**1. Introduction**

Chestnut (genus *Castanea*; family Fagaceae), which is a major nut crop in East Asia and Southern Europe, is unique among temperate nut crops because its seed is starchy rather than oily, making it ecologically and economically valuable [1,2]. The production of American chestnut, however, has been severely a ffected in North America by chestnut blight due to *Cryphonectria parasitica* [3,4]. As a perennial crop, chestnut is not only used as an important starch-based food product consumed by people living in rural areas [5,6], but also as a potential functional food because it is a rich source of bioactive compounds, including phenolics [5]. Additionally, chestnut seeds contain a considerable

abundance of γ-aminobutyric acid (GABA) [7], which is the key component of the GABA shunt crucial for carbon and nitrogen metabolism in plants [8]. Previous studies have also identified GABA-enriched functional foods [9,10]. However, there is still relatively little information regarding GABA functions affecting chestnut seed activities, including germination.

Seeds, including orthodox and recalcitrant seeds [11], play a major role in agriculture, serving as food and feed, as well as plant propagation units. Seed germination is influenced by internal metabolic changes [12,13] and a complex process in which starch-degrading α-amylase and various proteases are activated to decrease the total dry matter content [14]. During germination, starch and proteins are degraded into smaller molecules, such as soluble sugars and free amino acids [15], resulting in a significant increase in the free amino acid content [16]. Previous studies proved that GABA levels increase during the germination of barnyard millet [16] and wheat [17,18] seeds under normal conditions, as well as wheat seeds under saline conditions [18]. The accumulation of endogenous GABA in dry seeds facilitates early metabolic reorganization during germination [19]. Moreover, exogenous GABA reportedly a ffects the seed germination process in barley [20] and *Haloxylon ammodendron* [21] under normal conditions. It also modulates respiration during the germination of *H. ammodendron* seeds [21] and regulates H2O2 production in *Caragana intermedia* [22] and poplar [23] in response to salt stress. The application of exogenous GABA can also mitigate salt-mediated damages by enhancing starch catabolism and the use of sugars and amino acids [24] or by enhancing the antioxidant system to induce the accumulation of phenolic compounds during seed germination [25]. Additionally, an exogenous GABA treatment can delay the loss of titratable acidity and malate, thereby maintaining the quality of stored apple fruits [26]. These studies confirmed that GABA can a ffect seed germination or fruit storage by altering the metabolism of carbon and nitrogen, as well as reactive oxygen species (ROS).

Chestnut seeds are recalcitrant and exhibit dormancy [27,28], which di fferentiates them from most other recalcitrant seeds [29]. This dormancy increases the shelf-life of the seeds used as food and enables the seeds to survive unfavorable winter conditions so they can germinate in the following spring. The normal germination of chestnuts is vital for their use as rootstock seedlings during grafting [30]. Earlier reports indicated that the germination rates of 56 selected types and/or cultivars of European chestnut seeds were 17.6–86.6% [31] and 51.6–97.3% [32] in two regions in Turkey, depending on their dormancy-breaking time due to stratification temperatures [28,31,32]. Similar to orthodox seeds, the dormancy of recalcitrant seeds appears to be induced by abscisic acid (ABA) [29,33], which is abundant in chestnut seed coats [34]. In contrast, the contents of the ABA antagonist GA3 are high in embryos and cotyledons [34]. Thus, di fferent exogenous treatments have been used to increase germination rates, including an H2O2 treatment [35] and the application of GA3 [30,34]. Notably, an earlier investigation demonstrated that GABA is a major amino component associated with the high accumulation of several amino acids during chestnut seed germination [36], suggesting there is a close relationship between GABA and the recalcitrant chestnut seeds.

