During the postharvest storage of horticultural commodities, temperature and/or atmospheric conditions are adjusted so that ethylene production and respiratory rates are reduced, and ripening/senescence is delayed, resulting in the preservation of nutritional and sensory quality (Table 1). LT storage of mulberry leaves in air preserves color, while enhancing GABA accumulation (Li E. et al., 2018), as is often found in the vegetative organs of many plant species (Shelp et al., 2021). This result might be attributed to the elevated activity of glutamate (Glu) decarboxylase (GAD) and limited activity of the catabolic enzyme GABA transaminase (GABA-T) (Figure 1). In contrast, the LT storage of zucchini fruit promotes GABA catabolism, without causing its accumulation (Palma et al., 2014). The loss of GABA and the increase in GABA-T activity is more substantive in a chilling-tolerant cultivar than a chilling-sensitive cultivar, suggesting that GABA catabolism replenishes the tricarboxylic acid cycle (TCAC) to generate reducing equivalents and energy that could alleviate oxidative damage (Shelp et al., 2021). The authors have interpreted the increase in diamine oxidase (DAO) activity and putrescine (Put) accumulation as support for the involvement of Put catabolism in GABA production and the alleviation of chilling injury (Figure 1). Conditioning at 15°C prior to LT storage improves the tolerance in the chilling-sensitive zucchini cultivar by decreasing the GABA level and increasing the ATP level and activities of enzymatic antioxidants (peroxidase, catalase) (Carvajal et al., 2015). Improved chilling tolerance in peaches by hot water treatment prior to LT storage is associated with membrane stability (as indicated by less electrolyte leakage and lower malondialdehyde accumulation), and the maintenance of high levels of amino acid (including GABA and proline), polyamines (PAs) and radical scavenging capacity (phenols) (Wang L. et al., 2021).

Anoxia preserves the quality of drying green tea leaves and of soybean sprouts stored at room temperature (RT) and promotes GABA accumulation and the diversion of succinic semialdehyde (SSA) from succinate to γ-hydroxybutyrate (GHB) (Allan et al., 2003; Table 1 and Figure 1). Several mechanisms could account for the accumulation of GABA: calmodulin (CaM) activation of CsGAD1; elevated expression of CsGAD2; oxidation of Put/proline; and, feedback inhibition of GABA transaminase (CsGABA-T) activity (Mei et al., 2016; Liao et al., 2017; Shelp et al., 2021; Table 1). Complete inhibition of DAO activity by aminoguanidine (4–11 h of treatment) suggests that about 25% of the GABA is derived from the PA degradation pathways (Liao et al., 2017; Figure 1). However, this interpretation can be challenged. Based upon the increasing accumulation of Put with aminoguanidine over the same time period, we estimate that Put degradation would account for only 3% of the anoxia-induced rate of GABA accumulation. Notably, the spermidine (Spd) pool also decreases at an estimated rate of approximately 3% of the rate for GABA accumulation, suggesting that the terminal oxidation of Spd can substitute for the terminal oxidation of Put (Shelp et al., 2012b). This re-assessment of the published data is consistent with our recent interpretation of the ΔGABA/ΔPut stoichiometry published for wheat roots treated simultaneously with salinity and aminoguanidine (Shelp et al., 2021). While increasing DAO activity seems contrary to our interpretation of the metabolite data, it could reflect an “anticipation response” to the return to normoxia, as proposed for alanine transaminase and glutamate dehydrogenase (Limami et al., 2008).

The quality of green tea leaves at RT and of broccoli florets at LT is preserved under anaerobic conditions imposed by either anoxia or elevated CO₂ (Hansen et al., 2001; Chen et al., 2018; Table 1). However, the accumulation of GABA and succinate, and the depletion of Glu is more rapid with CO₂ than with N₂, whereas the accumulation of alanine is faster with N₂. There is greater Put and NADH accumulation, and less NADPH, citrate, 2-oxoglutarate (OG) and fumarate accumulation with CO₂ than air. Thus, the GABA shunt is more active with CO₂, but the inhibition of the TCAC and mitochondrial electron transport chain (mETC) occurs more quickly with N₂ (Chen et al., 2018). While the storage of mushrooms with 100% CO₂ at LT also stimulates the production of Glu-derived GABA, storage with N₂ stimulates the production of both Glu-and PA-derived GABA (Chen et al., 2020), perhaps due in part to protein degradation. Notably, AbGAD, unlike most plant GADs, does not possess a CaM-binding domain, and therefore its activity is likely to be stimulated by cytosolic acidification only.

