Kazuhito Akama

Efficient transformation of Arabidopsis thaliana: comparison of the efficiencies with various organs, plant ecotypes and Agrobacterium strains

SummaryThe efficiency of Agrobacterium-mediated transformation of Arabidopsis thaliana was compared with different organs, Arabidopsis ecotypes, and Agrobacterium strains. Efficiency of shoot regeneration was examined using hypocotyl, cotyledon and root explants prepared from young seedlings. Hypocotyl expiants had the highest regeneration efficiency in all of the four Arabidopsis ecotypes tested, when based on a tissue culture system of callus-inducing medium (CIM: Valvekens et al. 1988) and shoot-inducing medium (SIM: Feldmann and Marks 1986). Histochemical analysis using the ß-glucuronidase (GUS) reporter gene showed that the gusA gene expression increased as the period of preincubation on CIM was extended, suggesting that dividing cells are susceptible to Agrobacterium infection. In order to obtain transgenic shoots, hypocotyl explants preincubated for 7 or 8 days on CIM were infected with Agrobacterium containing a binary vector which carries two drug-resistant genes as selection markers, and transferred to SIM for selection of transformed shoots. Of four Arabidopsis ecotypes and of three Agrobacterium strains examined, Wassilewskija ecotype and EHA101 strain showed the highest efficiency of regeneration of transformed shoots. By combining the most efficient factors of preincubation period, Arabidopsis ecotype, tissue, and bacterial strain, we obtained a transformation efficiency of about 80–90%. Southern analysis of 124 transgenic plants showed that 44% had one copy of inserted T-DNA while the others had more than one copy.

Rice (Oryza sativa) contains a novel isoform of glutamate decarboxylase that lacks an authentic calmodulin-binding domain at the C-terminus.

We have isolated full-length cDNAs for two distinct isoforms of glutamate decarboxylase (GAD), designated OsGAD1 and OsGAD2 from a rice shoot cDNA library. Open reading frames found in OsGAD1 and OsGAD2 cDNAs encode putative proteins of 501 (56.7 kDa) and 500 amino acids (55.6 kDa), respectively. They show 69% identity to each other and 67-78% identity to dicotyledonous counterpart sequences determined so far. Comparative analysis of relevant genomic clones obtained from the rice genomic library with these cDNAs as probes demonstrated that the number and sizes of introns deduced for these two genes differ considerably. Interestingly, in the regions in the putative gene products corresponding to the C-terminal 30-amino-acid peptide known as the calmodulin-binding domain of plant GADs, OsGAD1 possesses a typical motif, while OsGAD2 contains several substitutions of amino acids that contribute strongly to the binding of calmodulin (CaM). An in vitro CaM-binding assay of these proteins over-expressed in Escherichia coli revealed that OsGAD1 can in fact bind specifically to bovine CaM but OsGAD2 cannot. RNA analysis showed that transcripts of OsGAD1 and OsGAD2 were present in all tissues examined, but their expression was differentially regulated, at least in roots and maturing seeds.

Efficient Agrobacterium-mediated transformation of Arabidopsis thaliana using the bar gene as selectable marker

SummaryWe have established an efficient Agrobacterium-mediated transformation procedure for Arabidopsis thaliana genotype C24 using the chimeric bialaphos resistance gene (bar) coding for phosphinothricin acetyltransferase (PAT). Hypocotyl explants from young seedlings cocultivated with agrobacteria carrying a bar gene were selected on shoot-inducing media containing different concentrations of phosphinothricin (PPT) which is an active component of bialaphos. We found that 20 mg/l of PPT completely inhibited the control explants from growing whereas the explants transformed with the bar gene gave rise to multiple shoots resistant to PPT after 3 weeks under the same selection conditions. The transformation system could also be applied to root explants. Resulting plantlets could produce viable seeds in vitro within 3 months after preparation of the explants. The stable inheritance of the resistance trait, the integration and expression of the bar gene in the progeny were confirmed by genetic tests, Southern analysis and PAT enzyme assay, respectively. In addition, the mature plants in soil showed tolerance to the herbicide Basta.

