A Novel Salt-tolerant l-myo-Inositol-1-phosphate Synthase from Porteresia coarctata (Roxb.) Tateoka, a Halophytic Wild Rice

l-myo-Inositol-1-phosphate synthase (EC 5.5.1.4, MIPS), an evolutionarily conserved enzyme protein, catalyzes the synthesis of inositol, which is implicated in a number of metabolic reactions in the biological kingdom. Here we report on the isolation of the gene (PINO1) for a novel salt-tolerant MIPS from the wild halophytic rice, Porteresia coarctata (Roxb.) Tateoka. Identity of the PINO1 gene was confirmed by functional complementation in a yeast inositol auxotrophic strain. Comparison of the nucleotide and deduced amino acid sequences of PINO1 with that of the homologous gene from Oryza sativa L. (RINO1) revealed distinct differences in a stretch of 37 amino acids, between amino acids 174 and 210. Purified bacterially expressed PINO1 protein demonstrated a salt-tolerant character in vitro compared with the salt-sensitive RINO1 protein as with those purified from the native source or an expressed salt-sensitive mutant PINO1 protein wherein amino acids 174–210 have been deleted. Analysis of the salt effect on oligomerization and tryptophan fluorescence of the RINO1 and PINO1 proteins revealed that the structure of PINO1 protein is stable toward salt environment. Furthermore, introgression of PINO1 rendered transgenic tobacco plants capable of growth in 200–300 mm NaCl with retention of ∼40–80% of the photosynthetic competence with concomitant increased inositol production compared with unstressed control. MIPS protein isolated from PINO1 transgenics showed salt-tolerant property in vitro confirming functional expression in planta of the PINO1 gene. To our knowledge, this is the first report of a salt-tolerant MIPS from any source.

Inositols are six-carbon cyclohexane hexitols found ubiqui-tously in the biological kingdom, and its metabolism plays a vital role in growth regulation, membrane biogenesis, osmotolerance, and in many other processes. As phosphorylated derivatives, its role as a phosphorus store and as a "second messenger" in signal transduction pathways has long been recognized. myo-Inositol, physiologically the most favored stereoisomer among the eight possible geometric isomers of inositol, also enters into an array of biochemical reactions having diverse functions in cellular metabolism both as free and conjugated and phosphorylated or methylated forms (1)(2)(3).
The primary enzyme for the synthesis of L-myo-inositol 1-phosphate from glucose 6-phosphate is L-myo-inositol-1-phosphate synthase (EC 5.5.1.4; referred to as MIPS), 1 which synthesizes L-myo-inositol 1-phosphate through an internal oxidoreduction reaction involving NAD ϩ . Free inositol is generated by dephosphorylation of the MIPS product by a specific Mg 2ϩ -dependent inositol-1-phosphate phosphatase (EC 3.1. 3.25). This mechanism is followed by all myo-inositolproducing organisms throughout the phylogenetic lines, and MIPS has been identified as an evolutionarily conserved protein (4). The structural gene coding for cytosolic MIPS, termed INO1, was first identified in yeast, Saccharomyces cerevisiae (5,6) and cloned by Klig and Henry (7). Subsequently, the complete nucleotide sequence of the full-length INO1 gene from S. cerevisiae was reported by Johnson and Henry (8). Until now, over 60 genes homologous to INO1 have been cloned and characterized from a wide variety of prokaryotic, archaeal, and eukaryotic sources (9 -12), and the conservation of a probable "core catalytic structure" among all has been proposed (12). The crystal structures of MIPS(s) from Saccharomyces and Mycobacterium have been worked out, providing evidence for the structural insight for the proposed reaction mechanism (13)(14)(15)(16)(17).
In addition to the cytosolic form of MIPS reported from a wide range of plant, animal, and other sources, an organellar form of the enzyme has been demonstrated in the chloroplasts of Pisum sp., Vigna radiata, Euglena gracilis, Oryza sativa, and Phaseolus sp. (18 -22). The chloroplastic form of the enzyme has been found to be similar to the cytosolic enzyme, with respect to biochemical and immunological properties (20), and was demonstrated to be regulated by light and salt (21). At least two molecular forms of the enzyme having ϳ80and ϳ60-kDa subunits have been identified in rice chloroplasts. Proteolytic processing of the ϳ80-kDa subunit to the ϳ60-kDa subunit followed by its phosphorylation have been identified as biochemical events resulting in activation of the chloroplastic MIPS during light and salt growth (23). Although the chloroplastic MIPS has been characterized at the enzyme protein level, identification of its structural gene is still to come.
In this communication, we report the cloning and bacterial expression of the INO1 gene from Porteresia coarctata (Roxb.) Tateoka (PINO1), a halophytic wild rice, biochemical and biophysical characterization of the gene product with special reference to its salt tolerance property, and comparison with the O. sativa INO1 gene (RINO1) product. Further experiments report functional expression of the RINO1 and PINO1 genes as salt-sensitive and salt-tolerant MIPS proteins, respectively, upon introgression into tobacco. Introgression and expression of PINO1 allows growth of the transgenic plants in salt environment with concomitant increased inositol production and maintenance of photosynthetic potential.
Purification of Cytosolic MIPS from Oryza and Porteresia Leaves-Cytosolic MIPS from leaves of O. sativa and P. coarctata were isolated following the procedure developed in this laboratory (20). Low speed supernatant from the leaf homogenate was subjected to 40 -70% (NH 4 ) 2 SO 4 precipitation. Dissolved precipitate was run through a column of Superose 12. Enzymatically active fractions were loaded onto a DEAE-Sephacel column and eluted with a linear 0 -0.25 M NH 4 Cl gradient. Enzymatically active fractions were further purified by filtration through a Bio-Gel A0.5 column. The final preparation was dialyzed in 20 mM Tris-HCl, pH 7.5, containing 10 mM ␤-ME (mercaptoethanol) and 10% glycerol.