Elevated CO₂ in air at LT improves chilling tolerance in cherimoya fruit (Merodio et al., 1998), and delays ripening/senescence in tomato and strawberry fruits (Deewatthanawong et al., 2010a,b; Blanch et al., 2012; Li D. et al., 2018; Table 1). These positive outcomes are typically accompanied by the accumulation of GABA and occasionally PAs, as well as limited flux of GABA-carbon through the GABA shunt into the TCAC and the mETC. Elevated CO₂, in combination with low O₂, delays the ripening/senescence of tomato fruit stored at RT (Makino et al., 2008; Mae et al., 2012), and longan (Zhou et al., 2016), strawberry (Pott et al., 2020) and “Empire” apple (Deewatthanawong and Watkins, 2010; Trobacher et al., 2013a; Deyman et al., 2014a,b; Brikis et al., 2018) fruits stored at LT. These findings have been attributed to the elevated generation of GABA from Glu, rather than PAs, and enhanced flux of GABA-carbon through the GABA shunt into a non-cyclic TCAC for generation of ATP (Shelp et al., 2012b; Brikis et al., 2018). Wang C. et al. (2014) previously suggested that Glu-derived GABA accumulation in melon roots can alleviate hypoxia damage by accelerating PA biosynthesis and conversion, as well as preventing PA degradation.

Some pome fruit are particularly sensitive to LT, CA storage (e.g., “Honeycrisp” apples and “Cold Snap” pears) (Chiu et al., 2015; Lum et al., 2016a,2017; Table 1). In these cases, a dramatic increase in the GABA level coincides with CA-or senescence-related injury and is likely due to the disruption of cellular compartmentation and the release of acidic vacuolar contents to the cytosol (Bown and Shelp, 2006). Interestingly, conditioning of “Honeycrisp” apples at 10°C improves the resistance to CA-related injury, decreases the GABA level and increases the ratios of NAD(P)H/NAD(P) ⁺ (Lum et al., 2016a).

Overall, these studies indicate that LT, CA-mediated improvements in the postharvest marketability of horticultural products is generally associated with the promotion of GABA biosynthesis and GABA shunt activity, with or without the accumulation of GABA. The onset of CA-or senescence-related injury during prolonged storage may also be associated with the accumulation of GABA. Discrepancies in data from the various studies might be explained by: pretreatment and conditioning of plant materials prior to storage; the use of different cultivars and single time point determinations, rather than time courses; excessive handling or wounding of plant materials prior to metabolite extraction; and, the use of non-saturating levels of Glu and inhibitory levels of GABA and pyruvate, respectively, in in vitro assays of GAD and GABA-TP activities [for examples, compare Zhou et al. (2016), Li D. et al. (2018), and Chen et al. (2020) with Snedden et al. (1995), Van Cauwenberghe and Shelp (1999), Clark et al. (2009a), and Shelp et al. (2012a)]. Interpretation could vary somewhat because: (i) it is not possible to directly assess the Ca²⁺/CaM stimulation of GAD activity in situ; (ii) the existence of a 2-OG-dependent plant GABA-T is questionable [for examples, compare Makino et al. (2008) and Deewatthanawong et al. (2010b) with Clark et al. (2009b), Koike et al. (2013), Trobacher et al. (2013a), Shimajiri et al. (2013), and Shelp et al. (2021)]; (iii) the expression of a gene does not establish that the encoded protein is operational; (iv) the understanding of precursor/product relations and flux is often incomplete (e.g., pool sizes alone do not indicate flux; Put accumulation does not establish greater contribution than Glu to GABA generation) (Shelp et al., 2012a); and, (v) there is often a failure to consider the importance of multiple isoforms of the GABA pathway enzymes (Shelp et al., 2012c,2021).

Gene sequences for the key steps in GABA metabolism in apple fruit have been identified, allowing elucidation of the biochemical properties and subcellular location of multiple isoforms of the encoded proteins (Table 2 and Figure 1). Three cytosolic GADs are present, but unlike MdGAD3, MdGAD1,2 are Ca²⁺/CaM-dependent and more sensitive to pH (Trobacher et al., 2013b). There are also two mitochondrial pyruvate/glyoxylate-dependent GABA transaminases (MdGABA-Ts, designated as GABA-TP), two mitochondrial NAD⁺-dependent SSADHs (MdSSADH1,2 or MdALDH5F1,2), and two NADPH-dependent glyoxylate/succinic semialdehyde reductases (MdGLYR1,2 or MdSSR1,2) with different subcellular locations (Trobacher et al., 2013a; Brikis et al., 2017, 2018; Zarei et al., 2017). MdGLYR1 is cytosolic, whereas MdGLYR2 is both plastidial and mitochondrial. Two of the six apple fruit FAD-dependent polyamine oxidases (MdPAO2,4) are peroxisomal and likely catalyze the back-conversion of Spm and Spd to Spd and Put, respectively (Brikis et al., 2018). Three of the five MdCuAOs identified are peroxisomal (MdCuAO1,4-5), but only one of these, MdCuAO1, has been shown to exclusively utilize diamines (diaminopropane, Put and cadaverine) as substrates (Zarei et al., 2015a; Brikis et al., 2018). A candidate plastidial diamine oxidase activity has not yet been identified. Two NAD⁺-dependent 4-aminobutanal dehydrogenases (ABALDH) exist in apple fruit (MdALDH10A8,9 or MdAMADH1,2): one is peroxisomal and the other plastidial (Zarei et al., 2015b,2016; Brikis et al., 2018).

