Plant Molecular Biology

, Volume 63, Issue 4, pp 505–517

Arabidopsis thaliana AtGpp1 and AtGpp2: two novel low molecular weight phosphatases involved in plant glycerol metabolism


  • José Antonio Caparrós-Martín
    • Instituto de Biología Molecular y Celular de Plantas Primo-YúferaUniversidad Politécnica de Valencia-CSIC
  • Sonja Reiland
    • Instituto de Biología Molecular y Celular de Plantas Primo-YúferaUniversidad Politécnica de Valencia-CSIC
  • Karl Köchert
    • Instituto de Biología Molecular y Celular de Plantas Primo-YúferaUniversidad Politécnica de Valencia-CSIC
  • Mari Cruz Cutanda
    • Instituto de Biología Molecular y Celular de Plantas Primo-YúferaUniversidad Politécnica de Valencia-CSIC
    • Instituto de Biología Molecular y Celular de Plantas Primo-YúferaUniversidad Politécnica de Valencia-CSIC

DOI: 10.1007/s11103-006-9104-0

Cite this article as:
Caparrós-Martín, J.A., Reiland, S., Köchert, K. et al. Plant Mol Biol (2007) 63: 505. doi:10.1007/s11103-006-9104-0


We have isolated two Arabidopsis thaliana genes, AtGpp1 and AtGpp2, showing homology with the yeast low molecular weight phosphatases GPP1 and GPP2, which have a high specificity for dl-glycerol-3-phosphate, and moreover homology with DOG1 and DOG2 that dephosphorylate 2-deoxyglucose-6-phosphate. Using a comparative genomic approach, the corresponding genes were identified as conceptual translated haloacid dehalogenase-like hydrolase proteins. AtGpp1 (gi 18416631) and AtGpp2 (gi 18423981), encode proteins that share 95% identity, with a predicted Mw of 33 and 27 kDa and a pI of 7.8 and 5.6, respectively. Both isoforms have a high specificity for dl-glycerol-3-phosphate, pH optima at 7.0, and Km in the range of 3.5–5.2 mM. AtGpp1 and AtGpp2 are expressed throughout development in all plant organs, most strongly in siliqua, and expression is not affected by osmotic, ionic or oxidative stress. A putative chloroplast transit peptide cTP-containing sequence is appended to the AtGpp1 N-terminus while AtGpp2, devoid of this tail, is predicted to be in the extraplastidial cytosol; this compartmenting was further confirmed by subcellular fractionation. An immunohystochemical localization study, using anti-AtGpp2 antibodies, indicates that the AtGpp proteins are mainly restricted to the meristem of immature flower and vascular elements of the root, shoot, leave, siliqua and developing embryo. Considerable immunoreaction was observed in the cytoplasm as well as in plastid compartments of distinct cells types from different heterotrophic Arabidopsis tissues, and particularly localised within phloem companion cells. Transgenic Arabidopsis plants, with gain of AtGpp2 function, show altered phosphatase activity rates and improved tolerance to salt, osmotic and oxidative stress.


ArabidopsisGlycerol metabolismGlycerol-3-phosphateGlycerol-3-phosphatasesTransgenic plantsStress


Glycerol is formed after lipase-catalysis of triglycerides to glycerol and free fatty acids or through reduction of dihydroxyacetone phosphate (DHAP) by dihydroxyacetone phosphate reductase (DHAPR), with NADH as reductant, and the subsequent glycerol-3-phosphate (G3P) dephosphorylation by glycerol-3-phosphatase (GPP). Reversibly, glycerol is converted to G3P, the precursor for the synthesis of glycerolipids, by kinase-phosphorylation and, through oxidation of G3P to dihydroxy-acetone phosphate by NAD+-dependent glycerol-3-phosphate dehydrogenase (GPD), returns to the glycolytic or gluconeogeneic pathways (Heldt 1997). Thereby, glycerophospholipids and triacylglycerides, which are major components of membranes and storage forms of energy in most living cells, are all constructed on the scaffold of a glycerol molecule.

While the function of GPD has been well characterised (Albertyn et al. 1994; Valadi et al. 2004), little attention was paid to the detection of specific G3P phosphatase activity (Tsuboi and Hudson 1956; Tsuboi et al. 1957; Gancedo et al. 1968) until the purification and characterisation of the two dl-glycerol-3-phosphatase isoenzymes GPP1 and GPP2 from Saccharomyces cerevisiae (Norbeck et al. 1996).

In addition to the central role of glycerol in the linkage of lipids, this polyalcohol has been also regarded as osmoregulator compatible osmolite, protecting cellular structures and preventing the cell collapse under osmotic stress (Brown 1978). In bakers’ yeast, glycerol concentration increased from below 0.05 molar at low salinity to about 1.2 molar at 0.9 M NaCl (Olz et al. 1993), and the osmotic control of glycerol production have been proposed to be exerted through G3P dephosphorilation by the specific activity of the two phosphatases GPP1 and GPP2 (Norbeck et al. 1996).

As in yeast, glycerol accumulation has multiple effects on plant cell metabolism related to the glycolytic flux (Aubert et al. 1994) and enhances resistance to a variety of abiotic stresses associated with dehydration (Eastmond 2004). Thereby, a similar specific activity, to the one previously identified in yeast (Tsuboi and Hudson 1956; Tsuboi et al. 1957; Gancedo et al. 1968), was also described in the partially purified dl-glycerol-1-phosphatase from the halotolerant alga Dunaliella salina, which indeed is involved in the rapid mechanism of osmotic adjustment in these algae (Sussman and Avron 1981). To get more insight about the glycerol anabolism and its function in plant cell, we have cloned and characterised two Arabidopsis thaliana isoforms AtGpp1 and AtGpp2 of the dl-glycerol-3-phosphatase. AtGpp1 and AtGpp2 show high homology with yeast phosphatases GPP1 and GPP2 (Norbeck et al. 1996), which have a high specificity for dl-glycerol-3-phosphate, as well as with DOG1 and DOG2 that dephosphorylate 2-deoxyglucose-6-phosphate (Rández-Gil et al. 1995). Both, AtGpp1 and AtGpp2 are constitutively expressed in all the Arabidopsis tissues and unaffected under abiotic stress. The AtGpp2 overexpression in transgenic Arabidopsis plants increases the specific dl-glycerol-3-phosphatase activity and improves the tolerance to salt, osmotic and oxidative stress.