Accumulation Mechanism of γ-Aminobutyric Acid in Tomatoes (Solanum lycopersicum L.) under Low O2 with and without CO2

The storage of ripe tomatoes in low-O(2) conditions with and without CO(2) promotes γ-aminobutyric acid (GABA) accumulation. The activities of glutamate decarboxylase (GAD) and α-ketoglutarate-dependent GABA transaminase (GABA-TK) were higher and lower, respectively, following storage under hypoxic (2.4 or 3.5% O(2)) or adjusted aerobic (11% O(2)) conditions compared to the activities in air for 7 days at 25 °C. GAD activity was consistent with the expression level of mRNA for GAD. The GABA concentration in tomatoes stored under hypoxic conditions and adjusted aerobic conditions was 60-90% higher than that when they are stored in air on the same day. These results demonstrate that upregulation of GAD activity and downregulation of GABA-TK activity cause GABA accumulation in tomatoes stored under low-O(2) conditions. Meanwhile, the effect of CO(2) on GABA accumulation is probably minimal.

Plant nuclear tRNAMet genes are ubiquitously interrupted by introns

We have isolated three independent clones for nuclear elongator tRNAMet genes from an Arabidopsis DNA library using a tRNAMet-specific probe generated by PCR. Each of the coding sequences for tRNAMet in these clones is identical and is interrupted by an identical 11 bp long intervening sequence at the same position in the anticodon loop of the tRNA. Their sequences differ at two positions from the intron in a soybean counterpart. Southern analysis of Arabidopsis DNA demonstrates that a gene family coding for tRNAMet is dispersed at at least eight loci in the genome. The unspliced precursor tRNAMet intermediate was detected by RNA analysis using an oligonucleotide probe complementary to the putative intron sequence. In order to know whether introns commonly interrupt plant tRNAMet genes, their coding sequences were PCR-amplified from the DNAs of eight phylogenetically separate plant species. All 53 sequences determined contain 10 to 13 bp long intervening sequences, always positioned one base downstream from the anticodon. They can all be potentially folded into the secondary structure characteristic for plant intron-containing precursor tRNAs. Surprisingly, GC residues are always present at the 5′-distal end of each intron.

Characterization of nuclear tRNATyr introns: their evolution from red algae to higher plants

We have previously isolated numerous intron‐containing nuclear tRNATyr genes derived from either monocotyledonous (Triticum) or dicotyledonous (Arabidopsis, Nicotiana) plants by screening the corresponding genomic phage libraries with a synthetic tRNATyr‐specific oligonucleotide. Here we have characterized additional tRNATyr genes from phylogenetically divergent plant species representing red algae (Champia), brown algae (Cystophyllum), green algae (Ulva), stonewort (Chara), liverwort (Marchantia), moss (Polytrichum), fern (Rumohra) and gymnosperms (Ginkgo) using amplification of the coding sequences from the corresponding genomic DNAs by polymerase chain reaction (PCR). All novel tRNATyr genes contain intervening sequences of variable sequence and length ranging in size from 11 to 21 bp. However, two features are conserved in all plant pre‐tRNATyr introns: they possess a uridine and less frequently an adenosine at the 5′ boundary and can adopt similar intron secondary structures in which an extended anticodon helix of 4–5 bp is formed by base‐pairing between nucleotides of the intron and the anticodon loop. In order to elucidate the potential role of the highly conserved uridine at the first intron position, we have replaced it by all other nucleosides in an Arabidopsis pre‐tRNATyr and have studied in wheat germ extract its effect on splicing and on conversion of U to Ψ in the GΨA anticodon. Furthermore, we discuss the putative acquisition of tRNATyr introns at an early step of evolution after the separation of Archaea and Eucarya.

Genetic manipulation of the γ‐aminobutyric acid (GABA) shunt in rice: overexpression of truncated glutamate decarboxylase (GAD2) and knockdown of γ‐aminobutyric acid transaminase (GABA‐T) lead to sustained and high levels of GABA accumulation in rice kernels

