198 rpm at midnight on day 12) encoded for the C3 typical Ala, Tt9871_4 (1,409 rpm at midnight on day 9) and Tt9871_6 (399 rpm at midnight on day 12) encoded for the C4-typical Ser (Supplemental Fig. S2). Read mapping on the contig level identified all isoforms as upregulated during CAM (Supplemental Table S3); however, it cannot be determined whether Tt63271 and Tt9871_8, Tt9871_4 and Tt9871_6 are alleles or recent duplicates.
Oxaloacetate (OAA) resulting from PEP carboxylation is reduced to malate by MALATE DEHYDROGENASE (MDH). The cytosolic MDH was unchanged but constitutively highly expressed in C3 conditions (941 rpm at midday on day 0) and during CAM (1,141 rpm, expression rank 110 at midday on day 12). Malate is stored as malic acid in the vacuole during the night (Cheffings et al., 1997). Genes encoding two malate channels of the ALUMINUM-ACTIVATED MALATE TRANSPORTER (ALMT) family were upregulated 6-fold and 28-fold at midnight, but reached only 4 and 19 rpm after upregulation, respectively. While below 1 rpm in abundance at any other day, transcripts encoding the TONOPLAST DICARBOXYLATE TRANSPORTER are detectable on day 12 (MD: 5 rpm, MN: 11 rpm). VHA-B3, encoding the subunit B3 of the vacuolar ATPase (V-ATPase), which sustains the electrochemical gradient during import of malic acid (White and Smith, 1989), is the major expressed subunit, being significantly upregulated 2-fold on day 12. Of the 11 other genes encoding for V-ATPase subunits, two were unaltered and eight genes were slightly but significantly downregulated during CAM (Supplemental Table S4). TYPE I PROTON-TRANSLOCATING PYROPHOSPHATASE is downregulated 2-fold and 3-fold at midday of days 9 and 12.
During the day, malic acid is released from the vacuole and, depending on the species, is believed to be decarboxylated by NADP-ME, NAD-ME, and/or PEP CARBOXYKINASE (PEPCK; Dittrich, 1976). In T. triangulare, the plastidial NADP-ME4 was upregulated 5-fold at midday to 780 rpm and the apparent cytosolic NADP-ME1 was upregulated 6-fold to 579 rpm, while NAD-ME1 was 2-fold upregulated to 173 rpm and PEPCK was not significantly upregulated. The T. triangulare contigs of NADP-ME extracted from the assembly with high-read mappings possess a target peptide and cluster with the plastidial AtNADP-ME4 with high bootstrap support (Supplemental Fig. S3; Supplemental Table S5). In contrast to the downregulation of one isoform of NADP-ME at midnight in M. crystallinum (Cushman et al., 2008b), both genes encoding NADP-ME isoforms were upregulated at both midday and midnight on both CAM days. Transcript amounts of the plastidial DICARBOXYLATE TRANSPORTERS were unaltered (DiT1, expression between 256 and 381 rpm at midday, DiT2 between 48 and 105 rpm), while all three mitochondrial DICARBOXYLATE CARRIERS were upregulated at midday on both CAM days (DIC1 15-fold to 787 rpm, DIC2 12-fold to 899 rpm, DIC3 14-fold to 310 rpm on day 12). The plastidial pyruvate importer BILE-ACID SODIUM SYMPORTER2 (BASS2) was upregulated at night on both CAM days (4-fold to 411 rpm on day 9). Pyruvate, produced by malate decarboxylation, is phosphorylated to PEP by PPDK encoded by a single gene (9-fold upregulated to 18,371 rpm at midnight on day 9) and fed back to gluconeogenic starch synthesis (Kluge and Osmond, 1971). Extraction, alignment, and quantification of two different contigs encoding for PPDK (Supplemental Table S6) revealed markedly increased transcript amounts during CAM exclusively for one contig, Tt26901, which encodes for a protein with a 77 amino acid shorter N terminus. At midnight on day 12 transcript amounts for Tt26901 were 59-fold higher than for the longer PPDK encoding contig Tt24575 (Supplemental Table S6). Transcript amounts of the PEP/PHOSPHATE TRANSLOCATOR (PPT), catalyzing the export of PEP to the cytosol, were downregulated during CAM (4-fold to 23 rpm at midnight on day 12).
Another important aspect of CAM photosynthesis is starch turnover and its connection to the carboxylation/decarboxylation cycle. During the night, starch is degraded, likely both via the phosphorolytic and hydrolytic pathways, to provide PEP for PEPC via glycolysis (Weise et al., 2011). Four genes encoding enzymes, required both for phosphorolytic and hydrolytic starch degradation according to Weise et al. (2011) and Streb and Zeeman (2012), LIMIT DEXTRINASE (LDA), ISOAMYLASE3 (ISA3), GLUCAN WATER DIKINASE (GWD1), and DISPROPORTIONATING ENZYME1 (DPE1), were upregulated at midday on both CAM days (e.g. on day 12: 4-fold to 178 rpm, 4-fold to 195 rpm, 10-fold to 995 rpm and 3-fold to 197 rpm, respectively). ISA3 is downregulated 3-fold at midnight on both CAM days and DPE1 is downregulated at midnight of day 12. Genes encoding enzymes specific for hydrolytic starch degradation, PHOSPHOGLUCAN PHOSPHATASE (abbreviated SEX4 for STARCH EXCESS4) and BETA-AMYLASE1 (BAM1) were upregulated at midday (5-fold to 863 rpm and 3-fold to 150 rpm, respectively) as well, while the gene encoding PHOSPHOGLUCAN, WATER DIKINASE (PWD) was downregulated at midnight on both CAM days. Of the two exporters, MALTOSE EXPORTER1 (MEX1) and PLASTIDIC GLC TRANSLOCATOR1 (GLT1) exporting the hydrolytic breakdown products maltose and Glc (Weber et al., 2000; Niittylä et al., 2004), only the gene encoding MEX1 was significantly upregulated 2-fold on both CAM days at midday. MEX1 and GLT1 were both downregulated at midnight on both CAM days. The cytosolic enzymes catalyzing the conversion of Glc and maltose to Glc-phosphates (HEXOKINASE2 [HXK2], DPE2, STARCH PHOSPHORYLASE2 [PHS2] and PHOSPHOGLUCOMUTASE3 [PGM3]) were all downregulated at midnight upon drought (e.g. on day 12: 2-fold to 78 rpm, 2-fold to 110 rpm and 3-fold to 111 rpm, respectively). Of the enzymes specific for phosphorolytic starch degradation, the gene encoding ALPHA-AMYLASE3 (AMY3) was upregulated (4-fold on day 12 to 702 rpm), while the plastidial PHS1 gene was downregulated (3-fold on day 12 to 252 rpm) at midnight on both CAM days. The Glc-6-phosphate (G6P) exporter (GPT) expression was highly induced at midday upon CAM induction (17-fold; 1,095 rpm on day 12) and downregulated at midnight on day 12 (5-fold to 20 rpm).
The genes encoding cytosolic glycolytic enzymes PHOSPHOGLUCOSE ISOMERASE (PGI, upregulated 2-fold to 171 rpm at midday), PHOSPHOFRUCTOKINASE (PFK, constitutive level at midday ranging from 275 to 361 rpm), FRU-BISPHOPHATE ALDOLASE (FBA, upregulated 5-fold at midday on day 12 to 1,247 rpm), TRIOSEPHOSPHATE ISOMERASE (TPI, constitutive level at midday ranging from 980 to 1,541 rpm), SUBUNIT C2 OF GLYCERALDEHYDE-3-PHOSPHATE DEHYDROGENASE (GAPC-2, upregulated 2-fold at midday to 2,649 rpm), PHOSPHOGLYCERATE MUTASE (PGlyM, upregulated at midday 4-fold to 675 rpm), and ENOLASE (upregulated 3-fold to 2,273 rpm) were of high abundance or more abundant in CAM at midday. The gene encoding PHOSPHOGLYCERATE MUTASE (PGK) is of lower abundance (2-fold at midday to 1,167 rpm on day 12). The glycolytic enzymes, whose transcripts were more abundant, produce 3-phosphoglycerate (3-PGA), which may enter the chloroplast via TRIOSE PHOSPHATE/PHOSPHATE TRANSLOCATOR (TPT, constitutive level ranging from 1,632 to 2,905 rpm). Triose-phosphates (3-PGA and GAP) resulting from Rubisco-based carbon fixation and from recycling the pyruvate out of the decarboxylation reaction can be stored as starch. The genes encoding starch precursor biosynthetic enzymes, PHOSPHOGLUCOMUTASE (PGM) and the LARGE SUBUNIT 4 OF ADP-GLC PYROPHOSPHORYLASE (APL4), were more abundant (6-fold to 1063 rpm and 9-fold to 3790 rpm at midday on day 12, respectively) and transcript levels of the GRANULE-BOUND STARCH SYNTHASES (GBSS) were unchanged while those of ISA1 and STARCH BRANCHING ENZYME3 (BE3) were significantly less abundant (4-fold to 66 rpm and 3-fold to 296 rpm at midday, respectively). The transcript levels of more abundant genes involved in carboxylation, decarboxylation, glycolysis, and gluconeogenesis were reduced to the levels of well-watered plants upon rewatering, except for GWD (3-fold upregulated at midday), GPT1 (4-fold upregulated at midday), and NADP-ME1 (3-fold upregulated at midday), which remained highly abundant on day 2 after rewatering.
Malate, Citrate, Soluble Sugars, and Starch
In C3 performing plants, malate levels were somewhat higher at midday than at midnight (Fig. 4, day 0, 4, and 2 d after rewatering). By contrast, in plants exhibiting CAM (Fig. 4, days 9 and 12; Supplemental Table S2), malate levels were up to 8-fold higher at midnight as compared to midday and up to 4-fold higher as compared to midnight during C3. The increased malate levels at midnight are consistent with the increased acidification measured at the end of the night (Fig. 1B) but did not reflect the full diurnal amplitude during the light/dark cycle. As midnight levels are higher than midday levels, malate decarboxylation may occur quite rapidly during the first 6 h of the light period (Fig. 4). Citrate showed a similar pattern to malate, as observed in other CAM species (Winter and Smith, 1996; Lüttge, 2002), with increased amounts during the night in CAM plants compared to C3 plants, albeit with a lower amplitude than malate. In agreement with the C3 plant Arabidopsis (Smith et al., 2004), T. triangulare leaves performing C3 photosynthesis showed high Glc, Fru, and Suc pool sizes at the middle of the day, reflecting high photosynthetic activity (Fig. 4). While Glc and Fru levels dropped at night, Suc levels in the middle of the night were high, probably due to Suc synthesis from starch during the night as reported previously (Chia et al., 2004; Smith et al., 2004). The amounts of Glc, Fru, and Suc during the day and night dropped in CAM conditions (days 9 and 12) and were accompanied by reduced amounts of starch. The total amount of starch showed a continuous decrease from days 4 to 9 to 12. Two days after rewatering, starch levels were increased to the amounts before drought, while Glc, Fru, and Suc pools were not fully restored.
Figure 4.
Levels of organic acids, soluble sugars and starch in T. triangulare in response to varying levels of water availability. Except for starch, all metabolites were measured by GC-MS and normalized to dry weight (DW) and internal ribitol standard (mean ± se, n = 3–4, asterisks indicate Student’s t test significance in comparison to day 0 at ***P < 0.001, **P < 0.01, and *P < 0.05). Starch was normalized to dry weight (mean ± se, n = 2–4). re2, 2 d after rewatering.
The photorespiratory metabolites glycerate and glycolate were significantly reduced in CAM conditions (Supplemental Fig. S4). During the day, glycerate was depleted 40.2-fold and 238.9-fold on days 9 and 12, and during the night, 3.6-fold and 16.7-fold. Glycolate was depleted during the day on days 9 and 12 (2.2-fold and 3.1-fold).
The contents of the putative compatible solute raffinose increased significantly during day (14.4-fold and 12.3-fold) and night (9.2-fold and 4.7 fold) on days 9 and 12 compared to day 0, and the increased amounts at night were not fully reverted 2 d after rewatering (Fig. 4). Well-known compatible solutes in C3 plants (Bohnert et al,, 1995; Hare et al., 1998), Pro and the sugar alcohol mannitol, were only enriched in individual plants of the biological replicates (Supplemental Fig. S5).
