Plant Mol Biol (2014) 85:429–441 DOI 10.1007/s11103-014-0195-8

Transgenic Arabidopsis expressing osmolyte glycine betaine synthesizing enzymes from halophilic methanogen promote tolerance to drought and salt stress Shu‑Jung Lai · Mei‑Chin Lai · Ren‑Jye Lee · Yu‑Hsuan Chen · Hungchen Emilie Yen 

Received: 2 November 2013 / Accepted: 17 April 2014 / Published online: 7 May 2014 © Springer Science+Business Media Dordrecht 2014

Abstract  Glycine betaine (betaine) has the highest cellular osmoprotective efficiency which does not accumulate in most glycophytes. The biosynthetic pathway for betaine in higher plants is derived from the oxidation of low-accumulating metabolite choline that limiting the ability of most plants to produce betaine. Halophilic methanoarchaeon Methanohalophilus portucalensis FDF1T is a model anaerobic methanogen to study the acclimation of water-deficit stresses which de novo synthesize betaine by the stepwise methylation of glycine, catalyzed by glycine sarcosine N-methyltransferase (GSMT) and sarcosine dimethylglycine N-methyltransferase. In this report, genes encoding these betaine biosynthesizing enzymes, Mpgsmt and Mpsdmt, were introduced into Arabidopsis. The homozygous Mpgsmt (G), Mpsdmt (S), and their cross, Mpgsmt and Mpsdmt (G × S) plants showed increased accumulation of betaine. Water loss from detached leaves was slower in G, S, and G × S lines than wild-type (WT). Pot-grown

transgenic plants showed better growth than WT after 9 days of withholding water or irrigating with 300 mM NaCl. G, S, G × S lines also maintained higher relative water content and photosystem II activity than WT under salt stress. This suggests heterologously expressed Mpgsmt and Mpsdmt could enhance tolerance to drought and salt stress in Arabidopsis. We also found a twofold increase in quaternary ammonium compounds in salt-stressed leaves of G lines, presumably due to the activation of GSMT activity by high salinity. This study demonstrates that introducing stress-activated enzymes is a way of avoiding the divergence of primary metabolites under normal growing conditions, while also providing protection in stressful environments.

MC Lai responsible for clone and seed stock requests. HE Yen responsible for contents and experimental procedures in this article.

Abbreviations GSMT Glycine sarcosine N-methyltransferase LC/MS Liquid chromatograph–mass spectrometry QAC Quaternary ammonium compound RWC Relative water content RWL Rate of water loss SDMT Sarcosine dimethylglycine N-methyltransferase

Electronic supplementary material  The online version of this article (doi:10.1007/s11103-014-0195-8) contains supplementary material, which is available to authorized users. S.-J. Lai · M.-C. Lai (*) · Y.-H. Chen · H. E. Yen (*)  Department of Life Sciences, National Chung Hsing University, Taichung 40227, Taiwan e-mail: [email protected] H. E. Yen e-mail: [email protected] R.-J. Lee  Instrument Center and Department of Chemistry, National Chung Hsing University, Taichung 40227, Taiwan

Keywords  Halophilic methanoarchaea · Glycine betaine · Glycine sarcosine N-methyltransferase (GSMT) · Sarcosine dimethylglycine N-methyltransferase (SDMT) · Salt tolerance · Transgenic Arabidopsis

Introduction Plants are frequently exposed to salt, drought, heat, chilling, and freezing conditions, which all lead to water-deficit stress. Both a limited water supply and exposure to a hyperosmotic environment result in cellular dehydration.

