J. Plant Physiol. 161. 467 – 477 (2004) http://www.elsevier-deutschland.de/jplhp

Transcriptome changes in foxtail millet genotypes at high salinity: Identification and characterization of a PHGPX gene specifically upregulated by NaCl in a salt-tolerant line Nese Sreenivasulu1, Manoela Miranda1, Harischandra Sripathy Prakash2, Ulrich Wobus1, Winfriede Weschke1 * 1

Institut für Pflanzengenetik und Kulturpflanzenforschung (IPK), 06466 Gatersleben, Germany

2

Department of Applied Botany, Mysore University, India

Received April 1, 2003 · Accepted July 1, 2003

Summary Using a macro array filter with 711 cDNA inserts representing 620 unigenes selected from a barley EST collection, we identified transcripts differentially expressed in salt (NaCl)-treated tolerant (cv. Prasad) and sensitive (cv. Lepakshi) seedlings of foxtail millet (Setaria italica L.). Transcripts of hydrogen peroxide scavenging enzymes such as phospholipid hydroperoxide glutathione peroxidase (PHGPX), ascorbate peroxidase (APX) and catalase 1 (CAT1) in addition to some genes of cellular metabolism were found to be especially up-regulated at high salinity in the tolerant line. To analyse this process at the protein level we examined protein expression patterns under various stress conditions. A 25 kD protein with a pI of 4.8 was found to be induced prominently under high salt concentrations (250 mmol/L). This salt-induced 25 kD protein has been purified and identified by peptide sequencing as PHGPX protein. The increase of the PHGPX protein level under salt stress in the tolerant line parallels the PHGPX mRNA results of array analysis but was more pronounced. We cloned and characterized the foxtail millet PHGPX cDNA, which shows 85 % and 95 % homology at the DNA and protein level, respectively, to one stress-induced member of the small barley PHGPX gene family encoding non-selenium glutathione peroxidases. As shown by Southern blot analysis, a small family of PHGPX genes exists in foxtail millet, too. The specific expression pattern of the PHGPX gene in salt-induced tolerant millet seedlings suggests that its product plays an important role in the defense reaction against salt-induced oxidative damage and that the characterized glutathione peroxidase is one of the components conferring resistance against salt to the tolerant foxtail millet cultivar. Key words: Foxtail millet (Setaria italica L.) – salt-induced oxidative stress – salt tolerance – cDNA array analysis – phospholipid hydroperoxide glutathione peroxidase (PHGPX)

* E-mail corresponding author: [email protected] 0176-1617/04/161/04-467 $ 30.00/0

468

Nese Sreenivasulu et al.

Introduction Foxtail millet (Setaria italica L.) is an important food crop grown in India, China, and Japan on salinity-prone lands as wells as during adverse conditions such as severe drought. It is also planted in Australia, North Africa, and South America for hay and silage. Increasing salinisation of agricultural land causes toxic effects in plant development, primarily at the seedling level. Hence, an improved understanding of acute adaptive and general protective mechanisms conferring enhanced salt tolerance to seedlings becomes an important issue in stress physiology and is necessary to ensure optimal growth and yield of crop plants. Salt stress results in alterations of plant metabolism, including reduced water potential, ion imbalance, and toxicity (Cramer et al. 1994, Bohnert and Jensen 1996, Hasegawa et al. 2000). During salt-induced oxidative stress, excess reactive oxygen species (ROS) were generated in chloroplasts mitochondria, and peroxisomes via a number of metabolic pathways such as photosynthesis, respiration, and photorespiration (Noctor and Foyer 1998). Because of chloroplast and mitochondrial membrane damage, the ROS is released into the cytoplasm. The presence of ROS, such as singlet oxygen (1O2*), superoxide (O2 – ), hydrogen peroxide (H2O2), and hydroxyl radicals (HO˙), in different compartments pose a threat to cells by unrestricted oxidation of various cellular molecules such as nucleic acids, proteins, and lipids (Del Río et al. 1991, Leprince et al. 2000). Deleterious effects of ROS in plants are controlled by antioxidative enzymes. O2 – is the primary source of ROS molecules, which is converted into H2O2 by the action of superoxide dismutase (SOD; E.C. 1.15.1.1). In order to quench the generated H2O2, plants evolved H2O2 scavenging antioxidative enzymes, such as classical peroxidases (POX; E.C. 1.11.1.7), ascorbate peroxidases (APX; E.C. 1.11.1.11), glutathione peroxidases (GPX; E.C. 1.11.1.9), and catalases (CAT; E.C. 1.11.16) (Halliwell and Gutteridge 1989, Noctor and Foyer 1998). The protection of the plant from reactive oxygen species seems to be one important component of the complex tolerance trait. H2O2 generated in glyoxysomes and peroxisomes is detoxified mainly by catalase (CAT; E.C. 1.11.16), while in other sub-cellular compartments (chloroplast, mitochondria, and cytoplasm) H2O2 is converted to H2O by APX and GPX. In animals, glutathione peroxidases are studied extensively, and various forms of GPX have been identified, including cytosolic GPX, plasma GPX, gastrointestinal GPX, and phospholipid hydroperoxide glutathione peroxidase (PHGPX). Even though all four groups reveal similarities in their primary structure, PHGPX differs from the other three by being a monomer and having the ability to interact with peroxidised lipids and complex lipids, which are integrated in bio-membranes. Therefore, PHGPX has been considered the main line of enzymatic defense against oxidative bio-membrane destruction in animal cells (Ursini et al. 1995). Although PHGPX has not been studied extensively in plants, some reports showed the

