In Experiment II, the respiration rate of the fruit under different CA conditions was monitored. For this, the outlet of each CA container was connected to a 2 L glass cuvette filled with approximately 600–750 g of fruit. In this way, the cuvettes received the same gaseous conditions as the other stored fruit. At the start of the respiration measurement, the glass cuvettes were disconnected from the flow-through system and all valves were closed. Concentrations of O₂ and CO₂ were measured immediately after closure and again after about 6 h of incubation; from this, the respiration rate was calculated. The calculation considered the O₂ consumption rate, cuvette volume, and fruit mass; the results were expressed in nmol O₂ kg^–1 s^–1. CO₂ and O₂ were measured using a headspace gas analyzer (Checkmate 3, PBI Dansensor, Ringstead, Denmark).

Fruit showed typical B. cinerea infection symptoms (e.g., brown spots or visible gray mycelium). Only fruit showing Botrytis rot symptoms (Feliziani and Romanazzi, 2016; Petrasch et al., 2019) were considered for spoilage calculation. Botrytis rot was the pivotal fungal pathogen throughout the experiment. Occasionally, some other fungal symptoms were observed, such as Rhizopus fruit rot, and these were not considered in the calculation. The spoilage incidence was expressed as the percentage of fruit per block affected by a fungal infection. The weight of an individual cup or punnet with fruit was recorded using an MS6002TS balance (Mettler-Toledo GmbH, Giessen, Germany) at harvest. Weight loss was expressed as the percentage of weight loss compared to the initial fruit weight, corrected for the weight of cups or punnets.

In Experiment I, firmness (limited compression) was measured using FirmTech FT7 (UP GmbH, Ibbenbüren, Germany). Firmness was expressed as the force-displacement (in g mm^–1) after about 1 mm of compression of the fruit shoulder without a calyx. Each replicate (cup) had 5 fruit. The same fruit were used for Brix measurement using a refractometer (PAL-1, Atago, Japan).

In Experiment II, fruit firmness (penetration) was measured using a universal testing machine (Zwick Z2.5/TS1S materials testing machine; Ulm, Germany). Fruit were cut in half lengthwise and compressed at the equator with a probe (ø 2.5 mm) at a constant plunger speed (2.5 mm s^–1) to a fixed penetration depth (5 mm). The maximum force (N) was recorded during the penetration of the fruit piece. Each replicate (punnet) had 15 strawberry fruit.

Fruit from one punnet (approximately 25 fruit) were cut into small pieces, immediately frozen in liquid N₂, and ground into powder. Part of the frozen powder was freeze-dried for sugar and organic acid extraction and also for expressing all nutrients as per gram dry weight. Frozen powder (0.3 g) was dissolved in 1.5 ml of water, thoroughly shaken, and centrifuged at 21,000 × g at 20°C for 15 min. The supernatant was used for pH measurement.

Frozen powder (0.2 g) was thawed on ice in darkness with 1 ml of ice-cold 3.3% meta-phosphoric acid. Samples were sonicated for 10 min, then centrifuged at 21,000 × g at 4°C for 10 min. The supernatant was filtered through a 0.45 μm cellulose filter and injected into an HPLC system consisting of a GS50 pump (Dionex), a 340S UV-VIS detector (Dionex), and a MIDAS autosampler (Spark Holland) equipped with a ProntoSIL 120-3 C18 AQ, 250 mm × 3 mm column (Knauer). The column was eluted with 400 μl L^–1 H₃PO₄ + 2.5 ml L^–1 MeOH + 0.1 mM EDTA in distilled water, followed by a wash step with 30% acetonitrile in distilled water at a flow rate of 0.35 ml min^–1 at 35°C. Ascorbic acid was detected at 243 nm. The system was calibrated using an authentic ascorbic acid standard (Acros Organics) prepared at 3.3% MPA.

