Biochemical Mechanism on GABA Accumulation During Fruit Development in Tomato

Plant Cell Physiol. 49(9): 1378–1389 (2008) doi:10.1093/pcp/pcn113, available online at www.pcp.oxfordjournals.org ß The Author 2008. Published by Oxford University Press on behalf of Japanese Society of Plant Physiologists. All rights reserved. For permissions, please email: [email protected] Biochemical Mechanism on GABA Accumulation During Fruit Development in Tomato

Takashi Akihiro 1, 5, Satoshi Koike 1, 5, Ryoji Tani 1, 5, Takehiro Tominaga 1, Shin Watanabe 1, Yoko Iijima 2, Koh Aoki 2, Daisuke Shibata 2, Hiroshi Ashihara 4, Chiaki Matsukura 1, Kazuhito Akama 3, Tatsuhito Fujimura 1 and Hiroshi Ezura 1, * 1 Graduate School of Life and Environmental Sciences, University of Tsukuba, Tsukuba, Ibaraki, 305-8572 Japan 2 Kazusa DNA Research Institute, Kazusa-Kamatari 2-6-7, Kisarazu, Chiba, 292-0818 Japan 3

Faculty of Life and Environmental Science, Shimane University, Matsue city, Shimane, 690-8504 Japan Downloaded from https://academic.oup.com/pcp/article/49/9/1378/1885677 by guest on 28 September 2021 4 Department of Advanced Bioscience, Graduate School of Humanities and Sciences, Ochanomizu University, Ohtsuka, Bunkyo-ku, Tokyo, 112-8610 Japan

A large amount of -aminobutyric acid (GABA) was -aminobutyric acid; GABA-T, GABA transaminase; found to accumulate in tomato (Solanum lycopersicum) fruits GABA-TK, a-ketoglutarate-dependent GABA transami- before the breaker stage. Shortly thereafter, GABA was nase; GABA-TP, pyruvate-dependent GABA transaminase; rapidly catabolized after the breaker stage. We screened the GAD, glutamate decarboxylase; GC–MS, gas chromato- GABA-rich tomato cultivar ‘DG03-9’ which did not show graphy–mass spectrometry; Glu, glutamate; PVDF, poly- rapid GABA catabolism after the breaker stage. Although vinylidene difluoride; RACE, rapid amplification of cDNA GABA hyperaccumulation and rapid catabolism in fruits is ends; RT–PCR, reverse transcription–PCR; SSADH, suc- well known, the mechanisms are not clearly understood. In cinic semialdehyde dehydrogenase. order to clarify these mechanisms, we performed comparative studies of ‘Micro-Tom’ and ‘DG03-9’ fruits for the analysis of gene expression levels, protein levels and enzymatic activity levels of GABA biosynthesis- and catabolism-related enzymes. During GABA accumulation, we found positive Introduction correlations among GABA contents and expression levels of SlGAD2 and SlGAD3. Both of these genes encode glutamate -Aminobutyric acid (GABA) is a four-carbon non- decarboxylase (GAD) which is a key enzyme of GABA protein amino acid and is widespread in bacteria, animals biosynthesis. During GABA catabolism, we found a strong and plants. In vertebrates, GABA is known to be a major correlation between GABA contents and enzyme activity of inhibitory neurotransmitter. GABA has the ability to a-ketoglutarate-dependent GABA transaminase (GABA- lower blood pressure in rats (Elliott and Hobbiger 1959, TK). The contents of glutamate and aspartate, which are Abe et al. 1995, Aoki et al. 2003, Hayakawa et al. 2004) synthesized from GABA and glutamate, respectively, and humans (Takahashi et al. 1961, Inoue et al. 2003, increased with elevation of GABA-TK enzymatic activity. Kajimoto et al. 2004, Noguchi et al. 2007), and is also GABA-TK is the major GABA transaminase form in animals capable of reducing stress (Abdou et al. 2006). Therefore, and appears to be a minor form in plants. In ‘DG03-9’ fruits, the effects of GABA on human health have been the GAD enzymatic activity was prolonged until the ripening subject of a substantial amount of attention in food stage, and GABA-TK activity was significantly low. Taken production. Many GABA-containing foods (germinated together, our results suggest that GAD and GABA-TK play brown rice, chocolate, wine, etc.) are currently available crucial roles in GABA accumulation and catabolism, in the marketplace. respectively, in tomato fruits. The presence of GABA in plants was first described in potato tubers in 1949 (Steward et al. 1949). To date, the widespread occurrence of GABA has been documented in Keywords: GABA — GABA-TK — Glutamate — Micro- many plants. GABA is rapidly produced in response to Tom. anaerobic conditions (Streeter and Thompson 1972), Abbreviations: AAT, aspartate transaminase; Asp, -radiation (Jaarma 1969), low pH (Lane and Stiller aspartic acid; CaM, calmodulin; DAF, days after flowering; 1970), low or high temperatures and darkness, and by DIG, digoxigenin; DTT, dithiothreitol; GABA, mechanical manipulation (Wallace et al. 1984). Various

5These authors contributed equally to this work. *Corresponding author: E-mail, [email protected]; Fax, þ81-29-853-7734.

1378 GABA-TK is crucial for GABA catabolism in tomato fruits 1379 functions of GABA in plants have been described, including (iii) succinic semialdehyde dehydrogenase (SSADH; EC involvement with the regulation of cytosolic pH (Snedden 1.2.1.16) which catalyzes the oxidation of succinic semi- et al. 1995), protection against oxidative stress (Bouche` aldehyde to succinate, which then enters the tricarboxylic et al. 2003), defense against insects (McLean et al. 2003, acid cycle. MacGregor et al. 2003) and the regulation of pollen The main purpose of this study was to clarify the tube growth and guidance (Palanivelu et al. 2003). The mechanism of GABA accumulation and catabolism in majority of scientific studies carried out with the aim of tomato fruits during fruit development. We performed a deciphering the functional roles of GABA have concen- comparative analysis of ‘DG03-9’ (GABA hyperaccumu- trated on stress-related and signaling roles. In addition lator) and ‘Micro-Tom’ (control) in terms of GABA and GABA-related metabolites, gene expression, protein and to these approaches, multiple roles of GABA functioning Downloaded from https://academic.oup.com/pcp/article/49/9/1378/1885677 by guest on 28 September 2021 as a metabolite have also been clarified (Breitkreuz et al. enzymatic activity. Upon completion of this study we have 1999, Bouche` and Fromm 2004, Studart-Guimara˜es et al. demonstrated that GAD and a-ketoglutarate-dependent 2007, Fait et al. 2008). In comparison with other GABA transaminase (GABA-TK) might play crucial roles species, GABA is relatively abundant in tomato (Solanum in GABA accumulation and catabolism, respectively, in the lycopersicum) fruits (Matumoto et al. 1997). GABA tomato fruits. and glutamate (Glu) are the most abundant amino acids in tomato fruits (Inaba et al. 1980, Rolin et al. Results 2000). GABA contents in cv ‘Kyouryoku-toukou’, ‘Cherry tomato’ and ‘Moneymaker’ tomato fruits increased Data mining, isolation and characterization of genes encoding after flowering, reached a maximum level during the GAD, GABA-T and SSADH protein, and their respective mature green stage and rapidly decreased after the genomic organization breaker stage (Inaba et al. 1980, Rolin et al. 2000, Carrari Although Gallego et al. (1995) isolated tomato ERT et al. 2006). Glu levels in tomato fruit increase when the D1 (X80840), a gene encoding a putative GAD protein, GABA levels decrease (Rolin et al. 2000, Carrari et al. other GABA shunt-related genes have not been previously 2006). We have screened the GABA-rich tomato cv isolated. Mining of the tomato expressed sequence tag ‘DG03-9’, which did not show rapid GABA catabolism (EST) database indicated that there are three GADs, three after the breaker stage, from a total of 61 varieties GABA-Ts and one SSADH cDNA available in the (Saito et al. 2008). To date, the molecular mechanism database (Table 1). These clones were successfully amplified of GABA accumulation and catabolism in tomato fruits is by reverse transcription–PCR (RT–PCR) with a set of gene- not well understood. GABA is synthesized from Glu and specific primer pairs (Supplementary Table S1) according to this reaction is catalyzed by glutamate decarboxylase the available sequence information. Total RNA isolated (GAD). GABA is catabolized through the GABA from ‘Micro-Tom’ fruit at 9 days after flowering (DAF) was shunt in the mitochondria, when it bypasses the first used at the template source for amplification of the genes. two steps of the tricarboxylic acid cycle. The GABA shunt is New accession numbers were assigned to the genes which composed of three enzymes: (i) GAD (EC:4.1.1.15); were isolated from ‘Micro-Tom’ (Table 1). (ii) GABA transaminase (GABA-T; EC 2.6.1.19) Since the amino acid sequences of SlGAD1 and ERT which converts GABA to succinic semialdehyde; and D1 are exactly identical (data not shown), it is possible that