Assay of MIPS Enzyme-The enzyme was assayed colorimetrically by the periodate oxidation method of Barnett et al. (24) with modifications and further corroboration by the inositol-1-phosphatase assay as described previously (19 -21). The amount of P i released from the MIPS reaction product upon periodate oxidation or inositol-1-phosphatase hydrolysis was estimated by the method of Chen et al. (25). Protein was estimated according to the method of Bradford (26) using bovine serum albumin as the standard.
Cloning and Sequencing of INO1 Gene from P. coarctata (PINO1) and O. sativa (RINO1)-A full-length cDNA for the MIPS gene has been obtained from Porteresia (PINO1) as well as Oryza (RINO1) leaf poly(A) RNA by reverse transcriptase-PCR. Total RNA was isolated from mature leaves of Oryza and Porteresia following the method of Ostrem et al. (27). Poly(A) RNA was isolated from the total RNA by the poly(A) tract mRNA isolation kit (Promega) following the manufacturer's instructions. 20 -30 ng of poly(A) RNA was used for first strand cDNA synthesis using Superscript II RNase H-reverse transcriptase (Invitrogen). cDNA thus synthesized was used as the template for PCR amplification of the INO1 gene. For cloning of the full-length cDNA of INO1 for Oryza (RINO1) and Porteresia (PINO1), sense (5Ј-3Ј) and antisense (3Ј-5Ј) oligonucleotide primers were designed based on the published RINO1 sequences (GenBank TM accession number AB 012107) with additional sites for NdeI and XhoI at the 5Ј and 3Ј ends, respectively. PCR amplification was done as follows: 94°C for 3 min, 94°C for 1.5 min, 55°C for 1.5 min, and 72°C for 2 min for 32 cycles; and 72°C for 10 min. The amplified product (ϳ1.5 kb) was eluted and purified from the gel and ligated overnight at 4°C to the pGEMT-Easy vector (Promega). The ligation mixture was used for transformation of high efficiency JM109 competent cells (Promega), and transformants were selected on ampicillin/IPTG/X-gal plates. Plasmid DNA was isolated from the transformants in large quantities, purified, and completely sequenced by contracting to the DNA Sequencing Facility at the University of Arizona, Tucson. The sequence data obtained for RINO1 were checked with the published sequences showing complete identity and that for PINO1 was compared through standard bioinformatic tools. The sequences of the PINO1 gene have been deposited in GenBank TM under accession number AF412340. 2 Complementation of RINO1 and PINO1 in Yeast Inositol Auxotrophic Strain FY250 -The procedure followed was that of Bacchawat and Mande (9). PCR-amplified RINO1 and PINO1 genes were cloned into pGEMT-T easy vector and were subsequently cloned downstream of a strong constitutively expressed yeast glyceraldehyde-3-phosphate dehydrogenase (GAPDH) promoter of a yeast multicopy expression vector p426 GAPDH within EcoRI/XhoI restriction sites. The yeast strain FY250 (MAT␣ trp1 his3 ura3 leu2 Ino1::HIS3, a gift from Dr. A. Bacchawat, IMTECH, Chandigarh, India) was transformed (28) with the p426 GAPDH containing RINO1 and PINO1 and with a p426 GAPDH without any insert as control. Transformed cells having the plasmid constructs were allowed to grow in the presence of 10 M inositol but lacking uracil and then gradually transferred to a medium lacking both inositol and uracil.
Generation of Deletion Mutants of PINO1 (⌬PINO1-1 and ⌬PINO1-2) -Two types of deletion mutants of PINO1 were generated. The first one, designated ⌬PINO1-1, was a catalytically inactive PINO1 in which a portion of the PINO1 gene between Asn-342 and Lys-361, a part of the stretch identified as the "core catalytic domain" of the MIPS enzyme protein (12) was deleted by in vitro mutagenesis using the Stratagene Quick-change mutagenesis kit following the manufacturer's protocol. The following primers were used for this purpose: 5Ј-TACAACCACTT-GGGAGCAGCGTGGTCGAT-3Ј and 5Ј-ATCGACCACGCTGCTCCCCA-AGTGGTTGTA-3Ј.
The second mutant of PINO1, designated ⌬PINO1-2, was generated by an internal deletion of the amino acids Trp-174 to Ser-210 of PINO1, a portion identified as non-identical with RINO1. To generate this mutant, two separate fragments, one 519 bp (upstream of the deleted region) and the other 906 bp (downstream of the deleted region) were separately PCR-amplified from a full-length PINO1 gene cloned into pGEMT-Easy using the following two pair of primers: 5Ј-ACATATGT-TCATCGAGAGCTTCCGCGT-3Ј and 5Ј-TCTGTCGACGGACTCCATG-TAAGGCCTAAGCT-3Ј for the upstream fragment, and 5Ј-TCTGTCG-ACAAGGACATCAAGGAGTTCAAGG-3Ј and 5Ј-CTCAAGCTTCTTGT-ACTCCAGGATCATGTTGTTCTCAGGG-3Ј for the downstream fragment. The forward and reverse primers for the upstream-and downstream-amplified fragments contain NdeI and HindIII as the mutated gene terminal restriction sites, whereas the reverse and forward primers for the upstream-and downstream-amplified fragments contain the SalI site. After the PCR amplification of these fragments separately, they were ligated in-frame in a tripartite ligation reaction using the SalI sites introduced in the primers to generate the full-length internally deleted mutant PINO1 gene termed ⌬PINO1-2.
Bacterial Overexpression of the cloned RINO1, PINO1, and ⌬PINO1 Genes-The cDNA for RINO1, PINO1, and the mutant genes ⌬PINO1-1 and ⌬PINO1-2 were subcloned into the NdeI/XhoI or HindIII sites of the expression vector pET-20b(ϩ). The resulting plasmids were introduced into the host strain E. coli BL-21 (DE3) by transformation (29). The bacteria were grown in LB medium at 37°C up to A 600 of 0.5 absorbance units and induced by 0.5 mM IPTG for 6 h. The bacteria were collected by centrifugation and lysed by sonication in a buffer containing 20 mM Tris-HCl, pH 7.5, 10 mM ␤-ME, 2 mM PMSF. The extracts were centri- fuged, and both the pellet and the supernatant were analyzed by 10% SDS-PAGE.
