Research article

Cold sensitivity in rice (Oryza sativa L.) is strongly correlated with a naturally occurring I99V mutation in the multifunctional glutathione transferase isoenzyme GSTZ2

Sang-Ic Kim, Virgilio C. Andaya, Thomas H. Tai


GSTZs [Zeta class GSTs (glutathione transferases)] are multifunctional enzymes that belong to a highly conserved subfamily of soluble GSTs found in species ranging from fungi and plants to animals. GSTZs are known to function as MAAIs [MAA (maleylacetoacetate) isomerases], which play a role in tyrosine catabolism by catalysing the isomerization of MAA to FAA (fumarylacetoacetate). As tyrosine metabolism in plants differs from animals, the significance of GSTZ/MAAI is unclear. In rice (Oryza sativa L.), a major QTL (quantitative trait locus) for seedling cold tolerance has been fine mapped to a region containing the genes OsGSTZ1 and OsGSTZ2. Sequencing of tolerant (ssp. japonica cv. M-202) and sensitive (ssp. indica cv. IR50) cultivars revealed two SNPs (single nucleotide polymorphisms) in OsGSTZ2 that result in amino acid differences (I99V and N184I). Recombinant OsGSTZ2 containing the Val99 residue found in IR50 had significantly reduced activity on MAA and DCA (dichloroacetic acid), but the Ile184 residue had no effect. The distribution of the SNP (c.295A>G) among various rice accessions indicates a significant association with chilling sensitivity in rice seedlings. The results of the present study show that naturally occurring OsGSTZ2 isoforms differ in their enzymatic properties, which may contribute to the differential response to chilling stress generally exhibited by the two major rice subspecies.

  • chilling tolerance
  • glutathione transferase Zeta
  • isoform
  • Oryza sativa
  • single nucleotide polymorphism


Soluble GSTs (glutathione transferases, EC belong to a superfamily of multifuctional enzymes that have similar tertiary structures and are primarily known for their ability to conjugate glutathione (GSH, γ-Glu-Cys-Gly) to various electrophilic xenobiotics, and to bind and sequester an array of hydrophobic compounds of endogenous and xenobiotic origin [1]. Members of this superfamily are organized into classes based on primary sequence similarities. GSTZ (Zeta class GST) was first identified through homology searches of ESTs (expressed sequence tags) for GST-like sequences [2]. GSTZ is found in species ranging from fungi and plants to humans and has been shown to be identical with MAAI [MAA (maleylacetoacetate) isomerase, EC] [3]. MAAI catalyses the GSH-dependent isomerization of MAA to FAA (fumarylacetoacetate), the penultimate step in tyrosine catabolism, an essential pathway in animals [1,4]. Various human disorders result from deficiencies in the enzymes of phenylalanine and tyrosine catabolism with the exception of MAAI/GSTZ [1]. In addition to MAAI activity, GSTZ also catalyses the biotransformation of α-halo acids such as DCA (dichloroacetic acid) which GSTZ oxygenates to glyoxylic acid [5].

The initial steps in the breakdown of tyrosine lead to the generation of homogentisate, the first committed intermediate, which is subsequently degraded to fumarate and acetoacetate via HGO (homogentisate dioxygenase), MAAI/GSTZ and FAH (FAA hydrolase). In plants, these enzymes are present and functional; however, the significance of this part of the tyrosine catabolic pathway is uncertain given that plants are able to regulate biosynthesis of aromatic amino acids and homogentisate is used in other key pathways [6]. Little is known about what, if any, other functional roles GSTZ/MAAI, which were first identified from transcripts that accumulated in senescing carnation petals [7], may play in plants. Arabidopsis thaliana AtGSTZ1 has been shown to have similar activity [8] and structure [9] to human GSTZ [2,4,10]. Overexpression of the OsGSTZ2 gene, which encodes an MAAI in rice (Oryza sativa L.), results in enhanced germination and seedling growth at low temperatures, suggesting a possible role for tyrosine catabolism in abiotic stress tolerance [11].

The two major subspecies of O. sativa are indica and japonica. In temperate regions, the majority of rice grown is of the japonica subspecies, which is generally more tolerant to low temperatures (≤15–20 °C) [12]. Prolonged exposure of rice seedlings to temperatures ≤10 °C results in wilting and necrosis of chilling-susceptible rice cultivars. Genetic analysis of a recombinant inbred line mapping population derived from a cross between cv. M-202, a cold-tolerant temperate japonica, and cv. IR50, a cold-sensitive indica, resulted in the identification of qCTS12, a major QTL (quantitative trait locus) on rice chromosome 12 for tolerance to low-temperature-induced wilting and necrosis [13].

