The Impact of Fruit Ripeness on the Lower Oxygen Limit, Chlorophyll Fluorescence and Fermentation Behavior in Apples

Dynamic Controlled Atmosphere-Chlorophyll Fluorescence storage (DCA-CF) uses a fluorescence-based measurement method to detect fermentation in apples (Malus × domestica BORKH.) caused by low-oxygen levels at an early stage. In recent years, it has been observed that individual apples of the same variety and origin can exhibit different fermentation behavior when stored under completely identical conditions. The causes of the different fermentation behavior must be found in order to be able to use DCA storage optimally. This study aimed to find the causes of the different fermentation behaviors of individual apples. Our results show that fruit ripeness can affect the lower oxygen limit (LOL), especially immediately after harvest, when the starch degradation in the fruit is not yet complete. A significant increase in the LOL was observed in ‘Elstar’ (2020: 0.3 kPa, 0.6 kPa, 0.9 kPa; 2021: 0.3 kPa, 0.4 kPa, 0.6 kPa). ‘Braeburn’ also exhibited this behavior regarding the LOL at a lower level. The LOL could not be identified for some of the fruit (varying from 12.5% to 41.7% of the examined apples) previously stored in Ultra Low Oxygen (ULO) storage for 4 months. Also, the chlorophyll content in the apple skin influences the fluorescence measurement method. Within 2 weeks, the chlorophyll content in the apple skin was halved. If the chlorophyll content drops, the reliability of the fluorescence measurement also decreases. It turned out that apples with an Fv/Fm < 0.7 were unsuitable for valid LOL identification.


Introduction
Dynamic Controlled Atmosphere (DCA) is increasingly used in modern long-term storage of apples (Malus × domestica BORKH.) because of proven benefits to fruit quality maintenance and shelf life.With DCA storage, the oxygen level is dynamically adjusted and set to values of ≤ 1 kPa O2 in the long term (Zanella et al. 2008;Gasser et al. 2008).However, there is a risk of fermentation at oxygen levels of ≤ 1 kPa O2 because the apples are stored Data Availability Statements The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request.Tim-Pascal Schlie Tim-Pascal.Schlie@hs-osnabrueck.de 1 Faculty of Agricultural Sciences and Landscape Architecture, Hochschule Osnabrück-University of Applied Sciences, Oldenburger Landstraße 24, 49090 Osnabrück, Germany 2 Chamber of Agriculture Lower Saxony, Fruit Research Center Jork, Moorende 53, 21635 Jork, Germany close to the lower oxygen limit (LOL).The LOL represents the oxygen level that is still tolerated by the apple without developing disorders of fermentation (Prange et al. 2002;Wright et al. 2008Wright et al. , 2012)).One option to monitor the fruit's fermentation behavior is the measurement of chlorophyll fluorescence (Prange et al. 2002;Wright et al. 2015).Other ways to identify the LOL are by determining fermentation products in the fruit and measuring the respiration quotient (RQ) (Thewes et al. 2020).The fluorescence parameter minimum fluorescence (Fo) is a sensitive indicator for low-oxygen stress.For this purpose, a representative fruit sample is selected on which the fluorescence is measured (DeLong et al. 2004;Prange et al. 2013).The fluorescence increases when the oxygen level drops below the LOL (Harris and Heber 1993;Wright et al. 2008).The cell metabolism changes from aerobic respiration to fermentation, which also changes photosynthetic activity in the fruit skin (chloroplasts) (Wright et al. 2010(Wright et al. , 2012)).The fermentative metabolism is thus made visible via an indirect method.In recent years, it has been observed in research and practice that individual apples of the same variety and origin can exhibit different fermentation behavior when stored under completely identical conditions.This makes it difficult to select a representative sample for the fluorescence measurement (Schlie et al. 2020).Against this background, the aim of this study was to find the causes for the different fermentation behavior of individual apples.If these causes are known, it should be possible to select a representative sample or to store only fruits with the same fermentation behavior together.