The e ffects of GABA on chestnut seed germination remain unclear. Thus, on the basis of our previous findings that GABA regulates stress responses [22,23] and adventitious root development [37], in this study, we treated Chinese chestnut (*Castanea mollissima*) cultivar "Yanshanzaofeng" seeds with GABA and vigabatrin (VGB; a specific GABA transaminase inhibitor) to investigate seed germination changes, as well as the central carbon/nitrogen metabolic activities, which demonstrated that both treatments inhibited chestnut seed germination and might by altering the balance of carbon and nitrogen metabolism, which would provide a better understanding for elucidating the role of GABA during the storage and germination of recalcitrant chestnut seeds.

#### **2. Materials and Methods**

#### *2.1. Plant Materials and Treatments*

Chinese chestnut (*Castanea mollissima*) cultivar "Yanshanzaofeng" seeds were collected from healthy trees growing in the core chestnut-producing region of Qianxi county, Tangshan city, Hebei province, China, after which they were stored at 0–1 ◦C until further use. Relatively uniform chestnut seeds were washed five times with sterile water and air-dried. The seeds were then soaked in sterile water (control/CK), 10 mM GABA (lab use only; Sigma-Aldrich, St. Louis, MO, USA), or 100 μM vigabatrin (VGB; lab use only; MedChem Express, Monmouth Junction, NJ, USA) for 15 h at 25 ◦C before they were placed evenly on a tray (33.5 cm × 26 cm × 11 cm) containing sterilized sand. A germination test was conducted in a climate chamber with a 16 h light (25 ◦C): 8 h dark (20 ◦C) cycle and 60% relative humidity. The seed germination rate was calculated, and the root length was measured after 2, 5, 8, 15, and 30 days. Five seeds per replicate were collected at 15 h before the treatment (t0; i.e., the time-point when the seed imbibition was initiated) and at 0, 48, and 120 h after sowing. After discarding the seed coats, the kernels were ground into a powder and stored at −80 ◦C for the subsequent physiological measurements. Each treatment was completed with three replicates, each comprising 50 seeds.

#### *2.2. Calculation of the Seed Germination Rate*

Seed radicle protrusion was used as the criterion for judging germination. Germinated seeds were counted at the designated treatment times, after which the germination rate was calculated based on the ratio of the number of germinated seeds and the total number of seeds in each treatment.

#### *2.3. Measurement of Reactive Oxygen Species*

The H2O2 content was measured as previously described [38]. Briefly, 0.1 g fresh powder were dissolved in 1 mL acetone and then thoroughly mixed on ice. After a centrifugation (8000× *g* for 10 min at 4 ◦C), the supernatant was mixed with a titanium sulfate solution and concentrated ammonia. After another centrifugation (4000× *g* for 10 min at 25 ◦C), the sediment was dissolved in concentrated sulfuric acid and incubated at room temperature for 5 min. The absorbance of the reaction solution was measured at 415 nm, and the H2O2 content was recorded as μmol/g fresh weight (FW).

#### *2.4. Measurement of Soluble Sugars and Starch*

The total soluble sugar content was measured with a commercial assay kit (Comin Biotechnology, Suzhou, China) based on the anthrone-sulfuric acid method as previously described [39]. Briefly, 0.1 g fresh powder were dissolved in 1 mL sterile water and heated for 10 min at 95 ◦C. After a centrifugation (8000× *g* for 10 min at 25 ◦C), the diluted supernatant was added to a reaction mixture comprising anthranone in ethyl acetate and a concentrated sulfuric acid solution and then heated for 10 min at 95 ◦C. The absorbance of the reaction solution was measured at 620 nm, and the soluble sugar content was recorded as % FW.

The starch content was quantified with a commercial assay kit (Comin Biotechnology) based on a previously described method [40]. Briefly, 0.1 g fresh powder were dissolved in 1 mL ethanol solution and heated at 80 ◦C for 30 min. The solution was centrifuged (3000× *g* for 5 min at 25 ◦C), and the sediment was dissolved in 0.5 mL sterile water and heated at 95 ◦C for 15 min. After cooling, 0.35 mL perchloric acid and 0.85 mL sterile water were added, and the resulting solution was thoroughly mixed and then centrifuged (3000× *g* for 10 min at 25 ◦C). A 50 μL aliquot of the solution was mixed with 250 μL reaction mixture (3.75 mL anthranone solution and 21.25 mL concentrated sulfuric acid) and then heated at 80 ◦C for 10 min. The absorbance of the reaction solution was measured at 620 nm, and the starch content was recorded as % FW.