The M-202 allele of qCTS12 confers tolerance to IR50, explaining approx. 40% of the phenotypic variance observed, and the locus has been fine mapped to a 55 kb region containing the tandemly arranged OsGSTZ1 and OsGSTZ2 genes [14]. Each gene consists of nine exons and eight introns encoding proteins of 212 and 230 amino acids (76% sequence identity) respectively [15]. The OsGSTZ genes are primarily expressed in leaves, with OsGSTZ1 constitutively expressed at approx. a 12-fold higher level than OsGSTZ2 [15,16]; however, OsGSTZ2 expression is strongly induced by jasmonate and various stresses [16].

In the present paper, we report the identification of two non-synonymous SNPs (single nucleotide polymorphisms) between the M-202 and IR50 alleles of OsGSTZ2. The two GSTZ2 isoforms were expressed in Escherichia coli, purified, and their ability to catalyse the isomerization of MAA to FAA and the oxygenation of DCA to glyoxylate was determined. A significant difference in activity on both substrates was observed, which was due to one of the amino acid substitutions (I99V). We hypothesize that this functional difference in the OsGSTZ2 isoforms explains, in part, the differential response observed between cold-tolerant and cold-sensitive rice cultivars.


Plant material

Rice (Oryza sativa L.) cv. M-202 (ssp. japonica) and cv. IR50 (ssp. indica) were the source of the DNA for genomic sequencing of OsGSTZ1 and OsGSTZ2, and of the mRNA for cDNA sequencing, gene expression analysis, and cloning and expression of the OsGSTZ2 isoforms.

Reagents and services

Reagents were obtained from Sigma–Aldrich. Primers were synthesized by IDT. Restriction enzymes were from New England Biolabs. DNA sequencing was performed by the University of California at Davis Sequencing Facility.

Genomic sequencing

OsGSTZ1 and OsGSTZ2 were sequenced from M-202 and IR50 genomic DNA isolated as described previously [17]. Primers were designed to amplify overlapping 1-kb fragments using the Nipponbare reference sequence [18]. Sequence data were obtained by direct sequencing of PCR products that were amplified in 50 μl reaction volumes consisting of 50 ng of genomic DNA template, 0.25 μM primers, 1×PCR buffer, 4 μl of 25 mM dNTPs and 1 unit of ExTaq polymerase (TaKaRa Bio USA). Four independent PCRs for each line were combined and then purified using a QIAquick PCR purification kit (Qiagen) prior to sequencing. Sequences were analysed using the Contig Express program of Vector NTI Advance 10 (Invitrogen). Sequences were compared with reference japonica (cv. Nipponbare) and indica (cv. 93–11) rice genome sequences from the Gramene database ( Multiple sequence alignments were generated using the ClustalW program (

Cloning and sequencing of cDNA and construction of plasmids

Total RNA samples were isolated from leaves of M-202 and IR50 using TRIzol® reagent (Invitrogen). First-strand cDNA was synthesized from 5 μg of total RNA by specific primers using the SuperScript™ III reverse transcriptase system (Invitrogen) according to the manufacturer's instructions. PCR products were cloned into PCR2.1-TOPO vector (Invitrogen) for sequencing. Specific primers for cDNA amplification for sequencing were as follows: cGSTZ1 (forward, 5′-CGGCTCACATTCACTCTCACCTCA-3′, and reverse, 5′- TGCTCATTTAACCCTTCAGGCCAG-3′), and cGSTZ2 (forward, 5′- ATCATCCCCAATTTTCTCTTCATTCATAAT-3′, and reverse, 5′- AGACATGTACTCATTCATTTGATGACAAATTC-3′). For recombinant protein expression, the GSTZ1 and GSTZ2 cDNAs were amplified with the gene-specific primers pGSTZ1 (forward, 5′-TCGCGGATCCGAATTCATGGCGGCGGCGGAGAAGAC-3′, and reverse, 5′-GACGGAGCTCGAATTCTCAGGATGAAGGTGCATCTG-3′) and pGSTZ2 (forward, 5′-TCGCGGATCCGAATTCATGGCGTCGTCAAAGCCAATC-3′, and reverse 5′-GACGGAGCTCGAATTCTTAGCATGAAGGTGCATCTGGCTG-3′). PCR products were cloned into the pET-28a expression vector (EMD Chemicals) using an In-fusion™ PCR cloning kit (Clontech). The recombinant proteins (Ile99Ile184 and Val99Asn184) were obtained by swapping the BamHI fragments of OsGSTZ2 cDNA of M-202 and IR50. Fidelity of the constructs was confirmed by sequencing.