Table 1 List of gene encode for GAD, GABA-T and SSADH protein Gene name Length Mol. wt Accession no. bp Amino acids (KDa) cv. Micro-Tom GAD SlGAD1 1783 502 56.7 BT012959 (AB359913) (ERTD1) (X80840) SlGAD2 1756 503 56.9 BT013106 (AB359914) SlGAD3 1606 484 56.7 TC190777 (AB359915) GABA-T Sl-GABA-T1 2005 515 56.7 TC170679 (AB359916) Sl-GABA-T2 1586 458 50.5 TC169925 (AB359917) Sl-GABA-T3 1786 520 57.2 TC173935 (AB359918) SSADH SlSSADH 1838 522 55.8 BT013982 (AB359919) 1380 GABA-TK is crucial for GABA catabolism in tomato fruits

SlGAD1 might be an allele of ERT D1. The putative active A IMG MG Yell Red site of SlGAD protein is strongly conserved (Tanase et al. 1979) and has a putative Ca2þ/calmodulin (CaM)-binding site at its C-terminus (Baum et al. 1993, Arazi et al. 1995) 1600 (Supplementary Fig. S1). A comparison of the deduced 1 amino acid sequences of these clones and Arabidopsis − 1200 AtPOP2 (AT3G22200; GABA transaminase) proteins also showed high homology (Supplementary Fig. S2A, B). 800 SlGABA-T proteins were nearly identical in the consensus mol 100 gFW pyridoxal phosphate-binding motif (Supplementary GABA contents Downloaded from https://academic.oup.com/pcp/article/49/9/1378/1885677 by guest on 28 September 2021 m 400 Fig. S2A). SlSSADH proteins also have strongly conserved putative aldehyde dehydrogenase motifs (Arazi et al. 1995) 0 (Supplementary Fig. S3). 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 DAF We performed Southern blot analysis in order to B estimate the copy numbers of GAD, GABA-T and SSADH DAF 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 genes in the tomato genome. Under highly stringent SlGAD1 conditions for hybridization and washing, at least six, three and one band were detected (Supplementary Fig. S4) SlGAD2 with GAD, GABA-T and SSADH gene-specific probes SlGAD3 (Supplementary Figs. S1, S2 and S3), respectively. SlGABA-T1 SlGABA-T2 Characterization of GABA and GABA metabolism-related enzymes during the development of tomato fruits SlGABA-T3 GABA contents in fruits increased after flowering and SlSSADH reached a maximum at 27 DAF (Fig. 1A). Sixty-seven SlE8 percent of the total amino acid content was GABA in fruits SlGAPDH at 27 DAF (data not shown). After 27 DAF, GABA contents in fruits rapidly decreased, as shown in cv rRNA ‘Kyouryoku-Toukou’, ‘Cherry tomato’ and ‘Moneymaker’ (Inaba et al. 1980, Rolin et al. 2000, Carrari et al. 2006). Fig. 1 Time-course changes of GABA contents and gene GABA contents in fruits at 45 DAF were approximately 1/9 expression of GABA shunt genes in ‘Micro-Tom’ fruits during of the GABA contents in the fruits at 27 DAF (Fig. 1A). various stages of fruit development. (A) ‘Micro-Tom’ fruits during the stages of fruit development. IMG, immature green stage; MG, The expression properties of GABA shunt genes in mature green stage; Yell, yellow stage; Red, red stage. Time-course ‘Micro-Tom’ developing fruits were analyzed with semi- changes of GABA contents in ‘Micro-Tom’ fruits. The presented quantitative RT–PCR using isoform-specific primer pairs. values for GABA contents represent the means from three Although three SlGAD genes were expressed in developing replications. Vertical bars represent the standard deviation. (B) fruits (Fig. 1B), they exhibited differential patterns of The expression of three GAD genes, three GABA-T genes and a SSADH gene in developing ‘Micro-Tom’ fruits was analyzed using gene expression during the various stages of fruit develop- semi-quantitative RT–PCR. The E8 gene was analyzed as a positive ment. The changes in SlGAD2 and SlGAD3 gene expres- control for fruits. The GAPDH gene was also analyzed as an sion in fruit were in good accordance with the respective internal control. The bottom panel represents an ethidium bromide- changes of GABA contents in fruits (Fig. 1A, B). The stained gel which demonstrates the equal loading of RNA in expression patterns of SlGAD2 and SlGAD3 and the each lane. ethylene-responsive gene (E8) showed opposite patterns of expression during the stages of fruit development. SlGAD1 gene expression increased after the breaker Characterization of the GABA-rich variety ‘DG03-9’ stage and the expression pattern for this gene was not Through a preliminary screening of 61 tomato well correlated with GABA contents during fruit ripening. varieties, we identified ‘DG03-9’ as a GABA-rich cultivar Three SlGABA-T genes and the SlSSADH gene were (Saito et al. 2008). The GABA content in ‘DG03-9’ fruits expressed in fruits; however, they also exhibited differ- at the ripening stage was the highest among all of the ential patterns during fruit maturation (Fig. 1B). The 61 tomato varieties. In order to clarify the time-course of changes in SlGABA-T and SlSSADH gene expression in changes in GABA contents in ‘DG03-9’ fruits, we measured fruits were not correlated with GABA contents in fruits GABA contents in ‘DG03-9’ fruits. The breaker stage in (Fig. 1A, B). ‘DG03-9’ fruits occurred at 27–30 DAF. ‘DG03-9’ fruits GABA-TK is crucial for GABA catabolism in tomato fruits 1381 showed GABA accumulation prior to the breaker stage. indicated that GAD enzymatic activity in ‘DG03-9’ fruits However, ‘DG03-9’ fruits did not show rapid catabolism of was prolonged after the breaker stage. GABA subsequent to the breaker stage (Fig. 2A). These mRNA levels of SlGABA-T genes did not correlate results indicate that the higher accumulation of GABA in with GABA contents in ‘Micro-Tom’ fruits (Fig. 1A, B). ‘DG03-9’ is caused by the slow degradation of GABA after Similar results were also obtained with measurements in the breaker stage of fruit development. ‘DG03-9’ fruits (Fig. 4A). In both varieties, the expression properties of GABA-T genes were similar (Fig. 4A), and Comparison of GABA and other GABA-related components pyruvate-dependent GABA transaminase (GABA-TP) in fruits of ‘Micro-Tom’ and the GABA-rich cultivar enzymatic activity did not correlate with GABA contents (Fig. 4B). GABA-TK enzymatic activity in ‘Micro-Tom’ ‘DG03-9’ Downloaded from https://academic.oup.com/pcp/article/49/9/1378/1885677 by guest on 28 September 2021 To characterize the ‘DG03-9’ fruits further, we fruits dramatically increased after the breaker stage; performed gas chromatography–mass spectrometry (GC– however, in contrast, GABA-TK enzymatic activity did MS) analysis as a means of measuring the GABA shunt and not increase in ‘DG03-9’ fruits until after the breaker stage tricarboxylic acid cycle components. We utilized this (Fig. 4C). A direct correlation between the increase of approach to perform a comparative analysis between GABA-TK enzyme activity and the decrease of GABA ‘Micro-Tom’ and ‘DG03-9’ fruits. GC–MS analysis was contents in ‘Micro-Tom’ fruits was detected after the unable to identify succinic CoA, oxaloacetate and succinic breaker stage. GABA-TK enzymatic activity was signifi- semialdehyde contents. Therefore, succinic semialdehyde cantly higher (4200 times) than GABA-TP enzymatic contents were measured by the modified GABase method activity (Fig. 4B, 4C). In contrast, GABA-TP enzymatic (Streeter and Thompson 1972). Succinic semialdehyde and activity in ‘Micro-Tom’ leaves was about five times higher succinic acid contents were not significantly different than GABA-TK enzymatic activity (Supplementary between both cultivars before and after the breaker stage. Fig. S6). These results indicated that GABA-TK plays an It is well known that GABA and citrate contents in tomato important role in the degradation of GABA in tomato fruits increase after the breaker stage (Rolin et al. 2000). fruits. Conversely, malate, Glu and aspartic acid (Asp) contents SlSSADH mRNA levels in ‘Micro-Tom’ fruits at increase after the breaker stage (Rolin et al. 2000, Roessner- mature green, yellow and red stages were approximately Tunali et al. 2003, Carrari et al. 2006, Carrari and Fernie 20% higher than in ‘DG03-9’ fruits (Fig. 4A). However, 2006, Mattoo et al. 2006, Mounet et al. 2007). Interestingly, SSADH enzymatic activities (Fig. 5B) and succinic semi- Asp and citrate contents did not increase significantly after aldehyde contents (Fig. 2c) in both varieties were not the breaker stage in ‘DG03-9’ fruits. In addition, Glu significantly different. contents in ‘DG03-9’ did not increase just after the breaker stage. Although malate contents in ‘Micro-Tom’ were Discussion approximately 2–3 times higher than in ‘DG03-9’, the tendency towards the reduction of malate at the red stage In comparison with other plants, tomato accumulates a was the same in both cultivars. These results suggest that large amount of GABA during the process of fruit ‘DG03-9’ has mutations not only in genes that are development. In tomato fruits, GABA contents reach a implicated in GABA degradation but also in genes that maximum level prior to the breaker stage and decrease are related to the synthesis or degradation of citrate, Glu rapidly thereafter. However, the process by which tomato and Asp after the breaker stage. can regulate dramatic changes in the GABA contents of fruits is not well known. In order to elucidate the Comparison between ‘Micro-Tom’ and ‘DG03-9’ fruits in mechanism of regulation of GABA contents in tomato gene expression levels and enzymatic activities of GABA fruits, we performed a comparative characterization shunt enzymes and in GAD protein levels between GABA-rich and GABA-standard tomato varieties We analyzed SlGAD, SlGABA-T and SlSSADH gene with respect to changes in gene expression, protein levels expression levels as a means to identify and characterize and enzymatic activities. differences in GABA shunt enzymes between ‘Micro-Tom’ There are two types of GABA-Ts which have been and ‘DG03-9’ fruits. SlGAD2 and SlGAD3 gene expression studied in various model systems. GABA-Ts use either levels in ‘Micro-Tom’ fruits were dramatically decreased a-ketoglutarate or pyruvate as amino acid acceptors in after the breaker stage (Fig. 1A). In sharp contrast, SlGAD2 order to produce Glu or alanine (Bouche` and Fromm 2004). and SlGAD3 transcripts were detected after the breaker In mammals, it appears that only GABA-TK is present stage in ‘DG03-9’ fruits (Fig. 3A). Similar results were (Bouche` and Fromm 2004), whereas in plants both GABA- obtained for the levels of GAD protein (Fig. 3B) and its TP and GABA-TK activities have been detected respective enzymatic activity (Fig. 3C). These results clearly (Streeter and Thompson 1972, Reggiani et al. 1988, 1382 GABA-TK is crucial for GABA catabolism in tomato fruits