Gamma-aminobutyric acid (GABA) is a non-protein amino acid commonly present in all organisms. Because cellular levels of GABA in plants are mainly regulated by synthesis (glutamate decarboxylase, GAD) and catabolism (GABA-transaminase, GABA-T), we attempted seed-specific manipulation of the GABA shunt to achieve stable GABA accumulation in rice. A truncated GAD2 sequence, one of five GAD genes, controlled by the glutelin (GluB-1) or rice embryo globulin promoters (REG) and GABA-T-based trigger sequences in RNA interference (RNAi) cassettes controlled by one of these promoters as well, was introduced into rice (cv. Koshihikari) to establish stable transgenic lines under herbicide selection using pyriminobac. T₁ and T₂ generations of rice lines displayed high GABA concentrations (2-100 mg/100 g grain). In analyses of two selected lines from the T₃ generation, there was a strong correlation between GABA level and the expression of truncated GAD2, whereas the inhibitory effect of GABA-T expression was relatively weak. In these two lines both with two T-DNA copies, their starch, amylose, and protein levels were slightly lower than non-transformed cv. Koshihikari. Free amino acid analysis of mature kernels of these lines demonstrated elevated levels of GABA (75-350 mg/100 g polished rice) and also high levels of several amino acids, such as Ala, Ser, and Val. Because these lines of seeds could sustain their GABA content after harvest (up to 6 months), the strategy in this study could lead to the accumulation GABA and for these to be sustained in the edible parts.

Splicing of Arabidopsis tRNAMet precursors in tobacco cell and wheat germ extracts

Intron-containing tRNA genes are exceptional within nuclear plant genomes. It appears that merely two tRNA gene families coding for tRNATyrGΨ A and elongator tRNAMetCmAU contain intervening sequences. We have previously investigated the features required by wheat germ splicing endonuclease for efficient and accurate intron excision from Arabidopsis pre-tRNATyr. Here we have studied the expression of an Arabidopsis elongator tRNAMet gene in two plant extracts of different origin. This gene was first transcribed either in HeLa or in tobacco cell nuclear extract and splicing of intron-containing tRNAMet precursors was then examined in wheat germ S23 extract and in the tobacco system. The results show that conversion of pre-tRNAMet to mature tRNA proceeds very efficiently in both plant extracts. In order to elucidate the potential role of specific nucleotides at the 3′ and 5′ splice sites and of a structured intron for pre-tRNAMet splicing in either extract, we have performed a systematic survey by mutational analyses. The results show that cytidine residues at intron-exon boundaries impair pre-tRNAMet splicing and that a highly structured intron is indispensable for pre-tRNAMet splicing. tRNA precursors with an extended anticodon stem of three to four base pairs are readily accepted as substrates by wheat and tobacco splicing endonuclease, whereas pre-tRNA molecules that can form an extended anticodon stem of only two putative base pairs are not spliced at all. An amber suppressor, generated from the intron-containing elongator tRNAMet gene, is efficiently processed and spliced in both plant extracts.

Differential subcellular localization, enzymatic properties and expression patterns of γ-aminobutyric acid transaminases (GABA-Ts) in rice (Oryza sativa).

γ-Aminobutyric acid transaminase (GABA-T) catalyzes the conversion of GABA to succinic semialdehyde. The rice (Oryza sativa) genome possesses four putative GABA-T genes, which exhibit high amino acid identity (73-82%) but differ in length of the N-terminal region. Transient expression of GABA-T-green fluorescent fusion proteins in onion epidermal cells demonstrated that two of the four enzymes were targeted to mitochondria, a third to chloroplasts, and the fourth to cytosol. Enzymatic analysis of three organelle-targeted GABA-Ts revealed that they used pyruvate and glyoxylate as amino acceptors and that two of the enzymes functioned in mitochondria and chloroplasts at similar levels of activity, whereas the second mitochondrial enzyme displayed very low activity. Transcriptional analysis demonstrated that two of the four genes were more highly expressed in the vegetative organs tested, but exhibited a different pattern during seed maturation. Together, these results suggest that members of the rice GABA-T gene family vary in many respects, such as intracellular targeting, enzymatic activity and regulation of gene expression.

Sequence analysis of three tRNAPhe nuclear genes and a mutated gene, and one gene for tRNAAla from Arabidopsis thaliana.

Three genes and one mutant gene for tRNAPhe (GAA) and one gene for tRNAAla (UGC) were isolated from a whole-cell DNA library of Arabidopsis thaliana. All three tRNAPhe genes are identical in their nucleotide sequence, but differ in their 5′ and 3′ flanking regions. The mutant tRNAPhe (GAA) gene differs from the other three genes by one nucleotide change from highly conserved G to C at the 57th nucleotide position. The primary structure of the first tRNAAla gene was also determined in this experiment.