Mapman and K-Means Clustering Indicate Multiple Layers of Response and Regulation
In order to understand alterations of mRNA amounts beyond the changes in the CAM-related genes described earlier (Fig. 3), three independent analyses were used to test general changes at the metabolic pathway and single gene level: (1) Mapman-based analysis using all values followed by Wilcoxon Rank Sum test (see “Materials and Methods”) for enrichment (Thimm et al., 2004; Supplemental Fig. S6), (2) k-means clustering of the significantly changed genes followed by gene ontology (GO) term enrichment analysis (Fig. 5), and (3) manual inspection of the 50 genes with the highest fold changes on day 12 (Tables I and II). In Mapman (Supplemental Fig. S6), the metabolism overview again indicated little to no gene expression changes on day 4 of drought and rewatered plants compared to well-watered samples and massive changes during days 9 and 12 of water-limitation (Supplemental Fig. S6). The changes included downregulation of genes involved in cell wall metabolism (Supplemental Dataset S4). Genes of cell wall proteins of all classes were markedly downregulated (q < 10−6 at midday), as were the biosynthesis genes for cell wall polymers (cellulose: q < 10−08; modifiers: q < 10−10 at midday; pectin esterases: q < 10−11; pectate lyase: q < 10−11, precursor synthesis: q < 10−7; hemicellulose: q < 10−6 at midnight). Downregulation of photosynthetic genes of the light reactions was visible on day 9 but more pronounced on day 12 of drought (q < 10−5). Consistent with the previous analysis (Fig. 3), starch turnover genes (q < 10−4) and genes of glycolysis (q < 10−3 on day 12) were strongly upregulated. Additional strong upregulation was limited to the genes in the raffinose synthesis pathway (q < 10−2), the committed step of Pro biosynthesis, and to the myoinositol oxidases (Tables I and II).
Figure 5.
K-means clustering of relative gene expression and selected enriched GO terms. Genes that were found to be differentially expressed in the middle of the day between one of the water-limited stages and day 0 (DESeq2, q < 0.01) were used for the k-means approach (6800 genes in total). For the full list of enriched GO terms, see Supplemental Dataset S4. Gray line, Expression of single genes; black line, average of all genes in cluster; re2, 2 d after rewatering.
Table I. Fifty most highly upregulated genes on day 12 after drought.
rpm, Reads per million (n = 3); MD, middle of the day; MN, middle of the night; log2-FC, log2-fold change in expression at given time point on day 12 compared to day 0; ns, not significantly DEG.
| Locus | Annotation (TAIR10) | Day 0 MD | Day 0 MN | Day 12 MD | Day 12 MN | Log2-FC (MD ; MN) |
|---|---|---|---|---|---|---|
| rpm | rpm | rpm | rpm | |||
| AT2G42600 | Phosphoenolpyruvate carboxylase 2 | 208.67 | 374.33 | 4,803.67 | 9,910.67 | 4.05 ; 4.7 |
| AT1G53310 | Phosphoenolpyruvate carboxylase 1 | 312.00 | 613.33 | 6,656.33 | 15,510.33 | 3.91 ; 4.61 |
| AT4G17030 | Expansin-like B1 | 0.00 | 0.00 | 4.67 | 1.67 | 4.91 ; 2.55 |
| AT4G26260 | Myoinositol oxygenase 4 | 1.00 | 0.00 | 373.33 | 29.00 | 7.36 ; 5.02 |
| AT1G14520 | Myoinositol oxygenase 1 | 3.33 | 1.33 | 1,290.00 | 112.67 | 7.25 ; 5.42 |
| AT2G19800 | Myoinositol oxygenase 2 | 1.67 | 0.67 | 451.00 | 37.00 | 6.6 ; 4.83 |
| AT1G55740 | Seed imbibition 1 | 9.33 | 46.67 | 605.33 | 2,236.33 | 5.41 ; 5.38 |
| AT3G08860 | PYRIMIDINE 4 | 3.33 | 44.33 | 139.67 | 23.33 | 4.68 ; ns |
| AT2G38400 | Ala:glyoxylate aminotransferase 3 | 2.67 | 30.00 | 103.67 | 17.67 | 4.54 ; ns |
| AT3G50980 | Dehydrin xero 1 | 0.00 | 0.00 | 7.67 | 0.00 | 4.9 ; ns |
| AT3G22840 | Chlorophyll A-B binding family protein | 17.67 | 15.00 | 4,029.67 | 17.33 | 6.74 ; ns |
| AT5G15250 | FTSH protease 6 | 6.67 | 2.67 | 764.33 | 3.00 | 6.18 ; ns |
| AT5G20110 | Dynein light chain type 1 family protein | 3.67 | 23.67 | 368.67 | 12.33 | 6.05 ; ns |
| AT4G27360 | Dynein light chain type 1 family protein | 19.67 | 32.33 | 654.33 | 54.33 | 4.62 ; 0.83 |
| AT3G29410 | Terpenoid cyclases/Protein prenyltransferases superfamily protein | 0.00 | 0.00 | 4.67 | 17.67 | 4.86 ; 5.96 |
| AT5G16020 | Gamete-expressed 3 | 0.33 | 1.33 | 41.00 | 3.67 | 5.75 ; 1.59 |
| AT1G80920 | Chaperone DnaJ-domain superfamily protein | 0.33 | 12.33 | 23.33 | 578.33 | 4.6 ; 5.46 |
| AT3G22740 | Homo-Cys S-methyltransferase 3 | 2.33 | 23.33 | 118.33 | 145.33 | 5.08 ; 2.68 |
| AT1G10060 | Branched-chain amino acid transaminase 1 | 1.00 | 4.67 | 31.33 | 11.00 | 4.63 ; ns |
| AT1G29900 | Carbamoyl phosphate synthetase B | 69.67 | 80.67 | 309.67 | 1,955.67 | 1.75 ; 4.56 |
| AT1G12740 | Cytochrome P450, family 87, subfamily A, polypeptide 2 | 4.00 | 86.33 | 290.67 | 10.67 | 4.47 ; -2.13 |
| AT1G52240 | RHO guanyl-nucleotide exchange factor 11 | 6.67 | 113.33 | 1,652.00 | 31.00 | 7.23 ; -1.69 |
| AT1G28480 | Thioredoxin superfamily protein | 1.00 | 0.67 | 2.00 | 37.00 | ns ; 5.76 |
| AT1G18100 | PEBP (phosphatidylethanolamine-binding protein) family protein | 1.00 | 1.00 | 34.67 | 59.00 | 4.52 ; 5.57 |
| AT1G03790 | SOMNUS | 0.00 | 0.67 | 19.67 | 4.33 | 5.15 ; 2.38 |
| AT1G54070 | Dormancy/auxin associated family protein | 0.00 | 1.67 | 2.33 | 75.33 | 2.5 ; 5.01 |
| AT1G80390 | Indole-3-acetic acid inducible 15 | 0.00 | 0.00 | 5.67 | 5.33 | 4.73 ; 5 |
| AT5G12840 | Nuclear factor Y, subunit A1 | 0.67 | 2.67 | 25.67 | 31.00 | 4.75 ; 3.56 |
| AT1G15330 | Cystathionine beta-synthase (CBS) protein | 0.00 | 0.00 | 1.33 | 5.00 | 3.54 ; 4.74 |
| AT3G21700 | Ras-related small GTP-binding family protein | 0.67 | 3.00 | 17.33 | 82.33 | 3.56 ; 4.61 |
| AT1G21000 | PLATZ transcription factor family protein | 0.33 | 25.33 | 18.33 | 135.00 | 4.48 ; 2.45 |
| AT3G63060 | EID1-like 3 | 0.00 | 0.00 | 22.00 | 13.33 | 5.88 ; 5.32 |
| AT3G48530 | SNF1-related protein kinase regulatory subunit gamma 1 | 0.67 | 31.00 | 39.67 | 275.00 | 5.05 ; 3.1 |
| AT5G56550 | Oxidative stress 3 | 1.33 | 4.33 | 18.33 | 310.33 | 3.03 ; 5.71 |
| AT5G47560 | Tonoplast dicarboxylate transporter | 0.00 | 0.00 | 5.33 | 10.67 | 4.99 ; 5.93 |
| AT1G32450 | Nitrate transporter 1.5 | 4.33 | 68.00 | 233.67 | 34.33 | 5.04 ; ns |
| AT4G21680 | NITRATE TRANSPORTER 1.8 | 2.00 | 20.33 | 101.33 | 16.67 | 5.02 ; ns |
| AT3G27250 | Unknown | 0.00 | 1.00 | 85.67 | 58.33 | 7.08 ; 5.63 |
| AT1G52720 | Unknown | 0.67 | 26.67 | 156.33 | 80.33 | 6.72 ; 1.63 |
| AT4G19390 | Uncharacterized protein family (UPF0114) | 6.33 | 49.33 | 1,234.33 | 276.00 | 6.58 ; 2.35 |
| AT3G03170 | Unknown | 0.00 | 0.67 | 4.33 | 66.00 | 3.56 ; 6.21 |
| AT3G48510 | Unknown | 2.67 | 1.33 | 78.33 | 106.00 | 4.16 ; 6.08 |
| AT4G26288 | Unknown | 0.00 | 0.33 | 2.00 | 32.67 | 3.02 ; 5.99 |
| AT5G40790 | Unknown | 1.67 | 0.33 | 38.33 | 48.00 | 3.96 ; 5.85 |
| AT1G15380 | Lactoylglutathione lyase/glyoxalase I family protein | 4.67 | 9.67 | 337.00 | 109.33 | 5.65 ; 3.42 |
| AT5G50360 | Unknown | 1.00 | 0.00 | 3.67 | 11.67 | ns ; 5.3 |
| AT2G44670 | Protein of unknown function (DUF581) | 0.33 | 2.33 | 3.33 | 65.67 | 1.92 ; 4.81 |
| AT5G02020 | Unknown | 1.67 | 3.33 | 7.33 | 96.67 | 1.74 ; 4.57 |
| AT1G27461 | Unknown | 0.00 | 0.00 | 0.00 | 3.67 | ns ; 4.51 |
| AT2G28780 | Unknown | 0.33 | 0.00 | 16.00 | 0.00 | 4.49 ; ns |
Table II. Fifty most highly downregulated genes on day 12 after drought.
rpm, Reads per million (n = 3); MD, middle of the day; MN, middle of the night; log2-FC, log2-fold change in expression at given time point on day 12 compared to day 0; ns, not significantly DEG.