13

430

A high salinity environment causes ion imbalance and hyperosmotic stress, which is one of the major problems affecting the world’s overall agricultural production (Hasegawa et al. 2000). Accumulation of low molecular weight organic compatible solutes (osmolytes) such as sugar alcohols, amino acid derivatives, and quaternary ammonium compounds (QAC) are a major strategy to overcome osmotic stress (Kempf and Bremer 1998; Rhodes and Hanson 1993). Many drought- or salt-tolerant plants actively accumulate compatible solutes to promote water flow into the cells, a process called osmotic adjustment (Flowers and Colmer 2008). Plants that can adjust osmotically under water-deficit stress typically show a slower rate of water loss (RWL), and maintain a high relative water content (RWC) compared to those that cannot. Compatible solutes also provide protection to offset toxic ions and free radicals. A QAC, such as glycine betaine (betaine), is a common osmolyte that plays a crucial function as an osmoprotectant in microorganisms (Roberts 2000; Roesser and Muller 2001; Sleator and Hill 2001), and plants (Chen and Murata 2008, 2011). The zwitterionic nature of betaine allows it to form hydration shells around macromolecules to stabilize their structures (Auton et al. 2011; Sleator and Hill 2001). Betaine protects cellular proteins/protein complexes from stress-induced inactivation (Allakhverdiev et al. 2003), and exogenous application of betaine to plants enhances their tolerance to abiotic stresses (Ma et al. 2006). All members of Chenopodiaceae, a plant family that thrives in arid and saline habitats, synthesize betaine from choline in a two-step oxidation process (Hanson et al. 1985). Many investigations report on enhanced tolerance to abiotic stresses by transforming genes that encode key enzymes in the biosynthesis of glycine betaine through the choline-oxidation pathway (for review, Chen and Murata 2008). Heterologous expression of bacterial genes encoding choline oxidase (Ahmad et al. 2008; Goel et al. 2011) and choline dehydrogenase (Quan et al. 2004) improves salt and drought tolerance in potato, tomato, and maize. Choline, the precursor to the choline-oxidation pathway, is not an abundant metabolite in plants (Rontein et al. 2002). The accumulation of betaine in transgenic Arabidopsis, Brassica napus, and tobacco expressing the bacterial choline oxidase gene, was found to be limited by the availability of choline (Huang et al. 2000). According to the results, the levels of betaine in transgenic plants are still far lower than the levels in betaine synthesizing halophytes. Nevertheless, a few-fold increase in betaine content is sufficient to enhance tolerance to water-deficit stresses. Another betaine biosynthesis pathway shown in Fig. 1 is derived from the three-step methylation of glycine to form sarcosine, N, N-dimethylglycine and betaine via glycine sarcosine N-methyltransferase (GSMT) and sarcosine

13

Plant Mol Biol (2014) 85:429–441

dimethylglycine N-methyltransferase (SDMT) or dimethylglycine N-methyltransferase (DMT), respectively (Lai et al. 1999; Roberts et al. 1992). This betaine de novo biosynthesis pathway was characterized in halophilic methanoarchaeon Methanohalophilus portucalensis FDF1T (Chen et al. 2009; Lai et al. 2006; Lai and Lai 2011) and halophilic bacteria Actinopolyspora halophila, Halorhodospira halochloris, Aphanothece halophytica, Synechococcus sp. WH8102 and Myxococcus xanthus (Kimura et al. 2010; Lu et al. 2006; Nyyssölä et al. 2000, 2001; Waditee et al. 2003). Waditee et al. (2005) co-expressed GSMT/DMT from A. halophytica in Arabidopsis and found these plants accumulated higher concentrations of betaine, and exhibited higher salinity tolerance than plants expressing choline-oxidizing enzymes. Similar results were observed in cyanobacteria Synechococcus (Waditee et al. 2005) and Anabaena (Singh et al. 2013; WaditeeSirisattha et al. 2012). Glycine is an abundant amino acid synthesized via the glycolate pathway in algae and higher plants. The glycolate pathway is a salvage pathway for 2-phosphoglycolate produced by oxygenase activity of Rubisco, a process called photorespiration. C3 plants have a high rate of photorespiration under water-deficit stress, and glycine can serve as a precursor for the synthesis of glutathione, which is involved in stress protection (Wingler et al. 2000). Based on carbon flow under water-deficit stress, glycine may have an advantage over choline as the precursor for genetically engineered betaine biosynthesis in C3 plants. We previously characterized complete gene cluster encoding genes for two betaine de novo biosynthesis enzymes GSMT and SDMT from halophilic methanoarchaeon M. portucalensis, and heterologously expressed them into E. coli. In vitro methyltransferase assays showed dramatic activating effects of sodium and potassium ions on the activity of recombinant MpGSMT. Conformational changes of substrate binding sites, and the formation of dimeric forms are two possible factors that contribute to the activation of MpGSMT (Lai and Lai 2011). On the contrary, the presence of Na+ and K+ inhibited the activity of GSMT in halotolerant cyanobacterium A. halophytica (Waditee et al. 2003). Together, with the fact that the activity of recombinant MpSDMT was not regulated by Na+, K+ or the end product betaine, we suggest that de novo betaine biosynthesis is regulated mainly at the first committed step of the pathway in M. portucalensis. Therefore, the de novo betaine synthesis is salt-induced and energyeconomic in this methanoarchaeon. In this study, GSMT and SDMT from M. portucalensis were heterologously expressed in Arabidopsis. The improved drought and salt tolerance of transgenic Arabidopsis presented in this study suggest that it is feasible to apply archaeal stress-tolerant traits to higher plants.