Materials and methods


The materials used for cloning were obtained From New England Biolabs (pMAL-c2X vector and Escherichia coli TB1 host for expression), Roche Applied Science (First Strand DNA synthesis kit for reverse transcriptase PCR (avian myeloblastis virus)), Sigma (GeneElute mammalian total RNA kit and REDTaq DNA polymerase), Sigma-Genosys (oligonucleotides), and Stratagene (pBluescript SK+ vector). The pSBETa helper vector was constructed at the Max-Planck Institut, Köln, Germany (Schenk et al. 1995).

Plant material and stress treatments

Arabidopsis thaliana ecotype Columbia was grown in the greenhouse at 25°C for 8 h in the dark and 16 h in light. For seedling stress, wild-type and transgenic surface-sterilized Arabidopsis seeds were sown in Petri dishes containing 25 ml MSS medium (MS (Murashige and Skoog, 1962) (Sigma, M5524) + 1% Agargel™ (Sigma, A-3301) + 3% sucrose), MSS + 100 mM NaCl or 20 mM LiCl for salt stress, MSS + 200 mM sorbitol for osmotic stress, MSS + 10 mM H2O2 for oxidative stress, and MSS medium supplemented with 20 mM G3P for the G3P treatments. Seedlings were grown for 12 days at 25°C under fluorescent light, 8 h dark and 16 h light.

Genome analysis

Comparative genomics were performed using programs such as BLAST (Altschul et al. 1990) and data bank resources from the NCBI. Protein domain families were generated with the ProDom program from the Swiss-Pro and TrEMBL sequence databases (Corpet et al. 2000). The Wisconsin Package software, Version 10.0-UNIX, from Genetic Computer Group (GCG; Madison, WI) was used for Pileup and Prettybox sequence alignment with gcg10.

cDNA cloning

Using peptide sequence motives shared between the yeast dl-glycerol-3-phosphatases (Norbeck et al. 1996), virtual clones were isolated by BLAST (Altschul et al. 1990) search screening from the predicted conceptual translated proteins of the A. thaliana genomic library. The corresponding homologous genes were cloned by reverse PCR using leaf total RNA as template for the first strand cDNA synthesis together with an oligo(dT) primer and avian myeloblastis virus reverse transcriptase. The first strand cDNA template was PCR amplified using REDTaq DNA polymerase and the 5′-forward and 3′-reverse gene-specific adapted primers 5′-CCGGGGATCCATGTTAACAACTCCGACAAGATTC-3′/5′-CGAGAAGCTTCCACCTTAGTTGTGTGAATCCTGG-3′; and 5′-CCGGGGATCCATGTCGAATCCTGCAGCCGTCACC-3′/5′-CGAGAAGCTTCCACCTTAGTTCGAATCTTCGAATG-3′ for AtGpp1 and AtGpp2, respectively. The PCR products were cloned as 894 bp (AtGpp1) or 720 bp (AtGpp2) BamH I/Hind III fragments into the pBluescript SK+ vector. Sequencing of the pBluescript SK+ clones revealed that the sequence of the AtGpp proteins is the same as that published in the GenBankTM. The cDNAs, containing the entire gene coding region of AtGpp2 or a truncate form of AtGpp1, lacking the 147 bp long chloroplast targeting signal peptide, were subcloned BamH I/Hind III into the pMAL-c2X expression vector (for truncate AtGpp1 the 5′-forward adapted primer 5′-CCGGGGATCCATGTCGACCCCTGCCGCCGCCGTC-3′ was used) and transformed into the expression strain E. coli TB1 for recombinant protein production. The cloning site used in the pMAL-c2X polylinker (locate downstream of the malE gene), adding vector-encode residues Ile-Ser-Glu-Phe-Gly-Ser fused between the factor Xa cleavage site and the NH2-terminal methionine residue of the cloned AtGpp proteins. To improve the expression of the eukaryotic AtGpp genes in the E. coli system, E. coli TB1 cells were co-transformed with pMAL-c2XAtGpp and the helper plasmid pSBETa. The positive co-transformed colonies were selected on 200 μg/ml ampicillin and 100 μg/ml kanamycin. The AtGpp coding regions were further PCR amplified using the 5′-forward and 3′-reverse gene-specific adapted primers 5′-CGAGAAGCTTCCACCATGTTAACAACTCCGAC-3′/5′-CCGGGGATCCTTAGTTGTGTGAATCCTGGAAGG-3′; and 5′-CGAGAAGCTTCCACCATGTCGAATCCTGCAGCCG-3′/5′-CCGGGGATCCTTAGTTCGAATCTTCGAATGGAGG-3′ for AtGpp1 and AtGpp2, respectively. The PCR products were cloned in sense and anti-sense orientations as Hind III/BamH I fragments, into plasmid pJIT 163 (Guerineau et al. 1992). The corresponding Kpn I/Xho I DNA fragments from the above constructions, containing the full length cDNA of AtGpp1 or AtGpp2 genes, flanked by the cauliflower mosaic virus (CaMV) 35S promoter with a duplicated enhancer and by the CaMV polyadenylation sequence, were finally cloned into the binary plant vector pBin19 (Bevan 1984) and used for transformation of Agrobacterium helper strain LBA 4404 (Hoekema et al. 1983) by high-voltage electroporation (Wen-Jun and Forde 1989). The positive transformed colonies were selected on 100 μg/ml kanamycin.