GABA Breaker A 1800 B IMG MG Yell Red 1 − 1375 Micro-Tom 950

mol· 100 gFW 475 m DG03-9 0 IMG MGYell Red

C Succinc semialdehyde Succinate Fumarate 1 1 1 −

− 20 − 40 10 ** ** ** 20 5 10 ** * ** mol· 100 gFW mol· 100 gFW mol· 100 gFW m m m 0 0 0 IMG MG Yell Red IMG MG Yell Red IMG MG Yell Red

Malate Citrate Isocitrate 1 1 1 40000 − − − 2000 4000 ** * ** ** ** * ** ** 1000 20000 ** 2000 Downloaded from https://academic.oup.com/pcp/article/49/9/1378/1885677 by guest on 28 September 2021 ** mol· 100 gFW mol· 100 gFW mol· 100 gFW m m m 0 0 0 IMG MG Yell Red IMG MG Yell Red IMG MG Yell Red

Glutamate Aspartic acid Alanine 1 1 1 − − − 2000 1000 ** 200

1000 500 100 * ** * * mol· 100 gFW mol· 100 gFW mol· 100 gFW m m m 0 0 0 IMG MG Yell Red IMG MG Yell Red IMG MG Yell Red

Aspartic acid D Isocitrate GDH AAT a-Ketoglutarate Cis-Aconitate Glutamate GAD

Citrate GABA-TK Succinyl-CoA GABA α-Ketogulutarate Pyruvate Oxaloacetate GABA-TP Succinic Succinate semialdehyde Malate SSADH Alanine Fumarate

Fig. 2 Comparison of GABA and GABA-related components in ‘Micro-Tom’ and ‘DG03-9’ fruits. (A) Time-course changes of GABA contents in ‘Micro-Tom’ (filled box) and ‘DG03-9’ (open box) fruits. The presented values for GABA contents represent the means from three replications. Vertical bars represent the standard deviation. The arrow indicates the day when the breaker stage initiated. (B) Development of ‘Micro-Tom’ and ‘DG03-9’ fruits (scale bar ¼ 1 cm). (C) Comparison of GABA and GABA-related components in ‘Micro- Tom’ and ‘DG03-9’ fruits during the stages of fruit development. The presented values for each component represent the means from three replications. Vertical bars represent the standard error. The level of significance was determined using the Student’s t-test (P50.05; P50.01). (D) Tricarboxylic acid cycle and GABA shunt pathway. IMG, immature green stage; MG, mature green stage; Yell, yellow stage; Red, red stage. GAD, glutamate decarboxylase; GABA-TK, a-ketoglutarate-dependent GABA aminotransferase; GABA-TP, pyruvate- dependent GABA aminotransferase; SSADH, succinic semialdehyde dehydrogenase; AAT, aspartate transaminase. GABA-TK is crucial for GABA catabolism in tomato fruits 1383

‘Micro-Tom’ ‘DG03-9’ A 400 300 SlGAD1 % 200 mRNA 100 0 IMG MG Yell Red IMG MG Yell Red 150

SlGAD2 Downloaded from https://academic.oup.com/pcp/article/49/9/1378/1885677 by guest on 28 September 2021 % 75 mRNA

0 IMG MG Yell Red IMG MG Yell Red 200 150 SlGAD3 % 100 mRNA 50 0 IMG MG Yell Red IMG MG Yell Red

B ‘Micro-Tom’ ‘DG03-9’ Rice seed protein

GAD Protein

IMG MG Yell Red IMG MG Yell Red

C ‘Micro-Tom’ ‘DG03-9’ )