presence of PHGPX cDNAs exhibiting significant homology to mammalian PHGPX in different species: Citrus sinesis (Holland et al. 1993), Nicotiana sylvestris (Criqui et al. 1992), Spinacia oleracea (Sugimoto et al. 1997), Arabidopsis thaliana (Sugimoto and Sakamoto 1997), Lycopersicon esculentum (Depége et al. 1998), Hordeum vulgare (Churin et al. 1999), and Oryza sativa (Li et al. 2000). Furthermore, PHGPX mRNA levels have been shown to increase in plant tissues under stresses, such as salinity (Gueta-Dahan et al. 1997), heavy metals (Sugimoto and Sakamoto 1997), herbicide resistance (Cummins et al. 1999), mechanical stimulation (Depège et al. 1998, 2000), and infection by viral or bacterial pathogens (Levine et al. 1994). Comparative expression analysis of the genes involved in the salt stress response of tolerant and sensitive cultivars treated with and without salt provide valuable knowledge to understand the genetic and physiological basis of salt tolerance. In the present study, we have used a barley cDNA macro array to monitor transcript abundance in tolerant and sensitive foxtail millet seedlings exposed to long-term high salinity stress. By macro array analysis, we identified PHGPX as one of the genes up-regulated in the tolerant cultivar under long-term salt exposure. Here we report the molecular cloning and characterization of a PHGPX cDNA as well as the isolation, purification, and characterization of PHGPX protein from salt-treated tolerant foxtail millet seedlings. Our comparative studies indicate that the PHGPX transcript level, especially enhanced in the tolerant foxtail millet cultivar under NaCl treatment, is correlated with the very prominent level of a 25 kDa protein in the tolerant line, which was identified by peptide sequencing as PHGPX.

Materials and Methods Plant material and salinity treatment Seeds of the foxtail millet cultivar Prasad (tolerant) and Lepakshi (sensitive) were provided from Andhra Pradesh Agriculture Experimental Station, Anantapur, India. The seeds were surface sterilised with 0.1% (w/v) sodium hypochloride solution for 5 min, thoroughly rinsed with distilled water, and allowed to germinate on filter paper in petri dishes (Sreenivasulu et al. 1999) containing Hoagland medium with 0 (control) and 250 mmol/L NaCl. The petri dishes were kept at 25 ˚C under aseptic conditions. Seedlings were harvested after 5 days of incubation in the dark. At 250 mmol/L NaCl, shoot growth of the salt-sensitive cultivar was extensively inhibited, whereas the salt-tolerant cultivar was able to develop shoots (Sreenivasulu et al. 2000).

cDNA array analysis Barley EST’s derived from cDNA libraries of developing caryopses, roots, and etiolated seedlings were used to create a cDNA array (Sreenivasulu et al. 2002). The array contains 711 amplified cDNA inserts representing 620 unigenes, including approximately 75 stressrelated genes. Synthesis of 33P-labelled cDNA based on polyA + -RNA

Salt-induced transcriptome changes in tolerant foxtail millet extracted from 35 µg of total RNA as well as array hybridization and evaluation procedures have been described in detail elsewhere (Sreenivasulu et al. 2002).

RT-PCR mediated cloning of PHGPX cDNAs Total RNA from salt-tolerant and salt-sensitive seedlings treated with 250 mmol/L NaCl was isolated as described by Heim et al. (1993). RTPCR was performed under standard conditions by reverse transcription of mRNA using MMLV reverse transcriptase and oligo d(T)16 primer, followed by PCR with two combinations of degenerated primers (forward: 5′-GTBAAYGTYGCHMARTGTG-3 /reverse 1: 5′-TTRTCA AYMARRAAYTTRGWRAAG-3′, reverse 2: 5′-CRGCYTTRAADCKWGT GC-3′), which were designed on conserved regions of known PHGPX genes. A cDNA library from salt-treated tolerant foxtail millet seedlings (5-day-old) was constructed using the λZAP2 express cDNA synthesis kit (Stratagene) and screened under low stringency conditions using the 33P-labelled cDNA fragment (316 bp) obtained by RT-PCR. Seven positive phages were isolated and sequenced.

Southern hybridization For Southern hybridizations, genomic DNA was isolated following Heim et al. (1996). Ten µg of DNA were digested with BamHI, EcoRI, HindIII, PvuII, and XhoI and separated on a 1 % agarose gel. The gel was blotted overnight onto a Hybond-N + nylon membrane (Amersham Braunschweig, Germany). The 316 bp RT-PCR fragment of the millet PHGPX was used as a probe after labeling with [α-32P] dCTP. Hybridizations were performed at 65 ˚C, and the membrane was washed twice with 2 × SSPE/0.1 % SDS, twice with 1 × SSC/0.1 % SDS, and once with 0.5 × SSC/0.1% SDS at 65 ˚C for 15 min each.

Northern hybridization Total RNA from control and NaCl (250 mmol/L) treated tolerant and sensitive seedlings (15 µg/lane) was separated on a 1 % denaturing agarose gel and blotted onto Hybond N + membranes (Amersham Pharmacia). Membranes were hybridized using the 32P-labelled PCR fragment (see above) as a probe according to Church and Gilbert (1984). The hybridization signals were quantified as previously described (Sreenivasulu et al. 2000).

Protein characterization Protein extraction and estimation of protein content Five-day-old seedlings (100 mg) of both cultivars grown at 250 mmol/L NaCl were ground to a fine powder with liquid nitrogen and homogenised with ice-cold 50 mmol/L Tris-HCl buffer (pH 7.4). The extracts were centrifuged for 20 min at 8,000 g at 4 ˚C. Protein concentration was determined by the dye-binding assay (Bradford 1976) using bovine serum albumin (BSA) as a standard.

SDS-PAGE analysis An equal amount of each protein sample (35 µg) was loaded on 12 – 15 % gradient SDS-PAGE gels. After running, the gels were stained with silver nitrate solution for protein detection.

469

Purification of the 25 kD protein Five-day-old seedlings (5 g) of the tolerant cultivar grown at 250 mmol/ L NaCl were homogenised in liquid nitrogen and extracted with 50 mmol/L Tris-HCl buffer (pH 7.4) containing 1 mmol/L ethylenediamine tetra-acetic acid (EDTA), 2 mmol/L phenylmethylsulphonyl fluoride (PMSF), 0.6 % insoluble polyvinylpyrrolidone (PVP), and 0.5 mol/L sucrose. Peroxidase purification was carried out following a modified version of the protocol described by Sreenivasulu et al. (1999). Ammonium sulphate (0.2 g per mL) was added to the crude extract get 35 – 40 % saturation. The precipitated pellet was dissolved in extraction buffer, further dialysed and analyzed for peroxidase activity. The homogenate containing peroxidase activity was filtered through a nylon sieve (40 µm) and centrifuged at 8,000 g for 10 min. The DEAESepharose (Sigma) column (bed: 10 × 2.5 cm) was pre-equilibrated with Tris-HCl buffer (50 mmol/L pH 7.2) and the supernatant containing peroxidases was subjected to ion-exchange chromatography. The sample was eluted using a NaCl step-gradient (0.2, 0.4, 0.6, 0.8 and 1.0 mol/L) and 1mL fractions were collected. The absorbance of each fraction was monitored at 280 nm and total peroxidase activity was determined (see Sreenivasulu et al. 1999). The fractions number 12 – 18 containing peroxidases were pooled, lyophilized, and subjected to FPLC (Superose) to achieve higher purity. The column was pre-equilibrated and eluted with PBS buffer (pH 7.4) as flow-through. The protein of the major peak was collected and concentrated.