Purification of recombinant proteins

Selected co-transformed E. coli strains, containing fusion plasmid pMAL-c2XAtGpp and the helper pSBET, were gown at 37°C to 2 × 108 cells/ml (A600∼0.5) in 1 l of rich broth + glucose & ampicillin + kanamycin (10 g tryptone, 5 g yeast extract, 5 g NaCl, 2 g glucose, autoclave; add sterile 200 μg/ml ampicillin and 100 μg/ml kanamycin), induced with 1 mM isopropyl-β-d-thiogalactoside (IPTG) and harvested 2 h after induction. Fusion proteins were released from the harvested cells by sonication in column buffer (20 mM Tris–HCl pH 7.4, 200 mM NaCl, and 1 mM EDTA) and collected after elution from the amylose resin with column buffer + 10 mM maltose. Proteins were separated by SDS-PAGE electrophoresis in 12% polyacrylamide gels (Schagger and von Jagow 1987), using prestained molecular weight standards from New England Biolabs.

Activity assays

The biochemical characterisation of the purified enzyme was assayed as previously described by Sussman and Avron (1981). The reaction mixture contained 20 mM Tris–HCl pH 7.0, 5 mM MgCl2, and 10 mM dl-glycerol-3-phosphate (Sigma, G2138) in a total volume of 1 ml. Approximately 25 μg/ml of purified AtGpp1 and 5 μg/ml of purified AtGpp2 were used in the enzymatic reactions. After 15, 30 and 60 min of incubation with the enzyme, samples of 0.3 ml were withdrawn and added to 0.7 ml phosphate determination mixture. The released inorganic phosphate was determined according to Ames (1966) and the reaction rate was calculated in relation to the amount of enzyme and time. Different substrate concentrations (0.5–32 mM) were used to receive the kinetic parameters Km and Vmax. Three further possible substrates for the two isoforms of AtGpp were added to a final concentration of 10 mM. To demonstrate the influence of the pH, reaction mixtures with the pH range of 5–8 were used. The stereospecificity was examined by using l-glycerol-3-phosphate (sn-glycerol-3-phosphate) (Sigma, G-7886) and the racemic substrate (dl-form). The experiments were done at least twice with values differing not significantly.

For the enzymatic activity assay in total protein extract of Arabidopsis seedlings, 12 days old seedlings were harvested, put in liquid nitrogen immediately and grinded to a fine powder. Per each gram of powder 4 ml of protein extraction buffer (20 mM Tris–HCl pH 7.0, 1 mM PMSF) were added. Afterwards the samples were centrifuged at 16,000 rpm with a Beckman JA-20 rotor for 30 min at 4°C to centrifuge down cell debris. The supernatant was then passed through a 0.45 μm syringe filter to remove all rests of cell debris that were not sedimented. Samples were stored at 4°C until be used. To assess phosphatase activity, 200 μg/ml of protein from total protein extracts of wild-type and transgenic Arabidopsis seedlings were used and incubated in reaction mixture at 32°C. As substrate 30 mM l-glycerol-3-phosphate was used whereas controls were prepared without substrate. After 5, 15, 30 and 45 min of incubation, aliquots of 150 μl were taken and put in 350 μl of ice-cold phosphate-determination-mix to stop the enzymatic reaction. The samples were then incubated at 45°C for 20 min and afterwards measured at 820 nm to determine the concentration of inorganic phosphate.

Glycerol determination

To determine the concentration of glycerol, the ENZYTEC™ fluid Glycerol determination kit (scil Diagnostics GmbH, Id-N°: 5360) was used. Distilled water was used as blank. Incubation was done at 25°C for 15 min with reagent #1 and for 25 min with reagent #2 at the same temperature. Absorbance was read at 340 nm.

Northern blots and hybridization

Northern analyses were performed using approximately 20 μg of total RNA per track. Four isolated DNA fragments were nick-translated in the presence of α-[32P]dCTP to be used as probes (Maniatis et al. 1982). Probes AtGpp1 and AtGpp2 cDNA, were two 894 bp and 720 bp Hind III/BamH I fragments, which contain the complete AtGpp1 and AtGpp2 coding regions, respectively; probes AtGpp1-3′-untranslated and AtGpp2-3′-untranslated, were two 300 bp EcoR I/Xho I fragments obtained by PCR amplification from the genomic 3′-untranslated region, using the 5′-forward and 3′-reverse gene-specific adapted primers 5′-CGAGGAATTCAACAACAAAGCTTTCAGGGCAACG-3′/5′-CCGGCTCGAGCAAGTTCTGTTAACGACATCCCGC-3′; 5′-CGAGGAATTCTAAACGATACAAACGTCTTCAAGG-3′/5′-CCGGCTCGAGGAAGGAATCAATTCAGCTATACCG-3′, for AtGpp1 and AtGpp2, respectively. Hybridization was performed in 3 × SSC, 0.05% PVP, 0.05 Ficoll, 1% SDS and 50 μg/ml ssDNA at 65°C. Filters were washed at high stringency (0.1 × SSC, 0.5% SDS at 65°C). No differences in the pattern of hybridization were observed by using the different probes. The experiments were done more than once and the data shown are representative.