1 40 −

mg protein 20 1 −

GAD activity 10

(nmol min 0 IMG MG Yell Red IMG MG Yell Red

Fig. 3 Comparison of mRNA levels, protein levels and enzymatic activity of GAD in ‘Micro-Tom’ and ‘DG03-9’ fruit. (A) Semi- quantitative RT–PCR analysis of SlGAD gene expression. Total RNA was isolated from tomato fruits and used as the template for synthesis of single-stranded cDNA. Gene-specific amplification was accomplished with PCR isoform-specific primer sets (Supplementary Table S1). In order to increase sensitivity, samples were fractionated by electrophoresis on a 1% agarose gel, transferred to a membrane, and subsequently probed with the same fragments that were labeled with [a-32P]dCTP. The intensity of expression levels was quantified by using the ‘QuantityOne’ software (PDI, Inc.) with reference to the intensity of the band of ‘Micro-Tom’ at the immature green stage (IMG). (B) Western blot analysis of GAD protein levels. Anti-OsGAD1 and OsGAD2 (1 : 1 mixed) primary antibodies (Akama and Takaiwa 2007) were hybridized with blotted total protein extracts (5 mg) from ‘Micro-Tom’ and ‘DG03-9’ fruits at the IMG, mature green (MG), yellow (Yell) and red (Red) stages. Rice seed proteins (15 mg) at 15 DAF were used as a positive control antigen. (C) Time-course changes of GAD enzymatic activity. Total proteins were extracted from ‘Micro-Tom’ (filled box) and ‘DG03-9’ (open box) fruits at the IMG, MG, Yell and Red stages. The presented values represent the means from three replications. The vertical bars represent the standard deviation.

Wallace et al. 1984, Shelp et al. 1995, Cauwenberghe et al. but not a-ketoglutarate as an amine donor (Cauwenberghe 2002). In Arabidopsis, only genes encoding GABA-TP have et al. 2002). The GABA contents in flowers of an been isolated (Cauwenberghe et al. 2002). In an in vitro Arabidopsis GABA-TP-disrupted mutant (pop2) were study, recombinant Arabidopsis GABA-TP used pyruvate 100-fold higher than those measured in wild-type plants 1384 GABA-TK is crucial for GABA catabolism in tomato fruits

A ‘Micro-Tom’ ‘DG03-9’ 100

SlGABA-T1 mRNA % 50

0 IMG MG Yell Red IMG MG Yell Red 100 Downloaded from https://academic.oup.com/pcp/article/49/9/1378/1885677 by guest on 28 September 2021 SlGABA-T2 % 50 mRNA

0 IMG MG Yell Red IMG MG Yell Red 200

SlGABA-T3 % 100 mRNA

0 IMG MG Yell Red IMG MG Yell Red

B ‘Micro-Tom’ ‘DG03-9’ )

1 0.016 −

0.012

mg protein 0.008 1 − 0.004 GABA-TP activity

(nmol min 0.000 IMG MG Yell Red IMG MG Yell Red

C ‘Micro-Tom’ ‘DG03-9’ )

1 16 −

mg protein 8 1 −

4 GABA-TK activity

(nmol min 0 IMG MG Yell Red IMG MG Yell Red

Fig. 4 Comparison of mRNA levels, protein levels and enzymatic activity of GABA-T in ‘Micro-Tom’ and ‘DG03-9’ fruit. (A) Semi- quantitative RT–PCR analysis of expression of SlGABA-T genes. Time-course changes of the mRNA levels of SlGABA-T genes in ‘Micro- Tom’ (filled box) and ‘DG03-9’ (open box) fruits during different developmental stages were determined. Relative intensities for gene expression levels of three SlGABA-T genes were quantified by using the ‘QuantityOne’ software (PDI, Inc.) with reference to the intensity of the band of ‘Micro-Tom’ at the immature green (IMG) stage. (B) Time-course changes of GABA-TP enzymatic activity. (C) Time-course changes of GABA-TK enzymatic activity. Total proteins were extracted from ‘Micro-Tom’ (filled box) and ‘DG03-9’ (open box) fruits at the IMG, mature green (MG), yellow (Yell) and red stages. Vertical bars represent the standard deviation. The arrow indicates the day when the breaker stage started.

(Palanivelu et al. 2003). GABA-TK activity has also been 1984, Shelp et al. 1995). In these cases, GABA-TP activities detected in soybean seeds, rice roots, tobacco leaves, radish were 1.2–19 times higher than GABA-TK activities. The mature leaves, tomato leaves and potato tubers (Streeter gene encoding GABA-TP of plants was identified; however, and Thompson 1972, Reggiani et al. 1988, Wallace et al. GABA-TK of plants remains to be identified. In this study, GABA-TK is crucial for GABA catabolism in tomato fruits 1385

A ‘Micro-Tom’ ‘DG03-9’

150

100 SlSSADH % mRNA 50

0 IMG MG Yell Red IMG MG Yell Red

B ‘Micro-Tom’ ‘DG03-9’ Downloaded from https://academic.oup.com/pcp/article/49/9/1378/1885677 by guest on 28 September 2021 )

1 0.35 − 0.30 0.25 0.20 mg protein 1

− 0.15 0.10 SSADH activity 0.05

(nmol min 0.00 IMG MG Yell Red IMG MG Yell Red

Fig. 5 Comparison of mRNA levels, protein levels and enzymatic activity of SSADH in ‘Micro-Tom’ and ‘DG03-9’ fruit. (A) Semi- quantitative RT–PCR analysis of SlSSADH gene expression. Time-course changes of SlSSADH mRNA levels in ‘Micro-Tom’ (filled box) and ‘DG03-9’ (open box) fruits during different development stages were identified with semi-quantitative RT–PCR. The intensity of gene expression levels of the SlSSADH gene was quantified by using the ‘QuantityOne’ software (PDI, Inc.) with reference to the intensity of the band of ‘Micro-Tom’ at the immature green stage. (B) Time-course of changes of SSADH enzymatic activity. The presented values represent the means from three replications. Vertical bars represent the standard deviation. The arrow indicates the day when the breaker stage started.