Glutathione peroxidase activity Glutathione-dependent peroxidase activity was determined only for the purified 25 kD protein fraction by adding 0.25 mmol/L β-NADPH, 3 mmol/L glutathione, 0.1 % (v/v) triton X-100 and 1.5 U glutathione reductase. The standard reaction mixture (1mL) was incubated at 30 ˚C for 5 min and NADPH oxidation was measured at 340 nm after adding peroxide substrate.

Amino acid sequencing The purified salt-induced protein was run on a 12 % SDS-PAGE (Laemmli 1970) and electroblotted onto a polyvinylidene difluoride (PVDF) membrane (Millipore) using the Multiphor II (LKB, Pharmacia) electrophoretic transfer apparatus according to the manufacturer’s protocol. 10 mmol/L 3-[cyclohexylamino]-1-propane-sulphonic acid (CAPS; pH 11) was used as transfer buffer. The membrane was stained by coomassie blue, and the part of the membrane containing the purified protein was subjected to the sequencer LF3400 (Beckman Instruments Fullerton, Ca, USA).

Estimation of malonaldehyde (MDA) concentration The MDA content of 5-day-old seedlings was determined by the thio barbituric acid (TBA) reaction as previously described (Sreenivasulu et al. 1999).

Results Macroarray-based stress-specific gene expression analysis The salt-tolerant foxtail millet line, Prasad, exhibits proper growth in the presence of 250 mmol/L NaCl, unlike the sensi-

470

Nese Sreenivasulu et al.

Table 1. Genes up- or down-regulated upon NaCl-treatment in seedlings of the salt tolerant S. italica variety Prasad. The genes were identified by cDNA macro array analysis. Ratios were calculated by dividing the normalized signal intensity higher under conditions of salt treatment (upper part of the table, up-regulated genes) or higher under control conditions (lower part of the table, down-regulated genes) by the respective lower signal intensity. The genes (ESTs) marked by a filled circle are up-regulated in both the tolerant and the sensitive variety. Standard deviations (SD) were calculated from the ratios estimated in three independent experiments. Significance of up- and down-regulation were checked by the t-test. Significance at the 0.05 level is visualized by one asterisk, significance at the 0.01 level by two asterisks. Non-significant values are labeled by «ns». ID, identification number of the ESTs. EST-ID

Putative function

Tolerant Ratio

Sensitive Ratio

2.14 ± 0.11 * 2.22 ± 0.06 * 2.22 ± 0.04 *

1.32 ± 0.30 ns 1.15 ± 0.21 ns 1.26 ± 0.13 ns

2.05 ± 0.06 * 1.99 ± 0.02 **

1.19 ± 0.11 ns 1.18 ± 0.06 ns

2.11 ± 0.11 *

1.20 ± 0.24 ns

2.04 ± 0.11 * 2.03 ± 0.01 **

2.09 ± 0.47 ns 1.81 ± 0.52 ns

2.05 ± 0.03 * 2.415 ± 0.04 *

1.21 ± 0.15 ns 1.19 ± 0.15 ns

2.450 ± 0.18 ns 2.015 ± 0.02 ** 2.01 ± 0.08 * 1.960 ± 0.00 **

2.16 ± 0.38 ns 1.86 ± 0.64 ns 1.11 ± 0.25 ns 1.69 ± 0.29 ns

1.29 ± 0.01 * 1.12 ± 0.01 * 1.04 ± 0.05 ns

2.13 ± 0.19 ns 2.09 ± 0.04 * 2.09 ± 0.01 **

1.40 ± 0.25 ns

2.20 ± 0.04 *

Genes up-regulated hydrogen peroxide scavenging enzyme HY09D05 HY10B17 HY03C01

glutathione peroxidase l-ascorbate peroxidase, cytosolic catalase protease inhibitor

HY05L03 HY09O22

alpha-amylase/trypsin inhibitor subtilisin-chymotripsin inhibitor Ion-channel

HY06G06

potassium channel protein 1 cellular metabolism

䊉 䊉

HY06N11 HY03K17 HY03O15 HY02G05

fructose-bisphosphate aldolase, cytosolic phosphoribosylformylglycinamidine cyclo-ligase dihydrodipicolinate synthase 2 carboxyphosphonoenolpyruvate mutase

HY05K13 HY09N04 HK04B02 HW01M06

cyclophilin argonaute protein Krüppel-like transcription factor unknown

other proteins 䊉 䊉 䊉

Genes down-regulated Chaperonins HY05F13 HY03B11 HY01L09

heat shock cognate 70 kda protein 1 heat shock cognate 70 kda protein heat shock cognate 70 kda protein 2 carbohydrate metabolism

HY01G12

cell wall invertase 2

tive line, Lepakshi. The salt concentration and the time point chosen for analysis of seedling growth were based on earlier studies (Sreenivasulu et al. 2000). To monitor NaCl-stress induced gene expression, mRNA samples from 5-day-old tolerant and sensitive foxtail millet seedlings grown upon 250 mmol/L NaCl and without salt treatment (control) were isolated and reverse transcribed into first-strand cDNA probes. Four different 33P-labelled second-strand cDNAs, two generated from tolerant (salt-treated, control) and two from sensitive seedlings (salt-treated, control) were hybridized to the barley cDNA array (Sreenivasulu et al. 2002). Fourteeen nonredundant sequences were found to be significantly up-regulated by two-fold or more in the salt-treated tolerant cultivar as

compared to the control (significant at the 0.05 level, one asterisk; significant at the 0.01 level, two asterisks; see Table 1). In the salt-sensitive cultivar, five of these transcripts were up-regulated to nearly the same extent (marked by filled circles in Table 1). However, up-regulation of these candidate genes was not confirmed by statistical significance (non-significant, ns). Ten of the sequences were not regulated to a reliable extent (see Table 1). In order to determine whether the transcripts up-regulated in the salt-tolerant cultivar belong to a particular class of genes, we associated expression data with functional classification. The genes (cDNA fragments on the macro array filter) only up-regulated in the salt-tolerant cultivar represent mainly hydrogen peroxide scavenging