Plant Mol Biol (2014) 85:429–441

431

Fig. 1  De novo betaine biosynthesis pathway a Enzymes catalyze betaine biosynthesis from glycine. GNMT glycine N-methyltransferase, GSMT glycine sarcosine N-methyltransferase, SDMT sarcosine dimethylglycine N-methyltransferase, DMT dimethylglycine N-methyltransferase. b LC/MS detection of standard solution containing 2 μg mL−1 of glycine (Gly), sarcosine (Sar), dimethylglycine (Dim), and betaine (Bet). The retention time (RT) and mass-to-charge ratio (m/z) is indicated on the top of each mass spectrum. c LC/MS detection of products catalyzed by recombinant MpSDMT. The activity of purified recombinant MpSDMT was assayed with either sarcosine (left panel) or dimethylglycine (right panel) as the substrate. Metabolite levels in the assay were calculated using peak areas and expressed as μg mL−1

Materials and methods Growth conditions and stress treatments Sterilized seeds of Arabidopsis thaliana (ecotype Columbia) were germinated in half-strength Murashige and Skoog (MS) agar medium in a walk-in growth room with a cycle of 14 h of light and 10 h dark at 22 °C with 50–60 % relative humidity. One-week-old seedlings were transferred to sterilized mixed soil (peat moss:perlite:vermiculite =  8:1:1; or peat moss: commercial potting soil (No. 2, King Root plant medium, Taiwan) = 2:1 and maintained in the same growing conditions until seed set. Unless otherwise noted, 0.5 × HYPONeX #2 fertilizer (major components 4 % ammonium, 4 % nitrate, 20 % potassium and 20 % phosphorus pentoxide in 1000x stock) was applied twice a week. For stress treatments during the seedling stage, the surface-sterilized seeds were placed on half-strength MS agar medium containing 0 or 100 mM NaCl, incubated in the

growth chamber, and seedling growth was observed for 2 weeks. For salt stress treatment during the adult stage, 6-week-old plants were irrigated with 20 mL of 150 mM or 300 mM NaCl solution daily for 9 days, while control plants were irrigated with 20 mL of deionized water daily. For drought stress treatment, the water supply was withheld for 9 days in 6-week-old plants maintained at 50–60 % relative humidity. Construction of transgenic Arabidopsis expressing Mpgsmt and Mpsdmt Open reading frames of gsmt (HQ634197) and sdmt (HQ634198) from M. portucalensis (Lai and Lai 2011) were cloned into the BamHI site of the binary vector pBI121 to be in-frame fused with β-glucuronidase gene (GUS). Plasmid pBI121-Mpgsmt and pBI121-Mpsdmt were then respectively introduced into Agrobacterium tumefaciens GV3101 via electroporation. Arabidopsis plants were transformed with A. tumefaciens carrying pBI121-Mpgsmt