Three Arabidopsis cDNA clones, AtHGO (U24625), AtFAH (U21503) and AtGSTZ1 (U21503) were obtained from the Arabidopsis Biological Resource Center ( Specific primers for PCR amplification were as follows: AtHGO (forward, 5′-TCGCGGATCCGAATTCATGGCGTTGCTGAAGTCTTT-3′, and reverse, 5′-GACGGAGCTCGAATTCTCAAGGCGGTGAAGGAACAA-3′). AtFAH (forward, 5′-TCGCGGATCCGAATTCATGGAAGAGAAGAAGAAGGA-3′, and reverse, 5′GACGGAGCTCGAATTCTTACTCCGAAGCTCCTGGTT-3′); and AtMAAI/GSTZ (forward, 5′-TCGCGGATCCGAATTCATGGCGAATTCCGGCGAAGA-3′, and reverse, 5′-GACGGAGCTCGAATTCTCAGATGGTGGAAGAAGGAG-3′). PCR products were directly cloned into the pET-28a expression vector using an In-fusion™ PCR cloning kit and were verified by sequencing as described above.

Expression of recombinant proteins in E. coli

The plasmid construct was transformed into the E. coli strain BL21(DE3) (EMD Chemicals). The E. coli cells were grown in 100 ml of LB (Luria–Bertani) broth until D600 reached 1.0. Expression of the recombinant protein was induced by addition of 0.3 mM IPTG (isopropyl β-D-thiogalactopyranoside) for 3 h at room temperature (24 °C). E. coli cells were lysed by sonication and the crude lysate was used for enzyme assays or for further purification. All of the recombinant proteins generated by using the pET28a expression vector had a His6 tag at the N-terminus. Purification of His6-tagged recombinant proteins was performed using the ProBond Kit™ (Invitrogen) according to the manufacturer's instructions. The purity of these proteins was assessed by Coomassie Blue staining of SDS/PAGE gels (Supplementary Figure S1 at

DCA-DC (DCA dechlorinating) activity assay

DCA-DC assays were performed as described previously [5]. All of the enzyme reactions were performed in microtitre plates. In brief, the purified recombinant GSTZ enzyme (1 μg) was incubated with 0.5 mM DCA and 1 mM glutathione in 100 μl of 0.1 M sodium phosphate buffer (pH 7.4) at 37 °C for 20 min. The reaction was stopped by adding 5 μl of trifluoroacetic acid. The reaction mixtures were placed on ice for 10 min and then centrifuged (16000 g at 4 °C) for 5 min to remove precipitated proteins. The supernatant (85 μl) was transferred to a new 96-well microplate and neutralized with 50 μl of 1 M NaOH. To each tube, 30 μl of 0.8 M sodium phosphate buffer (pH 6.8) and 40 μl of phenylhydrazine hydrochloride solution (100 mg in 15 ml of water) were added and the contents were mixed. The mixtures were kept at room temperature for 10 min and then placed on ice for 5 min. Chilled concentrated HCl (50 μl) and 40 μl of potassium ferricyanide solution (500 mg in 15 ml of water) was added to each sample. The samples were thoroughly mixed and then kept at room temperature for exactly 15 min; the absorbance was measured at 535 nm against water. Glyoxylic acid concentrations were quantified with a standard curve prepared under the same conditions (0–100 nM). Preliminary analysis showed that the enzyme activity was linear. The kinetics of the glutathione-dependent oxygenation of DCA was determined using 0.01–2 mM DCA for the DCA at 1 mM of glutathione. To measure the Km for glutathione, 0.5 mM DCA was used, and glutathione concentrations ranged from 0.01 to 2 mM. The enzyme activities are expressed as nmol of product formed per second per mg of protein (nkat·mg−1). Kinetic values were calculated using the program Kaleidagraph (Synergy Software).

The temperature and pH optimum of GSTZ was determined for the DCA-DC activity. For determination of the temperature optimum, enzyme assays were conducted at various temperatures (4, 10, 15, 22, 30, 37, 42, 50 and 60 °C) for 20 min in 0.1 M sodium phosphate buffer (pH 7.4). For determination of the pH optimum, enzyme assays were conducted at 37 °C for 20 min in 0.1 M citrate-sodium phosphate buffer (pH 4 and 5), sodium phosphate buffer (pH 6, 6.5, 7, 7.5 and 8) or Tris/HCl buffer (pH 8.5 and 9).