The frozen sample (0.3 g) was mixed with 1.5 ml of 50% methanol containing 1% formic acid extraction solvent. The supernatant was filtered through a 0.45 μm cellulose filter prior to injection into an HPLC system (Ultimate 3000, Dionex, Sunnyvale, CA, United States). Solvents used were (A) 0.1% trifluoroacetic acid in distilled water, and (B) 0.1% trifluoroacetic acid in HPLC-grade acetonitrile, establishing the following gradient: 5–28% of B for 0–35 min, 28–75% of B for 35–37 min, isocratic 75% of B for 37–40 min, 75–5% of B for 40–42 min, and isocratic 5% of B for 42–50 min, using a flow rate of 0.8 ml min^–1. Separation was achieved using a C18 column (150 mm × 3 mm, 3 μm; HyPURITY C18, Thermo Scientific, New York, NY, United States), and peaks were identified using a UV/vis detector at 520 nm using pelargonidin 3-glucoside as an authentic standard (Extrasynthese).

A total of 15 mg of freeze-dried tissue was mixed with 5 ml of 75% ethanol, followed by shaking and incubating in a water bath at 80°C for 20 min before centrifuging at 8,500 × g for 5 min at 4°C. A vacuum centrifuge (Savant SpeedVac SPD2010, Thermo Fisher Scientific, Germany) was used to dry 1 ml of the supernatant at 55°C for 2.5 h. The pellet was resuspended in 1 ml of distilled water and sonicated for 10 min, followed by centrifugation for 10 min at 4°C at 14,800 × g. Before analysis, the supernatant was diluted with distilled water 50 times for sugar and 5 times for organic acids.

Soluble sugar measurements were carried out using a High-Performance Anion Exchange Chromatography with Pulsed Amperometric Detection (HPAEC-PAD; Dionex ICS5000, Thermo Fisher Scientific, Germany) equipped with a CarboPac1 column (250 mm × 2 mm) eluted with 100 mM NaOH at a flow rate of 0.25 ml min^–1 at 25°C. Quantification was performed using glucose, fructose, and sucrose standards from Sigma-Aldrich.

Citric and malic acids in the extracts were analyzed using an IC system equipped with a Triathlon autosampler (Spark Holland), GS50 pump (Dionex), ED50A detector (Dionex) operating in the conductivity mode, and an ASRS ultra II 2 mm suppressor (Dionex). Anions were separated at 30°C on an IonPac AS11HC (250 mm × 2 mm) column (Dionex), using the following multistep gradients: 1 mM NaOH, 0 min; 1 mM NaOH, 1 min; 14 mM NaOH, 23 min; 30 mM NaOH, 31 min; 60 mM NaOH, 41 min at a flow rate of 0.38 ml min^–1. Quantification was performed using authentic DL-malic acid and citric acid standards from Sigma-Aldrich.

Treatment effects on all measured variables were tested using a one-way analysis of variance (ANOVA) at each time point during the experimental period. Experiment I was carried out with 3 blocks, and Experiment II was carried out with 2 blocks. In each block, replicated fruit were measured as an individual or a pooled sample, depending on variables. An average value of each block was used for statistical analysis. Homogeneity and normality of residuals in the ANOVA were tested using Bartlett’s test and Shapiro–Wilk test, respectively. Fisher’s protected least significant difference (LSD) test was used as a post hoc test. All statistical analyses were performed in Genstat (19th edition, VSN International Ltd., Hemel Hempstead, United Kingdom). All tests were conducted at α = 0.05.

During CA storage, the spoilage incidence of CA-treated fruit was lower compared to control-treated fruit (Figure 1A). During shelf life, the spoilage incidence increased dramatically in fruit from the ambient atmosphere and 10 kPa CO₂ + 10 kPa O₂ (Figure 1B). Fruit stored at 15 kPa CO₂ showed reduced spoilage incidence with increasing O₂ concentrations from 5 to 10 kPa. Under conditions of 10 kPa O₂, fruit treated with 15 and 20 kPa CO₂ showed the lowest spoilage incidence among all treatments. Overall Brix values of fruit showed a decreasing trend during CA storage but did not show a further change during the 3-day subsequent shelf life; fruit firmness did not change during CA storage but showed a slight decrease during the 3-day subsequent shelf-life period. The different treatments had no significant effects on the changes of Brix values and the firmness of fruit (Supplementary Figure 1).

The fruit respiration rate was higher in CA0 and the control conditions compared to the elevated CO₂ treatments (20 and 30 kPa), indicating that increased CO₂ concentrations suppressed respiration (Figure 2A). The respiration rate fluctuated during the first 5 days of storage, regardless of treatments. After day 9, the respiration rate started to increase. This increase was smaller and not significant for fruit stored under 30 kPa CO₂ conditions. There were no apparent differences in respiration related to the immediate or stepwise application of high CO₂ (20 and 30 kPa).