Solubilization and Purification of Expressed RINO1, PINO1, and ⌬PINO1 Gene Products-The pellet fractions containing the expressed proteins were solubilized in a buffer containing urea (8 M urea, 0.5 M NaCl, 20 mM Tris-HCl, pH 7.5, 10 mM ␤-ME, 2 mM PMSF) and kept for 30 min at room temperature. Solubilized samples were centrifuged at 15,000 rpm for 15 min. The supernatant was dialyzed serially (step dialysis with 8, 7, 6, 5, 4, 3, and 2 M urea and finally with urea-free buffer) to remove urea. Dialyzed expressed protein sample was purified by DEAE-Sephacel and Bio-Gel A0.5 as for the native enzyme described earlier (20).
PAGE and Western Blot Analysis-SDS-PAGE was performed according to Laemmli (30). For immunodetection, proteins were blotted onto polyvinylidene difluoride membrane, and the blot was probed with rabbit anti-MIPS antibody (1:1000) raised against purified recombinant MIPS of Entamoeba (31). Bound antibody was detected by chemiluminescence (Amersham Biosciences).
Fluorescence Spectroscopy-Tryptophan fluorescence spectra were recorded using a Hitachi F-4500 spectrofluorimeter. Protein solution (0.1 mg/ml in 20 mM Tris-HCl, pH 7.5) was taken in a quartz cuvette (4 ϫ 4 mm). Excitation wavelength was selected at 295 nm, and the emission was scanned at a speed of 240 nm/min from 310 to 400 nm by using excitation and emission slits at 5 nm each. Each spectrum was an average of three scans. Appropriate control buffer spectra were subtracted from sample spectra to generate the fluorescence spectra of the proteins. Wavelength of maximum emission for each spectrum was determined by derivative analysis using the instrument software.
Fluorescence Quenching Experiments-Tryptophan fluorescence was quenched by titrating the protein solution with 5 M acrylamide or potassium iodide. 2.0 ml of 0.1 mg/ml protein solution was taken in a 3-ml quartz cuvette containing a magnetic stir bar. Excitation and emission wavelength were set at 295 and 340 nm, respectively. Freshly prepared quencher (acrylamide or potassium iodide) was added in small aliquots in the cuvette, and after each addition the solution was stirred magnetically for 1 min, and thereafter the emission reading was taken. The fluorescence readings for all concentrations of the titrant were corrected for the "dilution effect" due to addition of the titrant. The readings were also corrected for the "inner filter effect" according to Equation 1, where F and F corr indicate the uncorrected and corrected fluorescence, and A ex and A em indicate the absorbance of the solution at the excitation and emission wavelengths, respectively. The quenching data were analyzed according to modified Stern-Volmer Equation 2 (32), where F o and F are the fluorescence intensities in absence and presence of the quencher; [Q] indicates the molar concentration of the quencher, and f a indicates the fraction of tryptophans accessible to the quencher.
The accessible fraction f a and the effective Stern-Volmer quenching constants (K SV ) eff are equal to f a . K SV values were obtained from the ordinate intercept and slope, respectively, of the linear portion of the Bis-ANS Fluorescence Assay-Purified recombinant RINO1 and PINO1 proteins were dialyzed against 2 liters of 20 mM Tris-HCl with 10 mM ␤-ME. 2 ml of dialyzed protein (0.1 mg/ml) in each was taken into a 3-ml quartz cuvette containing a magnetic stirrer. Excitation and emission wavelength were set at 370 and 490 nm, respectively. Freshly prepared Bis-ANS solution (372 M stock solution) was added in small aliquots to the cuvette, and after each addition the solution was stirred magnetically for 1 min.
Assay for Aggregation of RINO1 and PINO1 Proteins in Vitro under NaCl Environment-Purified bacterially expressed RINO1 and PINO1 proteins (0.875 M) were assayed for static light scattering in the absence and presence of different NaCl concentrations at 37°C in a Shimadzu-160A spectrophotometer. Optical density of the samples at 360 nm were measured over a period of 170 min at regular intervals where increased OD was indicative of aggregation of the protein samples (33).

Construction of Plant Expression Vectors and Tobacco
Transformation-Full-length clones of RINO1, PINO1, ⌬PINO1-1, and ⌬PINO1-2, as obtained, were subcloned at the XbaI/KpnI site of the plant expression vector pCAMBIA1301 (a gift from Prof. Akhilesh Tyagi, University of Delhi, South Campus) under the control of the constitutive cauliflower mosaic virus 35S promoter and nopaline synthase (Nos) termi-nator. The vector contains the hpt gene for hygromycin resistance and gus as the reporter gene. The resultant constructs SMSA9, SMSA10, 3KSMA1, and 3KSMA2 contain RINO1, PINO1, ⌬PINO1-1, and ⌬PINO1-2 genes respectively. All the constructs and the control plasmid pCAMBIA1301 without any insert were introduced into Agrobacterium tumefaciens LBA 4404 by the freeze-thaw method (34). Tobacco leaf discs were infected with A. tumefaciens containing SMSA9, SMSA10, 3KSMA1, 3KSMA2, and the control plasmid pCAMBIA1301 without any insert. After 3 days of co-cultivation the leaf discs were transferred to the regeneration medium supplemented with cefotaxim (250 mg/liter) and hygromycin (15 mg/liter). Cultures were maintained at 26°C and under continuous illumination provided by white fluorescent tubes. Shoot bud differentiation started after 14 -16 days of culture in the regeneration medium, which elongated into shoots within 30 -35 days. Shoots regenerated in the selection medium were transferred to hormone-free MS medium containing hygromycin (30 mg/liter) and cefotaxim (250 mg/liter) and allowed to develop roots in this medium. Ten putative transformed plants for each construct were screened for stable integration of gus, RINO1, PINO1, ⌬PINO1-1, and ⌬PINO1-2 genes. For histochemical gus assay, leaf segments and roots taken from the putative transformants (T 0 ) were incubated in a buffer containing 100 mM phosphate, 10 mM disodium EDTA, 0.5 mM ferro-and ferricyanide, and 0.1% X-glu dissolved in dimethylformamide at 37°C for 10 -12 h (35). For analysis of introgression of RINO1, PINO1, ⌬PINO1-1, and ⌬PINO1-2 genes by PCR, genomic DNA was isolated (36) from young leaves of both the transformed and untransformed T 0 plants, and PCR amplification of RINO1 and PINO1 genes was done following the conditions mentioned earlier. The transgenic T 0 plants were allowed to flower and set seeds by preventing cross-pollination. Seeds were collected from transformed plants and germinated on medium containing hygromycin (20 mg/liter). Seedlings (T 1 ) resistant to hygromycin were maintained to full growth at 25 Ϯ 2°C and a 16-h light and 8-h dark cycle in the culture room. Plants were analyzed for their photosynthetic efficiency by the Photosynthetic Efficiency Analyzer (Handy-PEA, Hansatech, UK). Photosynthetic performance of individual plants was calculated following the method of Strasser et al. (37). The analysis was done with the kind cooperation from Dr. A. K. Mishra, Utkal University, Bhubaneswar, India.