Plant Materials
The German major variety 'Elstar, PCP' and the 1-methylcyclopropene (1-MCP) incompatible variety 'Braeburn, Mariri Red' were used for the investigations.The apple trees were grafted on M 9 rootstocks and grown in the experimental orchards of Fruit Research Center Jork, Germany (53°31 0 N, 9°44 0 E).The trees were planted in 2013 ('Elstar') and 2005 ('Braeburn').The plant space was 3.50 m x 1.00 m.The apples were harvested on three different harvest dates listed in Table 1.Apples of 10 trees from a plot were harvested for each harvest date.Only the best quality fruits were selected.The harvest date was determined using the ripeness parameters firmness, starch index and total soluble solids (TSS) (further information see Sects."Statistical Analysis" and "Different Harvest Dates") and the ripeness forecast from the Fruit Research Center in Jork.The 'Elstar' fruits were harvested with a fruit size of 65-75 mm (harvest date 1) and 70-80 mm (harvest date 2 and 3).The fruit size of 'Braeburn' was 65-75 mm (harvest date 1-3).The skin color of both varieties had a red content between 50% and 75% (manual selection).One part of the apples was briefly stored under cold storage (2 °C; ± 0.5 °C) until the experiments started.Another part of the apples was stored under ultra-low oxygen (ULO) (1.2 kPa O2, <0.1kPa CO2, 2 °C ± 0.5 °C).

Chlorophyll Fluorescence and LOL
The chlorophyll fluorescence measurements were performed using the Mini-Apple-PAM TM system (Walz, Ef-feltrich, Germany) based on pulse amplitude modulation (PAM) technology.The Mini-Apple-PAM TM system is able to measure the fluorescence of individual fruits (sensorsamples distance < 0.5 cm).A red light (λmax = 625 nm; 0.1 µmol m -2 s -1 ) was used to excite Fo.The sensor is located next to the red LED.The fluorescence parameter Fo was used to identify the LOL.Fo values were normalized (Wright et al. 2008).The Fo baseline was previously recorded by measuring Fo at oxygen levels of ≥ 1.5 kPa.
The Fo baseline is needed to identify the increase in fluorescence clearly.The rise of Fo deviating from the Fo baseline made the LOL visible (Wright et al. 2012).Furthermore, a saturating pulse (max.4000 μmol m -2 s -1 ) was applied to determine maximum fluorescence (Fm).Fm was needed to calculate the maximum quantum yield of photosystem II (Fv/Fm = (Fm -Fo) /Fm) (Maxwell and Johnson 2000).Before the fluorescence measurements started, the apples were dark-adapted for 30 min.The fluorescence was measured at the equator of the fruit where the green-yellow ground color was dominant.

Experimental Design
The investigations were carried out in a storage room (2.5 °C; ± 0.5 °C).The Mini-Apple-PAM TM system was integrated into a gas-tight stainless steel box (140 cm × 60 cm × 50 cm), shown in Fig. 1.A total of 96 measurement points were available.These were distributed over four layers (block layout).A fan and a line system made of PVC tubes were integrated to ensure good airflow.In this way, a uniform air flow was ensured and it was verified that there was a uniform oxygen distribution in the stainless steel box by measuring the oxygen content at different levels (data not shown).The air atmosphere was reduced by supplying nitrogen (purity 99.8%).In the first step, the oxygen was reduced to about 3 kPa within 15 min.In the second step, the oxygen was reduced from 3 kPa to 0 kPa over 4-6 days.The oxygen concentration in the storage box was mea-  The experiments were split into two sections and carried out in several LOL test runs.Experiment A aimed to identify the LOL shortly after harvest and after several months of ULO storage (n = 96).The variants are shown in Table 2.The LOL test run took place 1 week after harvest for 'Elstar' and 3 weeks after harvest for 'Braeburn' (3 weeks of cooling before oxygen reduction) (Köpcke 2007).
Experiment B: Table 3 shows the test variants of experiment B. Regarding fermentation behavior and LOL, two The control was stored under regular air (2 °C; ± 0.5 °C) immediately after harvest.The delayed cold storage variant was exposed to room temperature (~20 °C) for 12 days.After that, the apples were stored like the control.Both variants were examined in one LOL test run, so the samples were exposed to low-oxygen conditions for the same amount of time.In this way, the fruit's own fermentation products were more comparable.