MAAI activity assay

MAAI activity was determined using a coupled reaction in the presence of excess FAH as described previously [6]. First, MAA was enzymatically synthesized from 200 μM homogentisic acid in 20 mM Tris/HCl (pH 7.8), using a crude preparation from E. coli expressing AtHGO. The reaction was incubated at 30 °C until the absorbance at 330 nm was stable, indicating complete ring cleavage of homogentisate. A crude extract from E. coli expressing AtFAH was added, followed by 1 mM glutathione and 1 μg of the recombinant GSTZ proteins. The MAAI activity was monitored 2 min after addition of GSTZ protein by measuring the decrease in absorbance at 330 nm. Preliminary analysis showed that enzyme activity was linear. MAAI activity is expressed as nkat·mg−1.

SNP genotyping and association analysis with seedling cold tolerance

To detect the OsGSTZ2 gene SNP (A>G) underlying the I99V substitution, an allelic discrimination assay was performed using the ABI 7300 Real-Time PCR system (Applied Biosystems). Primers (forward, 5′-TTTATCTGCTCCCCATTTGAATATGTGT-3′ and reverse, 5′-TGGAGAGGTTGGATGCTTGAAC-3; amplicon 84 bp) and Taqman probes (M-202, 5′VIC-CAGATTGCAAACATAGTTT-3′ and IR50, 5′FAM-CAGATTGCAAACGTAGTTT-3′; FAM is 6-carboxyfluorescein) were designed using Primer Express version 2.0 (Applied Biosystems). PCR and cycling conditions were conducted according to the manufacturer's protocol. A total of 76 rice germplasm accessions (Supplementary Table S1 at were assayed and default software was used to plot and automatically call genotypes. Seedling cold-tolerance phenotyping was performed essentially as described previously [14] using a Conviron PGR15 growth chamber (Controlled Environments Limited) for cold treatment of seedlings (9 °C constant temperature, 12 h photoperiod, and photosynthetic photon flux density of 500 μmol·m−2·s−1). Some accessions were scored for another study [19].


Sequencing and expression analysis of OsGSTZ genes

Sequencing of OsGSTZ1 and OsGSTZ2 genomic DNA and corresponding cDNAs from the cold-tolerant M-202 and cold-sensitive IR50 cultivars revealed a number of SNPs and small Indels (insertion/deletions) in both genes (Figure 1). In OsGSTZ1, seven sequence variations were detected including one insertion (CAGCTGCC) in the 5′-UTR (untranslated region) of the IR50 allele, two Indels and three SNPs in introns, and one SNP in the 3′-UTR. None of the sequence variations are predicted to affect the activity of the encoded protein. In OsGSTZ2, out of six SNPs between M-202 and IR50, three were found in the coding region and three were in introns. Two SNPs (A>G at position 295 and A>T at position 551 from the start codon) were predicted to result in a change of amino acids (I99V and N184I respectively), whereas one SNP (A>G at position 120 from the start codon) does not affect the translation. Thus there are at least two OsGSTZ2 isoforms (Ile99Asn184 in M-202 and Val99Ile184 in IR50). The two non-synonymous SNPs were also present in the reference japonica (cv. Nipponbare) and indica (cv. 93–11) rice genome sequences, suggesting that these OsGSTZ2 isoforms represent subspecies differences.

Figure 1 SNPs between M-202 and IR50 in OsGSTZ1 and OsGSTZ2

Grey boxes represent exons (nine) and solid lines represent introns (eight). Indels and SNPs are as shown. Two non-synonymous SNPs and the corresponding amino acid changes (I99V and N184I) are indicated.

Alignment of the OsGSTZ proteins with GSTZs identified in other species is shown in Figure 2. The residues at position 99 and 184 of OsGSTZ2 are not highly conserved in other species, and both are located in α-helices. I99V is located in α4, which is part of the proposed substrate-binding region (H-site) and N184I is located in α7.

Figure 2 Sequence alignments of OsGSTZs and GSTZ1 from other species

Arabidopsis thaliana (GenBank® accession numbers AC005312 and T16F16.18), Carnation (GenBank® accession number S33628), Homo sapiens (Human, GenBank® accession number AAB96392.1) and Caenorhabditis elegans (GenBank® accession number Y53G8B.1). The secondary structures shown are based on AtGSTZ1 as described previously [9]. α-Helices and β-strands are represented, and β-turns are marked with TT. Conserved amino acids (aa) are indicated as * (identical) and ‘.’ or ‘..’ (similar). Amino acid differences between M-202 and IR50 isoforms of OsGSTZ2 are highlighted, and the percentage sequence identity of each protein relative to OsGSTZ1 is as indicated.