The temporal patterns of specific metabolites have been compared to the expression of genes encoding the most biochemically relevant proteins in intact “Empire” apple fruit stored under LT and low O₂ with ambient or elevated CO₂ (0°C, 2.5 kPa O₂ and 0.03 or 5 kPa CO₂) (Brikis et al., 2018; Table 2). Five kPa CO₂ is known to elicit symptoms of external, but not internal, CO₂-induced injury in this cultivar within 16 weeks (Deyman et al., 2014b). Under LT, low-O₂ and ambient-CO₂ storage, there is a transient increase in amino acid availability, including Glu, early in the storage period (2–4 weeks), probably a reflection of protein hydrolysis (Brikis et al., 2018). This is accompanied by a rapid peak in the expression of alanine transaminase (MdAla-T), a marker of hypoxia (Cukrov et al., 2016), as well as in the pool of alanine (2–4 weeks), which decline slowly to a steady basal level (from 8 to 16 weeks) (Brikis et al., 2018). A rapid accumulation of GABA is also transient (2–4 weeks), but the pool size is approximately 60% of that for alanine, suggesting that the alanine is derived from both Ala-T and GABA-TP reactions. Notably, MdGAD1 expression increases linearly up to 12 weeks and then remains steady, whereas MdGAD2 expression increases up to only 4 weeks and then decreases, and MdGAD3 expression decreases over the storage period. Succinate does not accumulate, but the burst of GABA is followed by a much smaller transient increase in GHB (Figure 1). Nevertheless, MdSSADH1 expression increases up to 8 weeks and then slowly declines, whereas MdSSADH2 expression decreases over the storage period. The expression of MdSSR1 is transiently increased, peaking at 4–8 weeks, and may be correlated with GHB. While Put, Spd and spermine (Spm) represent potential precursors for GABA, their levels are only 1–5% of that for GABA. Furthermore, the Put level declines only slightly with the increase in GABA, while Spd accumulates slightly and Spm dramatically declines. The expression of MdPAO2,4, MdCuAO1 and MdALDH10A8,9 rapidly increases, peaking after 8, 2 and 4–8 weeks, respectively, whereas the expression of MdCuAO4-5 is transiently increased, peaking at 4 weeks.

With LT, low-O₂ and elevated-CO₂ storage, a pronounced transient peak of GABA is accompanied by a strong transient peak of succinate, and smaller transient peaks of GHB, Put, Spd and Spm (Brikis et al., 2018). With prolonged storage, only GABA and GHB exhibit subsequent increases. These changes are accompanied by minor, yet significant, increases in the expression of MdGAD1, MdCuAO1 and MdALDH10A8,9. Thus, the GABA pattern might be interpreted as a CO₂-induced shift from Glu/CaM-mediated stimulation/activation of GAD activity to H⁺-mediated stimulation of GAD activity (Trobacher et al., 2013b; Brikis et al., 2018). The patterns for succinate, GHB and Put might be explained by a combination of: elevated GABA production; differential effects of shifting redox balance on the activities of SSADH, TCAC, SSR, and ABALDH; and, limiting O₂ availability for DAO activity in bulky apple fruit (Shelp et al., 2012b; Brikis et al., 2018). Based on changing pool sizes, we can estimate the maximum rates of GABA and succinate synthesis to be ∼50 nmol g^–1 fresh mass (FM) wk^–1, and the maximum rates of GHB synthesis and Put/Spd depletion to be ∼0.2 and ∼1.5 nmol g^–1 FM wk^–1, respectively. Thus, the terminal oxidation of PAs and the direct decarboxylation of Glu can account for approximately 3 and 97%, respectively, of GABA synthesis. Moreover, only 3% of the SSA is diverted from succinate to GHB production. Overall, this study suggests that both genetic and biochemical mechanisms are involved in the metabolism of GABA in apple fruit stored under LT, CA conditions.