13

432

or pBI121-Mpsdmt using the floral-dip method (Clough and Bent, 1998). The seeds of self-pollinated transformants (T0) were harvested and placed on half-strength MS agar medium containing 50 μg mL−1 kanamycin to select transgenic lines (T1). The insertion of the transgene in three independent gsmt lines and six sdmt lines were confirmed by genomic PCR. Genomic DNA was extracted from 3-week-old plants using rapid plant genomic DNA isolation for PCR protocol according to Weigel and Glazebrook (2002). The insertion of the transgene was confirmed with primer pairs of Mpgsmt_BamHI-f: 5′-GGA TCC ATG AAC CAA TAC GGA AAA-3′ and Mpgsmt-r (500-523): 5′-AAT GGG CAT GCC TTG TCT CCA ATA-3′ to amplify a partial Mpgsmt fragment (530 bp), with primer pairs of Mpsdmt_NdeI-f: 5′-CAT ATG ATG TCT GAA AAC CAA AAA AC-3′ and Mpsdmt_XhoI-r: 5′-CTC GAG TTT TTT ACG TAA ATG GAA-3′ to amplify the full-length Mpsdmt (837 bp), respectively. Homozygous transgenic plants (T3 and T4 generation) were used for stress-tolerant analysis. For production of transgenic Arabidopsis expressing both gsmt and sdmt, homozygous gsmt T3 lines G’2_5_5 and G’2_5_6 were used as male parents and homozygous sdmt T4 lines S’1_3_35_2 and S’1_3_37_1 were used as female parents. The crossed plants remained growing for the future harvest of hybrid seeds. Transgenic plants expressing both gsmt and sdmt (G × S) were confirmed by RT-PCR and tested for stress tolerance. Reverse transcriptase polymerase chain reaction (RT‑PCR) Total RNA of 3-week-old wild-type (WT), homozygous gsmt (G) or sdmt (S), and G × S transgenic Arabidopsis leaves (0.05 g) were extracted by TRI reagent® (Applied Biosystems, Life Technologies, USA) as described previously (Lai and Lai 2011). Expression of gsmt and sdmt in transgenic Arabidopsis were detected by RT-PCR with primers Mpgsmt-forward (221–242): 5′-CGG TTC GTT TAC TTC AGG CAG G-3′ and Mpgsmt-reverse (449–470): 5′-TTC AGC AGC GCA TAG AAC TCA G-3′ for amplifying Mpgsmt (250 bp); and primers Mpsdmt-f (72–93): 5′-TTA CTT TAC CAT CTG GGG CGG C-3′ and Mpsdmtr (266–287): 5′-TTA AGA GCA ACG ACC TGA CAC C-3′ for amplifying Mpsdmt (215 bp). Expression of UBQ10 was used as an internal control with primers UBQ10-f: 5′-CTG CGT CTT CGT GGT GGT TTC TA-3′ and UBQ10r: 5′-GTC GAG TCA CTT TGC AGG CGT ATT A-3′. One μg total RNA was mixed with 1 μM corresponding reverse primers and incubated at 70 °C for 10 min, to denature the secondary structure of the RNA template. cDNA synthesis reactions were carried out for 1 h at 42 °C in 20 μL ImProm-II™ reaction buffer, 3 mM MgCl2, 0.5 mM of each dNTP, template mixtures as described above, and one reaction of ImProm-II™ reverse transcriptase (Promega, USA).

13

Plant Mol Biol (2014) 85:429–441

The reaction was stopped by incubation at 95 °C for 5 min. The PCR reactions with a final volume of 50 μL contained 1X PCR buffer (10 mM Tris–HCl pH 8.3, 50 mM KCl and 1.5 mM MgCl2), 0.4 mM dNTP, 1 μM forward and reverse primers as described above, 1 μL cDNA template, and 2.5 units of Takara Taq™ DNA polymerase (TAKARA BIO Inc., Japan). The PCR reaction was initiated at 94 °C for 5 min followed by 20 cycles of denaturation at 94 °C for 30 s, annealing at 50 °C for 30 s, and extension at 72 °C for 15 s. The RT-PCR products were analyzed by 1.5 % TAEagarose gel electrophoresis. GUS staining GUS staining of transgenic seedlings was used to detect heterologously expressed MpGSMT-GUS and MpSDMT-GUS proteins (Naleway, 1992). Two-week-old seedlings from WT, G, S and G × S transgenic Arabidopsis were treated with solution containing 2 mM X-gluc (5-bromo-4-chloro3-indolyl-beta-d-glucuronic acid), 50 mM Na2HPO4, and 0.1 % Triton X-100 at 37 °C for 24 h. Seedlings were then transferred to 90 % ethanol and incubated in the dark at room temperature for 24 h to discolor the chlorophyll. Stained seedlings were examined and photographed. Western blot Total protein was extracted from 10-day-old seedlings, separated by 10 % SDS-PAGE and transferred to PVDF membrane. The PVDF membrane was incubated with blocking solution (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 2 mM KH2PO4, 0.05 % Tween-20, 5 % nonfat milk, pH 7.4) at room temperature for 1 h. The PVDF membrane was further displaced to the solution containing anti-GUS antibody (1:2,500, Sigma) and incubated at room temperature for 1 h. After three 15 min washings with PBST buffer (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 2 mM KH2PO4, 0.05 % Tween-20, pH 7.4), the PVDF membrane was incubated with anti-rabbit-IgG antibody conjugated with horseradish peroxidase (1:2,500, Sigma) at room temperature for 1 h. Signals were detected by incubation with Western Lightning™ Chemiluminescence Reagent Plus (PerkinElmer, USA) for 5 min and scanned by Image Quant LAS 4000 mini system (GE Healthcare, USA). Yield of photosystem II For salt- and drought-stress treatments, the yield of photosystem II was measured with a Fluorescence Monitoring System (Hansatech, UK) according to Krall and Edwards (1992). During 9 days of treatment with salt stress, the yields of photosystem II (Fv/Fm’) was measured 3–4 h after watering with salt solution, under ambient light, in