Estimation of Inositol Content of Transgenic Plants-For isolation and estimation of inositol content of the transgenic plants, the method of Bieleski and Redgwell (38) was followed. Leaf tissue (ϳ500 mg) was homogenized with a mixture of methanol/chloroform/water/trichloroacetic acid (12:5:2:1) for isolation of total water-soluble sugars and polyols. Chloroform and water (1:1) was added to the homogenized sample for phase separation. The aqueous phase containing the bulk of the water soluble plant metabolites including inositol was lyophilized to complete dryness.
For gas chromatography analysis, the lyophilized samples were transferred quantitatively to glass reaction vials in pyridine and evaporated under high vacuum. All the samples were kept in a vacuum dessicator for 24 h over P 2 O 5 for complete removal of water. Samples were tetramethylsilane-derivatized (39) with Tri-Sil-Z (Pierce) and were run through gas-liquid chromatography in a Chemito 1000 gas chromatograph equipped with flame ionization detector. Gas chromatography conditions are as follows: 3% SP-2100 stationary phase (Supelco) supported on chromosorb-W (Sigma) packed in a 1.8-m (length) ϫ 2-mm (inner diameter) glass column with N 2 (flow rate 31 ml/min) as carrier gas, and the oven temperature programmed between 130 and 320°C at 10°C/min. Quantification was made against similar runs with authentic myo-inositol as standard.

Cloning and Sequencing of the Porteresia Gene (PINO1) for MIPS and Its Comparison with That from Oryza (RINO1)-
Cloning and sequencing of RINO1 and PINO1 genes have been described under "Experimental Procedures," and the complete nucleotide sequences of PINO1 are presented in Fig. 1A. Comparison of deduced amino acid sequences of PINO1 with those of RINO1 (Fig. 1B)  Such alterations make the PINO1 gene product as composed of a total of 512 amino acids, longer than that of RINO1 by 2 amino acids.
Functional Complementation of PINO1 in Yeast-For functional identification of the PINO1 gene, a complementation experiment was performed in which a yeast inositol auxotrophic strain, FY250, was used. The PINO1 PCR product was cloned into the yeast multicopy p426 GAPDH vector and transformed into the yeast strain FY250, along with appropriate controls as described under "Experimental Procedures." The FY250 strains, transformed with PINO1, RINO1 (as a positive control), or control vectors without any insert were grown in the presence or absence of inositol. Results presented in Fig. 2 show that although all three transformed FY250 strains grew well in presence of 10 M inositol and without uracil (Fig. 2, left  panel), only the FY250 strains transformed with p426 GAPDH having either RINO1 or PINO1 grew in the absence of both inositol and uracil (Fig. 2, right panel), providing evidence for functional complementation of inositol biosynthesis by PINO1, as with RINO1, in the auxotrophic yeast strain. The mutant ⌬PINO1-2 was similarly identified as complementing the inositol auxotroph, whereas ⌬PINO1-1 failed to do so (data not presented).
Bacterial Overexpression and Purification of the RINO1 and PINO1 Proteins-In order to characterize the purified gene products of PINO1 and RINO1, the corresponding PCR products were cloned into the bacterial expression vector pET-20b(ϩ), and induction of the gene products by IPTG was achieved as described under "Experimental Procedures." Both RINO1 and PINO1 were expressed as the expected ϳ60-kDa proteins predominantly in the particulate fraction as judged by SDS-PAGE of the induced cells. The particulate material was solubilized in 8 M urea buffer and further analyzed. The expressed RINO1 and PINO1 proteins were recovered in the soluble fraction both on SDS-PAGE and the corresponding Western blots (Fig. 3, A and B).
The expressed RINO1 and PINO1 proteins were purified to homogeneity following the procedure of DEAE-Sephacel chromatography and Bio-Gel A0.5 filtration as outlined under "Experimental Procedures." Results presented in Fig. 3 (C and D) show elution of both RINO1 (C) and PINO1 (D) proteins as near-symmetric protein peaks coincident with the MIPS activity, suggesting homogeneity of the purified proteins. Active fractions of the individual peaks show single protein bands on SDS-PAGE (Fig. 3, C and D, inset, left panel) and subsequent Western blots (Fig. 3, C and D, inset, right panel). The ⌬PINO1-1 and ⌬PINO1-2 gene(s) were similarly expressed, and the protein was purified to homogeneity through Western blot analysis and enzyme assay, respectively (data not presented).