Han et al. (2018) have monitored the expression of the GABA shunt enzymes and the levels of important metabolites in “Cripps Pink” apple fruit stored at RT in air for 70 days. The MdGADs exhibit different expression patterns, with MdGAD1 expression increasing gradually with time, MdGAD2 expression increasing until 30 days and then decreasing, and MdGAD3 expression decreasing. The expression of MdGABA-T1,2 and MdSSADH1 increases gradually from 0 to 30 days, peaking at the same time as the ethylene climacteric peak (30 days). Thus, the expression of MdGAD1, MdGAD2, MdGABA-T1,2, and MdSSADH1 in “Cripps Pink” apple fruit responds strongly under RT storage, essentially as in “Empire” apple fruit stored under LT, low O₂ and ambient or elevated CO₂ (Brikis et al., 2018). These findings, in conjunction with those of Brikis et al. (2018), lead us to conclude that the postharvest expression patterns for GABA shunt genes in apple fruits are more influenced by development, than by environment.

The temporal patterns for GABA (i.e., slow decrease of approximately 60% from 10 to 40 days, followed by a dramatic increase at 70 days, presumably due to fruit aging and cellular disintegration at the end of storage), succinate and malate (slow decrease of 40 and 20%, respectively, from 30 to 70 days) indicate that GABA does not accumulate under storage at RT, and that GABA is probably catabolized to succinate and malate (Han et al., 2018). The application of exogenous GABA increases the expression of MdGAD1, MdGAD2, MdGABA-T1,2 and MdSSADH1, restrains the decrease in malate and succinate levels, decreases respiration and ethylene production rates, and delays the ethylene production peak (Han et al., 2018; Table 3). Notably, the application of exogenous Ca²⁺ decreases the Glu level (before 30 days), increases the levels of GABA, succinate and malate (10–60 days) and expression of MdGAD1 (before 30 days), MdGAD2 (20–40 days), MdGABA-T1/2 (10–20 days) and MdSSADH (20–40 days), suppresses the respiration rate, and decreases the ethylene production peak (Han et al., 2021).

Together, these studies suggest that elevated endogenous GABA or exogenous GABA maintains the quality of apple fruit by stimulating the activity of the GABA shunt and the synthesis of malate, and delaying fruit ripening. Notably, the inhibition of ethylene-mediated ripening by 1-methylcyclopropene increases the GABA level in “Empire” and “Honeycrisp” apples and in “AC Harrow Crisp” pears stored under LT, CA conditions (Deyman et al., 2014a; Lum et al., 2016a; Flaherty et al., 2018). While the interaction between ethylene and GABA biosynthesis requires further study, exogenous GABA seems to elicit similar responses as LT, CA conditions.

The attenuation of LT injury in peach, banana, orange, pomegranate, persimmon and aonla fruits, as well as cut anthurium flowers, by exogenous GABA is evident from the preservation of membrane fluidity and stability (decrease in electrolyte leakage), which is accompanied by decreases in reactive oxygen species (ROS; e.g., hydrogen peroxide and superoxide radical), greater antioxidant and radical-scavenging capacities, the maintenance of intracellular ATP and NADH, and the accumulation of potential osmolytes (i.e., soluble sugars, PAs and proline) (Shang et al., 2011; Yang et al., 2011; Wang Y. et al., 2014; Aghdam et al., 2015, 2016a,2016b; Habibi et al., 2019, 2020; Nazoori et al., 2020; Niazi et al., 2021; Ali et al., 2022; Table 3).

Similar mechanisms are involved in: the delay of senescence and preservation of quality in LT-stored blueberries, cherries and mushrooms, RT-stored apples, and conditioned LT-stored apples (Ge et al., 2018; Han et al., 2018; Aghdam et al., 2019; Rabiei et al., 2019; Al Shoffe et al., 2021; Shekari et al., 2021); browning resistance and the preservation of quality in LT-stored pear and mango (Li et al., 2019; Rastegar et al., 2019); resistance against fungal infection in LT-or RT-stored pear, orange, strawberry and tomato fruits (Yu et al., 2014; Fu et al., 2017; Sheng et al., 2017; Yang et al., 2017); and resistance against various pathogens and browning in RT-stored walnut kernels (Ebrahimzadeh et al., 2019; Table 3). Notably, pathogen resistance is promoted by salicylate signaling and disease resistance proteins, and maintaining the integrity of the cell wall barrier (Yu et al., 2014; Fu et al., 2017; Yang et al., 2017; Gao et al., 2018a; Zhao et al., 2021; Table 3), and the loss of apple fruit acidity is retarded by accumulating malate and suppressing ethylene biosynthesis (Han et al., 2018; Table 3).