Subcellular fractionation

For chloroplasts isolation, leaves (4 g) were cut into small pieces and homogenized in 12 ml of a buffer containing 0.35 M sucrose, 1 mM PMSF, and 25 mM Hepes pH 7.0. Homogenates were filtered through a layer of Miracloth (Calbiochem) and centrifuged at 100 × g for 2 min at 4°C to separate nuclei and cell-wall material (Choe and Thimann 1975). The residue was discarded, and the supernatant was centrifuged at 4000 × g for 1 min at 4°C. Both supernatant and enriched chloroplast pellet fractions were analysed by immunoblot (Borrel et al. 1995). The intactness of the chloroplasts was verified by phase-contrast microscopy (Zeiss Axioskop 40 FL).

Generation of antibodies

Purified AtGpp2 protein was used to obtain anti-AtGpp2 immune serum. Rabbit immunization was carried out by three successive injections of 100 μg purified protein in 500 μl phosphate buffered saline (PBS) emulsified in equal volume of Freund’s complete adjuvant, as earlier described (Goday et al. 1994).

Immunoblot analysis

Protein was extracted from Arabidopsis leaves plant material (1 g) with a buffer containing 5 mM EDTA, 1 mM PMSF, 20 mM Tris–HCl pH 8.0 and centrifuged at 15,000 × g for 10 min at 4°C. Supernatants were transferred to new tubes and 1/10 volume of 20% w/v Trichloroacetic (Panreac, 252373) was added. After 30 min incubation on ice, samples were centrifuged at 15,000 × g for 15 min at 4°C. Then, pellets were washed with methanol 3 times and dried, afterwards were resuspended with 1 ml Laemmly, boiled at 95°C for 10 min and stored at −20°C. The supernatants were subjected to SDS-PAGE and immunoblotting. Proteins were separated by SDS-PAGE (12% acrylamide) and transferred to a nitrocellulose membrane (Protran BA83; Schleicher & Schuell, Dassel, Germany). Detection was carried out using the anti-AtGpp2 IgGs (diluted 1:1000). Goat anti-rabbit IgG coupled with alkaline phosphatase (Sigma, St. Louis, MO, USA) was used as secondary antibody. The experiments were done more than once and the data shown are representative.


For light microscopy immunohistochemistry, the IgGs obtained against the AtGpp2 protein were used. An all-purpose fixative (80% v/v ethanol; 3.5% v/v formaldehyde; 5% v/v acetic acid) was used for paraffin embedding. Sections from paraffin-embedded material were blocked with 3% goat serum in PBS (10 mM phosphate, 150 mM NaCl, pH 7.4) at 22°C for 30 min and incubated with anti-AtGpp2 serum (dilution 1:500), or pre-immune serum (dilution 1:500). Immunoreactivity was visualized by the avidin–biotin complex (Vectastain Elite ABC kit; Vector, Burlingame, CA, USA), using diaminobenzidine as substrate for peroxidase, in a Zeiss Axioskop 40 FL light microscope.

Plant transformation

Selected transformed Agrobacterium strains, containing fusion plasmid pBin19-pJITAtGpp, were gown at 37°C to 2 × 108 cells/ml (A600∼0.5) in 1 l of Lauria broth + kanamycin (10 g tryptone, 5 g yeast extract, 5 g NaCl, autoclave; add sterile 100 μg/ml kanamycin), and used for plant transformation. Arabidopsis thaliana adult plants (5 weeks old) were agro-infected by infiltration (Bechtold et al. 1993) and grown in the greenhouse to collect seeds. T1 seeds were growth in MSS medium supplemented with 100 mg/ml kanamycin (Sigma, K4378). A yield of one transformant for 1,000 seeds was obtained. Fifty independent kanamycin-resistant T1 plants were selected for each transformation, transferred to soil after 15 days and grown in a greenhouse to collect seeds. Ten lines for each transformation were selected, which segregated 3:1 in kanamycin, as expected for a single integration of the construction in the plant genome. From each of the selected lines, 10 T1 plants were grown and seeds were collected. T2 plants were segregated in MSS media supplemented with 100 mg/ml kanamycin. T2 plants were checked for transgene integration by PCR and Southern blot. Ten homozygous independent lines were selected for each sense and anti-sense T2 and used for phenotype characterization and stress treatments.


Isolation of A. thaliana dl-glycerol-3-phosphatase AtGpp1 and AtGpp2 genes

The complete sequence of the Arabidopsis genome (The Arabidopsis Genome Initiative 2000) has been used for the identification of the two uncharacterised A. thalianadl-glycerol-3-phosphatase genes. With the known budding yeast GPP1 and GPP2 protein sequences and using the BLAST program (Altschul et al. 1990) two putative A. thaliana, named AtGpp1 and AtGpp2, were identified in the comparative genome approach. AtGpp1 (298 amino acids; gi 18416631) is encoded on A. thaliana chromosome 4 and AtGpp2 (240 amino acids; gi 18423981) on chromosome 5. The two virtual isolated genes were cloned by RT-PCR and sequenced. In agreement with the NCBI reported annotations, AtGpp1 (locus tag: At4g25840) is a gene of 1943 bp length with five introns of 276, 190, 128, 116, and 82 bp length (at positions 262, 764, 1016, 1293 and 1490) and with an unspliced 3′-untranslated region and AtGpp2 (locus tag: At5g57440) is a gene of 1800 bp having five introns of 75, 99, 101, 108 and 272 bp length (at positions 522, 678, 926, 1089 and 1423) and an unspliced 3′-untranslated region. AtGpp1 and AtGpp2 have a deduced Mw of 33 and 27 kDa and pI of 7.8 and 5.6, respectively. Although the Arabidopsis genome contains homologous locus, others than At4g25840 and At5g57440, with similar scores and general function predicted phosphatase/phosphohexomutase (unknown At2g38740, hypothetical At1g56500, riboflavin kinase/FAD synthetase At4g21470, putative At4g39970, catalytic/hydrolase At3g48420 and catalytic/hydrolase/phosphoglycolate phosphatase At2g33255), we focused on the two loci sharing striking similarity, presuming a mimetic activity in Arabidopsis of that produced by yeast isoenzymes GPP1 and GPP2.