GST activity of recombinant OsGSTZ expressed in E. coli

To determine whether the M-202 and IR50 isoforms of OsGSTZ2 differed in their enzyme activity, we generated recombinant His6-tagged proteins by cloning the cDNAs of OsGSTZ1 and the two OsGSTZ2 into the pET-28a vector, and expressed and purified them from E. coli to analyse their enzyme activity. OsGSTZ1 and OsGSTZ2 are predicted to encode proteins of 23.95 and 23.64 kDa respectively. SDS/PAGE analysis of the recombinant GSTZs resulted in 25 kDa proteins, which are slightly bigger than the expected molecular mass, probably due to the His6 tag (Supplementary Figure S1). Consistent with the characterization of GSTZ from other organisms [1], OsGSTZs had little GSH-conjugating activity towards standard xenobiotic substrates such as CDNB (1-chloro-2,4-dinitrobenzene) and BITC (benzyl isothiocyanate), but did exhibit GSH-dependent DCA-DC activity. The specific activities of the OsGSTZs for DCA were determined, as shown in Table 1. The DCA-DC activities of recombinant OsGSTZ1 and OsGSTZ2M-202 were 8.3 and 19.9 nkat·mg−1 respectively. By comparison, the DCA-DC activity of OsGSTZ2IR50 was significantly reduced. The activity of the IR50 isoform (2.5 nkat·mg−1) was 8-fold less than that of the M-202 isoform. OsGSTZ2M-202 activity was similar to that of Arabidopsis thaliana AtGSTZ1 (20.6 nkat·mg−1), which was previously characterized by Dixon and Edwards [6] and confirmed in the present study.

View this table:
Table 1 DCA-DC and MAAI specific activities of purified recombinant OsGSTZ1, OsGSTZ2 isoforms and AtGSTZ1

All values represent means±S.D. for n = 3 independent determinations. *Indicates significant difference compared with OsGSTZ2M-202 at the level of P<0.01 (Student's t test).

MAAI activity of recombinant OsGSTZ

As MAA is the proposed natural substrate of GSTZ/MAAI, MAAI activity was examined to determine whether the OsGSTZ2 isoforms also differed in their ability to catalyse isomerization of MAA to FAA. Since MAA is not commercially available, the substrate was enzymatically synthesized from homogentisic acid using AtHGO as described previously [6]. Two of the Arabidopsis enzymes (AtHGO and AtFAH) were cloned into the pET-28a vector system and expressed in E. coli, and clarified extracts were used for the coupled assay to detect MAAI activity. OsGSTZ1 rapidly isomerized MAA (215 nkat·mg−1) in the presence of 2.5 mM GSH. OsGSTZ2M-202 displayed higher catalytic activity (481 nkat·mg−1) than OsGSTZ1. Consistent with DCA-DC activity, OsGSTZ2IR50 showed significantly lower activity (30.1 nkat·mg−1) towards MAA.

Determination of the effect of amino acid substitutions on DCA-DC and MAAI activity

The M-202 and IR50 isoforms of OsGSTZ2 differ by two amino acids. Two chimaeric recombinant proteins were generated, containing only one amino acid substitution each, in order to examine the individual contributions of the I99V and N184I substitutions to the difference in DCA-DC and MAAI activity between the two isoforms. Analysis of the chimaeric isoforms OsGSTZ2Ile99Ile184 and OsGSTZ2Val99Asn184 (Figure 3) revealed large reductions in specific activity of OsGSTZ2Val99Asn184 for both DCA and MAA substrates (3.5 nkat·mg−1 for DCA and 35.8 nkat·mg−1 for MAA). The activity exhibited by OsGSTZ2Val99Asn184 was comparable with that of OsGSTZ2IR50 (Table 1). The OsGSTZ2Ile99Ile184 isoform showed no significant change in its specific activity for those substrates (22.0 nkat·mg−1 for DCA and 461.6 nkat·mg−1 for MAA) compared with OsGSTZ2M-202. These results indicated that the substitution of Ile99 with Val99 is responsible for significant differences in enzymatic properties of the naturally occurring OsGSTZ2M-202 and OsGSTZ2IR50 isoforms.

Figure 3 DCA-DC and MAAI substrate saturation kinetics of recombinant OsGSTZ1 and the OsGSTZ2 isoforms

The amino acid substitutions are indicated for the two naturally occurring and the two chimaeric OsGSTZ2 proteins. Fixed co-substrate concentrations are as follows: (A) 0.5 mM DCA, (B and D) 1 mM GSH, and (C) 200 μM MAA. All values represent means±S.D. for n = 3 independent determinations.