| Locus | Annotation (TAIR10) | Day 0 MD | Day 0 MN | Day 12 MD | Day 12 MN | Log2-FC (MD ; MN) |
|---|---|---|---|---|---|---|
| rpm | rpm | rpm | rpm | |||
| AT5G20630 | Germin 3 | 2,449.67 | 648.33 | 2.00 | 1.67 | −9.13 ; −7.24 |
| AT1G72610 | Germin-like protein 1 | 932.67 | 274.67 | 1.00 | 1.00 | −8.55 ; −6.45 |
| AT1G64390 | Glycosyl hydrolase 9C2 | 177.00 | 80.00 | 0.67 | 1.67 | −7.45 ; −5.27 |
| AT4G21960 | Peroxidase superfamily protein | 1,529.33 | 8,040.00 | 10.67 | 349.67 | −7.09 ; −4.22 |
| AT4G11050 | Glycosyl hydrolase 9C3 | 70.67 | 31.67 | 0.00 | 0.33 | −7.06 ; −5.32 |
| AT5G22740 | Cellulose synthase-like A02 | 272.67 | 207.00 | 3.33 | 4.67 | −6.42 ; −5.27 |
| AT2G32990 | Glycosyl hydrolase 9B8 | 42.00 | 34.33 | 0.33 | 0.00 | −6.36 ; −6.19 |
| AT2G04780 | FASCICLIN-like arabinoogalactan 7 | 15.00 | 24.33 | 0.00 | 0.00 | −6.12 ; −6.27 |
| AT4G37450 | Arabinogalactan protein 18 | 10.33 | 22.67 | 0.00 | 0.00 | −5.64 ; −6.24 |
| AT2G35860 | FASCICLIN-like arabinogalactan protein 16 precursor | 94.33 | 125.67 | 2.67 | 1.33 | −5.11 ; −6.12 |
| AT5G03760 | Nucleotide-diphospho-sugar transferases superfamily protein | 61.00 | 64.33 | 1.00 | 0.67 | −5.75 ; −6.11 |
| AT2G37130 | Peroxidase superfamily protein | 21.00 | 7.67 | 0.00 | 0.00 | −6.03 ; −4.26 |
| AT3G11700 | FASCICLIN-like arabinogalactan protein 18 precursor | 129.33 | 195.67 | 4.67 | 2.67 | −5.04 ; −5.92 |
| AT1G02335 | germin-like protein subfamily 2 member 2 precursor | 9.33 | 16.33 | 0.00 | 0.00 | −5.47 ; −5.91 |
| AT1G67750 | Pectate lyase family protein | 15.00 | 53.33 | 0.00 | 0.67 | −5.82 ; −5.5 |
| AT4G13410 | Nucleotide-diphospho-sugar transferases superfamily protein | 27.67 | 23.00 | 0.67 | 0.00 | −5.65 ; −5.81 |
| AT3G28150 | TRICHOME BIREFRINGENCE-LIKE 22 | 35.67 | 12.00 | 0.33 | 0.00 | −5.79 ; −5.33 |
| AT3G53190 | Pectin lyase-like superfamily protein | 14.67 | 23.33 | 0.00 | 1.00 | −5.79 ; −4.07 |
| AT3G62020 | Germin-like protein 10 | 8.33 | 14.33 | 0.00 | 0.00 | −5.66 ; −5.65 |
| AT1G12090 | Extensin-like protein | 151.33 | 29.33 | 3.33 | 1.33 | −5.65 ; −4.22 |
| AT1G41830 | SKU5-similar 6 | 19.00 | 47.33 | 0.33 | 0.33 | −5.03 ; −5.65 |
| AT5G46890 | Bifunctional inhibitor/lipid-transfer protein/seed storage 2S albumin superfamily protein | 28.67 | 5.67 | 0.00 | 0.00 | −6.42 ; −3.84 |
| AT3G04290 | Li-tolerant lipase 1 | 235.00 | 817.67 | 1.00 | 13.00 | −6.39 ; −4.91 |
| AT1G76160 | SKU5 similar 5 | 37.33 | 89.33 | 0.33 | 1.00 | −6.13 ; −6.34 |
| AT5G33370 | GDSL-like Lipase/Acylhydrolase superfamily protein | 162.00 | 675.00 | 1.00 | 6.33 | −6.17 ; −5.57 |
| AT5G23940 | HXXXD-type acyl-transferase family protein | 95.00 | 148.33 | 1.33 | 11.67 | −6.13 ; −3.47 |
| AT3G16370 | GDSL-like Lipase/Acylhydrolase superfamily protein | 298.00 | 307.67 | 5.33 | 4.67 | −5.98 ; −5.78 |
| AT3G15850 | Fatty acid desaturase 5 | 20.00 | 9.00 | 0.00 | 2.00 | −5.91 ; −1.93 |
| AT1G21850 | SKU5 similar 8 | 17.33 | 35.33 | 0.33 | 0.67 | −5.17 ; −5.73 |
| AT4G28780 | GDSL-like Lipase/Acylhydrolase superfamily protein | 111.00 | 569.33 | 0.33 | 4.33 | −5.63 ; −5.47 |
| AT1G63710 | Cytochrome P450, family 86, subfamily A, polypeptide 7 | 45.00 | 193.33 | 1.00 | 1.00 | −5.27 ; −6.59 |
| AT3G10185 | Gibberellin-regulated family protein | 177.33 | 292.33 | 1.00 | 4.33 | −6.46 ; −4.94 |
| AT1G12570 | Glc-methanol-choline (GMC) oxidoreductase family protein | 88.67 | 59.33 | 1.00 | 2.67 | −6.09 ; −4.2 |
| AT2G45970 | Cytochrome P450, family 86, subfamily A, polypeptide 8 | 34.33 | 140.33 | 1.00 | 1.33 | −5.3 ; −6.07 |
| AT1G74670 | Gibberellin-regulated family protein | 19.67 | 28.00 | 0.00 | 0.33 | −6.03 ; −5.03 |
| AT1G61720 | NAD(P)-binding Rossmann-fold superfamily protein | 6.00 | 89.67 | 0.00 | 0.67 | −3.32 ; −5.82 |
| AT3G52500 | Eukaryotic aspartyl protease family protein | 43.33 | 35.67 | 1.33 | 0.33 | −5.01 ; −5.75 |
| AT5G44635 | Minichromosome maintenance (MCM2/3/5) family protein | 37.67 | 6.67 | 0.00 | 0.00 | −6.6 ; −4.12 |
| AT5G67100 | DNA-directed DNA polymerases | 21.67 | 9.67 | 0.33 | 1.00 | −5.93 ; −3.47 |
| AT5G46280 | Minichromosome maintenance (MCM2/3/5) family protein | 26.67 | 5.33 | 0.00 | 0.00 | −5.91 ; −3.97 |
| AT1G44900 | minichromosome maintenance (MCM2/3/5) family protein | 66.00 | 12.00 | 0.67 | 0.33 | −5.87 ; −3.62 |
| AT2G07690 | Minichromosome maintenance (MCM2/3/5) family protein | 73.67 | 22.67 | 1.33 | 5.67 | −5.74 ; −1.74 |
| AT4G02060 | Minichromosome maintenance (MCM2/3/5) family protein | 38.00 | 6.33 | 0.67 | 0.00 | −5.73 ; −3.52 |
| AT1G27040 | Major facilitator superfamily protein | 44.33 | 2.67 | 0.00 | 2.00 | −6.87 ; ns |
| AT5G62730 | Major facilitator superfamily protein | 40.00 | 3.00 | 0.00 | 3.00 | −6.49 ; ns |
| AT3G02500 | Unknown | 28.33 | 4.00 | 0.33 | 0.00 | −5.89 ; −3.27 |
| AT1G27930 | Protein of unknown function (DUF579) | 8.67 | 19.00 | 0.00 | 0.00 | −4.65 ; −5.79 |
| AT2G21100 | Disease resistance-responsive (dirigent-like protein) family protein | 16.33 | 49.67 | 0.67 | 0.33 | −4.04 ; −5.66 |
| AT4G17340 | Tonoplast intrinsic protein 2;2 | 321.33 | 125.00 | 3.67 | 31.00 | −5.74 ; ns |
| AT1G13260 | Related to ABI3/VP1 1 | 18.33 | 5.33 | 0.00 | 0.33 | −5.86 ; −3.45 |
K-means clustering grouped 6,800 significantly DEGs at midday compared to day 0 by transcriptional pattern into 12 clusters (Fig. 5). Six clusters contained 3,124 genes ascending (clusters 1–6, Fig. 5), and six clusters contained 3,676 genes descending with pronounced water-limitation (7–12, Fig. 5). The clusters with ascending genes included one cluster with gradual increase up to day 12 and only little recovery after 2 d of rewatering (cluster 1). This cluster was enriched in GO terms related to external stimulus, and response. clusters 2 and 3 show mostly stable expression between days 9 and 12 and the expression was fully (cluster 2) and partially (cluster 3) recovered upon 2 d of rewatering. The list of genes with fully restored abundance of cluster 2 is enriched in GO terms related to chloroplast organization, protein localization, and starch. The genes in cluster 3 recovered only partially to well-watered levels after rewatering, and this cluster showed significant enrichment of terms related to catabolic processes. Cluster 4 was only transiently upregulated on day 9 of water withholding but dropped back to background levels on day 12. It was enriched with GO terms relating to responses to signals, which include ethylene and mechanical stimulus. The expression of genes in clusters 5 and 6 peaked on day 12. Genes of cluster 5 only responded after 9 d of drought, increased sharply toward day 12, and mostly recovered after 2 d of rewatering and were enriched in GO terms related to ABA, fatty acid catabolism, and peroxisomes. The expression of genes in cluster 6 responded already after 4 d of drought, was only partially reverted and was enriched for mRNA related processes, development, and protein modification.
Similar to cluster 1, where genes were upregulated and stayed mostly up despite rewatering, cluster 7 contained genes that dropped in expression during days 9 and 12 of drought and barely recovered. Cluster 7 was enriched in genes related to growth, i.e. in GO terms referring to DNA replication, cell cycle, and ribosome genesis. Genes of cluster 8 were similarly downregulated during CAM conditions but fully recovered upon rewatering. This cluster is enriched in GO terms related to cell wall processes and cytoskeleton. Genes in cluster 9 only responded on day 9, further increased on day 12 of drought, and fully recovered after rewatering. This cluster was enriched in GO terms related to cell wall related processes, amino acid metabolism, and photosynthesis. Cluster 10 was the mirror cluster of cluster 4, which peaked on day 9, and included genes with peak downregulation on day 9. It was enriched in genes related to protein localization in the secretory pathway and the endomembrane system, including the protein population of the plasma membrane. Downregulation of genes in clusters 11 and 12 peaked on day 12. Genes in cluster 11 responded only after 12 d of drought and fully recovered after rewatering. This cluster was enriched in processes related to nucleotide metabolic processes. Expression of genes in cluster 12 began decreasing on day 4, continued to decrease until day 12, and after rewatering, were increased above well-watered levels. This overshoot cluster was enriched in genes related to posttranslational modification in particular phosphorylation and transport. Analogous k-means clustering of 7,563 significant DEGs at midnight revealed large overlap of major changes in comparison to day 0 at midday (Supplemental Fig. S7; Supplemental Dataset S5). Taken together, the clusters with elevated expression during drought mostly contained genes enriched in GO terms related to catabolism, RNA metabolism and also related to signaling. The clusters with reduced gene expression contained genes related to growth from photosynthesis over amino acid synthesis and cell wall processes to DNA replication and cell cycle.
To highlight the results of the analysis at the pathway level, the top 50 of significantly induced (Table I) and repressed (Table II) genes on day 12 of water limitation were analyzed. Of the fifty genes with the highest upregulation on day 12 in T. triangulare, 22 were upregulated by ABA in Arabidopsis (Table I). The CAM gene PPC was among the top fifty upregulated genes. Six genes were related to production of compatible solutes or their precursors including raffinose, beta-Ala, and sugar alcohols. One gene encoding a LATE EMBRYOGENESIS ABUNDANT (LEA) protein was among the top 50 induced genes as well as a heat inducible chaperone and the oxidative stress responsive OXIDATIVE STRESS3 (OXS3). Light protection was seemingly strengthened as evidenced by EARLY LIGHT INDUCIBLE1 (ELIP1), FTSH PROTEASE6 (FTSH6), and two dyneins, which may be involved in organelle repositioning (Heddad and Adamska, 2000; Hutin et al., 2003; Zelisko et al., 2005). Proteins encoded by 11 of the 50 most induced genes were involved in regulation: the SNF1 RELATED KINASE REGULATORY SUBUNIT GAMMA1 (KING1), the F-box protein EID1-LIKE3 (EDL3), RHO GUANYL-NUCLEOTIDE EXCHANGE FACTOR11, and four transcription factors including SOMNUS and a PLATZ transcription factor. There were also three transport proteins affected.
The fifty genes most reduced in expression on day 12 included 21 genes involved in cell wall synthesis, nine genes related to lipid metabolism and the vacuolar aquaporin TONOPLAST INTRINSIC PROTEIN2;2 (Table II). Downregulation of DNA replication via downregulation of the genes (MINICHROMOSOME MAINTENANCE2-3 and 5-7) encoding five of six subunits of the helicase and downregulation of the gene encoding DNA-directed DNA polymerase INCURVATA2 was also evident among the fifty most downregulated genes, indicating cell-cycle arrest.
Mediators of the Transcriptional Changes
K-means clustering identified a multilayer response with genes only transiently regulated on day 9 of drought and with genes with sustained change on both day 9 and day 12 of drought. To identify the candidate transcriptional regulators mediating the responses, genes encoding for putative transcription factors were extracted from PlnTFDB (Pérez-Rodríguez et al., 2010) and tested for differential expression on either day 9 or day 12 of water limitation relative to the well-watered state (day 0) and for cluster membership. Of 1,449 identified transcription factor genes, 582 (40%) were differentially regulated on either day 9 or day 12 or both compared to well-watered plants (Supplemental Dataset S6). Of the 582, a subset of 19 transcription factor genes belonged to cluster 4, the genes, which were only transiently upregulated on day 9 of drought (Table III). These included eight TFs of the APETALA2 AND ETHYLENE-RESPONSIVE ELEMENT BINDING PROTEINS (AP2-EREBP) class including C-REPEAT/DRE BINDING FACTORS (CBF2 and CBF3), which are known to be stress induced (Zou et al., 2011), REDOX RESPONSIVE TRANSCRIPTION FACTOR1 (RRTF1) and ETHYLENE RESPONSE FACTOR8 (ERF8). Both CBFs were shown to bind the drought responsive element DNA sequence (Liu et al., 1998). Two zinc finger transcription factors, SALT TOLERANCE ZINC FINGER (STZ) and SALT-INDUCIBLE ZINC FINGER2 (SFZ2), belonged to cluster 4. STZ was shown to function as a transcriptional repressor in response to drought and ABA (Sakamoto et al., 2000, 2004), while SFZ2 is induced by salt stress (Sun et al., 2007). In addition, three genes encoding NO APICAL MERISTEM (NAC) domain proteins, one GRAS, one bZIP, and three MYB-domain proteins made up the transcriptional part of cluster 4. The known functions of transcription factors in the transient group indicated that T. triangulare underwent a transient general stress response commonly observed in plants mediated primarily by the drought responsive element binding transcription factors.
Table III. Transcription factors of k-means cluster midday 4.