Salt-induced transcriptome changes in tolerant foxtail millet

471

enzymes as well as genes encoding proteinase inhibitors, a potassium channel protein, and proteins of either cellular metabolism or of unknown function. On the other hand, some transcripts, such as those of heat shock proteins and cell wall invertase, are down-regulated in the salt-sensitive line and down-regulation was also noticed in the tolerant line. The complete hybridization procedure was repeated twice using independently grown seedlings as starting material. The BLASTX2 results and results of functional annotation of the unigene set are available from our WWW server (http://pgrc.ipk-gatersleben.de/sreeni/b.html).

as compared to the sensitive cultivar under control conditions. Under salt treatment, PHGPX mRNA was expressed at nearly the same low level in the sensitive cultivar, whereas a significant increase of expression was found in tolerant seedlings. Thus, comparing cDNA array analysis and Northern blotting, nearly identical results were obtained (see Table 1 and Fig. 1). Similarly, APX transcripts were analysed by Northern hybridization before and were found to be up-regulated prominently under salt-treatment in the tolerant cultivar (Sreenivasulu et al. 2000), a result which can also be confirmed by the present array data (Table 1).

Expression analysis of hydrogen peroxide scavenging enzymes

Isolation and identification of a full-length cDNA coding for a PHGPX from millet

In this paper we focus on salt-mediated oxidative stressinduced phospholipid hydroperoxide glutathione peroxidase (PHGPX). As shown in Table 1, PHGPX transcript levels increased two-fold in the tolerant cultivar under high concentrations of NaCl (250 mmol/L), whereas no significant difference was found in the PHGPX expression level of salt-treated and control sensitive seedlings. We have evaluated the validity of the array results for PHGPX by using a millet PHGPX gene as probe in Northern blot analysis. Four different samples of total RNA were isolated from the same seedling material as used for the cDNA array analysis. This RNA was taken as template for RT-PCR amplification of a specific 316 bp fragment (see below) to be hybridized to Northern blots. As shown in Figure 1, the level of the PHGPX transcript is equal in the tolerant

To isolate PHGPX cDNA from millet, conserved domains of known PHGPX sequences from different species, including barley (Churin et al. 1999), together with sequence information regarding the barley PHGPX EST/cDNA spotted on our cDNA macroarray, were used to design degenerated primers for RT-PCR. Total RNA was isolated from seedlings of the salttolerant and the salt-sensitive cultivars grown upon 250 mmol/ L NaCl. The RNA probes were used as template for RT-PCR. Amplification was successful only with mRNA from tolerant seedlings (data not shown). After cloning of the RT-PCR fragments, sequence analyses showed that the fragments obtained with two different combinations of primers (see Material and Methods) represented the same PHGPX gene. The RT-PCR amplified millet cDNA fragment (316 bp in length) shares 85 % and 95 % homology with the PHGPX EST from barley at the nucleotide and amino acid level, respectively. This fragment was used to screen a cDNA library prepared from salt-treated (200 mmol/L) tolerant foxtail millet seedlings at low stringency. Sequence analysis of seven independent positive clones of different length revealed that all represent the same PHGPX gene. The longest cDNA was chosen and designated as SiPHGPX1. Alignment of the amino acid sequence of millet PHGPX1 to recently published putative plant PHGPX proteins is shown in Figure 2. Although computer analysis predicts a plastid targeting sequence in both the millet and the barley PHGPX12 gene (SiGPX1 and HvGPX12, respectively, in Fig. 2), we were unable to isolate sequences with additional nucleotides including a translation initiation Met codon at the extreme 5′-end of the cDNA. We also did not find an equivalent to the second barley glutathione peroxidase HvGPX15 (Fig. 2) in millet. Starting with the identified ATG (position 71 in Fig. 2), the predicted millet PHGPX1 protein consists of 237 amino acids with a calculated molecular mass of 25.7kD.

Figure 1. PHGPX transcript level in NaCl-treated ( + NaCl) and control ( – NaCl) seedlings of salt-tolerant (cv. Prasad) and salt-sensitive seedlings (cv. Lepakshi) of foxtail millet (Setaria italica L.). Fifteen µg total RNA were loaded per lane, electrophoresed, blotted, and hybridized with a 32P-labelled S. italica PHGPX probe.

Southern analysis of PGHPX genes To analyse the genomic organisation of millet PGHPX gene(s), genomic DNA from seedlings of the tolerant and sensi-

472

Figure 2.

Nese Sreenivasulu et al.

Salt-induced transcriptome changes in tolerant foxtail millet

473

Figure 3. Southern blot analysis of S. italica PHGPX sequences. Genomic DNA (10 µg) from foxtail millet seedlings of the tolerant and the sensitive genotype was digested with restriction endonucleases (HIBam HI; RI-Eco RI; HIII-HindIII; PII-PVUII; XI-XhoI), the DNA was subjected to electrophoresis and hybridized with a 32P-labelled S. italica PHGPX cDNA probe. Standard DNA markers are indicated in kilo bases (kb) at the right hand side.

tive cultivar were isolated, digested by different restriction enzymes and subjected to Southern blot analysis using the 32 P-labelled 316 bp RT-PCR fragment as a probe. Because none of the enzymes used to cut the genomic DNA can find any restriction sites within this fragment, the different fragments of nearly identical intensity found after blot hybridization within one lane of the Southern blot should represent different genomic loci of highly homologous genes (Fig. 3). Therefore, at least two nearly identical PHGPX genes can be expected within the millet genome. Furthermore, Southern blotting revealed no differences between the two varieties in the major bands, whereas a slight polymorphism is visible in the minor bands, which points to differences in the number of further possible members of the gene family lower in homology.