Fruit stored under CA0 and control conditions showed high spoilage incidence at the start of shelf life (Figure 2B). The spoilage incidence of CA0 overlapped with that of the control treatment, reaching about 58% spoilage at the end of the CA storage. In contrast, fruit stored under elevated CO₂ with 10 kPa O₂ treatments did not show any spoilage during the CA storage; during the subsequent shelf life at 12°C and ambient atmosphere, fruit spoilage did occur in these treatments. After the 3-day shelf life, fruit stored under CA30 had the lowest spoilage incidence (32%), followed by GCA30 (53%) and CA20 and GCA20 (over 60%). After 5 days of shelf life, the spoilage incidence was above 80% in all treatments.

During the CA storage and shelf life, the glucose content did not show a clear trend (Figure 3A), whereas the fructose content increased (Figure 3B), and the sucrose content decreased (Figure 3C). The total sugar content (glucose + fructose + sucrose) showed a downward trend for all treatments (Figure 3D). No apparent differences in the rate of change in individual sugars among treatments were observed. However, the sucrose content in fruit from both the static and stepwise 30 kPa CO₂ treatments showed a more rapid decrease in the CA storage than the sucrose level in fruit from other treatments (Figure 3E). This indicates that under high CO₂ conditions, there was a more pronounced conversion of sucrose into glucose and fructose.

Fruit pH increased from about 3.2 to 3.7 with minimal differences between treatments (Figure 4A). Fruit from the CA30 condition generally showed a slightly higher pH than the control and CA0 treatments. Citric acid did not show any changes throughout the CA storage and shelf life, and no treatment effects were observed (Figure 4B). The malic acid content in fruit was 2.2-fold lower than citric acid at harvest and showed a slight decrease after fruit were transferred to the shelf life condition in all treatments. No differences with respect to the CA treatments were observed (Figure 4C). The increase in pH during the experimental period can be ascribed to the decrease in the two major acids; fruit pH during the CA storage was lower than during the subsequent shelf life, regardless of atmospheric composition. In addition, high CO₂ treatments led to a higher pH compared to control and CA0 treatments after fruit were transferred to the ambient atmosphere (Figure 4D).

Softening was observed in fruit from control and CA0 treatments during the subsequent shelf life (Figure 5A). Fruit from high CO₂ treatments (except for CA30) retained their firmness during the shelf life and no promising effect of the stepwise elevation of CO₂ was observed. Fruit weight loss during the CA storage was minimal; in fact, fruit weight loss increased up to a maximum of 3% in the subsequent shelf life, and no differences were apparent between treatments (Figure 5B).

Pelargonidin-3-glucoside and pelargonidin-3-malonyl glucoside were the major anthocyanins in “Arabella” strawberry fruit, and the content of the former was lower (∼75%) than the content of the latter at harvest (Figures 6A,B). Overall, pelargonidin-3-glucoside content in fruit increased during the CA storage with decreasing trends during shelf life. There were no consistent differences between treatments. Similar trends were observed in the pelargonidin-3-malonyl glucoside content of fruit. Ascorbic acid levels of “Arabella” fruit showed slightly increasing trends in fruit from control and CA0 treatments throughout CA storage and shelf life. In contrast, the ascorbic acid level did not change in fruit treated with elevated CO₂ concentrations, where it was maintained around 10 mg g^–1 DW (Figure 6C). Collectively, until the end of shelf life, the antioxidants of fruit showed similar levels compared to fruit at harvest.