TF Family, Transcription factor family based on Pérez-Rodríguez et al. (2010).
| Locus | Annotation (TAIR10) | TF Family |
|---|---|---|
| AT1G12610 | Integrase-type DNA-binding superfamily protein | AP2-EREBP |
| AT1G19210 | Integrase-type DNA-binding superfamily protein | AP2-EREBP |
| AT1G33760 | Integrase-type DNA-binding superfamily protein | AP2-EREBP |
| AT1G53170 | Ethylene response factor 8 | AP2-EREBP |
| AT4G25470 | C-repeat/DRE binding factor 2 | AP2-EREBP |
| AT4G25480 | Dehydration response element B1A | AP2-EREBP |
| AT4G34410 | Redox responsive transcription factor 1 | AP2-EREBP |
| AT5G21960 | Integrase-type DNA-binding superfamily protein | AP2-EREBP |
| AT3G47640 | Basic helix-loop-helix (bHLH) DNA-binding superfamily protein | bHLH |
| AT2G21230 | Basic-Leu zipper (bZIP) transcription factor family protein | bZIP |
| AT1G27730 | Salt tolerance zinc finger | C2H2 |
| AT2G40140 | SALT-INDUCIBLE ZINC FINGER 2 | C3H |
| AT4G17230 | SCARECROW-like 13 | GRAS |
| AT1G18710 | Myb domain protein 47 | MYB |
| AT3G13540 | Myb domain protein 5 | MYB |
| AT1G01380 | Homeodomain-like superfamily protein | MYB-related |
| AT1G33060 | NAC 014 | NAC |
| AT3G49530 | NAC domain containing protein 62 | NAC |
| AT4G35580 | NAC transcription factor-like 9 | NAC |
Of the remaining 563 genes encoding transcription factor candidates, the top 25 upregulated were analyzed for known functions (Table IV). The ABA-responsive proliferation inhibitor SOMNUS (SOM) was the most highly upregulated transcription factor on day 12. The top 25 also included four genes encoding transcription factors known to be involved in ABA signaling, NUCLEAR FACTOR Y, SUBUNIT A1 (NF-YA1), NF-YA9, and HOMEOBOX7 (HB7) and one NAC-like transcription factor, NAC-LIKE, ACTIVATED BY AP3/PI (NAP), which acts upstream of ABA biosynthesis and promotes chlorophyll degradation (Yang et al., 2014). In addition, two transcription factors of the Orphans family, ETHYLENE RESPONSE2 (ETR2) and ETHYLENE INSENSITIVE4 (EIN4), which were shown to antagonistically control seed germination under salt stress (Wilson et al., 2014) as well as three heat shock factor family proteins were among the 25 most upregulated transcription factors. The sustained response of T. triangulare included ABA responsive transcription factors and growth associated regulators, while the transient response included the DRE responsive TFs of the CBF family.
Table IV. Twenty-five most highly upregulated transcription factors on day 12 after drought.
TF Family, Transcription factor family based on Pérez-Rodríguez et al. (2010); rpm, reads per million (n = 3); MD, middle of the day; MN, middle of the night; log2-FC, log2-fold change in expression at given time point on day 12 compared to day 0; ns, not significantly DEG.
| Locus | Annotation (TAIR10) | TF Family | Day 0 MD | Day 0 MN | Day 12 MD | Day 12 MN | Log2-FC (MD ; MN) |
|---|---|---|---|---|---|---|---|
| rpm | rpm | rpm | rpm | ||||
| AT1G03790 | SOMNUS | C3H | 0.00 | 0.67 | 19.67 | 4.33 | 5.15 ; 2.38 |
| AT5G12840 | Nuclear factor Y, subunit A1 | CCAAT | 0.67 | 2.67 | 25.67 | 31.00 | 4.75 ; 3.56 |
| AT1G21000 | PLATZ transcription factor family protein | PLATZ | 0.33 | 25.33 | 18.33 | 135.00 | 4.48 ; 2.45 |
| AT3G54320 | Integrase-type DNA-binding superfamily protein | AP2-EREBP | 1.00 | 1.33 | 3.33 | 27.00 | 1.4 ; 4.41 |
| AT5G67480 | BTB and TAZ domain protein 4 | TAZ | 19.67 | 71.00 | 461.33 | 129.00 | 4.15 ; 0.95 |
| AT2G46680 | Homeobox 7 | HB | 7.67 | 38.00 | 165.33 | 445.00 | 3.92 ; 3.56 |
| AT1G69490 | NAC-like, activated by AP3/PI | NAC | 8.67 | 76.00 | 206.33 | 366.00 | 3.73 ; 2.09 |
| AT1G52880 | NAC (No Apical Meristem) domain transcriptional regulator superfamily protein | NAC | 0.00 | 2.67 | 5.00 | 10.67 | 3.51 ; 1.8 |
| AT3G02550 | LOB domain-containing protein 41 | LOB | 0.00 | 0.00 | 0.00 | 1.33 | ns ; 3.3 |
| AT1G17590 | Nuclear factor Y, subunit A8 | CCAAT | 3.67 | 9.33 | 41.67 | 63.00 | 3.19 ; 2.75 |
| AT3G23150 | ETHYLENE RESPONSE 2 | Orphans | 12.33 | 10.67 | 86.00 | 94.00 | 2.47 ; 3.13 |
| AT3G04580 | ETHYLENE INSENSITIVE 4 | Orphans | 14.33 | 9.67 | 94.00 | 82.33 | 2.38 ; 3.13 |
| AT2G26150 | Heat shock transcription factor A2 | HSF | 4.67 | 6.00 | 9.00 | 47.00 | ns ; 3.06 |
| AT1G80840 | WRKY DNA-binding protein 40 | WRKY | 2.33 | 0.67 | 0.33 | 6.67 | −2.38 ; 3.04 |
| AT4G17900 | PLATZ transcription factor family protein | PLATZ | 8.33 | 38.67 | 91.67 | 55.33 | 3.01 ; ns |
| AT3G11580 | AP2/B3-like transcriptional factor family protein | ABI3VP1 | 0.33 | 1.33 | 4.33 | 1.67 | 3 ; ns |
| AT4G13980 | Winged-helix DNA-binding transcription factor family protein | HSF | 2.67 | 9.00 | 4.33 | 68.67 | ns ; 2.98 |
| AT3G24520 | Heat shock transcription factor C1 | HSF | 2.67 | 11.33 | 5.00 | 88.00 | ns ; 2.96 |
| AT3G20910 | Nuclear factor Y, subunit A9 | CCAAT | 11.00 | 25.67 | 72.67 | 181.00 | 2.32 ; 2.87 |
| AT4G02640 | bZIP transcription factor family protein | bZIP | 1.67 | 6.67 | 0.67 | 48.33 | ns ; 2.83 |
| AT2G25900 | Zinc finger C-x8-C-x5-C-x3-H type family protein | C3H | 12.67 | 12.67 | 119.00 | 45.33 | 2.83 ; 1.91 |
| AT3G15500 | NAC domain containing protein 3 | NAC | 0.67 | 8.00 | 8.33 | 21.00 | 2.82 ; 1.39 |
| AT1G32700 | PLATZ transcription factor family protein | PLATZ | 16.00 | 56.00 | 141.33 | 75.67 | 2.78 ; ns |
| AT4G39070 | B-box zinc finger family protein | Orphans | 1.00 | 3.00 | 1.67 | 22.00 | ns ; 2.77 |
| AT5G28770 | bZIP transcription factor family protein | bZIP | 2.67 | 11.00 | 2.00 | 74.00 | ns ; 2.77 |
To test if phytohormones other than ABA are also involved in the drought response, the overlap between genes significantly changed on day 12 at midday and genes specifically altered by the application of different hormones in Arabidopsis (Goda et al., 2008) was determined and statistically evaluated. Among the genes shown to be upregulated during hormone treatment of Arabidopsis with ABA, auxin, brassinosteroids, cytokinine, and ethylene, only the ABA-upregulated genes were enriched among the T. triangulare drought-inducible genes (Supplemental Fig. S8). A total of 434 ABA-regulated genes of Arabidopsis were drought-induced genes in T. triangulare and, accordingly, 541 ABA-regulated genes of Arabidopsis were repressed by drought in T. triangulare. The ABA-, auxin-, brassinosteroid-, and ethylene-downregulated genes were enriched among the T. triangulare drought-repressed genes. However, only few genes were shown to be repressed by auxin, brassinosteroids, and ethylene (90, 11, and 23, respectively). Thus, among the phytohormones, ABA made the major contribution to the control of drought-controlled genes and likely controls at least one-fifth of the genes differentially regulated under drought conditions in T. triangulare.
To visualize the ABA contribution to changes in general metabolism, the Mapman map of ABA induced changes in Arabidopsis was compared with those occurring in T. triangulare (Supplemental Fig. S9). The map of ABA-responsive genes replicates the induction of raffinose synthesis genes, and a mild reduction of photosynthesis genes but failed to result in the major changes in cell wall synthesis and changed photorespiration, sulfur metabolism, and secondary metabolism. The ABA signal detected in Arabidopsis at the level of metabolic gene expression is not fully congruent with the signals detected in T. triangulare, indicating additional layers of regulation in T. triangulare.
DISCUSSION
Facultative CAM
During CAM induction in T. triangulare, several key components of the CAM-cycle sensu stricto were transcriptionally upregulated such as genes encoding PEPC, NADP-ME, NAD-ME, and PPDK (Fig. 3). Other components, including transcripts of MDH and BCA3, remained unchanged but were constitutively highly expressed already prior to CAM induction. This is consistent with enzyme activity data for MDH (Holtum and Winter, 1982) and CA (Tsuzuki et al., 1982) in M. crystallinum operating in the C3 and CAM modes, that were already very high in the C3 state. Nonetheless, Cushman et al. (2008b) found additional upregulation of these enzymes during salt-induced CAM in M. crystallinum. In C4 photosynthesis, activities of the corresponding carboxylation and decarboxylation enzymes required for CAM are not temporally but spatially separated into different cell types (Sage, 2003). In C4 species of the genera Flaveria and Cleome, NAD-MDH is also not transcriptionally elevated relative to C3 sister species (Bräutigam et al., 2011; Gowik et al., 2011) and CA only in some, but not in all, cases (Bräutigam et al., 2014).
In contrast to the canonical scheme of the CAM-cycle with NADP-ME localized in the cytosol (Holtum et al. 2005 and references therein), we found two isoforms of NADP-ME to be upregulated at both midday and midnight during CAM in the cross species mapping (Fig. 3; Supplemental Dataset S3). The contig analysis indicates that indeed two contigs are highly expressed; however, both possess a target peptide (Emanuelsson et al., 2000) and cluster with the plastidial AtNADP-ME4 with high bootstrap support. (Supplemental Fig. S3). The putative plastidic malate importers of the DiT family are not upregulated (Supplemental Dataset S3). NAD-ME was also upregulated, albeit at lower absolute levels but concomitant with the required malate importer DIC (Supplemental Dataset S3). Taken together, the data suggest that multiple routes for decarboxylation are possible. Unlike in the transcriptomic analyses of C4 species (Bräutigam et al., 2011; Gowik et al., 2011), the transcriptomic evidence favors neither pathway in T. triangulare. It cannot be excluded that the low levels of CAM photosynthesis detected in T. triangulare (Winter and Holtum, 2014) are supported by multiple decarboxylation routes. Ideally our findings on the transcript levels of NAD(P)-ME would be confirmed by enzyme activity data, but so far all biochemical assays to measure activity in leaf material failed due to a suppression effect derived from an unknown constitute of the T. triangulare plant extract (Supplemental Fig. S10). Experimental differentiation of the biochemical pathway of malate decarboxylation in the light during CAM in T. triangulare will be part of future work.
PPDK was strongly upregulated during CAM-induction in T. triangulare (Fig. 3). In M. crystallinum, the plastidic location of PPDK was demonstrated by measurements of enzyme acvitity combined with studies of pyruvate metabolism of intact isolated chloroplasts (Winter et al., 1982; Holtum et al., 2005). However, immunogold labeling of PPDK in the CAM species Kalanchoë blossfeldiana (Kondo et al., 2001) also raises the possibility of cytosolic PEP regeneration. In support of this hypothesis, the contig Tt26901, which encodes PPDK with a truncated N terminus, showed markedly higher transcript amounts on days 9 and 12 compared to day 0 and compared to the longer contig, which encodes for a putative target peptide. The contigs possibly represent splicing variants of the same gene and encode for PPDKs dually targeted to the cytosol (Tt26901, truncated N terminus) or the chloroplast (Tt24575) as found in Arabidopsis (Parsley and Hibberd, 2006). In C4 species (NADP-ME and NAD-ME type), where the regeneration of PEP is invariably localized in the plastids, the transporters for pyruvate import (i.e. BASS2 and NHD1) and PEP export (i.e. PPT) are increased in expression compared to C3 sister plants (Bräutigam et al., 2008, 2014; Furumoto et al., 2011). Interestingly, while BASS2 was upregulated but lowly expressed at midnight, transcripts encoding transporters of the PPT family were found to be downregulated during the C3-CAM switch in T. triangulare (Supplemental Dataset S3). This contrasts with M. crystallinum where transcriptional induction of PPT was detected (Kore-eda et al., 2005). Transcripts matching the NHD1 gene were not detected. The absence of a target peptide in the major contig encoding for PPDK suggests that concomitant with the missing NHD1 and downregulated PPT, T. triangulare employs cytosolic PPDK. In addition to PPDK, both isoforms of PPDK regulatory protein, which promote circadian regulation of PPDK activity by dark-phased phosphorolytic inactivation (Dever et al., 2015), were slightly but significantly induced during CAM in T. triangulare (Supplemental Dataset S3).