Salt-specific induction of a 25 kD protein Proteins were extracted from tolerant seedlings exposed to salt (24 hours, 250 mmol/L NaCl), drought (24 hours, 20 %

Figure 4. Pattern of proteins extracted from 5-day-old tolerant foxtail millet seedlings grown under different types of stress. NaCl, treatment with 150 mmol/L NaCl; Drought, seedlings were grown in Hoagland medium during the first three days and treated with 20 % polyethylene glycol 8000 for 24 hours; Heat, 45 ˚C for 5 hours and Cold, 4 ˚C for 5 hours. Control, seedlings were grown in Hoagland medium without NaCl. The position of marker proteins is given at the left-hand side. The arrowhead marks the position of the 25 kD protein up-regulated in the tolerant variety under salt treatment.

polyethylene glycol 8000), higher temperature (45 ˚C, 5 hours), and cold (4 ˚C, 5 hours), and separated on a 12–15 % SDS-PAGE gradient gel. The protein patterns obtained after silver staining of the gradient gel (Fig. 4) indicated that a protein with an apparent molecular mass of 25 kD is specifically induced by NaCl stress. Low expression is also seen under

Figure 2. Amino acid sequence alignment of the foxtail millet (S. italica) PHGPX with other plant PHGPX sequences. Database searches were carried out using the BLAST program (Altschul et al. 1990) and sequence alignment was performed with the OMEGA 2.1 program. Amino acid sequences of PHGPX are aligned as follows: Setaria italica (SiGPX1: this paper Acc-no. AY 541694); Hordeum vulgare (HvGPX12 – Acc-no. CAB 59893; HvGPX15 – Acc-no. CAB 59894); Oryza sativa (OsGPX2 – Acc-no. BF 430853; OsGPX1 – Acc-no. CAC 17628); Pisum sativum (PsGPX1 – Acc-no. CAA 04142; PsGPX2- this paper EST-ID PSS08C08* [* The respective sequence is available from the IPK EST Database (CR-EST); http://pgrc.ipk-gatersleben.de/databases.php]); Arabidopsis thaliana (AtGPX2 – Acc-no. CAB 40757; AtGPX1 – Acc-no. AAC 09173). Residues common to all 9 proteins are accentuated by dark gray shading, and those shared by 6 to 9 proteins are shaded. The putative trans-membrane domains were predicted by TMHMM (Sonnhammer et al. 1998). Arrowheads indicate the predicted chloroplast cleavage sites. The boxed region represents the peptide sequence identified in the purified protein.

474

Nese Sreenivasulu et al. heat-stress conditions but not in drought- and cold-stressed seedlings (Fig. 4).

Purification of the salt-induced 25 kD protein Ammonium sulphate precipitation of the crude extract yielded an enriched fraction containing most of the peroxidases. The homogenate containing 25 kD protein was subjected to DEAE ion-exchange chromatography and a clear peak with high peroxidase activity identified in the 0.6 mol/L NaCl fraction (Fig. 5 a). Peak fractions were collected, pooled, lyophilized, and subjected to two-dimensional (2D) gel analysis. The pooled fraction contained proteins of 25 kD and 27 kD, respectively, with apparent isoelectric points of 4.8 and 3.7 (Fig. 5 b). This fraction was further subjected to Superose FPLC. The 25 kDa and 27 kD proteins were eluted at 141.28 min and 121.41 min, respectively (Fig. 5 c). Homogeneity of the salt-induced purified 25 kD protein was demonstrated by gel electrophoresis (inlet of Fig. 5 c). Furthermore, presence of glutathione activity (36 nmol min –1 mg –1) was demonstrated for the purified 25 kD protein.

Amino acid sequence analysis The purified 25 kD protein (pI 4.8) was subjected to N-terminal sequence analysis, but the N-terminal region was found to be blocked. Therefore, the protein was subjected to trypsin digestion. The generated peptides were isolated by reversedphase HPLC and sequenced. The amino acid sequence of the major well resolved peptide (VNVASQCGLTNSNYTELAQLYEK) showed 100 % homology to amino acids 106 – 128 of the non-selenium phospholipid hydroperoxide glutathione peroxidase from barley and rice (HvGPX12 and OsGPX2) and showed high homology to known non-selenium glutathione phospholipid hydroperoxide peroxidase sequences from other plant species as shown in Figure 2.

Effect of salinity on malonaldehyde content

Figure 5. Steps to purify the 25 kD protein up-regulated in salt-treated tolerant seedlings. (a) The ion-exchange column chromatography profile of proteins bound to DEAE-Sepharose and eluted by a NaCl concentration gradient. (b) Result of 2D-gel analysis of the fraction eluted at 0.6 mol/L NaCl showing two proteins with 27 kD (pI 3.7) and 25 kD (pI 4.8). Standard molecular weight markers are given in kD at the left-hand side. Standard pI markers are represented on top of the figure. (c) Superose FPLC profile of the purified protein fractions of 25 kD (pI 4.8) and 27 kD (pI 5.5); the two proteins are represented as peak at 121.41 and 141.28 min, respectively. The inlet shows a PAA gel analysis of the purified pI 4.8 fraction.

Malonaldehyde is an indicator of membrane damage. Therefore, we measured the MDA content of 5-day-old seedlings of the tolerant and sensitive cultivars grown at different NaCl concentrations and without NaCl (control). In both cultivars, increased salt concentration caused a significant increase in the MDA level. However, MDA accumulation was higher in the salt-susceptible line (Fig. 6). Up to 100 mmol/L NaCl, no dramatic difference was found in the MDA level of the two cultivars. At higher NaCl concentrations, the amount of MDA remained nearly constant in the tolerant cultivar (around 20 µmol at 100 mmol/L and 28 µmol at 250 mmol/L NaCl) but increased to more than double the amount (around 55 µmol) in seedlings of the sensitive cultivar exposed to 250 mmol/L NaCl.

Salt-induced transcriptome changes in tolerant foxtail millet

Figure 6. Malonaldehyde content of 5-day-old seedlings of control and salt-stressed (50, 100, 150, 200 and 250 mmol/L) seedlings of the salt-tolerant and the salt-sensitive S. italica genotype.