Biochemical Characterization of the Expressed RINO1 and PINO1 Gene Products-Both the purified and expressed RINO1 and PINO1 proteins were biochemically characterized and compared with the corresponding enzyme proteins isolated from native sources ( Table I). Estimates of K m and V max values for the substrate (Glc-6-P) and co-factor (NAD) were obtained with Bio-Gel A0.5 purified proteins using the Lineweaver-Burk plot. The purified recombinant RINO1 and PINO1 proteins show higher K m values for Glc-6-P than that for the corresponding purified native enzymes. The lower K m values for Glc-6-P for recombinant PINO1 protein suggest a higher substrate specificity compared with the recombinant RINO1 proteins. For both the proteins, the optimum temperature for enzyme activity was at 37°C, and at optimum pH for RINO1 and PINO1 recombinant proteins the activities were between 7.5 and 8.
The expressed RINO1 and the PINO1 proteins or the corresponding purified native MIPS enzymes differed greatly in their response to in vitro NaCl concentration (Fig. 4, A and B). As evident, both the native and the expressed RINO1 proteins exhibit concentration-dependent inhibition of the enzyme activity in vitro in the presence of increasing NaCl concentrations, the MIPS activity being completely abolished at 500 mM NaCl. In striking contrast to such situations, both the native and the expressed PINO1 proteins showed no inhibition of enzyme activity in vitro up to a concentration of 500 mM NaCl. Most interesting, although the bacterially expressed ⌬PINO1-2 gene product was catalytically active as the PINO1 gene product, it shows inhibition of the enzyme activity in the presence of increasing NaCl concentrations as in case of the RINO1 protein (Fig. 4B). The ⌬PINO1-1 gene product, on the other hand, although immunologically cross-reactive to the same anti-MIPS antibody, was catalytically inactive (data not presented).
Structural Basis of Salt Tolerance of the PINO1 Protein as Compared with the Salt-sensitive RINO1 Protein-In order to understand the nature of the structural changes in RINO1 protein causing in vitro inhibition of enzymatic activity due to addition of salt in contrast to the PINO1 protein (Fig. 4, A and  B), we performed gel permeation chromatography on Superose 12 of RINO1 and PINO1 proteins both in the presence and   Fig. 5 (A-D). It is seen that whereas the RINO1 protein is eluted as a single peak in the absence of salt (Fig. 5A), addition of 400 mM NaCl during chromatography leads to reduction in the original enzyme activity peak with concurrent appearance of high molecular weight enzymatically inactive fractions (Fig.  5B). The change in the enzyme elution pattern was judged by enzyme assay and detection of the MIPS protein by SDS-PAGE (Fig. 5, A and B, inset, left panel) and corresponding immunoblots of different fractions (Fig. 5, A and B, inset, right panel). This experiment suggests oligomerization of the RINO1 protein in the presence of 400 mM NaCl resulting in considerable decline in enzyme activity. In contrast to such situations, the PINO1 protein elutes at the same place as a native protein in the gel filtration column both in the presence and absence of 400 mM NaCl and without any change in enzymatic activity (Fig. 5, C and D). The salt-induced oligomerization of RINO1 protein was further verified in experiments described in Fig.  5E, where increasing aggregation of RINO1 protein in vitro at 37°C was noted with increasing NaCl concentration. In contrast, the PINO1 protein did not show any aggregation with increasing NaCl concentrations under identical conditions (Fig. 5F). Next, the tryptophan fluorescence spectra of the recombinant RINO1 and PINO1 proteins under different conditions were investigated (Fig. 6, A-E). In the absence of added salt, RINO1 protein shows significantly higher tryptophan fluorescence intensity than the PINO1 protein at the wavelength of maximum emission (336 nm). Tryptophan fluorescence intensity of RINO1 is quenched significantly in the presence of increasing NaCl concentrations, whereas that of PINO1 remains rather unaltered. It is also interesting to note that at a salt concentration of over 600 mM, the fluorescence intensities of both RINO1 and PINO1 become comparable (Fig. 6, A and B).
To understand further the differential behavior of RINO1 and PINO1 toward NaCl, we carried out tryptophan fluorescence quenching experiments where acrylamide and iodide were used as the complementary set of water-soluble quenchers. Acrylamide is a neutral quencher and is known to have the ability to penetrate into the protein interior. On the contrary, iodide is a negatively charged and highly hydrated bulky quencher having no ability to penetrate the protein interior, its quenching ability being mainly dependent on the location of the neighboring charged groups. We have analyzed our data by assuming varied and heterogeneous emissions from multiple tryptophans, and we report the quenching constant as the effective Stern-Volmer constant (K SV ) eff that represents the weighted average of the quenching constants of the individual tryptophan residues and may contain contributions from both static and dynamic quenching. Fig. 6C shows the acrylamide quenching data in the form of the modified Stern-Volmer plot. The quenching constant (K SV ) eff and the quenchable fractions f a are reported in Table II. Analysis of such data shows that the RINO1 protein in absence of added salt has a (K SV ) eff of 6.0 M Ϫ1 for acrylamide quenching. In the presence of 400 mM NaCl the (K SV ) eff of RINO1 decreased to 4.2. The (K SV ) eff of PINO1 remained constant around 3.4 M Ϫ1 both in the presence and the absence of added salt. Because the microenvironment of tryptophan residues of RINO1 and PINO1 proteins is very similar, the changes in the quenching constants in the absence of salt reflect changes in the outer surface of the proteins making penetration of acrylamide relatively easier in RINO1 than in PINO1. Additions of salts rearranges groups on the surface of RINO1 in such a way as to cause greater hindrance toward the penetration of acrylamide. For both RINO1 and PINO1, tryptophan fluorescence is 100% quenched by acrylamide.