Fruit Ripeness Measurements and Fermentation Analysis
The fruit ripeness was determined immediately after harvest (n = 40) and after each LOL test run (experiment A: n = 96; experiment B: n = 48).Some of the ripeness parameters are also quality parameters but are generally referred to as ripeness parameters in this study.The following parameters were examined.The firmness (kg cm -2 ) was measured on each fruit using the Fruit Texture Analyzer (Güss, Cape Town, South Africa).The measurements were taken at the fruit's equator on the transition zone from the green to red side of the fruit.Furthermore, after the LOL test runs, the firmness was also measured at the chlorophyll fluorescence measuring zone.The starch index (1-10) was assessed using the CTIFL scale (CTIFL = Center Technique Interprofessionnel des Fruits et Légumes; France) using iodine-potassium iodide (Vaysse 2002).The TSS (°Brix) The chlorophyll content was measured with a pigment analyzer (Control in Applied Physiology, Falkensee, Germany) on each individual fruit and reported as the Normalized Difference Vegetation Index (NDVI).The chlorophyll content was measured at the zone on the fruit, where later on, the chlorophyll fluorescence was measured during the LOL test run.The CLARUS 500 gas chromatography with headspace (Perkin-Elmer, Waltham, MA, USA) was used to perform the fermentation analysis.The filtered fruit juice (1 ml) was used to measure the fermentation products (acetaldehyde, ethanol).The measurements of fermentation products were carried block by block as a mixed sample (n = 4).

Statistical Analysis
All statistical analyses were performed using IBM SPSS Statistics (Version 26) (Chicago, IL, USA) and Microsoft Excel 2016 (Redmond, WA, USA).The interval scaled data were analyzed by analysis of variance (ANOVA) and Bonferroni test (p < 0.05).The Levene test checked the equal variances.If equal variances could not be assumed, a Welch-ANOVA was performed, followed by a Games-Howell test.The ordinal scaled data were analyzed by the Kruskal-Wallis test (p < 0.05).The block mean data were used to calculate Spearman-Rho's rank correlation.