Analysis of OsGSTZ kinetics and the effect of temperature and pH

To further characterize OsGSTZ1 and OsGSTZ2, steady-state kinetic experiments were conducted and the effects of temperature and pH on activity were examined. Kinetic analysis was conducted using recombinant OsGSTZ1 and OsGSTZ2M-202 as OsGSTZ2Ile99Ile184 properties were essentially identical with OsGSTZ2M-202 (Figure 3). Analysis could not be performed on the OsGSTZ2IR50 and OsGSTZ2Val99Asn184 isoforms due to their impaired activity. The steady-state kinetic parameters of OsGSTZ1 and OsGSTZ2M-202 are shown in Table 2. The Km of OsGSTZ2M-202 for GSH was 11-fold lower than OsGSTZ1. However, the Km of OsGSTZ2M-202 for DCA was similar to that of OsGSTZ1 and, whereas the Km for MAA was lower in OsGSTZ1 (0.15 mM) compared with OsGSTZ2M-202 (0.23 mM), the difference was not statistically significant. These results suggest that OsGSTZ2M-202 has a higher affinity for GSH than OsGSTZ1, and similar affinities for DCA and MAA. OsGSTZ2M-202 also exhibited higher Vmax values compared with OsGSTZ1. The Vmax of OsGSTZ2M-202 for GSH (20 nkat·mg−1 for DCA-DC and 500 nkat·mg−1 for MAAI) was approx. 2-fold higher than that of OsGSTZ1 (10 nkat·mg−1 for DCA-DC and 257 nkat·mg−1 for MAAI). The Vmax of OsGSTZ2M-202 for DCA and MAA (500 nkat·mg−1 and 1111 nkat·mg−1 respectively) was approx. 2-fold higher than that of OsGSTZ1 (257 nkat·mg−1 and 470 nkat·mg−1 respectively). As a control, we analysed the enzyme activity of AtGSTZ1. The kinetic parameters indicated that the enzyme activity of AtGSTZ1 was very similar to that of OsGST2M-202 (Table 2).

View this table:
Table 2 Steady-state kinetic parameters of recombinant OsGSTZ isozymes and AtGSTZ1

All values represent means±S.D. for n = 3 independent determinations. *Indicates significant difference compared with OsGSTZ2M-202 at the level of P<0.01 (Student's t test).

The effect of temperature and pH on the recombinant OsGSTZ1 and OsGSTZ2 were also examined using the substrate DCA. The maximum DCA-DC catalytic activity occurred between pH 7 and 7.4 for both OsGSTZ1 and OsGSTZ2M-202 (Supplementary Figure S2 at The optimum temperature for both enzymes was 37 °C, and these enzymes lost activity at temperatures higher than 50 °C and lower than 10 °C.

Association of the I99V allele with rice subspecies and seedling cold tolerance

Genotyping of the OsGSTZ2 I99V SNP (c.295A>G) was performed on 76 rice germplasm accessions (Supplementary Table S1). Out of 47 japonica accessions, only two accessions (4%) were found to have the G allele (Val99). In contrast, the G allele was present in 12 out of 29 indica accessions (41%) examined (Table 3). Since OsGSTZ2 is co-localized with qCTS12, a major QTL for seedling cold tolerance, we examined the association of this SNP with seedling cold tolerance. In general, the japonica accessions were cold tolerant whereas the indica accessions were cold sensitive. All of the G allele rice accessions (14 accessions) displayed significant cold tolerance, irrespective of their subspecies classification.

View this table:
Table 3 Comparison of O. sativa subspecies for seedling cold tolerance between germplasm accessions carrying the A compared with the G allele at I99V


We previously identified qCTS12, a major effect QTL for rice seedling cold tolerance, using a genetic mapping population derived from the cold-tolerant cv. M-202 and the cold-sensitive cv. IR50 [13], and fine mapped this QTL to a region containing the OsGSTZ1 and OsGSTZ2 genes [14]. To examine the possibility that OsGSTZs contribute to rice seedling cold tolerance, we sequenced the M-202 and IR50 alleles to search for functional polymorphisms. Of the polymorphisms detected in OsGSTZ2, two non-synonymous SNPs were predicted to change amino acids at positions 99 and 184 of OsGSTZ2, whereas none of the sequence variations detected in the OsGSTZ1 alleles were predicted to have functional significance. Recombinant proteins were produced in order to characterize the catalytic abilities of OsGSTZ1, OsGSTZ2M-202 and OsGSTZ2IR50. As expected, OsGSTZ1 and the two OsGSTZ2 isoforms exhibited little to no activity with standard GST substrates (CDNB and BITC), but activity with DCA and MAA was observed. Significantly, there was a substantial difference in activity of the M-202 and IR50 isoforms of OsGSTZ2 with DCA and MAA. Characterizing recombinant OsGSTZ2 proteins containing each amino acid difference individually confirmed that the SNP resulting in the I99V substitution was a functional polymorphism with regard to DCA dechlorination and MAA isomerization.