Hou et al. (2022) have shown that the fresh-cut process does not affect the organoleptic quality of carrots stored under LT for hours, though it appears to enhance GABA biosynthesis from both Glu and PAs (Table 4). This result is consistent with the previously reported impact of wounding/mechanical damage on GABA accumulation (Shelp et al., 2012a). Notably, the resistance to browning and bacterial pathogens in fresh-cut pear, apple and potato during prolonged LT storage is improved by both CA and exogenous GABA via the mechanisms described above (Gao et al., 2018a,b; Wang D. et al., 2021; Zhao et al., 2021).

In summary, the application of exogenous GABA to postharvest fruits, vegetables (including mushrooms), cut flowers, and walnuts delays senescence, attenuates chilling injury and fungal/bacterial-induced decay, and helps to preserve sensory and nutritional quality. GABA can promote activities of the GABA shunt, and the TCAC, antioxidant, secondary and phytohormone pathways, which in turn, reduce the stress-induced ROS level. However, the precise mechanisms whereby GABA interacts with other signaling molecules such as Ca²⁺, H₂O₂, PAs, salicylic acid, nitric oxide and melatonin, or with phytohormones such as ethylene, abscisic acid and auxin remain unknown (Bor and Turkan, 2019; Podlešákova et al., 2019; Seifikalhor et al., 2019; Suhel et al., 2022).

In cut flowers, excessive transpiration can result in a loss of turgor, premature wilting of flowers and leaves, and accelerate flower senescence. Water loss via the stomata can also result in a loss of FM and quality in leafy vegetables and immature green fruits. Therefore, it may be beneficial to manipulate endogenous GABA by applying exogenous GABA to restrict stomatal opening and prevent water loss (Xu et al., 2021a). Stomatal closure may also aid in preventing bacterial and fungal pathogens from entering leaves or fruits (Gahir et al., 2021). Thus, regulation of stomatal function may be a promising strategy for improving postharvest quality and safety of horticultural products (van Meeteren and Aliniaeifard, 2016).

Malate is the predominant organic acid in ripe apple fruit, and most of this is found in the vacuole. The transport of malate across the apple tonoplast is probably mediated by the apple ALMT9 (MdMa1) (Li et al., 2020; Table 2). Both the full-length protein, MdMa1, and its naturally occurring truncated protein, mdma1, localize to the tonoplast; when expressed in Xenopus laevis oocytes and Nicotiana benthamiana cells, MdMa1 mediates a malate-dependent inward-rectifying current, whereas the ma1-mediated transmembrane current is much weaker, indicating that ma1 has significantly lower malate transport activity than Ma1. RNA interference suppression of MdMa1 expression in “McIntosh” apple leaves, “Empire” apple fruit, and “Orin” apple calli significantly decreases the malate level. Notably, the most highly-related ortholog in Arabidopsis, ALMT9, transports mainly Cl^– into the vacuole, but is subject to negative regulation by cytosolic GABA (Bown and Shelp, 2016; Xu et al., 2021a,b). Thus, the application of GABA on apple fruit during LT storage could reduce malate accumulation and the acidity of apple fruits. Bai et al. (2015) have suggested that a major network of genes, including MdALMT9, is associated with the developmental regulation of apple fruit acidity in “Golden Delicious,” but such a network has not been investigated during the ripening period (Ban and Xu, 2020). It could have implication for breeding apples or other fruits in order to preserve or enhance their flavor during postharvest storage.

Tomato SlCAT9 encodes a tonoplast Glu/Asp/GABA exchanger and its expression increases in tomato fruit during ripening (Snowden et al., 2015; Table 2). Such an exchanger might provide a mechanism for remobilizing GABA from the vacuole during cellular Glu uptake (Chung et al., 1992). On the other hand, ripening-specific overexpression of SlCAT9 increases the accumulation of GABA, Glu and Asp by approximately 20-, one- and sixfold, respectively (Snowden et al., 2015). Notably, greater Glu and Asp accumulation in the vacuole contribute to umami taste development in tomato fruit during ripening (Takayama and Ezura, 2015). Elevated GABA accumulation in the vacuole of immature fruit might deter insect pests and pathogens, whereas lower GABA accumulation in tomato fruit during ripening might be beneficial for attracting insects and animals for successful seed dispersal (Takayama and Ezura, 2015; Shelp et al., 2021). The properties of apple CAT9 have not yet been characterized (Table 2), but they could have implications for altering the development, flavor and biotic resistance of apple fruits.