Different Harvest Dates
Table 4 shows the fruit ripeness of the apples immediately after harvest in 2020 and 2021.The three harvest dates differed essentially in terms of the starch index and TA.The firmness of 'Elstar' decreased with each subsequent harvest date, while the firmness of 'Braeburn' remained at a stable level.
There was a certain amount of time between the harvest and the start of the experiments (see Sect. "Experimental Design").This led to a further development of fruit ripening.Figure 2 shows the firmness (a to d) and starch index (e to h) of the varieties 'Elstar' and 'Braeburn', which were K The firmness of 'Braeburn' was comparatively high, from 10.1 kg cm -2 to 9.6 kg cm -2 .The apples from the ULO stor- age showed less firmness but not less than 8.6 kg cm -2 (harvest date 3 after ULO; 2021/22) (Fig. 2c,d).The starch indices of the three harvest dates, which were examined close to the harvest, differed significantly.The apples from the ULO storage were starch-free at the time of the LOL test runs (e to h).
Figure 3 shows the TSS (a to d) and TA (e to h) of 'Elstar' and 'Braeburn' after the LOL test runs.Overall, higher TSS were measured for both varieties in the 2020/21 season than in the second year of the study.The TA was highest on harvest date 1 and decreased with each subsequent harvest date.The acid degradation continued in the ULO storage.The apples in the 2021/22 season tended to show higher TA than the first year of the study.The significant differences can be seen in Fig. 4.
Figure 4 (a to h) shows the results of LOL test runs of the three harvest dates and ULO storage.The LOL of 'Elstar' (2020) increased from 0.25 kPa O2 to 0.6 kPa O2 within 2 weeks and up to 0.9 kPa O2 at harvest date 3.This gradual increase was also observed in 2021.The increases in LOL were significant in both years (Fig. 4a,b).The LOL of 'Braeburn' was identified after 3 weeks of pre-cooling under normal atmosphere cold storage.In 2020, the LOL of 0.27 kPa O2 (harvest date 1), 0.33 kPa O2 (harvest date 2) .46kPa O2 (harvest date 3) were identified.In 2021, the LOL of harvest dates 1 and 2 ("Braeburn") did not differ significantly (Fig. 4c,d).LOLs from 0.2 kPa O2 to 0.3 kPa O2 were identified in the ULO-stored apples.However, it should be noted that there were failures (LOL error cases), especially with the ULO apples.Figure 4 (f to h) shows the number of fruits where no LOL could be identified.'Elstar' failure ranged from 15.6% to 39.6% (2020/21) and 12.5% to 35.4% (2021/22) of the examined apples.A higher failure rate at 'Braeburn' (ULO) was seen in the first study year at harvest date 1 (16%) and in the second year (21.9% to 41.7% of the examined apples).
Figure 5 shows the chlorophyll content and Fv/Fm in the absence of low-oxygen stress (≥ 1.5 kPa).The chlorophyll content was measured at the point where the chlorophyll fluorescence had previously been measured.While an NDVI of 0.63 was measured on harvest date 1 (2020; 'Elstar'), the chlorophyll content decrease significantly to 0.3 within just 2 weeks (harvest date 2).In 2021, the chlorophyll content decreased significantly from 0.58 (harvest date 1) to 0.35 (harvest date 2).NDVI of -0.17 (2020) and 0.01 (2021) were measured in apples of harvest date 3.This rapid chlorophyll degradation in the fruit skin could also be observed in 'Braeburn' (2020: 0.61; 0.34; 0.17; 2021: 0.56; 0.26; 0.12).The decrease in chlorophyll content was significant.Only harvest date 3 ('Elstar', 2020) showed a NDVI of -0.17 after harvest and did not differ significantly from ULO apples (-0.18).Apples examined 1 week after harvest showed Fv/Fm of 0.76 to 0.70 (in the absence of low-oxygen stress).'Elstar' and 'Braeburn' of ULO storage showed significantly lower Fv/Fm of 0.56 to 0.55.

Delayed Cold Storage
With the delayed cold storage variant, advanced fruit ripening could be provoked.Table 5 shows the results of experiment B. The delayed cold storage resulted in differences in ripeness and quality.In particular, the differences were achieved in terms of firmness and TA.In both varieties, there are significant differences in the LOL, and Fv/Fm between the two variants examined.The delayed cold storage variant had a high numbers of apples where the LOL could not be identified (up to 35% of the examined fruits).
The two variants were tested in the same LOL test run.Because the fruits were treated in exactly the same lowoxygen stress atmosphere, the level of fermentation products was comparable (Table 6).The information on fermentation products must be considered in the context of K the specific storage time ≤ 1 kPa O2.Although both variants were exposed to low-oxygen conditions simultaneously, the variant delayed cold storage always showed significantly higher values for ethanol.In 2020 ('Elstar'), the ethanol values were twice as high as in the control.In 2021, ('Elstar') the ethanol values were even 5.5 times higher than in the control.In the case of the 'Braeburn', the values in the 2 test years were 1.5 and 2 times higher than in the 9.9 a 12.9 a 5.5 a 0.51 b 2 0 .5 6 a control.In contrast, there were no significant differences in the acetaldehyde level.1992).TSS also negatively correlated with Fv/Fm at 'Elstar' (rs = -0.77,p < 0.01) and 'Braeburn' (rs = -0.72,p < 0.01).