Structural analysis of human and Arabidopsis GSTZs has shown that GSTZ shares high structural similarity with other classes of GSTs [4,9]. As with other GSTZ proteins, OsGSTZ1 and OsGSTZ2 are highly conserved in the N-terminal glutathione-binding site (G-site) which contains the Zeta class signature motif (SSCX[WH]RVIAL) and the substrate-binding site (hydrophobic binding region, H-site) formed by residues from α4 and α6 helices and the N-terminal domain [9]. The basic residues in the H-site form a positively charged region at the active site, which enables the enzyme to bind the negatively charged carboxy moieties of the MAA substrate [9]. The residue Val99 found in the IR50 isoform results in significant reduction in OsGSTZ2 enzymatic activity with MAA and DCA. The amino acid substitution is located in the α4 helix that forms part of the H-site. Although both isoleucine and valine are neutral non-polar amino acids having similar hydrophobic properties, isoleucine is a bulkier residue. In humans, an isoleucine-to-valine substitution at position 104 caused alterations in the activity of human GSTP1–1 [20]. The specific mechanism by which the I99V change causes the significant reduction in OsGSTZ2 activity is not clear, although the substitution of a bulkier isoleucine residue with valine probably alters the geometry of the H-site [20].

Although both recombinant OsGSTZ1 and OsGSTZ2M-202 are active with the same substrates, they differ in their kinetics. Overall, OsGSTZ2 has higher activity than OsGSTZ1 with both MAA and DCA as substrates. OsGSTZ2 has a 10-fold higher affinity for GSH compared with OsGSTZ1, whereas the affinities of OsGSTZ2 for DCA and MAA are not significantly different to those of OsGSTZ1. OsGSTZ1 displays a relatively low enzyme activity and OsGSTZ1 is constitutively expressed. It is interesting to note that out of two GSTZs in Arabidopsis, only one appears to be transcribed at any significant level [8]. In addition, the GSTZ1 gene in the cultivated African rice species O. glaberrima has been reported to be disrupted by a major chromosomal rearrangement [15]. Our results on the different properties of OsGSTZ1 and OsGSTZ2, along with the reported differences in gene expression [16], and the observations from Arabidopsis and O. glaberrima, suggest that the two O. sativa genes are likely to have distinct, but perhaps overlapping, functional roles. It is possible that OsGSTZ1, which is a constitutively expressed enzyme with relatively lower activity, might perform a metabolic housekeeping function whereas OsGSTZ2, which has a higher activity and whose gene is induced by various hormones and stresses [16], may function primarily in the stress response. Thus, under low temperature conditions, OsGSTZ1 may not be able to compensate for the significant loss in activity exhibited by OsGSTZ2IR50, leading to severe chilling injury in the sensitive IR50 cultivar.

We have demonstrated that a single amino acid change from Ile99 to Val99 in OsGSTZ2 causes a significant reduction in its catalytic activity with known substrates and that the corresponding SNP (G allele) appears to be highly associated with seedling cold sensitivity within and across rice subpopulations. These results are consistent with previous molecular and genetic evidence implicating OsGSTZ2 in rice seedling cold tolerance [11,14]. The findings of the present study suggest several possibilities with regard to the role of OsGSTZ2 in response to low-temperature stress. Given the endogenous function of GSTZs in tyrosine catabolism, it may be that this pathway or one of its metabolites plays a role in preventing or causing chilling injury in plants such as rice. In humans, deficiencies in each of the enzymes involved in tyrosine catabolism result in diseases of various severity with the exception of GSTZ/MAAI [1]. Knockouts of the Gstz11 (MAAI) gene in mice are not lethal [21,22]; however, these mice are severely affected by increases in phenylalanine and tyrosine in their diets and over time develop conditions consistent with chronic accumulation of a toxic metabolite [1]. Furthermore, Gstz11 deficiency in mice causes oxidative stress and the activation of antioxidant-response pathways [22,23]. The Gstz11-knockout phenotype in mice suggests that alternative pathways for metabolizing MAA exist [21]. Indeed, both a non-enzymatic GSH-mediated isomerization of MAA to FAA [21] and a secondary mechanism involving GSH conjugation of MAA for export from the liver, resulting in massive depletion of liver GSH, have been reported [22]. We are unaware of the possibility of MAA accumulation in plants or whether there would be an associated toxicity. However, there are alternative anabolic pathways for homogentisate, such as the synthesis of vitamin E and plastoquinones [6], which might provide a bypass mechanism during non-stress conditions.