The data for the quenching of tryptophan fluorescence by iodide is shown in Fig. 6D. Iodide is able to quench only about 20% of tryptophan fluorescence of PINO1 at a (K SV ) eff of 1.7 M Ϫ1 , and 80% of the tryptophan is not quenchable by iodide. On the contrary, about 50% of total tryptophan fluorescence of RINO1 is quenched by iodide. The quenchable groups of RINO1 are of two types. About 10% of total fluorescence is quenched with a (K SV ) eff of 22 M Ϫ1 , whereas 40% is quenched with a (K SV ) eff of 1.9 M Ϫ1 , a value very similar to that of PINO1 (Table  II). Thus RINO1 has more tryptophan groups close to the surface compared with PINO1. The iodide quenching data thus reveal that although the protein interior in both RINO1 and PINO1 remains similar to each other, there is considerable difference in tryptophan accessibility due to the difference in exposition of ionic groups on the surface of the two proteins.
The differential conformational behavior of RINO1 and PINO1 proteins with added salts is also reflected in their surface hydrophobicity characteristics. Bis-ANS is a highly conformation-sensitive fluorescent probe that binds to the exposed hydrophobic sites on the protein surfaces. In the absence of salt, RINO1 shows higher surface hydrophobicity than PINO1 (Fig. 6E). In presence of 0.4 M NaCl, surface hydrophobicity of PINO1 is slightly (ϳ15%) reduced but that of RINO1 is reduced by more than 50%. This also shows that in the presence of salt, exposed hydrophobic groups in RINO1 get buried inside the globular structure, whereas in PINO1 such changes are minimal. Data presented in Fig. 6E also reflect that there are differences in the tertiary level of organization between RINO1 and PINO1 both in the presence and absence of salts.
Functional Introgression of RINO1, PINO1, ⌬PINO1-1, and ⌬PINO1-2 in Tobacco and the Phenotype of the Transgenic Plants-Experiments described in Fig. 4 suggested that the To ascertain whether such properties are retained by the individual genes on introgression and expression in the transformed plant system and to find out the effect of such functional introgression on the transformed system, suitable plant expression constructs of the RINO1, PINO1, ⌬PINO1-1, and ⌬PINO1-2 genes were made and the constructs used for tobacco plant transformation as detailed under "Experimental Procedures." Along with these, control plants were also raised where the plants were transformed with empty vector. Out of 20 -30 transgenic lines obtained for different constructs, transgenic lines (T 0 ) positive for both gus and the different transgene(s) were selected. Ten such independent transgenic lines for each construct were allowed to flower and set seeds. Seeds from such plants (T 0 ) were germinated in the presence of hygromycin, which displayed a simple (3hyg R :1hyg S ) segregation pattern in its progeny indicating single locus of insertion. Independent lines of transgenic plants (T 1 ) were analyzed for expression of the respective genes and the empty vector transformed, RINO1, PINO1, ⌬PINO1-1, and ⌬PINO1-2 transformed plants were analyzed and compared for their phenotype under salinity stress. The results of such experiments are depicted in Fig. 7 (A-E). In contrast, however, only the PINO1-transformed plants were able to grow between 200 and 300 mM NaCl in the growth media without appreciable loss of chlorophyll or growth vigor in comparison to others (Fig. 7A, c). The difference is most evident in plants grown at 300 mM NaCl, although with a less vigorous growth of the PINO1 transformed plants (Fig. 7A,  a-e).
The MIPS proteins isolated from the empty vector, RINO1, PINO1-, ⌬PINO1-1, and ⌬PINO1-2 transformed plants were isolated from the leaves and purified and tested for in vitro salt tolerance. Results of such experiments, described in Fig. 7C, clearly demonstrate that whereas the isolated enzymes from empty vector, RINO1, ⌬PINO1-1, and ⌬PINO1-2 transformed plants exhibit salt sensitivity as expected, the same from the PINO1 transformed plants show in vitro salt tolerance as the corresponding bacterially expressed MIPS proteins (Fig. 4), providing evidence for functional expression of the proteins in planta on introgression of the respective genes.
To ascertain the efficiency of the enzymatic production of inositol in vivo by the expressed MIPS under transgenic conditions, inositol content of the different transgenic plants grown in increasing concentrations of NaCl were also analyzed (Fig. 7D). It was found that the PINO1 transformed plants produced the highest amount of inositol (ϳ1.4 -0.7 mol/g fresh weight) when grown in growth media containing 0 -300 mM NaCl in comparison to the similarly grown empty vector transformed or ⌬PINO1-1 transformed (ϳ0.8 -0.1 mol/g fresh weight) or the RINO1 and ⌬PINO1-2 transformed (ϳ1.2-0.3 mol/g fresh weight) plants. Both RINO1 and the ⌬PINO1-2 plants produce approximately ϳ1.5-3-fold higher inositol over the empty vector or ⌬PINO1-1 transformed plants. About a 2-7-fold increase in the inositol content was noticed in PINO1 transformed plants grown in 300 mM NaCl compared with the RINO1, ⌬PINO1-2, ⌬PINO1-1, and empty vector transformed plants, respectively, grown under similar conditions.
The photosynthetic competence of the control and the transformed plants was also monitored (Fig. 7E). Such results show that the photosynthetic efficiency remains unchanged in all four types of plants grown at 0 or 100 mM NaCl. Reduction of photosynthetic competence to ϳ20% was noticed in the empty vector or ⌬PINO1-1 transformed plants growing in 200 mM NaCl, whereas the RINO1, ⌬PINO1-2, or the PINO1 transformed plants grown at 200 mM NaCl retain ϳ60 -80% photosynthetic competence in comparison to the corresponding unstressed plants. Drastic reduction of photosynthetic competence was recorded for the empty vector transformed control or ⌬PINO1-1 plants grown at 300 mM NaCl, whereas RINO1, ⌬PINO1-2, and PINO1 transformed plants show ϳ20 -50% photosynthetic competence compared with corresponding unstressed plants at this NaCl concentration. The PINO1 transformed plants showed more than ϳ2-fold higher photosynthetic competence over the two others at 300 mM NaCl. These results suggest considerable protection of the photosynthetic apparatus in PINO1 transgenic plants, with similar protection in RINO1 and ⌬PINO1-2 transformed plants albeit to a different degree, resulting in much less growth retardation of the transgenic plants in saline media in comparison to the empty vector transformed or ⌬PINO1-1 transformed ones.