Relationship Between Ripeness and Fluorescence
The TA and Fv/Fm also correlated significantly in both varieties ('Elstar': rs = 0.85, p < 0.01; 'Braeburn': rs = 0.77, p < 0.01).According to Cohen (1992), these are also strong effects.In summary, the progressive ripeness of the fruits leads to a significant decrease in Fv/Fm.
Figure 7 shows the relationship between the fluorescence parameters Fo and the chlorophyll content (a) and between Fv/Fm and chlorophyll content (b).The degradation of the chlorophyll content in the apple skin has an impact on both, Fo and Fv/Fm.The chlorophyll content correlates significantly with Fo in 'Elstar' (rs = 0.83, p < 0.01) and 'Braeburn' (rs = 0.51, p < 0.01).Furthermore, low chlorophyll levels lead to low Fv/Fm values.The correlation coefficient for 'Elstar' is rs = 0.78 (p < 0.01) and for 'Braeburn' rs = -0.77K 7 Relationship between Fo and the chlorophyll content (a), Fv/Fm and the chlorophyll content (b) using Spearman-Rho's rank correlation; correlation coefficient rs in 'Elstar' (blue) und 'Braeburn' (black).Data represent means, n = 48 (Fo = minimum fluorescence; Fv/Fm = maximum quantum yield of photosystem II) a b (p < 0.01).According to Cohen (1992), this are strong effects.