Sequence comparisons

Arabidopsis thaliana AtGpp1 and AtGpp2 share a striking similarity (93%, sequence identity of 87%). The two plant proteins show high similarity (38–39%, sequence identity of 24–25%) to S. cerevisiaedl-glycerol-3-phosphatases: GPP1 and GPP2 (gi 2507038 and gi 731470) (Norbeck et al. 1996) and (35–36%, sequence identity of 26–28%) to 2-deoxyglucose-6-phosphatases DOG1 and DOG2 (gi 6321834 and gi 6321833) (Rández-Gil et al. 1995). This homology is ubiquitously distributed, although the greatest similarity occurs particularly in a terminal domain expanded along 50 amino acids residues (Fig. 1). Interestingly, several identical randomly distributed amino acids (Thr64, Ala120, Glu137, Ser154, Arg193, Leu221, Asp233, Val237, Ala253 and Asp266, of AtGpp1) are shared between the Arabidopsis AtGpp and yeast GPP dl-glycerol-3-phosphatases but not with DOG 2-deoxyglucose-6-phosphatases, while AtGpp1 Ile258 and GPP2 Ile195 are substituted by the conserved Val residue in AtGpp2 and GPP1. A putative chloroplast transit peptide cTP-containing sequence (49 amino acids long), predicted with an output score of 0.807, is appended to the AtGpp1 N-terminus while AtGpp2, devoid of this tail, is predicted to be cytoplasmic (Emanuelsson et al. 2000). Without the cTP-signal, AtGpp1 and AtGpp2 have similar deduced average mass Mw (27.6 and 26.8 kDa, respectively) and same theoretical isoelectric point pI (5.6).
Fig. 1

Atgpp1 and Atgpp2 alignment with homologous proteins. The predicted amino acid sequence of A. thaliana Atgpp1 and Atgpp2 (gi 18416631 and gi 18423981) was compared with homologous representatives from S. cerevisiaedl-glycerol-3-phosphatase GPP1 and GPP2 (gi 2507038 and gi 731470) and 2-deoxyglucose-6-phosphatases DOG1 and DOG2 (gi 6321834 and gi 6321833). Residues are in black boxes if three of six residues at a position are identical. Conserved residues are in grey boxes. Residues Thr64, Ala120, Glu137, Ser154, Arg193, Leu221, Asp233, Val237, Ala253, Asp266 and Ile258 (of AtGpp1) are labelled with asterisks. The putative chloroplast transit peptide cTP-containing sequence (49 amino acids long) in Atgpp1 is underlined

Expression and purification of AtGpp1 and AtGpp2 proteins

The putative AtGpp1 (excluding the chloroplast targeting signal peptide) and AtGpp2 genes were expressed as MBP fusion proteins and purified from the corresponding E. coli clones by amylose affinity chromatography (Fig. 2).
Fig. 2

Purification of Atgpp1 and Atgpp2 proteins. SDS-PAGE of purified MBP-fusion proteins MBP-Atgpp1 (1) and MBP-Atgpp2 (2) by amylose affinity chromatography. Mw, prestained molecular weight marker in kDa. Purified MBP-fusion proteins are indicated by an arrow head and possible MBP and Atgpp forms, from unstable MBP-Atgpp, marked by asterisks

AtGpp1 and AtGpp2 enzymatic activity

The putative AtGpp1 and AtGpp2 isoenzymes were investigated for dl-glycerol-3-phosphatase activity. Table 1 represents the in vitro phosphatase activity of purified AtGpp1 and AtGpp2 on various organic phosphomonoesters. The activity of both plant proteins is preferentially detectable on dl-glycerol-3-phosphate, poorer activity is observed on glucose-6-phosphate (<3%), fructose-6-phosphate (<12%), or 2-deoxyglucose-6-phosphate (<2%). The apparent Km of AtGpp1 and AtGpp2 for dl-glycerol-3-phosphate were 5.2 and 3.5 and the Vmax 3 and 23 nmol min−1, respectively (Fig. 3A). Both isoforms developed optimal dl-glycerol-3-phosphatase activity at neutral pH 7 (Fig. 3B). AtGpp1 as well as AtGpp2 showed higher stereospecificity for the l- enantiomorph than for the dl-racemate (Fig. 3C, D).
Table 1

Phosphatase activity of purified Atgpp1 and Atgpp2 on various organic phosphomonoesters


Relative enzyme activity















The reaction mixture contained: 10 mM Tris–HCl (pH 7.0); 5 mM MgCl2; 10 mM substrate (four different phosphomonoesters) and 50 μg/ml of the purified protein. Reaction temperature: 32°C. For each assay the enzyme activity with dl-glycerol-3-phosphate was set to 100. The corresponding activities were the following (in m units/mg protein): Atgpp1 (3.91) and Atgpp2 (12.8)
Fig. 3

Atgpp1 and Atgpp2 phosphatase activity profile. (A) The apparent Km and Vmax of the two dl-glycerol-3-phosphatases for dl -glycerol-3-phosphate. The reaction mixture contained: 20 mM Tris–HCl (pH 7.0); 5 mM MgCl2; the indicated concentration of dl-glycerol-3-phosphate; 24.36 μg/ml of Atgpp1 (- -) and 5.15 μg/ml of Atgpp2 (—), respectively. Reaction temperature: 34°C. Atgpp1: Km = 5.2 mM, Vmax = 3 nmol min−1; Atgpp2: Km = 3.5 mM, Vmax = 23 nmol min−1. (B) Influence of pH on the dl-glycerol-3-phosphatase activity of Atgpp1 (—) and Atgpp2 (- -). (C, D) Stereospecificity of Atgpp1 (C) and Atgpp2 (D) on dl-glycerol-3-phosphate (—) and l-glycerol-3-phosphate (- -)