Influence of Inositol on in Vitro Growth of Tobacco Plants in Salt Medium-Following the results depicted in Fig. 7, the growth pattern of wild tobacco plants under salt environment in the presence of increased inositol concentrations in the growth media were analyzed. Plants grown in the presence of increasing concentrations of NaCl (0 -300 mM) in MS medium supplemented with increased inositol (2 or 5 mM) show better growth and photosynthetic efficiency than the plants grown in normal MS medium with 0.55 mM inositol (Fig. 8A). Photosynthetic efficiency remains unchanged in all types of plants grown at 0 or 100 mM NaCl. Reduction of photosynthetic competence to ϳ20% or less was noticed in wild plants growing in 200 -300 mM NaCl, whereas plants grown similarly but with increased inositol (2-5 mM) retain ϳ70 -40% photosynthetic competence in comparison to the corresponding unstressed plants (Fig. 8B). DISCUSSION Although studies during the last decade have enhanced our understanding of plant responses to a stressful environment, there are still many pieces to the puzzle that elude mechanistic insights. One aspect, separate from the current emphasis on stress sensing and signaling, deals with questions about evolutionary changes of protein sequence and structure that may be adaptive to functioning under stressed conditions. Are the proteins, or at least some of them, crucial for proper functioning under stress in stress-tolerant species specifically designed to function in a cellular environment characterized by tolerance to various abiotic stress factors? In case of salt tolerance, a multigenic trait, the quest for an answer to such an intriguing question resides in a genomic and proteomic comparison of the glycophytes with their halophytic models, considerations that have prompted analysis of Arabidopsis genes with those of Thellungiella, its halophytic relative (40). A similar comparison can be presumed between O. sativa L., the cultivated rice, and its halophytic wild relative, P. coarctata (Roxb.) Tateoka, which grows abundantly in the saline coastal region, and the system can very well be used for bio-prospecting of halotolerant homologues of rice genes.
To carry out an analysis that investigates such questions has been the primary objective of the present work. We have chosen the enzyme MIPS, which is conserved across evolutionarily diverse taxa (4,12), produces inositol, and is known to function in varied biochemical activities and also during stress in both prokaryotic and eukaryotic organisms (3,(41)(42)(43)(44)(45). Among the 65 INO1 genes known (12), the INO1 from Archaeoglobus fulgidis (11) has been shown to code for a thermotolerant protein, the only stress-tolerant MIPS so far reported. The present report of a salt-tolerant MIPS from P. coarctata is the only one known from any eukaryotic source until now.
Analysis of the nucleotide sequences of RINO1 and PINO1 genes establish that the PINO1 gene differs from the RINO1 gene considerably with respect to its organization between amino acids 173 and 320. The organization is characterized by deletion, addition, conservative substitution, and rearrangement with two additional amino acids thereby making the PINO1 gene longer by two amino acids as enumerated (Fig.  1B). A similar situation has been reported for the allene oxide cyclase homologue ("mangrin") from the mangrove, Bruguiera sexangula (44), wherein a stretch of 70 amino acid residues, different in its organization from the allene oxide cyclase from Arabidopsis or Lycopersicum, conferred salt tolerance to E. coli, yeast, and tobacco cells. Furthermore, whether the PINO1 gene is the only MIPS-coding gene in Porteresia or a duplicated paralogue of INO1 having a different organization has yet to be resolved. Functional identification of the PINO1 gene was made by using a yeast strain FY250, auxotrophic for inositol by means of an insertion of HIS3 in the INO1 open reading frame (9), a method adopted by others as well (9,45). Overexpression of both RINO1 and PINO1 was achieved through pET-20b(ϩ) bacterial expression vector. The protein was expressed initially in the insoluble fraction but could be solubilized by urea followed by renaturation in its active form through slow removal of urea. Both the RINO1 and PINO1 expressed proteins were purified to apparent homogeneity by procedures established earlier (Fig. 3) along with the ⌬PINO1-1 and ⌬PINO1-2 expressed proteins. Biochemical characterization of the RINO1 and the PINO1 expressed enzyme proteins (Table I) reveal comparable properties with the other known MIPS proteins (4). However, one striking difference between RINO1 and PINO1 was their response toward the NaCl effect in vitro. Both native and bacterially expressed PINO1 proteins turned out to be tolerant to NaCl up to 500 mM in vitro in striking contrast to the corresponding RINO1 protein (Fig. 4, A and B). However, the expressed ⌬PINO1-2 protein, the truncated PINO1 protein where the stretch between aa 174 and 210 has been deleted by in vitro mutagenesis, shows salt sensitivity, the RINO1 protein, although fully active catalytically in contrast to the ⌬PINO1-1, which turned out to be catalytically inactive (Fig. 4B). Such experiments suggest that this amino acid stretch is the likely functional domain for conferring salt stability in the PINO1 protein.