we have detected a significantly higher level of GABA-TK and GABA-TK enzymatic activity did not increase after the activity in tomato fruits after the breaker stage (Fig. 4C). breaker stage (Fig. 3). These results indicated that main- Since GABA-TK enzymatic activity was also detected in a taining GAD enzymatic activity after the breaker stage different tomato variety, cv ‘House Momotaro’ (data not resulted in the hyperaccumulation of GABA in ‘DG03-9’ shown), these data demonstrate that high GABA-TK fruits. Taken together, these data support the conclusion enzymatic activity is common in tomato fruits. that GAD and GABA-TK enzymatic activity play a crucial The observed increase in GABA-TK enzymatic activity in role in GABA biosynthesis and degradation in tomato ‘Micro-Tom’ fruits was well correlated with the reduction of fruits. GABA content (Figs. 2A, 4C). GABA-TK enzymatic ‘DG03-9’ is the Beta (B) mutant which contains a activity in the GABA-rich cultivar ‘DG03-9’ was very low mutation in the locus encoding lycopene b-cyclase (Ronen (Fig. 4C). Collectively, these data indicate that GABA-TK et al. 2000). The B mutant in tomato has the phenotype of might play a crucial role in the catabolism of GABA in orange fruits due to the accumulation of b-carotene (Ronen tomato fruits. et al. 2000). Therefore, the color of ‘DG03-9’ fruits GABA contents in tomato fruits were in good remained orange during the maturation stage (Fig. 2B) agreement with the levels of SlGAD2 and SlGAD3 gene and the colors of ‘Micro-Tom’ and ‘DG03-9’ fruits during expression, GAD protein level and GAD enzymatic the red stage were significantly different from each other activity. In ‘Micro-Tom’ fruits, SlGAD2 and SlGAD3 (Fig. 2B). The breaker stage occurred normally in ‘DG03-9’ mRNA levels, GAD protein level and GAD enzymatic at 28–34 DAF, and fruit softening also occurred (data not activity were dramatically reduced after the breaker stage show). These results indicated that the difference between (Fig. 3). In contrast, GAD enzymatic activity was detected ‘Micro-Tom’ and ‘DG03-9’ in GABA catabolism at the in ‘DG03-9’ fruits even after the breaker stage (Fig. 3). ripening stage was not caused by a difference in the period GAD enzymatic activity in ‘Micro-Tom’ fruits was regu- of fruit development. In a previous study, we evaluated fruit lated in an opposite manner to GABA-TK during different GABA content of seven b-carotene-rich cultivars harboring developmental stages. In contrast, GAD enzymatic activity the B locus (Saito et al. 2008). However, the fruit GABA in ‘DG03-9’ fruits did not decrease after the breaker stage levels of those cultivars were almost similar to or low 1386 GABA-TK is crucial for GABA catabolism in tomato fruits compared with other ordinary cultivars, suggesting that the added and vigorously mixed for 10 min. Samples were centrifuged GABA-rich trait of ‘DG03-9’ is not genetically linked to the again at 10,000g for 20 min at 48C. The supernatant from this centrifugation step was removed and 400 ml of diethyl ether was B locus. added and then vigorously mixed for 10 min and centrifuged at Increases in Glu and Asp contents after the breaker 10,000 r.p.m. for 10 min at 48C. The supernatant from this stage have been well characterized in previous studies centrifugation step was removed and left to stand under a draft of (Roline et al. 2000, Roessner-Tunali et al. 2003, Carrari air for 30 min for complete evaporation of ether. The ‘GABase’ et al. 2006, Carrari and Fernie 2006, Mattoo et al. 2006, assay for GABA was performed using the method described by Jakoby (1962) with slight modifications. The ‘GABase’ assay Mounet et al. 2007). However, GC–MS analysis indicated monitors the reduction of NADP to NADPH spectroscopically at that the content of Glu and Asp in ‘DG03-9’ fruits did not 340 nm, pH 8.6 at 258C, as a function of time using GABA as a increase after the breaker stage. Asp was synthesized from substrate. The ‘GABase’ assay for succinate semialdehyde was Downloaded from https://academic.oup.com/pcp/article/49/9/1378/1885677 by guest on 28 September 2021 Glu catalyzed by aspartate transaminase (AAT, EC 2.6.1.1). performed using the method as described by Streeter and AAT catalyzes the reversible reaction of transamination Thompson (1972). between Asp and 2-oxoglutarate to generate Glu and Isolation of GABA shunt enzymes genes oxaloacetate. AAT plays a key role in the metabolic The international tomato genome sequencing project regulation of carbon and nitrogen metabolism in all (Mu¨ eller et al. 2005) is accumulating genome and mRNA organisms (Torre et al. 2006). It is possible that the lower information which can be used through the Sol genomic Glu and Asp contents in ‘DG03-9’ fruits relative to ‘Micro- Network (http://www.sgn.cornell.edu/index.pl), TIGER data- Tom’ fruits after the breaker stage may result from lower base (http://compbio.dfci.harvard.edu/tgi/cgi-bin/tgi/Blast/index. GABA-TK enzymatic activity in ‘DG03-9’ fruits. In turn, cgi), KaFTom (http://www.pgb.kazusa.or.jp/kaftom/) and MiBase (http://www.pgb.kazusa.or.jp/kaftom/index.html). Nine this would result in decreased Glu accumulation, and cDNA sequences of putative GABA shunt-related enzyme genes eventually a reduction in the accumulation of Asp in were identified from these databases, and gene-specific sets of ‘DG03-9’ fruits. To clarify the detailed roles of GABA-TK primer pairs were designed accordingly to their respective in GABA catabolism and Glu accumulation in fruits will sequences. require further work including the isolation and character- cDNAs of SlGAD1 and SlGAD2 containing open reading ization of gene(s) encoding GABA-TK and the character- frames were obtained by RT–PCR using gene-specific sets of primer pairs. Total RNA from ‘Micro-Tom’ fruits at 9 DAF was ization of mutants deficient in this gene. isolated using the RNeasy maxi kit (Qiagen, Valencia, CA, USA). A10mg aliquot of total RNA was reverse transcribed using the Materials and Methods First strand cDNA synthesis kit (TAKARA SHUZO CO. LTD., Otsu, Shiga, Japan) with an oligo(dT) primer according to the manufacturer’s instructions. PCR was performed using the Plant materials and growth conditions oligonucleotide primers (Supplementary Fig. S1). The partial The S. lycopersicum L. cv. Micro-Tom germinated tomato sequence of SlGAD3 was obtained by RT–PCR, and a 50 and 30 seedlings were grown in a growth chamber with a day/night –2 –1 region of SlGAD3 was subsequently obtained by RACE (rapid photoperiod of 16/8 h at 60.5 mmol photons m s at 258C). amplification of cDNA ends) using the Smart Race cDNA Seedlings were supplied with a standard nutrient solution (Otsuka amplification kit (TAKARA SHUZO CO. LTD., Otsu, Shiga, House Nos. 1 and 2, Otsuka Chemical Co., Osaka Japan). The Japan) according to the manufacturer’s instructions. S. lycopersicum L. cv. ‘DG03-9’ germinated tomato seedlings For PCR amplification of the cDNAs, the samples were were grown in a glass house at the Gene Research Center, the initially denatured at 948C for 3 min, and then for 1 min in University of Tsukuba, Japan, in May 2007. Developing fruits of subsequent cycles. Primer annealing and extension reactions were ‘Micro-Tom’ and ‘DG03-9’ were collected at 12, 24, 36 and carried out at 50 and 728C for 60 and 90 s each, respectively. After 45 DAF. ‘Micro-Tom’ tomato fruit at 14–18 DAF and ‘DG03-9’ 35 cycles, the amplified DNA fragments were subcloned into the fruit at 10–14 DAF which were not fully expanded and still green pGEM-T Easy vector (Promega, Madison, WI, USA) and were defined as ‘immature green’. ‘Micro-Tom’ and ‘DG03-9’ fruit subsequently sequenced. at 22–26 DAF which were fully expanded and green were defined as ‘mature green’. ‘Micro-Tom’ fruit at 27–34 DAF and ‘DG03-9’ Genomic DNA extraction and Southern blot analysis fruit at 28–34 DAF which are fully expanded were defined as For extraction of plant genomic DNA, approximately 4 g of ‘yellow’. At this stage, there is yellowish plaque on the ‘DG03-9’ fresh leaves were frozen with liquid nitrogen and ground in a pericarp. ‘Micro-Tom’ fruit at 42–45 DAF and ‘DG03-9’ fruit mortar. Genomic DNA was extracted with the DNeasy Plant Maxi at 43–47 DAF which were fully expanded were defined as ‘red’. Kit (Qiagen, Valencia, CA, USA), and cDNA fragments were At this stage, the ‘DG03-9’ pericarp was orange. labeled with digoxigenin (DIG)-dUTP (Roche Diagnostics, Vienna, Austria) according to the manufacturer’s instructions. Extraction and measurement of GABA and succinic semialdehyde For Southern blot analysis, genomic DNA was digested with contents EcoRI, EcoRV, XhoI, DraI and HindIII under the condi- A 50 mg aliquot of fresh fruits was homogenized in 8% (w/v) tions specified by the supplier. The digested genomic DNAs trichloroacetic acid with a Tissuelyzer (Qiagen, Valencia, CA, were electrophoresed in 0.8% agarose gels and transferred USA) at 2,000 r.p.m. for 1 min at room temperature. Samples were to a positively charged nylon membrane (GE Healthcare, then centrifuged at 10,00 0g for 20 min at 48C. The supernatant Buckinghamshire, UK) by capillary transfer. Transferred mem- was transferred to a new tube and 400 ml of pure diethyl ether was branes were then hybridized with the DIG-labeled DNA probe for GABA-TK is crucial for GABA catabolism in tomato fruits 1387