Expression of AtGpp1 and AtGpp2 genes

Figure 4 illustrates the expression of the two A. thaliana genes in different tissues and under stress conditions. The AtGpp1 and AtGpp2 pattern of expression was examined in roots, shoots, leaves, flower and developing siliqua. Transcripts of both genes were detectable on Northern blots of RNA from all these organs. The highest expression was found in siliqua and the lowest in root (Fig. 4A). AtGpp1 and AtGpp2 expression was also examined in seedling subjected to salt (100 mM NaCl and 20 mM LiCl), osmotic (200 mM sorbitol), or oxidative (10 mM H2O2) stress. Between them, no obvious differences were observed in their expression patterns (Fig. 4B). In all blot experiments, two bands were detected that correspond in size to the AtGpp1 and AtGpp2 transcripts, yet in root samples an odd band appeared in the upper part of the blot.
Fig. 4

Expression of Atgpp1 and Atgpp2 mRNA during development and under stress treatment. Northern blot of Arabidopsis RNA probed with radiolabelled Atgpp2 cDNA. (A) From different tissues: root (1), shoot (2), leaf (3), flower (4), siliqua (5). (B) From 12-day-old seedling cultivated on control MSS medium (1), MSS + 100 mM NaCl (2), MSS + 200 mM sorbitol (3), MSS + 20 mM LiCl (4), and MSS + 10 mM H2O2 (5). Ethidium bromide (EtBr)-stained RNA was used as loading control

Subcellular distribution of AtGpp proteins

To confirm the predicted intracellular location of the AtGpp proteins, their immunodetection in tissue extract and subcellular fractions of cells from Arabidopsis leaves was performed. The fractions obtained were analysed by western blotting using antibodies raised against Arabidopsis thaliana AtGpp2, that also cross-react with AtGpp1 (result not shown). Figure 5 shows that AtGpp proteins are found as a band that migrated according to the predicted molecular mass in the extract of Arabidopsis leaves. The antiserum used also identifies an identical sized band in the cytosolic and chloroplast-enriched fraction.
Fig. 5

Immunodetection of Atgpp proteins in subcellular fractions of Arabidopsis leaf. Immunoblot of leaf protein extracts: full extract (1), cytosol (2), and chloroplast enriched fraction (3), using anti-Atgpp2 antibodies. Mw, protein markers in kDa. Atgpp is indicated with an arrow head

Spatial pattern of AtGpp proteins accumulation in Arabidopsis plant

To determine the spatial and temporal pattern of AtGpp proteins localisation, different Arabidopsis tissues were analysed by imunocytochemistry. Figure 6 shows paraffin-embedded sections of Arabidopsis organs incubated with anti-AtGpp2 antibodies. During embryogenesis, brown staining was mostly restricted to the embryo cotyledons and hypocotyls and, particularly, in parenchyma cells chloroplast and cytoplasm of vascular bundle companion cells from the bicarpellary ovary capsule (Fig. 6A–C). As can be seen in the flower oblique sections, high levels of protein are detected in the cytoplasm of vascular tissue cells of sepal, petal, stamen filament and stamen microspore tetrads, and in chloroplasts of mesophyll cells and fascicle companion cells from sepals (Fig. 6E–G). AtGpp protein accumulation was also observed in floral peduncle, leaf, shoot and root; restricted to cytoplasm of differentiated cells from vascular elements, mesophyll and cortex cells chloroplast, and in the companion cells of sieve elements (Fig. 6I–K, M–O, Q–S, U–W). The specificity of the reaction was confirmed by a lack of appreciable reaction of preimmune serum (Fig. 6D, H, L, P, T, X). In summary, AtGpp protein accumulation is ubiquitous, most strongly in the phloem companion cells of vascular tissue. The immunodetection moreover confirms the predicted intracellular distribution, in cytoplasm and plastid compartments of different cell types. Although the anti-AtGpp2 antigenic cross-reaction hinders the AtGpp proteins discrimination, the putative chloroplast transit peptide cTP-containing sequence at its N-terminus, leads to the presumption that the AtGpp1 location most likely is restricted to plastid organelles.
Fig. 6

Immunolocalisation of Atgpp proteins. Paraffin-embedded sections of different organs of A. thaliana were incubated with anti-Atgpp2 antibodies and an avidin–biotin-peroxidase detection system. Dark brown staining indicates Atgpp antibody-specific reaction. Transverse sections of: (A) Mature siliqua, (B, C, D) siliqua capsule, (E) wild-type developed flower, (F) stamen, (G, H) sepal, (I, J, K, L) floral peduncle, (M, N, O, P) leaf, (Q, R, S, T) stem, (U, V, W, X) root. (D, H, L, P, T, X) control reactions using preimmune serum. Abbreviations: A, anther; Cc, companion cell; Ch, chloroplast; Co, cortex; E, epidermis; G, gynoecium; M, microspore; P, pith; Pa, palisade mesophyll; Pe, petal; Ph, phloem; S, embryo; Se, sepal; Si, siliqua; St, stamen; Ue, upper epidermis; V, vascular bundle; X, xylem. Magnifications: ×50 in (A, E, I, M, Q, U); ×200 in (B, F, J, N, R, V); ×500 in (C, D, G, H, K, L, O, P, S, T, W, X)