To search for an explanation for the structural basis of such contrasting characteristics between the RINO1 and PINO1 proteins, several experiments were performed. Gel filtration and aggregation assays of the RINO1 and PINO1 proteins as influenced by NaCl revealed that the RINO1 protein undergoes oligomerization in the presence of NaCl with concomitant loss of enzyme activity, whereas the PINO1 protein remains unaltered under similar conditions (Fig. 5). To probe into the mechanism of such salt-mediated aggregation and loss of activity, this was followed by a series of experiments in which the effect of salt on the tryptophan fluorescence of RINO1 and PINO1 proteins was monitored (Fig. 6). Progressive decrease of fluorescence intensity of RINO1 protein with increasing salt concentration (Fig. 6A) indicates structural alterations of the saltsensitive protein. However, the emission maximum of RINO1 remains invariant as a function of increasing salt concentration, suggesting that the tryptophan environment remains unchanged. Because tryptophan residues usually remain buried within the globular structure, the salt-induced changes do not interrupt the tryptophan microenvironment. It probably moves other protein segments closer to tryptophan to facilitate energy transfer and hence reduce intensity. The structure of PINO1 protein is stable to the addition of salts. Because salts screen electrostatic interactions, there is considerable difference in the exposition of charged residues on the outer surface of RINO1 and PINO1. Furthermore, because oligomerization would affect mainly the protein surface and not the globular interior, this helps explain the changes in the acrylamide quenching constants of RINO1 with an increase in salt concentration (Table II). The penetration of the acrylamide into the globular interior through the surface is hindered due to protein-protein interaction in the oligomer, although the microenvironment of tryptophan remained unaltered. The salt sensitivity of RINO1 and its absence in PINO1 can thus be attributed to their difference in ionic environments prevailing on their surface and as evidenced from the iodide quenching data, due also to difference in hydrophobicity close to the surface.
Some preliminary structural comparisons were done on the RINO1 and PINO1 proteins after threading the two sequences on the yeast MIPS structure worked out recently (13,14) and available in the Protein Data Bank. The models generated for RINO1 and PINO1 were compared with the yeast MIPS mon-omer that showed close similarity with the yeast MIPS structure. However, when the two models for RINO1 and PINO1 proteins were compared with each other, a striking difference was observed in the helix following the region where the RINO1 and PINO1 amino acid sequences differed the most (aa 174 -210). The helix is continuous in the case of RINO1 but is disrupted in PINO1 (data not presented). That this break in the helix might be responsible for the structural integrity of the multimeric PINO1 protein under salt environment is evidenced by the salt sensitivity of the ⌬PINO1-2 protein (Fig. 4B). Further detailed work will be required to elucidate the structurefunction relationship of the two proteins in terms of salt sensitivity vis à vis salt tolerance.
By having established the salt-tolerant property of PINO1, an obvious question to us was its probable functional expression in planta on introgression and analysis of the transgenic plants thus generated. Transgenic tobacco plants were raised by introgression of RINO1, PINO1, ⌬PINO1-1, and ⌬PINO1-2 gene(s), and their growth in the presence of increasing NaCl concentrations was compared with that of empty vector transformed plants (Fig. 7). A noteworthy observation was the growth performance of the PINO1 transformed plants that showed near-normal phenotype with marginal growth reduction between 200 and 300 mM NaCl. MIPS protein(s) isolated from the variously transformed plants confirmed their characteristic salt sensitivity (RINO1 and ⌬PINO1-2) or tolerance property (PINO1) in vitro (Fig. 7C), providing evidence for functional expression in planta of the different transgenes following introgression. Moreover, the specific activity of the isolated PINO1 or RINO1 proteins corresponds to that of the bacterially expressed proteins, with the PINO1 protein being enzymatically more active. Because the inherent MIPS of the tobacco plant is a salt-sensitive one, the empty vector transformed control plants and ⌬PINO1-1 (catalytically dead mutant of PINO1) produced decreased amounts of inositol in increasing NaCl in the media. The sustained production of higher levels of inositol by the PINO1 transformed plants during growth between 0 and 200 mM NaCl reflects the uninhibited functioning of the PINO1 gene product under such growth conditions. Even at 300 mM NaCl, these plants produced the highest amount of inositol in comparison to both RINO1 and the ⌬PINO1-2 transformed plants which in turn produce 1.5-2-fold higher inositol compared with the control plants or the catalytically dead mutant of PINO1 (⌬PINO1-1) transformed plants. A comparison of the photosynthetic efficiency (Fig. 7E) and the inositol content (Fig. 7D) of the transformed plants reveals a fair correlation between the two.
If increased synthesis and accumulation of inositol are responsible for better growth and photosynthetic efficiency of the PINO1 transformed plants under salinity stress, growth of wild untransformed plants under salinity stress is expected to be facilitated by external supply of inositol to the growth medium. In order to verify such a possibility, wild tobacco plants were grown in vitro under increasing NaCl concentrations containing elevated levels of inositol in the growth media resulting in similar observations, thus confirming the role of inositol in salt tolerance (Fig. 8). Increased synthesis of inositol and its transport has been shown previously to play a major role in sequestration of sodium ions and maintenance of photosynthetic competence during salt stress (42,46). A protective role for inositol in maintaining the photosynthetic efficiency of the plants, as demonstrated previously for other osmolytes like glycine-betaine or proline (43,(47)(48), may also be inferred from such results.
Metabolic engineering of various pathways for production of osmolytes including polyols during salt stress have been at-tempted by a number of investigators (43,(47)(48). Regulation of inositol pathway in relation to salt stress has also been studied earlier (21,42). In a facultative halophyte-like Mesembryamthemum crystallinum, a coordinate transcriptional induction of INO1 and the inositol methyltransferase (IMT) gene(s) under salt environment for production of both inositol and its methylated derivative, pinitol has been documented (49). Furthermore, overexpression of MIPS coded by TUR1 cDNA from Spirodela polyrrhiza in Arabidopsis led to an increased free inositol pool (50). Because both inositol and pinitol are known to act as osmoprotectants during salinity stress, in systems lacking the IMT gene and thus unable to produce pinitol, a mechanism for increased synthesis of inositol under salt environment may be provided by a salt-tolerant MIPS. The experiments described herein increase the possibility of raising salt-tolerant plants by manipulation of the evolutionarily con- served inositol pathway through introduction of a salt-tolerant PINO1 gene. Furthermore, in contrast to many other systems where metabolite limitations and growth compromise due to inositol depletion (49,51) may impose restrictions on metabolic engineering for production of different osmolytes (47) for raising salt-tolerant plants, inositol requirements for normal growth of the plant is satisfied even during stress by the salttolerant MIPS in addition to its proposed function as an osmolyte as enumerated here.