Discussion Plant growth is greatly affected by high salinity. Tolerant plants respond and adapt to survive under salt-stress conditions by induction of stress gene expression. Several hundreds of genes have been shown by micro-array analysis to be induced by high-salinity in barley as well as in a salt-tolerant rice line (Ozturk et al. 2002, Kawasaki et al. 2001). Based on the well-known synteny of cereal genomes, we explored the ability of barley-cDNA-based arrays to examine the expression patterns in previously characterized (Sreenivasulu et al. 1999, 2000) foxtail millet genotypes differing in salt tolerance. The expression of more than 620 unigenes, including 75 stress-induced genes, was studied. Altogether, the obtained results demonstrate that barley macro-arrays are a very useful tool to analyze gene expression in a heterologous system like foxtail millet. Comparison of expression profiles between salt-tolerant (Prasad) and salt-sensitive (Lepakshi) genotypes under control and high salt-stress conditions allowed us to identify key genes possibly involved in tolerance capacity of line Prasad. Among the genes up-regulated under salt stress in the tolerant line are antioxidant enzymes (PHGPX, cytosolic APX and CAT1), protease inibitor genes (α-amylase/trypsin inhibitor, subtilisin-chymotrypsin inhibitor), a gene encoding the lysine biosynthesis enzyme dihydrodipicolinate synthase 2, and a potassium channel protein-encoding gene (Table 1). Potassium channels play an important role in potassium uptake and membrane potential maintenance (Vranová et al. 2001). Up-regulation of the potassium channel protein transcript in the salt-tolerant line during long-term salt exposure might be indicative of an increased ability of the tolerant cells to sequestrate Na + ions into vacuoles by maintaining a high cytosolic K + : Na + ratio, which is important for plant growth during salt-stressed conditions (Glenn et al. 1999). This suggestion is consistent with our previous findings that the salt-tolerant foxtail millet genotype maintains a constant

475

Na + ion level in the plant even at high external NaCl concentrations (Sreenivasulu et al. 2000). Other genes whose products protect plant cells against long-term salt exposure are associated with ROS scavenging and encode protease inhibitors and stress-induced proteins (Hasegawa et al. 2000). The finding that protease inhibitor transcripts (alpha-amylase/trypsin inhibitor and subtilisin-chymotrypsin inhibitor) are up-regulated in the tolerant genotype under prolonged salt treatments is compatible with the expression profiling results reported by Kawasaki et al. (2001) for salt-tolerant Pokkali rice. Furthermore, we identified a salt-induced 23 kD trypsin inhibitor protein (pI 6.2) in the tolerant foxtail millet genotype (N. Sreenivasulu, unpublished data). Trypsin inhibitors were reported to be induced by salt treatment, aluminium, desiccation, ABA, wounding, and fungal infection (Cordero et al. 1994, Lam et al. 1999, Richards et al. 1998). Over expression of a barley trypsin inhibitor in wheat has been shown to increase insect resistance (Altpeter et al. 1999). Up-regulation of antioxidant transcripts (cytosolic APX, peroxisomal catalase and plastidic GPX) protect the tolerant cultivar against hydrogen peroxide during long-term salt exposure, which supports our earlier results (Sreenivasulu et al. 1999, 2000) and further strengthens our notion that scavenging reactive oxygen species at different cellular compartments in individual lines provides remarkable tolerance. Over-expression of cytosolic ascorbate peroxidase in transgenic plants leads to higher tolerance against abiotic stress treatments (Smirnoff 2000). We further focused our analyses on a phospholipid hydroperoxide glutathione peroxidase gene (SiPHGPX), which was the protein most prominently induced by salt treatment in the tolerant foxtail millet line (Fig. 4). In mammalian cells, PHGPX is considered to be the key enzyme involved in scavenging hydroxyl radicals and hydrogen peroxide (Ursini et al. 1995). Related plant sequences have only recently been found (see Introduction), and they serve a similar function. PHGPX enzymes generally contain three conserved domains: NVASQ(C/X)G, ILAFPCNQF, and IKWNF(S/T)DFL(V/I)DK. The presence of selenocysteine (X), glutamic acid (Q), and tryptophan (W) residues were shown to be critical for the catalytic activity of mammalian PHGPX (Chu et al. 1993). However, in plant PHGPX proteins, selenocysteine is replaced by cystein. This replacement results in a drastic decrease of enzyme activity as compared to the mammalian enzyme (Eshdat et al. 1997). Recently, Hazebrouck et al. (2000) substituted cystein with selenocysteine in citrus PHGPX by genetic engineering and demonstrated a 4 fold enhancement of enzyme activity after expression of the modified gene in E. coli. Based on the alignment of many plant PHGPX amino acid sequences, at least two classes can be discerned (see Fig. 2), most probably representing plastidic (the longer cDNAs) and cytosolic isoforms. Among the plastidic PHGPX class, the sequences of barley and rice contain two methionine residues at 65 residues distance, whereas the downstream Met is not found in the dicot sequences from pea and Arabidopsis. In barley, Churin et al. (1999) described

476

Nese Sreenivasulu et al.