Controlled atmosphere storage retains product quality via reducing O₂ and increasing CO₂ concentrations relative to the ambient atmosphere. Generally, the lower the O₂ and the higher the CO₂ concentrations, the more the storage life can be extended by delaying senescence and ripening of commodities (Rama and Narasimham, 2003). However, based on our observations, this concept may not fully apply to strawberry fruit. In Experiment I, moderately decreased O₂ concentrations (5–10 kPa) combined with 15 kPa CO₂ greatly decreased the spoilage incidence of “Sonsation” strawberry fruit, which was observed only after fruit were transferred to post-storage shelf life (Figure 1B). Notably, 10 kPa O₂ clearly performed better than 5 or 7.5 kPa O₂. Generally, when high CO₂ is applied to strawberry fruit, the risk of initiating fermentation increases, which leads to the accumulation of acetaldehyde and ethanol, resulting in off-flavors (Rama and Narasimham, 2003; Kanellis et al., 2009). This could be prevented and/or attenuated by introducing more O₂ into the environment. In contrast, Ke et al. (1991) found that decreasing concentrations of O₂ (down to 0.25% O₂ balanced with N₂ only) led to reduced spoilage incidence in “Selva” strawberry fruit at both 0 and 5°C and did not affect the soluble solids content (brix), pH, or titratable acidity after 10-day CA treatments compared to air-treated fruit. However, this study also reported a higher ethanol content of fruit under this treatment compared to fruit from other treatments. It is unclear if these beneficial effects of ultralow O₂ persisted during the shelf life, as authors only did quality measurements at the end of the storage period.

We observed that increased CO₂ concentrations (10–20 kPa) combined with 10 kPa O₂ drastically lowered the spoilage incidence of fruit. Fruit firmness and Brix did not exhibit differences during CA storage and shelf life among treatments (Supplementary Figure 1). Application of 20 kPa CO₂ in air did not induce the fermentation of “Selva” strawberry fruit and ultimately resulted in a lower spoilage incidence (Ke et al., 1991). This shows that in order to benefit from CA storage, optimal O₂ concentrations might not be lower than 10 kPa and CO₂ should not be lower than 15 kPa.

Therefore, in Experiment II, we referred to the optimal 10 kPa O₂ CA conditions for “Sonsation” strawberry fruit, which showed a lower spoilage incidence without affecting fruit firmness and Brix, and applied that in combination with different concentrations of CO₂ to “Arabella” strawberry fruit. Both conditions with 10 kPa O₂ and high (20 and 30 kPa) CO₂ significantly suppressed spoilage of “Arabella” strawberry fruit. This confirms the results with “Sonsation” fruit, indicating the two cultivars may respond to high CO₂ conditions similarly.

Notably, in Experiment II, the condition of 10 kPa O₂ alone (Figure 2B, CA0) did not suppress the spoilage development compared to the control condition at the end of the CA storage. Combing the results of two experiments suggests that high CO₂ combined with reduced O₂ could be effective for reducing spoilage. When high CO₂ is applied to strawberry fruit, the concentration of O₂ should not be too low; otherwise, the additive effects of high CO₂ and low O₂ may induce fermentation.

CO₂ concentrations higher than 30 kPa can cause faster softening, berry discoloration, and the production of off-flavors, leading to quality loss. Elevating CO₂ concentrations in a stepwise manner may improve the adaption of products to high CO₂ compared to static CO₂ treatments such that the storage life, fruit firmness, color, nutritional compounds, and antioxidants are retained (Falagán et al., 2020). Raising gas concentrations over a 3- or 7-day period to reach final concentrations of 10 kPa CO₂ and 5 kPa O₂ resulted in a lower CO₂ production peak and reduced spoilage incidence in blueberries when compared to static CA or ambient atmospheric conditions (Falagán et al., 2020).

In this study, the respiration rate (Experiment II) of fruit under high CO₂ treatments was reduced throughout the storage period compared to fruit from low CO₂ treatments (Figure 2A). However, the pattern was not affected by applying high CO₂ in a stepwise manner. Similarly, the spoilage incidence was greatly suppressed by high CO₂ but was not affected by the stepwise application. The suppressed respiration caused by high CO₂ is due to the inhibition of succinate dehydrogenase that catalyzes the conversion of succinate to fumarate at the tricarboxylic acid (Kanellis et al., 2009). Malic acid, as the upstream metabolite of succinate and fumarate, was reduced by high CO₂ (20 kPa CO₂ in air) (Fernández-Trujillo et al., 1999; Ponce-Valadez et al., 2005). The observed decreasing trend of malic acid, especially in fruit treated with elevated CO₂ (Figure 4C), is in line with previous findings.