Plant Mol Biol (2014) 85:429–441

the growth room. Each value was an average of at least 15 measurements from three individual plants. The Fm’ value was determined by a saturating flash for the duration of 1 s and defined as light-adapted maximal fluorescence; Fv was the variable component of fluorescence. Relative water content (RWC) and rate of water loss (RWL) The 6-week-old fully expanded leaves were used for measurement of RWC and RWL. For measuring RWC, leaves from WT, G, S, and G × S lines were cut and their fresh weight (FW) was determined. The leaves were soaked in water for 2 h and their turgid weight (TW) was determined. The leaves were dried at 37 °C for 24 h and their dry weight (DW) was measured. RWC of leaves was calculated by the equation RWC (%) = [(FW−DW)/(TW−DW)]  × 100 % (Smart and Bingham 1974). The RWC was taken as an average from triplicate measurements of each plant line. For measuring RWL, three leaves from each plant line were cut, immediately weighed as a fresh weight of 0 (FW0) and set as 100 %. Detached leaves were then placed on a bench under laboratory conditions and weighed at various time points (FWt). The remaining water content was expressed as (FWt/FW0) × 100 %. The RWL was determined by the slope of changes in FW over the first 4 h. Colorimetric method for betaine determination Total amounts of betaine were measured by a modified colorimetric method (Grieve and Grattan 1983; Lai et al. 1995). Total extracts of the WT and the three transgenic lines were obtained from salt-treated leaves of 6-week-old Arabidopsis. Samples were dried at 65 °C for 1 day and ground with mortar and pestle. Dried, finely ground plant materials (50 mg) were shaken with 1 mL deionized water at 25 °C for 24 h. Total extracts were harvested by centrifugation with 11,300g at 25 °C for 5 min. Twenty-five microliters of the total extracts were diluted 1:1 with 2 N H2SO4 and cooled in ice water for 1 h. Betaine in the extracts were precipitated as periodide crystals with 20 μL of cold KI-I2 solution at 4 °C for 16 h. The supernatant was discarded after centrifugation with 10,000g at 0 °C for 15 min. Periodide crystals were dissolved in 900 μL of 1,2-dichloroethane for 2.5 h at room temperature and the absorbance was measured at 365 nm. A standard curve was established using a betaine concentration of 2.5–15.0 μg. Each sample was measured at least in triplicate. Liquid chromatograph–mass spectrometry Reaction products sarcosine, dimethylglycine, and betaine of recombinant MpSDMT were detected by liquid

433

chromatograph/mass spectrometry (LC/MS). An enzyme assay was carried out as described by Lai and Lai (2011). Reaction products were de-salted and vacuum-dried by SpeedVac (SC100, Savant). The samples were dissolved in distilled water and then separated with Shodex ODP2 HP-2D (2.1 × 150 mm, 5 μm) LC column for LC/MS analysis. The isocratic elution condition was carried out under mobile A (0.1 % formic acid in water)/mobile B (methanol) = 9/1 (v/v) with 0.12 mL min−1 flow rate for 15 min at 40 °C. The sample analysis was performed on an Accela liquid chromatographic system (Thermo Scientific, San José, USA) coupled to LTQ-Orbitrap XL MS (Thermo Scientific Bremen, Germany). Retention time of standard glycine, sarcosine, dimethylglycine, and betaine were approximately 2.8–2.9 min. Identical metabolites were confirmed by comparison with the retention times and mass spectra of authentic compounds. Metabolite levels were quantified by referring to the peak area in the specific chromatogram to that of the corresponding standard. The amount of glycine, sarcosine, dimethylglycine, and betaine were quantified with 2 μg mL−1 of standard solution. The accumulation of glycine, sarcosine, dimethylglycine, and betaine in WT and transgenic lines were detected via LC/MS. Leaves of pot-grown 6-week-old plants were collected for osmolyte extraction. Metabolites were extracted from dried, well-ground samples as described in a previous section. Betaine was detected by high resolution LC/MS with an exact molecular mass of 118.0863 dalton to eliminate isomer interference.