three cDNA isoforms suggested to be targeted to the cytosol, plastids, and peroxisomes. Whereas the first two isoforms were found to be up-regulated by stress, the latter one was down-regulated. In the cDNA-based protein-prediction of our salt-induced millet PHGPX, a methionine residue is found at position 63 (position 71 of the alignment in Fig. 2) in frame, corresponding to the start codon of the barley plastidic isoform. In addition, the most probably incomplete region upstream of this Met codon represents, as predicted by computer analyses, a transit peptide. We were unable to identify in our millet cDNA library or by PCR approaches another shorter cDNA corresponding to a presumably cytosolic isoform, but our Southern analyses (Fig. 3) suggest a small gene family also in the millet genome. The two Met codons in a message like HvGPX2 may allow the synthesis of proteins with and without a transit peptide but no data in favour of this speculation are available. Up-regulation of SiPHGPX under salt stress has been detected at the RNA level (Table 1, Fig. 1), and most prominently at the protein level (Fig. 4). These findings are consistent with other studies reporting up-regulated PHGPX protein levels in salt-stressed citrus (Beeor-Tzahar et al. 1995, Gueta-Dahan et al. 1997) and increased cDNA levels in barley (Churin et al. 1999) and pea (Hernández et al. 2000) after salt treatment. Mittova et al. (2002) demonstrated an increase of PHGPX levels in chloroplasts of a salt-treated tolerant tomato cultivar. This data together strongly suggest that induction of PHGPX is at least part of an antioxidant defense mechanism in different plant species, including millet, leading to long-term salttolerance. Our protein studies detected one especially prominent protein of 25 kD (Figs. 4, 5) identified as PHGPX by peptide sequencing and induced in foxtail millet seedlings during long-term high-concentration NaCl stress. The second, less prominent, protein of 27 kD (Fig. 5 b) turned out to be ascorbate peroxidase (N. Sreenivasulu, unpublished results). It is important to note that the high levels of the 25 kD PHGPX-protein were only found under high salt conditions in the tolerant line but were much less induced in the sensitive line and under heat stress. The protein was undetectable after exposure to drought and low temperatures (see Fig. 4). Since the expression differences are clearly pronounced at the protein level (cf. Figs. 1 and 4) both transcriptional and posttranscriptional regulation are implicated. Our data do not allow a decision on whether the characterized cDNA SiPHGPX1 represents the message of the 25 kD protein or of another isoform. However, the complete identity of the 23-residues-long peptide with a respective peptide predicted from the SiPHGPX1cDNA, together with the presence of an in frame ATG codon at the 3′-side of the transit peptide allowing theoretical translation of a 25 kD protein without transit peptide, does not exclude the identity hypothesis. In other species, both cytosolic and plastidic PHGPX isoforms have been found to be induced by salt stress (see above).

SiPHGPX is thought to play an important role in the protection of cells against oxidative stress by detoxifying organic peroxides and lipid peroxides to counteract salt-mediated oxidative membrane damage in 5-day-old seedlings of salt tolerant foxtail cultivar. This conclusion is strengthened by lower lipid peroxidation (lower MDA values, Fig. 6) recorded in tolerant seedlings under high NaCl concentrations. The lower MDA value provides evidence for increased membrane integrity and thus better survival of tolerant plants under NaCl treatment. In vivo modification of PHGPX gene expression in specific cell compartments in transgenic plants using a saltstress inducible promoter followed by measurements of proper substrate/electron donors will be necessary to clarify the exact role of the SiPHGPX gene in conferring salt tolerance. Acknowledgements. We are grateful to Dr. Wolfgang Michalek for providing barley ESTs. We thank Dr. Christian Horstmann for help in protein purification and Angela Stegmann and Elsa Fessel for excellent technical assistance. We also thank Dr. Armein Meister for his help in statistical analysis. This work was supported by grants from Deutscher Akademischer Austauschdienst (Award No. A/98/00061) and IPK internal funds.

References Altpeter F, Diaz I, McAuslane H, Gaddour K, Carbonero P, Vasil IK (1999) Increased insect resistance in transgenic wheat stably expressing trypsin inhibitor CMC. Mol Breed 5: 53 – 63 Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ (1990) Basic local alignment search tool. J Mol Biol 215: 403 – 410 Beeor-Tzahar T, Ben-Hayyim G, Holland D, Faltin Z, Eshdat Y (1995) A stress-associated citrus protein is a distinct plant phospholipid hydroperoxide glutathione peroxidase. FEBS Lett 366: 151–155 Bradford MN (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of proteindye binding. Analyt Biochem 72: 248 – 254 Bohnert HJ, Jensen RG (1996) Metabolic engineering for increased salt tolerance. The next step. Aust J Plant Physiol 23: 661– 667 Chu FF, Doroshow JH, Esworthy RS (1993) Expression, characterization and tissue distribution of a new cellular selenium-dependent glutathione peroxidase, GSHPx-GI. J Biol Chem 268: 2571– 2576 Church GM, Gilbert W (1984) Genomic sequencing. Proc Natl Acad Sci USA 81: 1991–1995 Churin Y, Schilling S, Börner T (1999) A gene family encoding glutathione peroxidase homologues in Hordeum vulgare (barley). FEBS Lett 459: 33 – 38 Cordero MJ, Raventos D, San-Segundo B (1994) Expression of a maize proteinase inhibitor gene is induced in response to wounding and fungal infection: systemic wound response of a monocot gene. Plant J 6: 141–150 Cramer GR, Alberico GJ, Schmidt C (1994) Salt tolerance is not associated with the sodium accumulation of two maize hybrids. Aust J Plant Physiol 21: 675 – 692 Criqui MC, Jamet E, Parmentier Y, Marbach J, Durr A, Fleck J (1992) Isolation and characterization of plant cDNA showing homology to animal glutathione peroxidases. Plant Mol Biol 18: 623 – 627