Transgenic overexpression of AtGpp genes

Using Agrobacterium-mediated transfer, the AtGpp genes were stably integrated, in sense and antisense orientation in the plant genome and expressed in transgenic plants. T1 plants were checked for transgene integration by PCR and Southern blot (not shown). Ten sense and anti-sense homozygous T2 transgenic independent lines were chosen for further analysis. The in vivo constitutive AtGpp phosphatase activity was investigated in seedling extracts from Arabidopsis AtGpp-transgenic lines and, to scrutinise its expected correlation with glycerol concentration, the latter was determined as well. Most of the sense AtGpp-transgenic lines incremented their phosphatase activity (Fig. 7), particularly obvious in the case of transgenic-AtGpp2 (Fig. 7B) but, surprisingly, not their glycerol accumulation (Fig. 7C, D). The increased enzymatic activity observed between the transgenic lines correlates with their tolerance to G3P. This improved tolerance is observed during germination and development. Sense AtGpp2-transgenic plants developed roots and true leaves under stress conditions (Fig. 8B), while sense AtGpp1-transgenic mainly remained at the cotyledon stage (Fig. 8A). Likewise, the sense AtGpp2-transgenic plants show higher germination rates than the wild-type and AtGpp1-transgenic under salt, osmotic and oxidative stress treatment, (Fig. 8C, D), although the observed protective effect is not extended on plant growth (not shown). Though, compared with the wild-type, the level of phosphatase activity in the analyzed anti-sense lines was slightly lower, we were yet not able to show significant differences (results not shown).
Fig. 7

Compared phosphatase activity and glycerol content between Arabidopsis wild-type and transgenic AtGpp seedlings. (A) l-glycerol-3-phosphate hydrolytic activity and glycerol content in seedlings extract of Arabidopsis wild-type (0) and sense transgenic AtGpp1 independent lines (1 to 7). (B) l-glycerol-3-phosphate hydrolytic activity and glycerol content in seedlings extract of Arabidopsis wild-type (0) and sense transgenic AtGpp2 independent lines (1 to 7). (C) Glycerol content in the same extract used in (A). (D) Glycerol content in the same extract used in (B). Data are the mean value calculated from three different experiments. Error bars correspond to the standard error
Fig. 8

Effect of AtGpp constitutive expression on Arabidopsis stress tolerance. (A) Arabidopsis wild-type (0) and transgenic T2 homozygous AtGpp1 independent lines (1 to 7) on medium MSS + 20 mM G3P, after 15 days’ culture. (B) Arabidopsis wild-type (0) and transgenic Thomozygous AtGpp2 independent lines (1 to 7) on medium MSS + 20 mM G3P, after 15 days’ culture. (C) Germination rate of Arabidopsis wild-type (0) and transgenic T2 homozygous AtGpp1 independent lines (1 to 7) on medium MSS + 100 mM NaCl (■), MSS + 200 mM sorbitol (□), and MSS + 10 mM H2O2 (░). (D) Germination rate of Arabidopsis wild-type (0) and transgenic T2 homozygous AtGpp2 independent lines (1 to 7) on medium MSS + 100 mM NaCl (■), MSS + 200 mM sorbitol (□), and MSS + 10 mM H2O2 (░). In (C) and (D), values are expressed as the percentage of radicle emergence of 100 seeds after 7 days of culture


We have identified and characterised two Arabidopsis thalianaAtGpp1 and AtGpp2 genes that encode dl-glycerol-3-phosphatases. AtGpp1 and AtGpp2 proteins share a strong similarity, and both show striking homology with the yeast dl-glycerol-3-phosphatases GPP1 and GPP2 (Norbeck et al. 1996) and 2-deoxyglucose-6-phosphatases DOG1 and DOG2 (Rández-Gil et al. 1995), in particular at the terminal domain. AtGpp1 is enlarged by a 5′-tail putative chloroplast transit peptide cTP-containing sequence that presume its plastid location, while AtGpp2 is predicted to be cytoplasmic (Emanuelsson et al. 2000).

Without the cTP-signal, AtGpp1 and AtGpp2 have similar average masses and the same theoretical isoelectrical point, and both purified isoenzymes show in vitro dl-glycerol-3-phosphatase specific activity. Less activity is observed on other organic phosphomonoesters like glucose-6-phosphate and fructose-6-phosphate. As with the Saccharomyces counterparts GPP1 and GPP2 (Norbeck et al. 1996), particularly amazing is the low affinity for 2-deoxyglucose-6-phosphate shown by the two Arabidopsis phosphatases, despite their shared homology with DOG1 and DOG2 (Rández-Gil et al. 1995). Interestingly, amino acids Thr64, Ala120, Glu137, Ser154, Arg193, Leu221, Asp233, Val237, Ala253 and Asp266, of AtGpp1 are conserved between Arabidopsis and yeast GPP dl-glycerol-3-phosphatases but not in DOG 2-deoxyglucose-6-phosphatases. Also, like the yeast dl-glycerol-3-phosphatases, both plant isoforms developed optimal activity at neutral pH 7 but, showed higher stereospecificity for the l-enantiomorph rather than for the dl-racemate, as observed in the halotolerant alga Dunaliella salina (Sussman and Avron 1981). AtGpp2 and GPP1 show more enzymatic activity than their respective isoforms and, interestingly, in both proteins conserved Val residues substitute the AtGpp1 Ile258 and GPP2 Ile195.

AtGpp1 and AtGpp2 are expressed in all Arabidopsis organs; the higher mRNA accumulation occurs in developing siliqua and the lower in root. Similar pattern of expression in all tissues have been shown for other Arabidopsis glycerol metabolism related genes [glycerol kinase GLI1 (Eastmond 2004), glycerol-3-phosphate dehydrogenase GPDH (Shen et al. 2003)] and with transcripts particularly abundant in developing siliqua [plastidic glycerol-3-phosphate dehydrogenase AtGPDH (Wei et al. 2001), glycerol-3-phosphate acyltransferase AtGPAT1 (Zheng et al. 2003), and polyol transporter AtPLT5 (Klepek et al. 2005)].