Results Identification of intermediates in the betaine de novo biosynthetic pathway Halophilic methanoarchaeon M. portucalensis de novo synthesizes betaine through successive methylation of glycine, catalyzed by S-adenosylmethionine (SAM)-dependent GSMT and SDMT (Fig. 1a). Thin-layer Chromatography was previously used to identify intermediates and products of MpGSMT and MpSDMT in M. portucalensis (Lai et al. 1999), and E. coli (Lai and Lai 2011). Given the need to identify these chemically related compounds in complex plant matrices in subsequent experiments, a method based on LC/MS was developed. Standard solutions containing equal molarities of glycine, sarcosine, dimethylglycine, and betaine were analyzed. Four compounds with molecular masses of 76, 90, 104, and 118 daltons were detected, corresponding to glycine, sarcosine, dimethylglycine, and betaine, respectively, at a retention time of 2.8–2.9 min (Fig. 1b). We further tested this system using reaction products of a methyltransferase assay catalyzed by purified

13

434

Plant Mol Biol (2014) 85:429–441

Fig. 2  Characterization of G, S, and G × S transgenic lines. a Confirmation of T-DNA insertion of Mpgsmt and Mpsdmt by genomic PCR. Genomic DNA was isolated from three gsmt (lane 1–3: G′2_1, G′2_4, G′2_5) and six sdmt (lane 1–6: S′1_5, S’1_6, S′1_7, S′1_2, S’1_3, S′1_4) T2 lines as a template for PCR. Arrows indicate the positions of expected PCR products of gsmt and sdmt as 550 bp and 834 bp, respectively. b RT-PCR analysis of gene expression of Mpgsmt and Mpsdmt in 3-week-old Arabidopsis leaves. Arrows indicate the positions of expected products of gsmt and sdmt as 250 and 215 bp, respectively. The expression of UBQ10 was used as an internal control. c Accumulation of GSMT-GUS and SDMT-GUS in

transgenic Arabidopsis. Western blot analysis of total protein probed with anti-GUS antibody. The arrow indicates the position of GSMTGUS and SDMT-GUS, both with a molecular mass of approximately 100 kDa. Coomassie blue-stained gel was used as a loading control (bottom). d GUS staining of G, S, and G × S plant lines. Two-weekold seedlings of transgenic lines were stained for GUS activity. Blue coloration was detected in all parts of plant. Gus staining for homozygous G and S lines and G × S line were performed by two separate experiments. e High-resolution LC/MS detection of betaine in WT, G, S, and G × S lines. Betaine levels were expressed as μmol g−1DW

MpSDMT. When sarcosine was used as the substrate, products dimethylglycine and betaine were detected (Fig. 1c). The amount of product was estimated to be about one-tenth of the amount of the substrate (dimethylglycine/sarcosine or betaine/dimethylglycine). When dimethylglycine was used as the substrate, the product betaine was detected. The substrate to product ratio was also 10:1 (Fig. 1c). These results show that LC/MS is applicable for identification and quantification of three methylated metabolites derived from glycine.

promoter, using a kanamycin resistant gene as a selection marker. The insertion of this transgene in three Mpgsmt and six Mpsdmt kanamycin-resistant T2 lines were confirmed by genomic PCR (Fig. 2a). T2 seeds with a 3:1 (kan resistant: kan sensitive) segregation ratio, and the succeeding T3 seeds with 100 % kan resistance (n > 500), were therefore regarded as homozygous lines. The homozygous gsmt (G) lines were crossed with homozygous sdmt (S) lines to obtain transgenic plants carrying both gsmt and sdmt (G × S). Gene expression of G, S, and G × S transgenic plants were detected by RT-PCR and the results showed these archaeal genes could be functionally transcribed in higher plants (Fig. 2b). At the protein level, GSMT-GUS and SDMT-GUS fusion proteins were detected by Western blotting (Fig. 2c), and histochemical staining of GUS (Fig. 2d). Betaine was detected in G, S, and G × S transgenic plants with the greatest accumulation found in the G × S line (Fig. 2e). Interestingly,