It is also possible that the role of OsGSTZ2 in cold tolerance involves isomerization of an endogenous substrate other than MAA. Dixon et al. [24] have noted that there are several plant secondary metabolic pathways that require isomerization activities for which specific enzymes have yet to be identified. Biosynthetic pathways for coumarins [25], carotenoids [26] and phytochrome [27] are among those in which GSTs could potentially provide isomerase functions [24]. Alternatively, given the wide array of catalytic functions that can be performed by GSTZs [1,4], the contribution of OsGSTZ2 to cold tolerance may be based on an activity other than isomerization. Differences in metabolic profiles between tolerant and sensitive rice seedlings undergoing cold stress may provide insight into the function of OsGSTZ2. We are currently developing transgenic and near-isogenic lines containing the qCTS12 locus from M-202 in the genetic background of IR50, which may be used for such an approach.

Another potential strategy for identifying substrates of OsGSTZ2 is based on the ability of GSTs to bind ligands with high affinity. This attribute has previously been exploited to identify ligands that interact with members of the Tau class GSTs under in vitro and in vivo conditions [28,29]. A ligand-fishing approach has also been employed to investigate the function of GSTLs (Lambda class GSTs) from wheat [30]. Recombinant GSTLs were engineered with a Strep tag (Trp-Ser-His-Pro-Gln-Phe-Glu-Lys), expressed, affinity-immobilized and incubated with plant extracts. After washing, associated ligands were eluted by solvent extraction and identified by MS. Comparison of the ligand profiles obtained from in vitro extracts and from in planta expression of the tagged wheat GSTL1 protein supported the notion that physiological binding partners could be obtained using this method [30]. Using a ligand-fishing approach with OsGSTZ proteins may provide valuable insight into the existence of alternative endogenous substrates, as well as possible functional differences between the GSTZ1 and GSTZ2 isoenzymes.

In summary, two naturally occurring GSTZ2 isoforms in rice differ dramatically in their activity with known substrates due to a single amino acid difference (I99V). Analysis of several rice cultivars shows that seedling cold sensitivity is highly associated with the SNP allele corresponding to the Val99-containing isoform, which exhibits significantly lower activity. The results of the present study suggest that the reduced activity of this enzyme plays a role in the chilling injury sustained by sensitive rice seedlings. Although confirmation of the involvement of OsGSTZ2 in cold tolerance awaits definitive genetic and molecular evidence, the findings of the present study support a possible biochemical basis for some of the variation in cold tolerance observed among rice cultivars. Further characterization of this enzyme, including the investigation of other potential endogenous substrates, may provide critical insight into the complex nature of cold tolerance in rice and other plants.


Sang-Ic Kim, Virgilio Andaya and Thomas Tai designed the experiments and carried out the research. Sang-Ic Kim and Thomas Tai were responsible for writing the manuscript.


This work was supported by the USDA-ARS (United States Department of Agriculture-Agricultural Research Service) CRIS (Current Research Information System) Project [grant number 5306-21000-017-00D (to T.H.T.)].


We thank Erin Easlon, Hoi Man Mak and Celia Takachi for greenhouse assistance, and Peter Colowit for greenhouse and technical assistance. We thank Dr C. Andaya and Dr G. Eizenga for useful comments on improving the manuscript prior to submission. Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture. USDA is an equal opportunity provider and employer.

Abbreviations: At, Arabidopsis thaliana; BITC, benzyl isothiocyanate; CDNB, 1-chloro-2,4-dinitrobenzene; DCA, dichloroacetic acid; DCA-DC, DCA dechlorinating; FAA, fumarylacetoacetate; FAH, FAA hydrolase; GST, glutathione transferase; GSTL, Lambda class GST; GSTZ, Zeta class GST; HGO, homogentisate dioxygenase; Indel, insertion/deletion; MAA, maleylacetoacetate; MAAI, MAA isomerase; Os, Oryza sativa; QTL, quantitative trait locus; SNP, single nucleotide polymorphism; UTR, untranslated region


View Abstract