While yeast GPP1 is unaffected by changes of external osmolarity the GPP2 intracellular concentration increases, contributing to the concentration of the glycerol osmoregulator during the hyperosmotic stress (Norbeck et al. 1996). However, neither AtGpp1 nor AtGpp2 genes seem to be induced in seedling under osmotic, ionic or oxidative stress, revealing a role for plant glycerol metabolism independent of serving to adjust the intracellular osmolarity.

The subcellular location of the AtGpp proteins is not surprising, since glycerol metabolism is ubiquitously distributed to attend the required phospholipid and glycerolipid compartmented biosynthesis and both, cytosolic and plastidic isoforms of the DHAPR/GPDH have been found (Gee et al. 1988; Wei et al. 2001). It could be assumed that AtGpp1 activity is involved in the modulation of the plastidic G3P/Pi equilibrium and, perhaps, in keeping the glycerol rate for optimal photochemical activity and conformation of the photosystem I PSI-particles (Ren et al. 2006); while AtGpp2 would account for the cytosolic maintenance of a homeostatic Pi balance. However, the particularly differentiated immunodetection localised to phloem companion cells is quite intriguing. Although sucrose is commonly accepted as the unique carrier of photoassimilated CO2 in Arabidopsis, where the phloem loading is catalyzed by the companion cell-specific transporter AtSUC2 (Truernit and Sauer 1995; Stadler and Sauer 1996), a new class of plasma membrane-localized H+-symporter AtPLT5, exhibiting unusual substrate specificity for myo-inositol, glycerol or ribose, has been recently described in most Arabidopsis tissues (Klepek et al. 2005). Indeed, it is tempting to consider glycerol as possible physiological substrate of Arabidopsis hypothetical AtPLTs companion cell-polyol transporters (Ramsperger-Gleixner et al. 2004), as additional carbohydrate for long distance translocation, since plant cells can also utilize glycerol as a carbon source (Aubert et al. 1994).

Neither AtGpp1 nor AtGpp2 overexpression enhanced significantly the presumed glycerol accumulation in transgenic Arabidopsis, indicating that the phosphatase activity would have a driving role in glycerol biosynthesis but not be rate-limiting to the plant glycerol-producing pathway, as has been previously described in yeast, where the flux from G3P to glycerol, produced by GPP1 and GPP2 genes, could be coordinated with that to glycerolipids and the G3P shuttle (Pahlman et al. 2001). Consequently, the effect of constitutive altered expression of AtGpp on Arabidopsis abiotic stress tolerance, did not improve the expected protection by the osmolite glycerol beyond germination, as reported in germinating seedlings of the glycerol-insensitive Arabidopsis mutant gli1, lacking glycerol kinase (Eastmond 2004), illustrating the low osmodependent channelling of carbon towards this polyol production under stress. Similar low protection was observed to reactive oxygen species (ROS), although a role of the glycerol metabolism in oxidative stress protection, associated with the GPP1 and GPP2 activity, was pointing out in yeast (Pahlman et al. 2001). Interestingly, the transgenic activity allows the growth on media with G3P, unveiling a possible function on cytosolic and chloroplastic G3P dissimilation exerted by AtGpp2 and AtGpp1, respectively. Cytoplasmic G3P accumulation perturbs the glycolytic cell metabolism by inhibition of glucose-6-phosphate isomerase activity, and the concomitant flowing back of carbon flux from triose phosphates to glucose-6-phosphate (Aubert et al. 1994). Moreover, G3P uptake is driven by the efflux of Pi (Elvin et al. 1985) and it could be very well that AtGpp modulate the G3P supply, for glycolysis, phospholipids biosynthesis and oxidative phosphorilation, preventing damage from homeostatic unbalances of inorganic phosphate (Pi). The finding of T-DNA KO or RNAi lines could provide additional evidence as to the function of these proteins in plants.

The cloning of the AtGpp biosynthetic enzymes completes the Arabidopsis glycerol anabolic pathway from DHAP. Their ubiquitous compartmentalisation would be in accordance with the required availability of glycerol, to satisfy the myriad of reaction steps in which this backbone molecule is involved. It is expected that the functional validation and molecular characterization of these novel low molecular weight phosphatases, will contribute to further unravel the mechanism of cell homeostasis, gaining more insight into the complex regulation of plants glycerol metabolism.


We acknowledge Professors Eduardo Primo-Yúfera (UPV, Valencia, Spain), Montserrat Pagès (CID-CSIC, Barcelona, Spain), Thomas Kupke (Lehrstuhl für Mikrobielle Genetik, Universität Tübingen, Germany), Francisco Montero and Olga Botella (Universidad de Castilla-La Mancha, Albacete, Spain) for their suggestions and warm support. We also thank the advice and provision of plasmid pJIT163 by Dr. Phil Mullineaux (John Innes Centre, Norwich, UK), pBin19 by Dr. Mike Bevan (John Innes Centre, Norwich, UK), and pSBETa by Dr. Florence Vignols and Yves Meyer (University of Perpignan, France); the antibodies production technical assistance by Dr. Concepción Cervera and Juan Carlos Moreno (UPV, Valencia, Spain); and the computer software help by Alexis González-Policarpo and Ramón Nogales-Rangel. This work was funded by the Research Project BIO2006-10138 from the MEC-FEDER of Spain.

Copyright information

© Springer Science+Business Media B.V. 2006