Heterologous expression of Mpgsmt and Mpsdmt in Arabidopsis The ORFs gsmt and sdmt from M. portucalensis were cloned separately into the binary vector pBI121, with an inframe fusion of GUS at the 3′ end, which was expressed under the control of the cauliflower mosaic virus 35S

13

Plant Mol Biol (2014) 85:429–441

435

Fig. 3  LC/MS identification of glycine and its methylated derivatives. Fully expanded leaves of 6-week-old wildtype (WT), or those expressing Mpgsmt (G) or Mpsdmt (S), or both Mpgsmt and Mpsdmt (G × S) were extracted by water. Water-soluble metabolites glycine (Gly), sarcosine (Sar), and the dimethylglycine isomer (Iso*) were detected by LC/MS. Betaine (Bet) and its structural isomer valine (Val) were detected by high-resolution LC/MS. The retention time (RT) is indicated on the top of each mass spectrum

transgenic lines expressing sdmt also accumulated a low level of betaine, suggesting that substrates of SDMT, sarcosine, dimethylglycine, or both, were present in Arabidopsis. The LC/MS system was then used to determine the levels of endogenous methylated glycine (sarcosine, dimethylglycine, and betaine) in WT and three transgenic lines (Fig.  3). In addition to glycine, sarcosine was detected in rosette leaves of WT plants, while no detectable amounts of dimethylglycine or betaine were found (Fig. 3, WT). One structural isomer of dimethylglycine with a retention time of 3.32 min and one isomer of betaine with a retention time of 3.19 min was detected (Fig. 3, WT). Based on the molecular masses, we predicted the isomer of dimethylglycine to be isovaline, and valine to be the isomer of betaine. The identity of valine was confirmed by tandem MS (data not shown). We then used high resolution LC/MS with an exact molecular mass of 118.0863 dalton for betaine detection to eliminate interference by the isomer valine. Analysis of the levels of glycine, sarcosine, dimethylglycine, and betaine in transgenic plants revealed that all three lines accumulated betaine (Fig. 3, G, S, and G × S), with the highest betaine accumulation in G × S lines. The amount of the precursor glycine and sarcosine decreased accordingly in transgenic lines compared to WT. The presence of sarcosine in Arabidopsis leaves served as the initial substrate for MpSDMT that led to the accumulation of betaine in S lines. Interestingly, betaine was also detected in G lines. In combination with the fact that dimethylglycine was undetectable in all samples, we suspected that

Arabidopsis leaves possess an endogenous, unknown enzyme with dimethylglycine N-methyltransferase (DMT) activity that converts dimethylglycine into betaine. The results showed that MpGSMT and MpSDMT, acting either separately or synergistically, were able to catalyze the biosynthesis of betaine in Arabidopsis. We used homozygous G and S T3 transgenic lines and G × S F1 progenies for subsequent stress tolerance experiments. Expression of Mpgsmt and Mpsdmt increases drought and salt tolerance Transgenic plants expressing either Mpgsmt, Mpsdmt, or both have no apparently distinctive phenotype under wellwatered and optimal fertilized conditions. To examine the water balance under water-deficit stress, the RWL of WT, G, S, and G × S plants was measured. Fully expanded leaves were excised and monitored for weight loss. Upon detachment, stomates rapidly close, such that the RWL largely depends on the ability of cell to retain water. A rapid water loss from turgid leaves was observed within hours after detachment (Fig. S1). The RWL was about 10–12 % per hour in WT, G and S lines, but significantly lower at 6 % per hour in G × S (Fig. 4a). When plants were grown under suboptimal conditions (i.e. with no fertilizer subsidies), the RWL of the WT was still about 12 % per hour while the RWL of G and S lines decreased to 10 and 8 % per hour, respectively, where G × S has the lowest rate at about 6 % per hour (Fig. 4a). A limited supply of mineral

13

436

Fig. 4  Drought tolerance tests a Leaves of 6-week-old plants from each line were cut and weighed at various time points. RWL was calculated between 1 and 4 h in triplicate. Bars labeled with an asterisk indicate significant differences between WT and transgenic lines by a t test at P