During the evolution of C4, the change of a single amino acid residue in the N-terminal region of PEPC, namely from Ala to Ser, was shown to increase the enzyme’s kinetic efficiency compared to C3 sister plants (Svensson et al., 1997; Bläsing et al., 2000; Svensson et al., 2003; Gowik et al., 2006). It was recently shown that in C4 and CAM species within the Caryophyllales, to which the Talinaceae belong, the isogenes of a recurrently recruited gene lineage encoding PEPC1, namely ppc1-E1c, almost exclusively encodes the C4-like Ser-780 (Christin et al., 2014). In the same study, in response to reduced irrigation, the expression of ppc1-E1c was upregulated at night in Portulaca oleracea, a C4 species with facultative CAM (Christin et al., 2014). The two PPC contigs, Tt9871_4 and Tt9871_6, that exhibit the C4-like Ser-780 (Supplemental Fig. S2A; Supplemental Table S3) indeed cluster together with ppc1-E1c isoforms of Talinaceae and P. oleracea (Supplemental Fig. S2B). However, highest transcript levels and the most drastic elevation in abundance during CAM induction was found for contig Tt63271, which encodes for the C3-like Ala-780 and clusters in the ppc1-E1b clade. Low-level CAM photosynthesis apparently employs different PEPC genes compared to C4 photosynthesis or fully expressed CAM photosynthesis. At night, PEPC is activated through phosphorylation via PEPC kinase (Carter et al., 1991), the activity of which is regulated transcriptionally through a circadian oscillator (Hartwell et al., 1996, 1999; Borland et al., 1999; Taybi et al., 2000). Phosphorylated PEPC is less sensitive to feedback inhibition by malate (Hartwell et al., 1996; Taybi et al., 2000, 2004). Consistent with the circadian regulation and strong upregulation of PEPC kinase at night during CAM induction in M. crystallinum (Hartwell et al., 1999; Cushman et al., 2008b; Dever et al., 2015), PEPC kinase was most markedly upregulated at midnight during CAM in T. triangulare (expression rank 300 on day 12, Supplemental Dataset S3), enabling increased PEPC activity at night. In contrast to C4 plants, PEPC of CAM plants needs to be efficiently switched off during the light through feedback inhibition by malate exported from the vacuole to prevent futile cycling of CO2 (Borland et al., 1999). All of the PEPC isoforms identified as upregulated in Talinum carry an Arg residue at position 890 (Supplemental Fig. S2; Supplemental Table S3; Paulus et al., 2013) and are thus likely feedback inhibited both by malate export during the day and once vacuolar storage capacity is exceeded during the night.
The upregulation of CAM-cycle genes was accompanied by elevated malate and citrate levels during the night and reduced malate levels during the day (Fig. 4) in agreement with increase in nocturnal acidity (Fig. 1). Photorespiration is not inactivated during the decarboxylation phase of CAM (Niewiadomska and Borland, 2007; Lüttge, 2011), and we observed a mostly stable expression of photorespiratory genes during the C3-CAM transition (Supplemental Table S7). Reduced steady-state metabolite pools of glycerate and glycolate (Supplemental Fig. S4), as well as the reduced sugar levels, were likely a consequence of reduced overall photosynthetic rate. Nocturnal carbon assimilation rates during CAM were estimated to be reduced to approx. 5% of CO2 fixation in the light in the C3 mode (Winter and Holtum, 2014). Hence, the associated metabolite fluxes during CAM are expected to be low.
The carboxylation reaction during the night depends on degrading starch for PEP production. The degradation can be either phosphorolytic or hydrolytic (Häusler et al., 2000; Weise et al., 2011). Some, but not all, transcripts encoding these two alternate starch breakdown pathways were upregulated (Fig. 3). Phosphorolytic starch breakdown preserves some of the energy released during breakage of the glycosidic bonds in starch and therefore consumes less ATP for activation of the released Glc residues. ATP, ADP, and/or free phosphate may balance hydrolytic and phosphorolytic starch breakdown depending on the ATP demand of the cytosol. Using starch and malate contents at the end of the night and the end of the day, the amount of starch turnover was calculated as sufficient to support the level of photosynthesis during CAM (Supplemental Table S8). Efficient starch turnover apparently required elevated expression of both starch synthesis and starch degradation genes, as well as sufficient branching of the starch to allow rapid degradation (Fig. 3). Based on strong upregulation of GPT, which was among the top 50 upregulated genes at midday, and the upregulation of MEX1 (Fig. 3), glycolysis likely operated in the cytosol, which was confirmed by the transcriptional upregulation of all but one of the cytosolic glycolytic enzymes (Fig. 3). Taken together, inducible CAM required upregulation of key CAM-cycle genes, elevation of starch turnover by changes in selected genes and increased expression of cytosolic glycolytic genes. While starch turnover capacity is increased (Fig. 3), the amount of starch is reduced (Fig. 4), which may be explained by lowered photosynthetic capacity during CAM (Winter and Holtum, 2014). In contrast to T. triangulare, CAM can substantially contribute to carbon gain in M. crystallinum (Winter and Holtum, 2007), and following salinity-induced CAM, starch accumulation is increased and directly correlates with acidification (Haider et al., 2012).
The rapid induction of CAM-related gene expression and the coexpression of ABA-related signaling networks upon drought stress points to a direct connection to the ABA signaling network as shown for Kalanchoë blossfeldiana (Taybi et al., 1995) and M. crystallinum (Taybi and Cushman, 2002). Control of CAM induction may be exerted either directly by ABA-mediated phosphorylation and activation of transcription factors or by downstream effectors (Table IV).
Drought Response and Carbon Resource Utilization
Beyond CAM photosynthesis, additional processes contribute to the adaptation of T. triangulare to drought, such as a classical drought response and management of carbon resources. Water limitation leads to stomatal closure during the day (Herrera et al., 1991; Winter and Holtum, 2014), which reduces the photosynthetic carbon assimilation and hence pool sizes of soluble sugars such as Suc, Glc, and Fru during the day and the night (Fig. 4). Under severe drought stress, part of these remaining soluble sugars may act as compatible solutes, retain water in the cells and protect cellular structures against damage. In contrast to the overall loss of these sugars, substantial accumulation of raffinose was observed (Fig. 4) likely triggered by transcriptional upregulation of raffinose biosynthesis (Supplemental Fig. S6; Table I). During cold acclimation in Arabidopsis (Nägele and Heyer, 2013) and Ajuga reptans (Findling et al., 2015), nonaqueous fractionations localized raffinose to variable extents within the vacuole, cytosol and chloroplast stroma. At this point we do not know whether raffinose accumulating in drought-stressed T. triangulare functions as a compatible solute in the cytosol, supports osmotic pressure maintenance in the vacuole, or has protective functions for membranes and/or proteins within the plastid or cytosol (Nishizawa et al., 2008). Although the gene encoding PYRROLINE-5-CARBOXYLATE SYNTHASE (P5CS), which catalyzes the committed step for Pro synthesis, was upregulated during CAM (expression rank 57 on day 12 at midday, Supplemental Dataset S1), Pro accumulated only in some, but not all, plants tested (Supplemental Fig. S5). Similarly, levels of the sugar alcohols glycerol, mannitol, myoinositol, and sorbitol, which could potentially act as compatible solutes, were not significantly altered during drought, although mannitol accumulated in individual plants (Supplemental Figs. S5 and S11). The discrepancy between transcript and metabolite measurements is likely based on the fact that the transcriptome precedes the metabolome. Although genes were upregulated in all replicates (Supplemental Dataset S1), the metabolic status may be delayed.
Stomatal closure reduces the consumption of absorbed light energy in photosynthesis, potentially leading to short-term redox stress. Among the top 50 upregulated genes, four were involved in light stress, e.g. the gene of ELIP1 (Hutin et al., 2003), which is ABA induced in Arabidopsis (Table I). Genes encoding components of light stress-induced oxidation of fatty acids in the peroxisomes were enriched in cluster 3, representing induced genes that do not quite reach well-watered levels after rewatering (Fig. 5). Thus, despite leaf-angle changes and leaf rolling (Taisma and Herrera, 2003; Fig. 1) that reduce light exposure, a transcriptional upregulation of light-protection mechanisms is seemingly required to aid leaf functioning. Additionally, the increase of raffinose, shown to act as a scavenger of hydroxyl radicals in Arabidopsis (Nishizawa et al., 2008), might play a role in supporting leaf functioning.
Under drought, increased use of resources such as fatty acids and proteins was induced (Fig. 5), which might explain the slight increase in free amino acids (Noctor and Foyer, 2000; Supplemental Fig. S11). Catabolism of fatty acids through peroxisomal processes was induced by day 9 already, maintained on day 12, and only partially reversed upon rewatering. Genes involved in processes related to mRNA turnover were also increased during drought (Fig. 5, cluster 6). These induced catabolic processes were, however, not associated with permanent damage as the transcriptome returned to prestress status upon rewatering (Fig. 2). Most genes encoding anabolic processes such as photosynthesis and amino acid biosynthesis and transport were transcriptionally downregulated (Fig. 5; Supplemental Fig. S6), but returned to prestress levels after rewatering. For example, cell wall biosynthesis was downregulated and completely reversed upon rewatering (Fig. 5, clusters 8 and 9; Supplemental Fig. S6). Finally, genes involved in replication and the cell cycle were among the most downregulated genes overall on day 12 (Table II), and processes related to replication and ribosome biogenesis were among those, which were downregulated upon drought but did not return to well-watered levels after rewatering (Fig. 5, cluster 7). Apparently, after 2 d, the expression levels of genes involved in primary production are increased however transcript levels of growth-related genes (e.g. genes involved in replication, translation, and cell division) have not yet completely recovered. This might be due to the fact that the full resource availability has not been restored within two days reflected by the still altered starch and metabolite pools (Fig. 2).
Regulation of the Coordinated Drought Response
Clustering differential gene expression indicated that the response of T. triangulare is multilayered since clusters respond and recover at different time points to different degrees (Fig. 3). Initially, drought stress induced a general stress response (clusters 4 and 9, Fig. 5) and a specific ABA-mediated response (cluster 5, Fig. 5; Supplemental Fig. S8). A conceptual model integrating the transcriptomic data is depicted in Figure 6. The general stress response included the CBFs, which are known to bind the drought responsive element (Zhu, 2002). However, their upregulation was transient and no longer detectable on day 12 of water limitation (Supplemental Dataset S6). During this initial stress response, our data indicate changes of the endomembrane system (cluster 9, Fig. 5). In contrast, the ABA-mediated response was further increased on day 12 (cluster 5, Fig. 5). The transcriptomic response also indicated a feed-forward loop with higher expression of KING1, which activates ABA signaling, and EDL3, which mediates ABA signaling (Table I). At the same time, the expression of a PP2C was increased (expression rank 424 on day 12 at midnight, Supplemental Dataset S1), which is required to attenuate the signaling pathway once the stress signal is removed (Umezawa et al., 2009). Hence, T. triangulare plants are transcriptionally enabled to both sustain the ABA signal to the transcriptome and also to switch it off once the stress is removed upon rewatering. The ABA signal is likely overlaid with a carbon starvation signal. This additional signal may explain why several clusters did not recover to well-watered expression levels although soil moisture was restored. Catabolism-related processes only recovered partially (Fig. 5, clusters 3 and 5), and replication and ribosome biogenesis related genes recovered only marginally (Fig. 5, cluster 7).
Figure 6.
Proposed regulation of the coordinated drought response in T. triangulare. Shown are genes discussed in the text. A general stress signal was mediated by transcription factors, which were transiently upregulated on day 9 (CBF2, C-repeat/DRE binding factor 2; CBF3, C-repeat/DRE binding factor 3; ERF8, ethylene response factor 8; RRTF1, redox responsive transcription factor 1; STZ, salt tolerance zinc finger) and triggered a yet unclear response. The sustained response to persistent drought was primarily mediated via ABA, the activity of which was maintained through a feed-forward loop (KING1, SNF1-related protein kinase regulatory subunit gamma 1; EDL3, EID1-like 3), through ABA-responsive transcription factors (SOM, SOMNUS; HB-7, homeobox 7; NAP, NAC-like, activated by AP3/PI; NF-YA1, nuclear factor Y subunit A1; NF-YA9, nuclear factor Y subunit A9) and besides the induction of CAM (see Fig. 3 for details; PPC1, phosphoenolpyruvate carboxylase 1; PPDK, Pyruvate phosphate dikinase, PEP/pyruvate binding domain; NADP-ME1, NADP-malic enzyme 1) included a classical drought response, i.e. synthesis of compatible solutes (RS1, Raffinose synthase 1; MIOX1, myoinositol oxygenase 1; OXS3, oxidative stress 3) and light protection (ELIP1, Early light inducible protein 1; FTSH6, FTSH protease 6). At the same time, growth was repressed through downregulation of the replication cassette (MCMs, minichromosome maintenance; ICU2, incurvata 2) and processes related to photosynthesis, amino acid synthesis, and cell wall synthesis (see Fig. 5 and Supplemental Fig. S4) allowing T. triangulare to engage in a state of quiescence during drought. Log2-FC, Absolute maximum log2-fold changes in gene expression compared to day 0 at midday or midnight.
The candidate transcription factors downstream of the initial signal, which is probably at least in part ABA mediated, are those that are significantly upregulated with a large magnitude. They include highly induced ABA-dependent transcription factors, such as HB7, NF-YA1 and 9, or NAP (Table IV), as well as the phosphorylation-controlled ABA-dependent transcription factors not detectable in the transcriptome analysis. Stalled growth, which is part of the response, may also at least in part be dependent on ABA as SOM is highly induced upon drought (Table I; Lim et al., 2013). SOM is required to control germination cessation in seeds downstream of the signal perception cascade (Kim et al., 2008) and thus likely controls effector genes required for growth. The trehalose-phosphate and SnRK1 mediated signaling (Baena-Gonzalez et al., 2007) does not significantly overlap with the signals observed in T. triangulare.