16 h at 458C. Hybridization, stringency washes and detection were was set to 2008C. The time of flight mass spectrometer was a performed following the instructions outlined in the DIG DNA Pegasus III MS system (Leco, Michigan, USA) with an electron labeling kit (Roche Diagnostics, Vienna, Austria), with a final impact ionization source set to 2508C. Mass spectra were wash in 1 SSC, 1% SDS at 608C. Membranes were then exposed monitored with an acquisition rate of 20 spectra s–1 in the mass to X-ray film for 5–180 min. range m/z ¼ 82–500. The contents of most metabolites were quantified with the standard calibration curves for each com- Expression analysis of GABA shunt enzyme genes by semi- pound. The peak area ratios of standard compounds at various quantitative RT–PCR concentrations to the internal standard (ribitol) were used for the Semi-quantitative RT–PCR was carried out essentially as calibration curves. previously described by Akihiro et al. (2005, 2006). Total RNAs were extracted from tomato plants using the RNeasy Maxi kit Enzyme extraction and assays of GABA shunt-related enzymes

(Qiagen, Valencia, CA, USA) and were digested with DNase I A 50 g aliquot of fresh tomato fruit was homogenized Downloaded from https://academic.oup.com/pcp/article/49/9/1378/1885677 by guest on 28 September 2021 (NipponGene, Chiyodaku, Tokyo, Japan) according to the with a Waring blender in a 5-fold volume of ice-cold extraction manufacturer’s instructions. The quality of total RNAs was buffer [0.1 M Tris–HCl pH 7.0, 10 mM dithiothreitol (DTT), 5 mM evaluated by the Agilent 2100 bioanalyzer (Agilent, Santa Clare, EDTA, 1 mM pyridoxal-5-phosphate and 1% (w/v) insoluble California, USA). Since the RNA integrity number (Schroeder polyvinylpyrrolidone] (Rolin et al. 2000) For measurement et al. 2006) of all samples exceeded 7.9, these RNAs were of GAD and GABA-TK activities, the homogenates subsequently used for cDNA synthesis. A 15 mg aliquot of total were centrifuged at 10,000g for 10 min at 48C, and the pellet was RNAs was used to synthesize single-stranded cDNA using the discarded. The extract was desalted using Sephadex G-50 First strand cDNA synthesis kit (TAKARA SHUZO CO. LTD., (GE Healthcare, UK) that was previously equilibrated in the Otsu, Shiga, Japan) according to the manufacturer’s instructions. extraction buffer. A 500 ml aliquot of effluent was added RT–PCR was performed semi-quantitatively with sets of gene- to microcon concentrators (Millipore, Danvers, Massachusetts, specific primers (Supplementary Table S1). For PCR ampli- USA) and reduced in volume to 100 ml (hereafter, this solution is fication, the cDNA was denatured at 948C for 2 min in the first referred as the ‘crude protein’). cycle, and then for 1 min in subsequent cycles. Primer annealing The GAD activity was measured as glutamate-dependent and extension reactions were carried out at 508C(SlGAD2 and GABA production (as described by Akama et al 2007). A 100 ml SlSSR1), 598C(SlSSADH), 608C(SlGAD3, SlGABA-T2 and aliquot of the crude protein was used for the assay of GAD activity SlGABA-T3), 668C(SlGAD1 and SlGABA-T3)or728C for 30 s in a 500 ml reaction mixture [100 mM Bis-Tris–HCl (pH 7.0), each. The cycle numbers for SlGAD1, SlGAD2, SlGAD3, SlGABA- 0.5 mM pyridoxal-5-phosphate, 1 mM DTT, 5 mM Glu, 0.1 mM T1, SlGABA-T2, SlGABA-T3, SlSSADH, SlSSR1, SlSSR2 and bovine CaM (Sigma-Aldrich, Missouri) and 0.5 mM CaCl2]. SlE8 were: 20, 20, 18, 23, 26, 23, 26, 18, 20 and 20, respectively. GABA contents were determined enzymatically using a commer- The PCR products were fractionated on a 1% (v/w) agarose gel, cial ‘GABase’ preparation (Sigma-Aldrich, Missouri, USA) þ transferred onto a Hybond N membrane (GE Healthcare, (Jakoby 1962). 32 Buckinghamshire, UK) and probed with [a- P]dCTP-labeled For measurement of GABA-TP and SSADH activity the DNA (Muromachi Chemicals Inc., Nerima, Tokyo, Japan) using protein extraction and concentration were carried out using the the Bca Best labeling kit (TAKARA SHUZO CO. LTD., Otsu, above-mentioned method, However, GABA-TP and SSADH Shiga, Japan) according to the manufacturer’s instructions. enzymatic activities were quite low or undetectable. Therefore, the amount of starting material and the methods of protein GC–MS analysis concentration were changed as follows: 50 g samples of fresh Tomato fruits (50 mg) were homogenized in liquid nitrogen tomato fruits were homogenized with a Waring blender in a 5-fold using a mortar and pestle. Each sample was dissolved with volume of ice-cold extraction buffer. The homogenate was methanol and chloroform (250 ml each). After 225 ml of 0.295 mM centrifuged at 10,000g for 10 min at 48C, and the pellet was ribitol solution was added to the samples as an internal standard, discarded. Ammonium sulfate was added [final concentration of they were vigorously mixed. These extracts were centrifuged at 60% (v/w)] to the supernatant. The sample was subsequently 15,000 r.p.m. for 10 min at room temperature, and 200 ml of the mixed for 30 min at 48C and then centrifuged at 10,000g for supernatant was filtered using Amicon Ultrafree-MC (Millipore, 30 min at 48C. The pellet was dissolved in 1 ml of the extraction Danvers, Massachusetts, USA). A part of the flow-through buffer. The buffer of the solution was changed to SSADH reaction fraction (80 ml) was evaporated to dryness using a centrifuge buffer [100 mM sodium pyrophosphate buffer (pH 9.0), 14 mM evaporator (EYELA centrifugal evaporator CVE-3100, EYELA, 2-mercaptoethanol, 0.5 mM NAD, 0.5 mM succinic semialdehyde Chuo, Tokyo, Japan). For methylation, 40 ml of methoxylamine (modified from Shelp et al. (1995)], which is the same as the hydrochloride (20 mg ml–1 pyridine) was added to the samples GABA-TK reaction buffer or the GABA-TP reaction buffer and incubated for 90 min at 378C. Trimethylsilylation was [100 mM Tris–HCl buffer (pH 9.0), 20 mM pyridoxal-5-phosphate, performed by addition of 50 mlofN-methyl-N-(trimethylsilyl)- 2 mM GABA and 10 mM a-ketoglutarate or 10 mM pyruvate trifluoroacetamide (MSTFA) solution for 30 min at 378C. (Shelp et al. 1995)], with Sephadex G-50 (GE Healthcare, A GC 6890 (Agilent Technologies, Santa Clare, California, Buckinghamshire, UK) which had been equilibrated with new USA) was operated under electronic pressure control and equipped buffer. with a split/splitless capillary inlet. A 1 ml aliquot of each sample The GABA-TK activity was measured as GABA-dependent was injected in the splitless mode with the injection temperature set Glu production (Shelp et al. 1995). The buffer-exchanged solution to 2508C. A 30 m DB-17ms column (J&W Scientific, Santa Clare, was incubated at 378C for 12 h and then boiled for 3 min. Glu California, 0.25 mm ID, 0.25 mm film thickness) was used. Helium contents were measured using the L-glutamate assay kit (Yamasa was used as a flow gas at 1.0 ml min–1. Separation was achieved Co., Chyoshi, Chiba, Japan) (Kusakabe et al. 1983) according to with a temperature program of 708C for 5 min, then ramped at the manufacturer’s instructions. GABA-TP activity was measured 158C min–1 to 3108C. The transfer line to the mass spectrometer as for GABA-dependent succinic semialdehyde production 1388 GABA-TK is crucial for GABA catabolism in tomato fruits