In strawberry fruit, sucrose is degraded via the invertase pathway to form fructose and glucose (Kanellis et al., 2009; Durán-Soria et al., 2020). We found that fructose and glucose in strawberry fruit slightly increased during the experimental period in all CA treatments, whereas sucrose decreased. High CO₂ treatments seemed to induce a more significant conversion from sucrose to glucose and, in particular, fructose (Figure 3E). Bang et al. (2019) also observed increases in glucose and fructose but a decrease in sucrose content in strawberry fruit after 1 day of treatment with 30 kPa CO₂ at 25°C compared to the ambient atmosphere condition. The transcriptome analysis indicated that the invertase inhibitor was downregulated under 30 kPa CO₂, triggering invertase activity; therefore, more glucose and fructose are synthesized in fruit.

A decreasing trend in firmness in control and CA0-treated fruit was observed, especially after the fruit were returned to the ambient atmosphere (Figure 5A). The firmness of fruit from high CO₂ treatments was maintained at approximately the same value at the end of shelf life compared to the firmness at harvest. The different firmness behaviors between 0 kPa CO₂ and high CO₂ corroborate that CO₂ treatments inhibit firmness loss of strawberry fruit in a non-reversible manner (Harker et al., 2000). Fruit treated with static 30 kPa CO₂ tended to be softer compared to fruit from other high CO₂ treatments. This could be due to the adverse effect of long-term storage under static 30 kPa CO₂ as it was not observed under stepwise high CO₂ treatments.

Increased firmness in strawberry fruit under elevated CO₂ is likely due to reinforcement of cell-to-cell bonding, which is associated with the increase in pH of the apoplast (Harker et al., 2000). According to our finding, the pH of “Arabella” strawberry fruit increased throughout the whole experimental period due to a reduction in organic acids, especially malic acid (Supplementary Figure 2), which is in line with previous findings (Ke et al., 1991; Holcroft and Kader, 1999; Blanch et al., 2015). Both H⁺ and HCO₃^– are produced by the solubilization of CO₂, which could affect pH (Lucas, 1979; Bown, 1985). The presence of H⁺ decreased pH, whereas the uptake of HCO₃^– into cells increased apoplastic pH due to the presence of an OH^– efflux instead of an H⁺ influx transport system (Laing and Browse, 1985). This increase in pH of the apoplast possibly enables Ca²⁺, rather than H⁺, as the ion species that binds to negatively charged carboxyl groups of the cell wall (Harker et al., 2000) to promote linking of neighboring pectin polymers through the egg-box model (Demarty et al., 1984). This idea is in line with our observations. We observed that fruit firmness increased during high CO₂ treatment, whereas it slightly decreased after fruit were transferred to the ambient atmosphere (Figure 5A). In contrast, firmness continuously declined in fruit in the control and CA0 treatments.

The differences in the anthocyanin content of fruit from different treatments during CA storage and shelf life were not consistent, presumably due to biological variation in samples. The ascorbic acid content in fruit did not change over the experiment, and no differences between the control and different CA treatments were apparent. From these aspects, our findings in “Arabella” strawberry fruit are different from most studies which found firmness loss, discoloration, and antioxidant losses in strawberry fruit treated with elevated CO₂ (Gil et al., 1997; Lee and Kader, 2000; Shin et al., 2008; Li et al., 2019).

Application of high CO₂ might induce fermentation. The extent of fermentation with increased CO₂ concentrations is cultivar-dependent in strawberry fruit, and some cultivars do not produce fermentation metabolites at all (Fernández-Trujillo et al., 1999; Watkins et al., 1999; Pelayo-Zaldívar et al., 2007). As the ratio of CO₂ production and O₂ consumption in fruit approached 1, both under 20 and 30 kPa CO₂ (data not shown), we assume that fermentative metabolism was not activated during CA storage. Therefore, “Arabella” strawberry fruit seem relatively tolerant to elevated CO₂. This could explain that although high CO₂ lowered the spoilage incidence, other quality properties, such as firmness, anthocyanins, and ascorbic acid, were hardly affected. The lower spoilage incidence not only resulted from the physiological effects of high CO₂, as we discussed above but may also be due to the direct suppression of high CO₂ on B. cinerea growth. In vitro inhibition tests showed that the B. cinerea colony diameter decreased with elevated CO₂ from 0 to 20 kPa, and it could not grow when CO₂ concentrations were higher than 30 kPa (Garcia-Gimeno et al., 2002). After fruit were transferred to the ambient atmosphere condition, the inhibitory effect was relieved, but the residual effect of elevated CO₂ still existed, which was observed in both experiments and the two cultivars studied in our research.