CONCLUSION
This study showed that the response of T. triangulare to drought stress was associated with profound and reversible changes in the transcriptome. Known key enzymes of the CAM-cycle and some of those connected to carbohydrate metabolism were upregulated, as were genes encoding functions associated with a classical drought response, including light protection and the synthesis of compatible solutes. Genes related to growth, photosynthesis, the synthesis of cell walls, and amino acids were downregulated in response to drought while catabolic processes were upregulated. A set of transcription factors likely to be involved in mediating these responses was identified. T. triangulare is easy to grow, flowers and produces seeds within few weeks after germination, and rapidly induces CAM in response to drought stress, thereby exhibiting important attributes of a potential model system to study key components of a drought response that includes CAM and its signaling components under stress. Research on T. triangulare may in the future contribute to engineering drought coping strategies such as CAM via synthetic biology into C3 crop plants with the goal to improve their performance in water-limited environments (Yang et al., 2015).
MATERIALS AND METHODS
Plant Material and Growth Conditions
Talinum triangulare plants were grown in Miracle-Gro Potting Mix (Miracle-Gro) in “Short-One” treepots, 1.6 l (Stuewe and Sons). The experiment was initiated with 28-d-old plants in a controlled environment chamber (Environmental Growth Chambers) maintained under 12 h light (30°C, 37% relative humidity)/12 h dark (22°C) cycles. Photon flux density at leaf level was 425 μmol m−2 s−1. Irrigation was withheld on day 1 and recommenced on day 14. Leaves were harvested when plants were well-watered as well as after 4, 9, and 12 d of water deprivation and watered for two days following the drought period.
Titratable Acidity
Mature leaves of the same plants used for RNASeq and metabolite profiling were harvested at the end of the light and dark periods, respectively, and frozen in liquid nitrogen after measurement of fresh weight. Organic acids were extracted by sequentially boiling leaves in 50% methanol and water. Titratable acidity was determined by measuring the volume of 5 mm KOH required to neutralize the aqueous extract to pH 6.5.
Metabolite Profiling
Lyophilized leaf material was extracted for metabolite analysis by gas chromatography-mass spectrometry (GC-MS) according to Fiehn et al. (2000) using a 7200 GC-QTOF (Agilent). Data analysis was conducted with the Mass Hunter Software (Agilent). For relative quantification, all metabolite peak areas were normalized to the peak area of the internal standard ribitol added prior to extraction.
RNA Extraction, Preparation, and Sequencing of Illumina Libraries
The topmost mature unshaded leaves (of approximately 3–4.5 cm length) of T. triangulare were harvested in the middle of the light or the middle of the dark period and immediately frozen in liquid nitrogen. RNA was isolated from ground tissue using the GeneMatrix Universal RNA Purification Kit (EURx Ltd.). Residues of DNA were removed with DNase (New England Biolabs). RNA integrity, sequencing library, and fragment size were analyzed on a 2100 Bioanalyzer (Agilent). Libraries were prepared using the TruSeq RNA Sample Prep Kit v2 (Illumina) and quantified with a Qubit 2.0 (Invitrogen). Samples were multiplexed with 12 libraries per lane and sequenced in single-end mode (Rapid Run, 150 bp read length) on an Illumina HiSEquation 2000 platform, yielding ∼14 million reads per library.
Contig Assembly and Sequence Alignment
To substantiate cross-species mapping against the sequenced reference genome of Arabidopsis, Illumina libraries from representative samples (day 0, day 9, and 2 d after rewatering) were pooled and paired-end sequenced (Rapid Run, 100 bp read length) on an Illumina HiSEquation 2000 platform for assembly of contigs. Trinity assembly of the reads using Trinity (Grabherr et al., 2011) in default mode yielded 105,520 contigs, which were collapsed to 39,781 putative open reading frames using Transdecoder (transdecoder.github.io) and manual removal of duplicated contigs. Remapping placed 81% of the original 150bp reads on the ORFs. Contigs were annotated via BlastX implemented in BLAST+ (Camacho et al., 2009) against peptide sequences of the Beta vulgaris reference genome RefBeet-1.1 (Dohm et al., 2014) and peptide sequences of a minimal set of the TAIR10 release of the Arabidopsis genome (http://www.arabidopsis.org/). Protein sequences were aligned using Clustal Omega at www.ebi.ac.uk (Sievers et al., 2011). Phylogenetic analyses were performed with the PhyML tool (Guindon et al., 2010) implemented in SeaView 4 (Gouy et al., 2010) in default mode after Gblocks creation with the least stringent parameters.
Read Mapping and Gene Expression Profiling
Illumina reads were aligned to a minimal set of coding sequences of the TAIR 10 release of the Arabidopsis genome (http://www.arabidopsis.org/) using BLAT (Kent, 2002) in protein space. The minimal transcriptome was obtained as described in Gowik et al. (2011) and contains 21,869 nuclear encoded protein-coding genes. The best BLAT hit for each read was determined by (1) lowest e-value and (2) highest bit score. Raw read counts were transformed to reads per million (rpm) to normalize for the number of reads available at each sampling stage. Cross-species mapping takes advantage of the completeness and annotation of the Arabidopsis genome and overcomes the limitations of transcriptome assembly (Franssen et al., 2011; Schliesky et al., 2012). To quantify abundance of contigs, the 150 bp single-end reads were mapped onto the assembled contigs using the RNA-Seq analysis tool implemented in the CLC genomics workbench version 8.5 (www.clcbio.com) in default mode.
Data Analysis
Data analysis was performed using R statistics software (R version 3.2.1 provided by the CRAN project, http://www.R-project.org). Differential gene expression was analyzed with the DESeq2 package (Love et al., 2014) in default mode. Automatic independent filtering retained 16,766 genes with nonzero total read count. A significance threshold of 0.01 was applied after p-value adjustment with false discovery rate via Benjamini-Hochberg correction (Benjamini and Hochberg, 1995). Log2 expression ratios calculated by DESeq2 were used for downstream analysis. For k-means clustering, transcript levels were log2-transformed and scaled via calculation of z-scores by gene. One k-means clustering was performed for each daytime of sampling (MD and MN) by R statistics software with 6,800 (MD) and 7,563 (MN) genes that were determined as differentially expressed genes (DEG) compared to well-watered (day 0), respectively. Enrichment for GO terms of biological processes within k-means clusters was performed via Fisher’s Exact Test implemented in the R-package topGO (Alexa et al., 2006; including DEG as the background set, significance level of alpha = 0.01). All log2-fold changes of water-limited days compared to day 0 were used for pathway analysis in Mapman (Thimm et al., 2004). Wilcoxon Rank Sum test implemented in Mapman was used to test for enrichment and corrected for multiple hypothesis testing via Benjamini-Hochberg correction (Benjamini and Hochberg, 1995). Arabidopsis response to treatments with the phytohormones ABA, 1-aminocyclopropane-1-carboxylic acid (a precursor of ethylene), brassinolide (a brassinosteroid), indole-3-acetic acid (i.e. auxin), and zeatin (a cytokinine) were extracted from AtGenExpress and are based on microarrays by Goda et al. (2008). Response to hormone treatment (n = 3) compared to mock treatment (n = 3) was called significant at q < 0.01 (corrected after Benjamini and Hochberg [1995]). Overlap of T. triangulare reponse to drought with Arabidopsis response to hormone treatments was calculated using Fisher’s Exact Test (p < 0.001). All log2-fold changes (treatment vs. mock) of Arabidopsis response to ABA treatment (Goda et al., 2008) were used for pathway analysis in Mapman. Data were annotated with publicly available information for Arabidopsis (TAIR10, http://www.arabidopsis.org/). Full quantitative data with annotation is available as Supplemental Dataset S1.
Accession Number
The read data have been submitted to the National Center for Biotechnology Information Gene Expression Omnibus under accession number GSE70601 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?token=wryvugcifxexnml&acc=GSE70601).
Supplemental Data
The following supplemental materials are available.
Supplemental Data Set S1. Quantitative information and annotation for all reads mapped onto the reference genome of Arabidopsis.
Supplemental Data Set S2. Quantitative information and annotation for all reads mapped onto the Talinum contigs.
Supplemental Data Set S3. Levels of all putative CAM genes sensu stricto and sensu lato.
Supplemental Data Set S4. Enriched Mapman categories for all plants harvested 4, 9, and 12 d after water deprivation and 2 d after rewatering (re2) at the middle of the day (MD) or night (MN).
Supplemental Data Set S5. Full list of enriched GO terms in k-means clusters (Fig. 5; Supplemental Fig. S7).
Supplemental Data Set S6. Transcription factors differentially expressed on day 9 and/or day 12.
Supplemental Figure S1. Changes in leaf transcriptomes and metabolomes under varying levels of water availability in the middle of the night (analogous to Fig. 1, B and C).
Supplemental Figure S2. T. triangulare PEPCs show characteristics of both C3 and C4 plants.
Supplemental Figure S3. Talinum NADP-ME with highest transcript are predicted to be localized in the plastid.
Supplemental Figure S4. Levels of photorespiratory metabolites.
Supplemental Figure S5. Levels of compatible solutes.
Supplemental Figure S6. Mapman overview of metabolism for all drought samples (day 4, day 9, and day 12 of water deprivation and 2 d after rewatering [re2] compared to day 0).
Supplemental Figure S7. K-means clustering of relative gene expression at the middle of the night.
Supplemental Figure S8. Overlap of differentially expressed genes during CAM in T. triangulare with hormone treated leaves of Arabidopsis (Goda et al., 2008).
Supplemental Figure S9. Mapman overview of metabolism for ABA response in Arabidopsis.
Supplemental Figure S10. Experiments to assay activity of NADP-ME (panels 1–4) or MDH (panel 5) with leaf extracts of T. triangulare (“Talinum extract”) or recombinant NADP-ME from Z. mays (ZmNADP-ME, provided by Anastasiia Bovdilova, group of Veronica G. Maurino).
Supplemental Figure S11. Levels of additional metabolites.
Supplemental Table S1. mRNA-Seq statistics.
Supplemental Table S2. Metabolite profiling statistics.
Supplemental Table S3. Expression levels of T. triangulare contigs encoding PEPC.
Supplemental Table S4. Expression levels of all T. triangulare contigs encoding for subunits of vacuolar ATP synthases.
Supplemental Table S5. Expression levels of T. triangulare contigs encoding NADP-ME.
Supplemental Table S6. Expression levels of two T. triangulare contigs encoding PPDK.
Supplemental Table S7. Expression levels of genes encoding for photorespiratory enzymes.
Supplemental Table S8. Estimated partitioning of carbon in organic acids and starch on day 12.
Supplementary Material
Acknowledgments
We acknowledge the excellent technical assistance of M. Graf, K. Weber, and E. Klemp for GC-MS measurements, S. Kurz for help with Illumina library preparation, and A. Virgo for titration of organic acids.
Glossary
- CAM
Crassulacean acid metabolism
- ABA
abscisic acid
- PEPC
PHOSPHOENOLPYRUVATE CARBOXYLASE
- DEG
differentially expressed gene
- PEP
phosphoenolpyruvate
- NADP-ME
NADP-MALIC ENZYME
- NAD-ME
NAD-MALIC ENZYME
- PPDK
PYRUVATE, ORTHOPHOSPHATE DIKINASE
- CA
CARBONIC ANHYDRASE
- OAA
oxaloacetate
- MDH
MALATE DEHYDROGENASE
- ALMT
ALUMINUM-ACTIVATED MALATE TRANSPORTER
- PEPCK
PEP CARBOXYKINASE
- PPT
PEP/PHOSPHATE TRANSLOCATOR
- GO
gene ontology
- GC-MS
gas chromatography-mass spectrometry
Footnotes
This work was supported by the International Graduate Program for Plant Science (iGRAD-Plant; to D.B.) and grants from Deutsche Forschungsgemeinschaft (EXC 1028; IRTG 1525 to A.P.M.W.). K.W. was supported by the Smithsonian Tropical Research Institute.
This article is available without a subscription.