(Shelp et al. 1995). SSADH activity was measured as succinic Akama, K. and Takaiwa, F. (2007) C-terminal extension of rice glutamate semialdehyde-dependent succinate production (Beutler 1989). The decarboxylase (OsGAD2) functions as an autoinhibitory domain and succinate content was measured using the succinate assay kit overexpression of a truncated mutant results in the accumulation of (Megazyme, Wicklow, Ireland) according to the manufacturer’s extremely high levels of GABA in plant cells. J. Exp. Bot. 58: 2699–2707. Akihiro, T., Mizuno, K. and Fujimura, T. (2005) Gene expression of ADP- instructions. glucose pyrophosphorylase and starch contents in rice cultured cells are cooperatively regulated by sucrose and ABA. Plant Cell Physiol. 46: Western blot analysis 937–946. A15mg aliquot of tomato fruit proteins was separated by Akihiro, T., Umezawa, T., Ueki, C., Lobna, B.M., Mizuno, K., Ohta, M. SDS–PAGE on polyacrylamide gels [10–20% (w/v) gradient gel]; and Fujimura, T. (2006) Genome wide cDNA-AFLP analysis of genes rapidly induced by combined sucrose and ABA treatment in rice cultured e-PAGEL (Atto Co., Bunkyo, Tokyo, Japan). The proteins were cells. FEBS Lett. 580: 5947–5952. transferred to polyvinylidene difluoride (PVDF) membranes Aoki, H., Furuya, Y., Endo, Y. and Fujimoto, K. (2003) Effect of (Hybond-P, GE Healthcare, Buckinghamshire, UK). The mem- -aminobutyric acid-enriched tempeh-like fermented soybean (GABA- Downloaded from https://academic.oup.com/pcp/article/49/9/1378/1885677 by guest on 28 September 2021 brane was initially blocked for 1 h at room temperature with 5% Tempeh) on the blood pressure of spontaneously hypertensive rats. Blocking One-P solution (Nakarai Tesque, Inc., Chyu-kyo, Kyoto, Biosci. Biotechnol. Biochem. 67: 1806–1808. Japan) in TBS-T [10 mM Tris-buffered saline with 0.05% (v/w) Arazi, T., Baum, G., Snedden, W.A., Shelp, B.J. and Fromm, H. (1995) Tween-20]. The membrane was then incubated for 1 h at 48C Molecular and biochemical analysis of calmodulin interactions with the with the primary antibody. Subsequent to rinsing with TBS-T, calmodulin-binding domain of plant glutamate decarboxylase. Plant the membrane was incubated for 1 h at room temperature Physiol. 108: 551–561. Baum, G., Chen, Y., Arazi, T., Takatsuji, H. and Fromm, H. (1993) A plant with a horseradish peroxidase-labeled goat anti-rabbit secondary glutamate decarboxylase containing a calmodulin binding domain. antibody (Roche Diagnostics, Vienna, Austria) diluted 1 : 2,000. Cloning, sequence, and functional analysis. J. Biol. Chem. 268: The immunoblots were developed using chemiluminescence 19610–19617. (ChemilumiOne; Nakarai Tesque, Inc., Chyu-kyo, Kyoto, Japan) Beutler, H.O. (1989) Succinate. In Methods of Enzymatic Analysis, 3rd edn, and visualized with the aid of a digital imaging system (LAS-1000; Vol. VII. Edited by Bergmeyer, H.U. pp. 25-33. VCH publishers, Fujifilm, Minatoku, Tokyo, Japan). The primary antibodies that Cambridge. were used in the study were mouse anti-rice OsGAD1 and Bouche` , N., Fait, A., Bouchez, D., Møller, S.G. and Fromm, H. (2003) OsGAD2 (Akama and Takaiwa 2007). Mitochondrial succinic-semialdehyde dehydrogenase of the -aminobutyrate shunt is required to restrict levels of reactive oxygen intermediates in plants. Proc. Natl Acad. Sci. USA 100: 6843–6848. Supplementary material Bouche` , N. and Fromm, H. (2004) GABA in plants; just a metabolite? Supplementary material mentioned in the article is Trends Plant Sci. 9: 110–115. available to online subscribers at the journal website Breitkreuz, K.E., Shelp, B.J., Fischer, W.N., Schwacke, R. and Rentsch, D. www.pcp.oxfordjournals.org. (1999) Identification and characterization of GABA, proline and quaternary ammonium compound transporters from Arabidopsis thali- ana. FEBS Lett. 450: 280–284. Funding Carrari, F., Baxter, C., Usadel, B., Urbanczyk-Wochniak, E., Zanor, M., et al. (2006) Integrated analysis of metabolite and transcript levels reveals the metabolic shifts that underlie tomato fruit development and highlight The Research and Development Program for New regulatory aspects of metabolic network behavior. Plant Physiol. 142: Bio-industry Initiatives (BRAIN). 1380–1396. Carrari, F. and Fernie, A.R. (2006) Metabolic regulation underlying tomato fruit development. J. Exp. Bot. 57: 1883–1897. Acknowledgments Cauwenberghe, O.R.V., Makhmoudova, A., McLean, M.D., Clark, S.M. and Shelp, B.J. (2002) Plant pyruvate-dependent gamma-aminobutyrate transaminase: identification of an Arabidopsis cDNA and its expression The authors thank Mr. Syuji Inai of Nippon Del Monte in Escherichia coli. Can. J. Bot. 80: 933–941. Corporation for providing the ‘DG03-9’ seeds. We also thank Elliott, C.A.K. and Hobbiger, F. (1959) Gamma aminobutyric acid: Dr. Christophe Rothan (INRA), Ms. Izumi Oshima, Mr. Kenji circulatory and respiratory effects in different species: re-investigation Higuchi, Mr, Noriyuki Hirooka and all laboratory members of the anti-strychine action in mice. J. Physiol. 146: 70–84. for help, advice and discussion. ‘Micro-Tom’ seeds (ID: Fait, A., Fromm, H., Walter, D., Galili, G. and Fernie, A.R. (2008) TOMJPF00001) were provided by the Gene Research Center, Highway or byway: the metabolic role of the GABA shunt in plants. University of Tsukuba, through the National Bio-Resource Project Trends Plant Sci. 13: 14–19. Gallego, P.P., Whotton, L., Picton, S., Grierson, D. and Gray, J.E. (1995) A (NBRP) of the MEXT, Japan. role for glutamate decarboxylase during tomato ripening: the characterization of a cDNA encoding a putative glutamate decarboxylase with References a calmodulin-binding site. Plant. Mol. Biol. 27: 1143–1151. Hayakawa, K., Kimura, M., Kasaha, K., Matsumoto, K., Sansawa, H. and Yamori, Y. (2004) Effect of -aminobutyric acid-enriched dairy produc- Abe, Y., Umemura, S., Sugimoto, K., Hirawa, N., Kato, Y., tion on the blood pressure of spontaneously hypertensive and normo- Yokoyama, N., Yokoyama, T., Iwai, J. and Ishii, M. (1995) Effect of tensice Wistar–Kyoto rats. Br. J. Nutr. 92: 411–417. green tea rich in -aminobutyric acid on blood pressure of Dahl salt Inaba, A., Yamamoto, T., Ito, T. and Nakamura, R. (1980) Changes in the sensitive rats. Am. J. Hypertens. 92: 411–417. concentrations of free amino acids and soluble nucleotide in attached and Abdou, A.M., Higashiguchi, S., Horie, K., Kim, M., Hatta, H. and detached tomato fruits during ripening. J. Jpn. Soc. Hort. Sci. 49: Yokogoshi, H. (2006) Relaxation and immunity enhancement effects of 435–441. -aminobutyric acid (GABA) administration in humans. BioFactors 26: Inoue, K., Shirai, T., Ochiai, H., Kasao, M., Hayakawa, K., Kimura, M. 201–208. and Sansawa, H. (2003) Blood-pressure-lowering effect of a novel GABA-TK is crucial for GABA catabolism in tomato fruits 1389