Salt-induced transcriptome changes in tolerant foxtail millet Cummins I, Cole DJ, Edwards R (1999) A role for glutathione transferases functioning as glutathione peroxidases in resistance to multiple herbicides in black-grass. Plant J 18: 285 – 292 Del Río LA, Sevilla F, Scandalio LM, Plama JM (1991) Nutritional effect and expression of superoxide dismutases: Induction and gene expression diagnostics, prospective protection against oxygen toxicity. Free Rad Res Communic 12: 819 – 828 Depège N, Drevet J, Boyer N (1998) Molecular cloning and characterization of tomato cDNAs encoding glutathione peroxidase-like proteins. Euro J Biochem 253: 445 – 451 Depège N, Varenne M, Boyer N (2000) Induction of oxidative stress and GPX-like protein activation in tomato plants after mechanical stimulation. Physiol Plant 110: 209 – 214 Eshdat Y, Holland D, Faltin Z, Ben-Hayyim G (1997) Plant glutathione peroxidases. Physiol Plant 100: 234 – 240 Glenn E, Brown JJ, Blumwald E (1999) Salt tolerance and crop potential of halophytes. Crit Rev Plant Sci 18: 227– 255 Gueta-Dahan Y, Yaniv Z, Zilinskas BA, Ben-Hayyim G (1997) Salt and oxidative stress: similar and specific responses and their relation to salt tolerance in citrus. Planta 203: 460 – 469 Halliwell B, Gutteridge JMC (1989) Free radicals in biology and medicine. Oxford University Press, London Hasegawa B, Bressan A, Zhu JK, Bohnert HJ (2000) Plant cellular and molecular responses to high salinity. Ann Rev Plant Physiol Plant Mol Biol 51: 463 – 499 Hazebrouck S, Camoin L, Faltin Z, Strosberg AD, Eshdat Y (2000) Substituting selenocysteine for catalytic cysteine 41 enhances enzymatic activity of plant phospholipid hydroperoxide glutathione peroxidase expressed in Escherichia coli. J Biol Chem 275: 28715 – 28721 Heim U, Weber H, Bäumlein H, Wobus U (1993) A sucrose-synthase gene of Vicia faba L.: Expression pattern in developing seeds in relation to starch synthesis and metabolic regulation. Planta 191: 394 – 401 Heim U, Weber H, Wobus U (1996) Cloning and characterization of full-length cDNA encoding sucrose phosphate synthase from faba bean. Gene 178: 201– 203 Hernández JA, Jiménez A, Mullineaux P, Sevilla F (2000) Tolerance of pea (Pisum sativum L.) to long-term salt stress is associated with induction of antioxidant defenses. Plant Cell Environ 23: 853 – 862 Holland D, Ben-Hayyim G, Faltin Z, Camoin L, Strosberg AD, Eshdat Y (1993) Molecular characterization of salt-stress-associated protein in citrus: protein and cDNA sequence homology to mammalian glutathione peroxidase. Plant Mol Biol 21: 923 – 927 Kawasaki S, Borchert C, Deyholos M, Wang H, Brazille S, Kawai K, Galbraith D, Bohnert HJ (2001) Gene expression profiles during the initial phase of salt stress in rice. Plant Cell 13: 889 – 905 Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophase T4. Nature 227: 680 – 685 Lam JM, Pwee KH, Sun WQ, Chua YL, Wang XJ (1999) Enzyme-stabilizing activity of seed trypsin inhibitors during dessication. Plant Sci 142: 209 – 218 Leprince O, Harren FJ, Buitink J, Alberda M, Hoekstra FA (2000) Metabolic dysfunction and unabated respiration precede the loss of membrane integrity during dehydration of germinating radicles. Plant Physiol 122: 597– 608

477

Levine A, Tenhaken R, Dixon R, Lamb C (1994) H2O2 from oxidative burst orchestrates the plant hypersensitive disease resistance response. Cell 79: 583 – 593 Li W, Feng H, Fan J, Zhang R, Zhao N, Liu J (2000) Molecular cloning and expression of a phospholipid hydroperoxide glutathione peroxidase homolog in Oryza sativa. Biochem Biophys Acta 1493: 225 – 230 Mittova V, Tal M, Volokita M, Guy M (2002) Salt stress induces up-regulation of an efficient chloroplast antioxidant system in the salt-tolerant and wild tomato species Lycopersicon pennellii but not in the cultivated species. Physiol Plant 115: 393 – 400 Noctor G, Foyer C (1998) Ascorbate and glutathione: keeping active oxygen under control. Ann Rev Plant Physiol Plant Mol Biol 49: 249 – 279 Ozturk ZN, Talamé V, Deyholos M, Michalowski CB, Galbraith DW, Gozukirmizi N, Tuberosa R, Bohnert HJ (2002) Monitoring large-scale changes in transcript abundance in drought- and salt-stressed barley. Plant Mol Biol 48: 551– 573 Richards KD, Schott EJ, Sharma YK, Davis KR, Gardner RC (1998) Aluminium induces oxidative stress genes in Arabidopsis thaliana L. Plant Physiol 116: 409 – 418 Singha S, Choudhuri MA (1990) Effect of salinity (NaCl) stress on H2O2 metabolism in Vigna and Oryzya seedlings. Biochem Physiol Pflanzen 186: 69–74 Smirnoff N (1993) The role of active oxygen in the response of plants to water deficit and desiccation. New Phytol 125: 27– 58 Smirnoff N (2000) Ascorbic acid: Metabolism and functions of a multifaceted molecule. Curr Opin Plant Biol 3: 229 – 235 Sonnhammer ELL, Eddy SR, Birney E, Bateman A, Durbin R (1998) Pfam: multiple sequence alignments and HMM-profiles of protein domains. Nucl Acids Res 26: 320 – 322 Sreenivasulu N, Ramanjulu S, Ramachandra-Kini K, Prakash HS, Shetty HS, Savithri HS, Sudhakar C (1999) Total peroxidase activity and peroxidase isoforms as modified by salt stress in two cultivars of fox-tail millet with differential salt tolerance. Plant Sci 141: 1– 9 Sreenivasulu N, Grimm B, Wobus U, Weschke W (2000) Differential response of antioxidant compounds to salinity stress in salt-tolerant and salt-sensitive seedlings of foxtail millet (Setaria italica). Physiol Plant 109: 435 – 442 Sreenivasulu N, Altschmied L, Panitz R, Hähnel U, Michalek W, Weschke W, Wobus U (2002) Identification of genes specifically expressed in maternal and filial tissues of barley caryopses: a cDNA array analysis. Mol Genet Genom 266: 758–767 Sugimoto M, Sakamoto W (1997) Putative phospholipid hydroperoxide glutathione peroxidase gene from Arabidopsis thaliana induced by oxidative stress. Genes Genetic Systematics 72: 311– 316 Sugimoto M, Furui S, Suzuki Y (1997) Molecular cloning and characterization of a cDNA encoding putative phospholipid hydroperoxide glutathione peroxidase from spinach. Biosci Biotech Biochem 61: 1379–1381 Ursini F, Maiorino M, Brigelius-Flohé R, Aumann KD, Roveri A, Schomburg D, Flohé L (1995) Diversity of glutathione peroxidase. Methods Enzym 252: 38 – 53 Vranová E, Tahtiharju S, Sriprang R, Willekens H, Heino P, Palva ET, Inzé D, VanCamp W (2001) The AKT3 potassium channel protein interacts with the AtPP2CA protein phosphatase 2C. J Exp Bot 52: 181–182

.