References
- Alexa A, Rahnenführer J, Lengauer T (2006) Improved scoring of functional groups from gene expression data by decorrelating GO graph structure. Bioinformatics 22: 1600–1607 [DOI] [PubMed] [Google Scholar]
- Ashton AR, Burnell JN, Furbank RT, Jenkins CLD, Hatch MD (1990) 3 enzymes of C4 photosynthesis. Enzymes of Primary Metabolism 3: 39 [Google Scholar]
- Baena-González E, Rolland F, Thevelein JM, Sheen J (2007) A central integrator of transcription networks in plant stress and energy signalling. Nature 448: 938–942 [DOI] [PubMed] [Google Scholar]
- Bartels D, Sunkar R (2005) Drought and salt tolerance in plants. Crit Rev Plant Sci 24: 23–58 [Google Scholar]
- Benjamini Y, Hochberg Y (1995) Controlling the false discovery rate - a practical and powerful approach to multiple testing. J R Stat Soc, B 57: 289–300 [Google Scholar]
- Bläsing OE, Westhoff P, Svensson P (2000) Evolution of C4 phosphoenolpyruvate carboxylase in Flaveria, a conserved serine residue in the carboxyl-terminal part of the enzyme is a major determinant for C4-specific characteristics. J Biol Chem 275: 27917–27923 [DOI] [PubMed] [Google Scholar]
- Bohnert HJ, Nelson DE, Jensen RG (1995) Adaptations to environmental stresses. Plant Cell 7: 1099–1111 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Borland AM, Barrera Zambrano VA, Ceusters J, Shorrock K (2011) The photosynthetic plasticity of crassulacean acid metabolism: an evolutionary innovation for sustainable productivity in a changing world. New Phytol 191: 619–633 [DOI] [PubMed] [Google Scholar]
- Borland AM, Hartwell J, Jenkins GI, Wilkins MB, Nimmo HG (1999) Metabolite control overrides circadian regulation of phosphoenolpyruvate carboxylase kinase and CO2 fixation in crassulacean acid metabolism. Plant Physiol 121: 889–896 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bray EA. (1997) Plant responses to water deficit. Trends Plant Sci 2: 48–54 [Google Scholar]
- Bräutigam A, Hoffmann-Benning S, Weber AP, Weber APM (2008) Comparative proteomics of chloroplast envelopes from C3 and C4 plants reveals specific adaptations of the plastid envelope to C4 photosynthesis and candidate proteins required for maintaining C4 metabolite fluxes. Plant Physiol 148: 568–579 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bräutigam A, Kajala K, Wullenweber J, Sommer M, Gagneul D, Weber KL, Carr KM, Gowik U, Mass J, Lercher MJ, et al. (2011) An mRNA blueprint for C4 photosynthesis derived from comparative transcriptomics of closely related C3 and C4 species. Plant Physiol 155: 142–156 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bräutigam A, Schliesky S, Külahoglu C, Osborne CP, Weber APM (2014) Towards an integrative model of C4 photosynthetic subtypes: insights from comparative transcriptome analysis of NAD-ME, NADP-ME, and PEP-CK C4 species. J Exp Bot 65: 3579–3593 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Camacho C, Coulouris G, Avagyan V, Ma N, Papadopoulos J, Bealer K, Madden TL (2009) BLAST+: architecture and applications. BMC Bioinformatics 10: 421. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Carter PJ, Nimmo HG, Fewson CA, Wilkins MB (1991) Circadian rhythms in the activity of a plant protein kinase. EMBO J 10: 2063–2068 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chaves MM, Maroco J, Pereira JS (2003) Understanding plant responses to drought - from genes to the whole plant. Funct Plant Biol 30: 239–264 [DOI] [PubMed] [Google Scholar]
- Cheffings CM, Pantoja O, Ashcroft FM, Smith JA (1997) Malate transport and vacuolar ion channels in CAM plants. J Exp Bot 48: 623–631 [DOI] [PubMed] [Google Scholar]
- Chia T, Thorneycroft D, Chapple A, Messerli G, Chen J, Zeeman SC, Smith SM, Smith AM (2004) A cytosolic glucosyltransferase is required for conversion of starch to sucrose in Arabidopsis leaves at night. Plant J 37: 853–863 [DOI] [PubMed] [Google Scholar]
- Christin P-A, Arakaki M, Osborne CP, Bräutigam A, Sage RF, Hibberd JM, Kelly S, Covshoff S, Wong GK-S, Hancock L, et al. (2014) Shared origins of a key enzyme during the evolution of C4 and CAM metabolism. J Exp Bot 65: 3609–3621 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cushman JC, Agarie S, Albion RL, Elliot SM, Taybi T, Borland AM (2008a) Isolation and characterization of mutants of common ice plant deficient in crassulacean acid metabolism. Plant Physiol 147: 228–238 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cushman JC, Tillett RL, Wood JA, Branco JM, Schlauch KA (2008b) Large-scale mRNA expression profiling in the common ice plant, Mesembryanthemum crystallinum, performing C3 photosynthesis and Crassulacean acid metabolism (CAM). J Exp Bot 59: 1875–1894 [DOI] [PubMed] [Google Scholar]
- Cutler SR, Rodriguez PL, Finkelstein RR, Abrams SR (2010) Abscisic acid: emergence of a core signaling network. Annu Rev Plant Biol 61: 651–679 [DOI] [PubMed] [Google Scholar]
- Dever LV, Boxall SF, Kneřová J, Hartwell J (2015) Transgenic perturbation of the decarboxylation phase of Crassulacean acid metabolism alters physiology and metabolism but has only a small effect on growth. Plant Physiol 167: 44–59 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Dittrich P. (1976) Nicotinamide adenine dinucleotide-specific “malic” enzyme in Kalanchoë daigremontiana and other plants exhibiting crassulacean acid metabolism. Plant Physiol 57: 310–314 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Dohm JC, Minoche AE, Holtgräwe D, Capella-Gutiérrez S, Zakrzewski F, Tafer H, Rupp O, Sörensen TR, Stracke R, Reinhardt R, et al. (2014) The genome of the recently domesticated crop plant sugar beet (Beta vulgaris). Nature 505: 546–549 [DOI] [PubMed] [Google Scholar]
- Elbein AD, Pan YT, Pastuszak I, Carroll D (2003) New insights on trehalose: a multifunctional molecule. Glycobiology 13: 17R–27R [DOI] [PubMed] [Google Scholar]
- Emanuelsson O, Nielsen H, Brunak S, von Heijne G (2000) Predicting subcellular localization of proteins based on their N-terminal amino acid sequence. J Mol Biol 300: 1005–1016 [DOI] [PubMed] [Google Scholar]
- Fiehn O, Kopka J, Dörmann P, Altmann T, Trethewey RN, Willmitzer L (2000) Metabolite profiling for plant functional genomics. Nat Biotechnol 18: 1157–1161 [DOI] [PubMed] [Google Scholar]
- Findling S, Zanger K, Krueger S, Lohaus G (2015) Subcellular distribution of raffinose oligosaccharides and other metabolites in summer and winter leaves of Ajuga reptans (Lamiaceae). Planta 241: 229–241 [DOI] [PubMed] [Google Scholar]
- Flexas J, Bota J, Galmés J, Medrano H, Ribas-Carbó M (2006) Keeping a positive carbon balance under adverse conditions: responses of photosynthesis and respiration to water stress. Physiol Plant 127: 343–352 [Google Scholar]
- Franssen SU, Shrestha RP, Bräutigam A, Bornberg-Bauer E, Weber APM (2011) Comprehensive transcriptome analysis of the highly complex Pisum sativum genome using next generation sequencing. BMC Genomics 12: 227–242 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Furumoto T, Yamaguchi T, Ohshima-Ichie Y, Nakamura M, Tsuchida-Iwata Y, Shimamura M, Ohnishi J, Hata S, Gowik U, Westhoff P, et al. (2011) A plastidial sodium-dependent pyruvate transporter. Nature 476: 472–475 [DOI] [PubMed] [Google Scholar]
- Goda H, Sasaki E, Akiyama K, Maruyama-Nakashita A, Nakabayashi K, Li W, Ogawa M, Yamauchi Y, Preston J, Aoki K, et al. (2008) The AtGenExpress hormone and chemical treatment data set: experimental design, data evaluation, model data analysis and data access. Plant J 55: 526–542 [DOI] [PubMed] [Google Scholar]
- Gouy M, Guindon S, Gascuel O (2010) SeaView version 4: A multiplatform graphical user interface for sequence alignment and phylogenetic tree building. Mol Biol Evol 27: 221–224 [DOI] [PubMed] [Google Scholar]
- Gowik U, Bräutigam A, Weber KL, Weber APM, Westhoff P (2011) Evolution of C4 photosynthesis in the genus Flaveria: how many and which genes does it take to make C4? Plant Cell 23: 2087–2105 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gowik U, Engelmann S, Bläsing OE, Raghavendra AS, Westhoff P (2006) Evolution of C(4) phosphoenolpyruvate carboxylase in the genus Alternanthera: gene families and the enzymatic characteristics of the C(4) isozyme and its orthologues in C(3) and C(3)/C(4) Alternantheras. Planta 223: 359–368 [DOI] [PubMed] [Google Scholar]
- Grabherr MG, Haas BJ, Yassour M, Levin JZ, Thompson DA, Amit I, Adiconis X, Fan L, Raychowdhury R, Zeng Q, et al. (2011) Full-length transcriptome assembly from RNA-Seq data without a reference genome. Nat Biotechnol 29: 644–652 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Guindon S, Dufayard J-F, Lefort V, Anisimova M, Hordijk W, Gascuel O (2010) New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of PhyML 3.0. Syst Biol 59: 307–321 [DOI] [PubMed] [Google Scholar]
- Haider MS, Barnes JD, Cushman JC, Borland AM (2012) A CAM- and starch-deficient mutant of the facultative CAM species Mesembryanthemum crystallinum reconciles sink demands by repartitioning carbon during acclimation to salinity. J Exp Bot 63: 1985–1996 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hare PD, Cress WA, Van Staden J (1998) Dissecting the roles of osmolyte accumulation during stress. Plant Cell Environ 21: 535–553 [Google Scholar]
- Harris D, Martin C (1991) Plasticity in the degree of CAM-cycling and its relationship to drought stress in 5 species of Talinum (Portulacaceae). Oecologia 86: 575–584 [DOI] [PubMed] [Google Scholar]
- Hartwell J, Gill A, Nimmo GA, Wilkins MB, Jenkins GI, Nimmo HG (1999) Phosphoenolpyruvate carboxylase kinase is a novel protein kinase regulated at the level of expression. Plant J 20: 333–342 [DOI] [PubMed] [Google Scholar]
- Hartwell J, Smith LH, Wilkins MB, Jenkins GI, Nimmo HG (1996) Higher plant phosphoenolpyruvate carboxylase kinase is regulated at the level of translatable mRNA in response to light or a circadian rhythm. Plant J 10: 1071–1078 [Google Scholar]
- Häusler RE, Baur B, Scharte J, Teichmann T, Eicks M, Fischer KL, Flügge U-I, Schubert S, Weber A, Fischer K (2000) Plastidic metabolite transporters and their physiological functions in the inducible crassulacean acid metabolism plant Mesembryanthemum crystallinum. Plant J 24: 285–296 [DOI] [PubMed] [Google Scholar]
- Heddad M, Adamska I (2000) Light stress-regulated two-helix proteins in Arabidopsis thaliana related to the chlorophyll a/b-binding gene family. Proc Natl Acad Sci USA 97: 3741–3746 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Herrera A. (1999) Effects of photoperiod and drought on the induction of Crassulacean acid metabolism and the reproduction of plants of Talinum triangulare. Can J Bot 77: 404–409 [Google Scholar]
- Herrera A. (2009) Crassulacean acid metabolism and fitness under water deficit stress: if not for carbon gain, what is facultative CAM good for? Ann Bot (Lond) 103: 645–653 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Herrera A, Ballestrini C, Montes E (2015) What is the potential for dark CO2 fixation in the facultative crassulacean acid metabolism species Talinum triangulare? J Plant Physiol 174: 55–61 [DOI] [PubMed] [Google Scholar]
- Herrera A, Delgado J, Paraguatey I (1991) Occurrence of inducible crassulacean acid metabolism in leaves of Talinum triangulare (Portulacaceae). J Exp Bot 42: 493–499 [Google Scholar]
- Holtum JAM, Smith JAC, Neuhaus HE (2005) Intracellular transport and pathways of carbon flow in plants with crassulacean acid metabolism. Funct Plant Biol 32: 429–449 [DOI] [PubMed] [Google Scholar]
- Holtum JAM, Winter K (1982) Activity of enzymes of carbon metabolism during the induction of Crassulacean acid metabolism in Mesembryanthemum crystallinum L. Planta 155: 8–16 [DOI] [PubMed] [Google Scholar]
- Hutin C, Nussaume L, Moise N, Moya I, Kloppstech K, Havaux M (2003) Early light-induced proteins protect Arabidopsis from photooxidative stress. Proc Natl Acad Sci USA 100: 4921–4926 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ingram J, Bartels D (1996) The molecular basis of dehydration tolerance in plants. Annu Rev Plant Physiol Plant Mol Biol 47: 377–403 [DOI] [PubMed] [Google Scholar]
- Kent WJ. (2002) BLAT - The BLAST-like alignment tool. Genome Res 12: 656–664 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kim DH, Yamaguchi S, Lim S, Oh E, Park J, Hanada A, Kamiya Y, Choi G (2008) SOMNUS, a CCCH-type zinc finger protein in Arabidopsis, negatively regulates light-dependent seed germination downstream of PIL5. Plant Cell 20: 1260–1277 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kluge M, Osmond CB (1971) Pyruvate Pi dikinase in crassulacean acid metabolism. Naturwissenschaften 58: 414–4155110404 [Google Scholar]