fermented milk containing gamma-aminobutyric acid (GABA) in mild Reggiani, R., Cantu´ , C.A., Brambilla, I. and Betani, A. (1988) hypertensives. Eur. J. Clin. Nutr. 57: 490–495. Accumulation and interconversion of amino acids in rice roots under Jaarma, M. (1969) Comparison of chemical changes in potato tubers anoxia. Plant Cell Physiol. 29: 981–987. induced by -irradiation and by chemical treatment. Acta Chim. Scand. Rolin, D., Baldet, P., Just, D., Chevalier, C., Biran, M. and Raymond, P. 23: 3435–3442. (2000) NMR study of low subcellular pH during the development of Jakoby, W.B. (1962) Enzymes of gamma-aminobutyrate metabolism cherry tomato fruit. Aust. J. Plant Physiol. 27: 61–69. (bacterial). Methods Enzymol. 5: 765–778. Roessner-Tunali, U., Hegemann, B., Lytovchenko, A., Carrari, F., Kajimoto, O., Hirata, H., Nakagawa, S., Kajimoto, Y., Hayakawa, K. and Bruedigam, C., Granot, D. and Fernie, A.R. (2003) Metabolic profiling Kimura, M. (2004) Hypotensive effect of fermented milk containing of transgenic tomato plants overexpressing hexokinase reveals that the -aminobutyric acid (GABA) in subjects with high normal blood influence of hexose phosphorylation diminishes during fruit development. pressure (in Japanese). Nippon Shokuhin Kagaku Kogaku Kaishi 51: Plant Physiol. 133: 84–99. 79–86. Ronen, G., Carmel-Goren, L., Zamir, D. and Hirschberg, J. (2000) An Kusakabe, H., Midorikawa, Y., Fujishima, T., Kuninaka, A. and alternative pathway to b-carotene formation in plant chromoplasts Downloaded from https://academic.oup.com/pcp/article/49/9/1378/1885677 by guest on 28 September 2021 Yoshino, H. (1983) Purification and properties of a new enzyme, L- discovered by map-based cloning of Beta and old-gold color mutations in glutamate oxidase, from Streptomyces sp. X-119-6 grown on wheat bran. tomato. Proc. Natl Acad. Sci. USA 97: 11102–11107. Agric. Biol. Chem. 47: 1323–1328. Saito, T., Matsukura, C., Sugiyama, M., Watahiki, A., Ohshima, I., et al. Lane, T.R. and Stiller, M. (1970) Glutamic acid decarboxylation in (2008) Screening for -aminobutyric acid (GABA)-rich tomato varieties. Chlorella. Plant Physiol. 45: 558–562. J. Jpn. Soc. Hort. Sci. 77: 242–250. MacGregor, K.B., Shelp, B.J., Peiris, S.E. and Bown, A.W. (2003) Schroeder, A, Mu¨ eller, O., Stocker, S., Salowsky, R., Leiber, M., Overexpression of glutamate decarboxylase in transgenic tobacco Gassmann, M., Lightfoot, S., Menzel, W., Granzow, M. and Ragg, T. deters feeding by phytophagous insect larvae. J. Chem. Ecol. 29: (2006) The RIN: an RNA integrity number for assigning integrity values 2177–2182. to RNA measurements. BMC. Mol. Biol. 7: 3. Mattoo, K., Sobolve, A.P., Neelam, A., Goyal, R.K., Handa, A.K. and Shelp, B.J., Walton, C.S., Snedden, W.A., Tuin, L.G., Oresnik, I.J. and Layzell, D.B. (1995) Gaba shunt in developing soybean seed is associated Segre, L. (2006) Nuclear magnetic resonance spectroscopy-based with hypoxia. Physiol. Plant. 94: 219–228. metabolite profiling of transgenic tomato fruit engineered to accumulate Snedden, W.A., Arazi, T., Fromm, H. and Shelp, B.J. (1995) Calcium/ spermidine and spermine reveals enhanced anabolic and nitrogen–carbon calmodulin activation of soybean glutamate decarboxylase. Plant Physiol. interactions. Plant Physiol. 142: 1759–1770. 108: 543–549. Matumoto, Y., Ohno, K. and Hiraoka, Y. (1997) Studies on the utilization Steward, F.C., Thompson, J.F. and Dent, C.E. (1949) -Aminobutyric acid: of functional food materials containing high levels of gamma-aminobu- a constituent of potato tubers? Science 110: 439–440. tyric acid (Part 1). Ehime Kougi Kenkyu Houkoku 35: 97–100 Streeter, J.G. and Thompson, J.F. (1972) In vivo and in vitro studies on (in Japanese). -aminobutyric acid metabolism with the radish plant (Raphanus sativus McLean, M.D., Yevtushenko, D.P., Deschene, D., Van, Cauwenberghe. O. L.). Plant Physiol. 49: 579–584. R., Makhmoudova, A, Potter, J.W., Bown, A.W. and Shelp, B.J. (2003) Studart-Guimara˜es, C., Fait, A., Nunes-Nesi, A., Carrari, F., Usadel, B. Overexpression of glutamate decarboxylase in transgenic tobacco plants and Fernie, A.R. (2007) Reduced expression of succinyl-coenzyme A confers resistance to the northern root-knot nematode. Mol. Breed. 11: ligase can be compensated for by up-regulation of the gamma- 277–285. aminobutyrate shunt in illuminated tomato leaves. Plant Physiol. 145: Mounet, F., Lemaire-Chamley, M., Maucourt, M., Cabasson, C., 626–639. Giraude, J.L., et al. (2007) Quantitative metabolic profiles of tomato Takahashi, H., Sumi, M. and Koshino, F. (1961) Effect of gamma- flesh and seeds during fruit development: complementary analysis with aminobutyric acid (GABA) on normotensive or hypertensive rats and ANN and PCA. Metabolomics 3: 273–288. men. Jpn. J. Physiol. 11: 89–95. Mu¨ eller, L.A., Tanksley, S.D., Giovannoni, J.J., Eck, J.V., Stack, S., et al. Tanase, S., Kojima, H. and Morino, Y. (1979) Pyridoxal 50-phosphate (2005) The tomato sequencing project, the first cornerstone of binding site of pig heart alanine aminotransferase. Biochemistry 18: the international Solanaceae project (SOL). Comp. Funct. Genom. 6: 3002–3007. 153–158. Torre, F.D.L., Santis, L.D., Suarez, M.F., Crespillo, R. and Canovas, F.M. Noguchi, T., Nakamura, K., Nagai, T., Katsuda, S. and Koga, H. (2007) (2006) Identification and functional analysis of a prokaryotic-type Antihypertensive effects of GABA-enriched potato snacks in sponta- aspartate aminotransferase: implications for plant amino acid metabo- neously hypertensive rats. Nippon Shokuhin Kagaku Kogaku Kaishi 54: lism. Plant J. 46: 414–425. 75–81. Wallace, W., Secor, J. and Schrader, L.E. (1984) Rapid accumulation of Palanivelu, R., Brass, L., Edlund, A.F. and Preuss, D. (2003) Pollen tube -aminobutyric acid and alanine in soybean leaves in response to an growth and guidance is regulated by POP2,anArabidopsis gene that abrupt transfer to lower temperature, darkness, or meachnical manipula- controls GABA levels. Cell. 114: 47–59. tion. Plant Physiol. 75: 170–175.

(Received May 17, 2008; Accepted August 6, 2008)