#### *2.5. Measurement of Soluble Proteins and Total Amino Acids*

The total soluble protein content was measured with a Bradford Protein Assay Kit (Sangon Biotech, Shanghai, China) based on the Coomassie brilliant blue method. Briefly, 0.1 g fresh powder were suspended in 5 mL phosphate buffer (pH 7.2) and centrifuged (8000× *g* for 10 min at 4 ◦C). The diluted supernatant was mixed with Coomassie brilliant blue R-250, after which the absorbance was measured at 595 nm. The total soluble protein content was recorded as mg/g FW.

The total amino acid content was measured with a commercial assay kit (Comin Biotechnology) based on a previously described method [41]. Briefly, 0.1 g fresh powder were dissolved in 1 mL glacial acetic acid, mixed thoroughly, and boiled for 15 min. After a centrifugation (8000× *g* for 5 min at 4 ◦C), the supernatant was added to a reaction solution consisting of 100 μL sodium acetate-glacial acetic acid, 100 μL ninhydrin, and 10 μL ascorbic acid. After boiling for 15 min, the absorbance of the reaction solution was measured at 570 nm, and the total amino acid content was recorded as mg/g FW.

#### *2.6. Measurement of Organic Acids and Amino Acids*

To analyze the organic acid and amino acid contents, 0.4 g fresh powder were suspended in 2 mL sterile water and boiled at 100 ◦C for 90 min. After a centrifugation, the supernatant was filtered (0.45 μm pore size). The organic acid and amino acid contents of the filtrate were analyzed by high-performance liquid chromatography with the 2695 Separations Module (Waters, Milford, MA, USA) as previously described [37,42]. All assays were completed with three biological replicates per treatment.

Organic acids: The citrate, succinate, and malate contents of the prepared extract were measured at 214 nm with the Waters 2487 UV detector, a Thermo ODS HYPERSIL (4.6 mm × 200 mm) chromatographic column, and 3% CH3OH + 97% H2O as the mobile phase (flow rate, 0.8 mL/min). The organic acid contents were recorded as mg/g FW.

Amino acids: The prepared extract was analyzed with a pre-column derivatized with the AccQ Fluor reagent. The mobile phase consisted of 140 mM NaAc solution (containing 17 mM triethylamine, pH 4.95) and a 60% aqueous acetonitrile solution. A Waters 2475 fluorescence detector and an AccQ Tag amino acid analysis column (3.9 mm × 150 mm) were used. The separate amino acids were detected at excitation and emission wavelengths of 250 and 395 nm, respectively. The amino acid contents were recorded as μg/g FW.

GABA: The prepared extract was mixed with acetonitrile containing 1% 2,4-dinitrofluorobenzene, after which an equal volume of NaHCO3 (pH 9) buffer was added, and the solution was incubated for 1 h at 60 ◦C. The solution contents were separated on a C18 SunFire column (4.6 mm × 250 mm), and the eluted products were detected with the Waters 2487 UV detector. The column temperature was maintained at 35 ◦C, and the mobile phase comprised phosphate buffer (pH 7), water, and acetonitrile, with a flow rate of 0.1 mL/min. The separated GABA was detected at 360 nm. The GABA content was recorded as μg/g FW.

## *2.7. Statistical Analysis*

Data were compared and analyzed with ANOVA (analysis of variance), and multiple comparisons were made with SPSS 16.0 (SPSS, Chicago, IL, USA). Differences were scored as significant at the *p* < 0.05 or *p* < 0.01 levels. A principal component analysis (PCA) was performed with the command prcomp() in R (http://www.r-project.org/), as previously described [43,44].