- Kollist H, Nuhkat M, Roelfsema MRG (2014) Closing gaps: linking elements that control stomatal movement. New Phytol 203: 44–62 [DOI] [PubMed] [Google Scholar]
- Kondo A, Nose A, Ueno O (2001) Coordinated accumulation of the chloroplastic and cytosolic pyruvate, Pi dikinases with enhanced expression of CAM in Kalanchoë blossfeldiana. Physiol Plant 111: 116–122 [Google Scholar]
- Kore-eda S, Noake C, Ohishi M, Ohnishi J, Cushman JC (2005) Transcriptional profiles of organellar metabolite transporters during induction of Crassulacean acid metabolism in Mesembryanthemum crystallinum. Funct Plant Biol 32: 451–466 [DOI] [PubMed] [Google Scholar]
- Lawlor DW, Cornic G (2002) Photosynthetic carbon assimilation and associated metabolism in relation to water deficits in higher plants. Plant Cell Environ 25: 275–294 [DOI] [PubMed] [Google Scholar]
- Lim S, Park J, Lee N, Jeong J, Toh S, Watanabe A, Kim J, Kang H, Kim DH, Kawakami N, et al. (2013) ABA-insensitive3, ABA-insensitive5, and DELLAs interact to activate the expression of SOMNUS and other high-temperature-inducible genes in imbibed seeds in Arabidopsis. Plant Cell 25: 4863–4878 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Liu Q, Kasuga M, Sakuma Y, Abe H, Miura S, Yamaguchi-Shinozaki K, Shinozaki K (1998) Two transcription factors, DREB1 and DREB2, with an EREBP/AP2 DNA binding domain separate two cellular signal transduction pathways in drought- and low-temperature-responsive gene expression, respectively, in Arabidopsis. Plant Cell 10: 1391–1406 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lobell DB, Gourdji SM (2012) The influence of climate change on global crop productivity. Plant Physiol 160: 1686–1697 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Love MI, Huber W, Anders S (2014) Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 15: 550–570 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lüttge U. (2002) CO2-concentrating: consequences in crassulacean acid metabolism. J Exp Bot 53: 2131–2142 [DOI] [PubMed] [Google Scholar]
- Lüttge U. (2011) Photorespiration in phase III of crassulacean acid metabolism: evolutionary and ecophysiological implications. Prog Bot 72: 371–384 [Google Scholar]
- Nägele T, Heyer AG (2013) Approximating subcellular organisation of carbohydrate metabolism during cold acclimation in different natural accessions of Arabidopsis thaliana. New Phytol 198: 777–787 [DOI] [PubMed] [Google Scholar]
- Niewiadomska E, Borland AM (2007) Crassulacean acid metabolism: a cause or consequence of oxidative stress in planta? Prog Bot 69: 247–266 [Google Scholar]
- Niittylä T, Messerli G, Trevisan M, Chen J, Smith AM, Zeeman SC (2004) A previously unknown maltose transporter essential for starch degradation in leaves. Science 303: 87–89 [DOI] [PubMed] [Google Scholar]
- Nishizawa A, Yabuta Y, Shigeoka S (2008) Galactinol and raffinose constitute a novel function to protect plants from oxidative damage. Plant Physiol 147: 1251–1263 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Noctor G, Foyer CH (2000) Homeostasis of adenylate status during photosynthesis in a fluctuating environment. J Exp Bot 51: 347–356 [DOI] [PubMed] [Google Scholar]
- Osmond CB. (1978) Crassulacean acid metabolism: a curiosity in context. Annu Rev Plant Physiol 29: 379–414 [Google Scholar]
- Park S-Y, Fung P, Nishimura N, Jensen DR, Fujii H, Zhao Y, Lumba S, Santiago J, Rodrigues A, Chow T-FF, et al. (2009) Abscisic acid inhibits type 2C protein phosphatases via the PYR/PYL family of START proteins. Science 324: 1068–1071 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Parsley K, Hibberd JM (2006) The Arabidopsis PPDK gene is transcribed from two promoters to produce differentially expressed transcripts responsible for cytosolic and plastidic proteins. Plant Mol Biol 62: 339–349 [DOI] [PubMed] [Google Scholar]
- Paulus JK, Schlieper D, Groth G (2013) Greater efficiency of photosynthetic carbon fixation due to single amino-acid substitution. Nat Commun 4: 1518–1524 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Pérez-Rodríguez P, Riaño-Pachón DM, Corrêa LGG, Rensing SA, Kersten B, Mueller-Roeber B (2010) PlnTFDB: updated content and new features of the plant transcription factor database. Nucleic Acids Res 38: D822–D827 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sage RF. (2003) The evolution of C4 photosynthesis. New Phytol 161: 341–370 [DOI] [PubMed] [Google Scholar]
- Sakamoto H, Araki T, Meshi T, Iwabuchi M (2000) Expression of a subset of the Arabidopsis Cys(2)/His(2)-type zinc-finger protein gene family under water stress. Gene 248: 23–32 [DOI] [PubMed] [Google Scholar]
- Sakamoto H, Maruyama K, Sakuma Y, Meshi T, Iwabuchi M, Shinozaki K, Yamaguchi-Shinozaki K (2004) Arabidopsis Cys2/His2-type zinc-finger proteins function as transcription repressors under drought, cold, and high-salinity stress conditions. Plant Physiol 136: 2734–2746 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Schliesky S, Gowik U, Weber APM, Bräutigam A (2012) RNA-Seq assembly - Are we there yet? Front Plant Sci 3: 220. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sievers F, Wilm A, Dineen D, Gibson TJ, Karplus K, Li W, Lopez R, McWilliam H, Remmert M, Söding J, et al. (2011) Fast, scalable generation of high-quality protein multiple sequence alignments using Clustal Omega. Mol Syst Biol 7: 539–544 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Smith SM, Fulton DC, Chia T, Thorneycroft D, Chapple A, Dunstan H, Hylton C, Zeeman SC, Smith AM (2004) Diurnal changes in the transcriptome encoding enzymes of starch metabolism provide evidence for both transcriptional and posttranscriptional regulation of starch metabolism in Arabidopsis leaves. Plant Physiol 136: 2687–2699 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Streb S, Zeeman SC (2012) Starch metabolism in Arabidopsis. The Arabidopsis Book 10: doi/10.199/tab.e0160 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sun J, Jiang H, Xu Y, Li H, Wu X, Xie Q, Li C (2007) The CCCH-type zinc finger proteins AtSZF1 and AtSZF2 regulate salt stress responses in Arabidopsis. Plant Cell Physiol 48: 1148–1158 [DOI] [PubMed] [Google Scholar]
- Svensson P, Bläsing OE, Westhoff P (1997) Evolution of the enzymatic characteristics of C4 phosphoenolpyruvate carboxylase--a comparison of the orthologous PPCA phosphoenolpyruvate carboxylases of Flaveria trinervia (C4) and Flaveria pringlei (C3). Eur J Biochem 246: 452–460 [DOI] [PubMed] [Google Scholar]
- Svensson P, Bläsing OE, Westhoff P (2003) Evolution of C4 phosphoenolpyruvate carboxylase. Arch Biochem Biophys 414: 180–188 [DOI] [PubMed] [Google Scholar]
- Taisma MA, Herrera A (1998) A relationship between fecundity, survival, and the operation of crassulacean acid metabolism in Talinum triangulare. Can J Bot 76: 1908–1915 [Google Scholar]
- Taisma MA, Herrera A (2003) Drought under natural conditions affects leaf properties, induces CAM and promotes reproduction in plants of Talinum triangulare. Interciencia 28: 292–297 [Google Scholar]
- Taybi T, Cushman JC (2002) Abscisic acid signaling and protein synthesis requirements for phosphoenolpyruvate carboxylase transcript induction in the common ice plant. J Plant Physiol 159: 1235–1243 [Google Scholar]
- Taybi T, Nimmo HG, Borland AM (2004) Expression of phosphoenolpyruvate carboxylase and phosphoenolpyruvate carboxylase kinase genes. Implications for genotypic capacity and phenotypic plasticity in the expression of crassulacean acid metabolism. Plant Physiol 135: 587–598 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Taybi T, Patil S, Chollet R, Cushman JC (2000) A minimal serine/threonine protein kinase circadianly regulates phosphoenolpyruvate carboxylase activity in crassulacean acid metabolism-induced leaves of the common ice plant. Plant Physiol 123: 1471–1482 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Taybi T, Sotta B, Gehrig H, Guclu S, Kluge M, Brulfert J (1995) Differential effects of abscisic acid on phosphoenolpyruvate carboxylase and CAM operation in Kalanchoë blossfeldiana. Bot Acta 108: 240–246 [Google Scholar]
- Thimm O, Bläsing O, Gibon Y, Nagel A, Meyer S, Krüger P, Selbig J, Müller LA, Rhee SY, Stitt M (2004) MAPMAN: a user-driven tool to display genomics data sets onto diagrams of metabolic pathways and other biological processes. Plant J 37: 914–939 [DOI] [PubMed] [Google Scholar]
- Tsuzuki M, Miyachi S, Winter K, Edwards GE (1982) Localization of carbonic anhydrase in crassulacean acid metabolism plants. Plant Sci Lett 24: 211–218 [Google Scholar]
- Umezawa T, Sugiyama N, Mizoguchi M, Hayashi S, Myouga F, Yamaguchi-Shinozaki K, Ishihama Y, Hirayama T, Shinozaki K (2009) Type 2C protein phosphatases directly regulate abscisic acid-activated protein kinases in Arabidopsis. Proc Natl Acad Sci USA 106: 17588–17593 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Weber A, Servaites JC, Geiger DR, Kofler H, Hille D, Gröner F, Hebbeker U, Flügge U-I (2000) Identification, purification, and molecular cloning of a putative plastidic glucose translocator. Plant Cell 12: 787–802 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Weise SE, van Wijk KJ, Sharkey TD (2011) The role of transitory starch in C(3), CAM, and C(4) metabolism and opportunities for engineering leaf starch accumulation. J Exp Bot 62: 3109–3118 [DOI] [PubMed] [Google Scholar]
- White PJ, Smith JA (1989) Proton and anion transport at the tonoplast in crassulacean-acid-metabolism plants: specificity of the malate-influx system in Kalanchoë daigremontiana. Planta 179: 265–274 [DOI] [PubMed] [Google Scholar]
- Wilson RL, Kim H, Bakshi A, Binder BM (2014) The ethylene receptors ETHYLENE RESPONSE1 and ETHYLENE RESPONSE2 have contrasting roles in seed germination of Arabidopsis during salt stress. Plant Physiol 165: 1353–1366 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Winter K, Foster JG, Edwards GE, Holtum JAM (1982) Intracellular localization of enzymes of carbon metabolism in Mesembryanthemum crystallinum exhibiting C3 photosynthesic characteristics or performing Crassulacean acid metabolism. Plant Physiol 69: 300–307 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Winter K, Garcia M, Holtum JAM (2008) On the nature of facultative and constitutive CAM: environmental and developmental control of CAM expression during early growth of Clusia, Kalanchöe, and Opuntia. J Exp Bot 59: 1829–1840 [DOI] [PubMed] [Google Scholar]
- Winter K, Holtum JAM (2007) Environment or development? Lifetime net CO2 exchange and control of the expression of Crassulacean acid metabolism in Mesembryanthemum crystallinum. Plant Physiol 143: 98–107 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Winter K, Holtum JAM (2014) Facultative crassulacean acid metabolism (CAM) plants: powerful tools for unravelling the functional elements of CAM photosynthesis. J Exp Bot 65: 3425–3441 [DOI] [PubMed] [Google Scholar]
- Winter K, Smith JAC (1996) An introduction to crassulacean acid metabolism. Biochemical principles and ecological diversity. In Crassulacean Acid Metabolism. Springer, Berlin Germany, pp 1–13 [Google Scholar]
- Winter K, von Willert DJ (1972) NaCl-induzierter Crassulaceansäure-stoffwechsel bei Mesembryanthemum crystallinum. Z Pflanzenphysiol 67: 166–170 [Google Scholar]
- Yang J, Worley E, Udvardi M (2014) A NAP-AAO3 regulatory module promotes chlorophyll degradation via ABA biosynthesis in Arabidopsis leaves. Plant Cell 26: 4862–4874 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Yang X, Cushman JC, Borland AM, Edwards EJ, Wullschleger SD, Tuskan GA, Owen NA, Griffiths H, Smith JAC, De Paoli HC, et al. (2015) A roadmap for research on crassulacean acid metabolism (CAM) to enhance sustainable food and bioenergy production in a hotter, drier world. New Phytol 207: 491–504 [DOI] [PubMed] [Google Scholar]
- Zelisko A, García-Lorenzo M, Jackowski G, Jansson S, Funk C (2005) AtFtsH6 is involved in the degradation of the light-harvesting complex II during high-light acclimation and senescence. Proc Natl Acad Sci USA 102: 13699–13704 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zhu J-K. (2002) Salt and drought stress signal transduction in plants. Annu Rev Plant Biol 53: 247–273 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zou C, Sun K, Mackaluso JD, Seddon AE, Jin R, Thomashow MF, Shiu S-H (2011) Cis-regulatory code of stress-responsive transcription in Arabidopsis thaliana. Proc Natl Acad Sci USA 108: 14992–14997 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.