Discussion
In the recent history of fruit storage, chlorophyll fluorescence has been used as a sensory parameter to reveal physiological stress (Prange et al. 2013).The well-known fluorescence system in DCA storage is Harvest-Watch.With this system, six to eight fruits are measured with one sensor and the fluorescence is output as a normalized mean (Wright et al. 2008).The disadvantage is that no statement can be made about an individual fruit (Schlie et al. 2022).However, previous research indicated that there might be differences in fermentation behavior from fruit to fruit of the same variety and origin (Köpcke 2014;Schlie et al. 2020).
One of the reasons for the different behavior of fruits could be the ripeness.Apples harvested immature with a starch index of around 1 (harvest date 1) showed a low LOL of 0.2 to 0.3 kPa O2.However, if the harvest occurred 2 and 2 weeks later, the LOL increased significantly.Gasser et al. (2008) also observed differences in LOL at different harvest dates.Furthermore, Gasser and von Arx (2015) were able to show that the LOL was also dependent on the variety and could vary from year to year.Interestingly, the LOL increased in our study in the first few weeks of the postharvest phase, in which the climacteric processes occur.This phase is characterized by physiological changes and increased metabolic activity, such as the increase in ethylene, hydrolysis of starch and the degradation of chlorophyll (Busatto et al. 2017;Paul et al. 2011;Wright et al. 2011).
The increased metabolic activity likely resulted in a higher respiratory rate, leading to grown oxygen need and LOL (Wright et al. 2012).Furthermore, it cannot be ruled out that other factors could also have influenced the LOL, such as CO2 gas diffusion through the fruit tissue (de Oliveira Anese et al. 2016) or the fruit's hormone balance (Pérez-Llorca et al. 2019).Whether a low LOL could be capped by early harvest with consistent DCA storage remains to be examined.Due to the increased fruit respiration, we suspect that capping the LOL in the climacteric phase is impossible.Using a dynamic oxygen level based on an online LOL answer could optimize DCA storage.However, there is a need for further research.The investigations with ULO apples showed that the LOL can fall during storage.Wright et al. (2012) suggest that apples may develop some adaptation to low-oxygen levels during long-term storage.However, the LOL data of the ULO apples from the present study should be interpreted with caution.The LOL could not be identified for part of the fruit.There were essentially two reasons for this.The first reason was fluctuating Fo signals at oxygen levels of ≥ 1.5 kPa O2.However, reliable identification of the LOL requires a clear Fo-baseline as a reference (Wright et al. 2012).The second reason was that the Fo signal did not increase despite oxygen conditions of 0 kPa O2.Fermentation products were detected in subsequent analyses of the affected individual fruits (data not shown).Weak or faulty fluorescence signals can indicate chlorophyll degradation or the beginning of senescence.The NDVI of 'Elstar' and 'Braeburn' was reduced by half within 2 weeks (harvest date 1 to 2).Some variants of 'Elstar' that were stored in the ULO even showed negative NDVI values.This indicates that the chlorophyll content has almost completely degraded because negative values imply water (Tucker and Sellers 1986).Identifying the LOL on individual fruits led to LOL error cases being noticed.The ripening of the fruit was explicitly promoted in the delayed cold storage variant.The delayed cold storage variant at 'Elstar' resulted in increased LOL error cases.In both varieties of the delayed cold storage variant, significantly higher ethanol levels were detected compared to the control.However, the apples were exposed to low-oxygen conditions simultaneously (individual storage time for each LOL test run).The treatment of room temperature before cold storage accelerated the metabolism in the fruit, which also increased the accumulation of products.
In addition to Fo, the fluorescence parameter Fv/Fm was also recorded.Fv/Fm was determined at the beginning of the LOL test runs before low-oxygen stress was induced, so that the fruit's general stress perception could be determined.An unstressed plant leaf has an Fv/Fm value of 0.75 to 0.85 (Maxwell and Johnson 2000;Murchie and Lawson 2013).Song et al. (1997) found Fv/Fm values of 0.75 to 0.7 in unstressed apple fruits.Wright et al. (2012) measured on the sun-exposed side of 'Honeycrisp' a lower Fv/Fm (0.65) than on the shaded side (0.74).Our investigations found Fv/Fm values of 0.77 to 0.7 in apples examined 1 and 3 weeks after harvest, respectively.However, apples that had previously been in ULO storage for 4 months showed Fv/Fm values of < 0.6 ('Elstar') and < 0.7 ('Braeburn'), although low-oxygen stress was not present at the time of measurement.The view of the correlation analysis shows clearly a relationship between Fv/Fm and the chlorophyll content, as well as a correlation between Fo and the chlorophyll content.Furthermore, we were able to show that the various fruit ripening parameters (firmness, strength index, TSS, TA) correlate with Fv/Fm.Also, Song et al. (1997) were able to prove that there is a relationship between firmness and Fv/Fm as well as between starch index and Fv/Fm.Our results showed when chlorophyll degradation and fruit ripening had reached a critical point, the reliability of the fluorescence measurement became weaker and, thus, probably also the LOL.Therefore, we propose considering the parameter Fv/Fm when selecting a representative sample in the future.Apples with Fv/Fm < 0.7 should not be used for fluorescence measurement.The Fv/Fm parameter could be a helpful tool to detect the LOL error cases before DCA storage.

Conclusion
In the present study, the fermentation behavior of apples was described by measuring the LOL and the fruit's own fermentative products.Our results support the hypothesis that apples of the same variety and origin can have a different LOL with the same oxygen concentration in fruit storage.Based on individual fruit measurements, it was shown that the LOL was influenced by the fruit ripeness.In order to optimize DCA storage, only uniform fruits with the same ripeness should be stored together in one storage room.This also includes speedy harvesting and immediate storage of the apples.For implementation in large storage houses, this means more intensive coordination between farmers and storage house clerks.In addition, the chlorophyll content in the fruit skin was important.A low chlorophyll content led to weaker, sometimes fluctuating fluorescence signals, which had a massive impact on the reliability of the fluorescence method.In particular, in the case of apples with advanced fruit ripening, high numbers occurred in which the LOL could not be identified.Unfavorable fruit ripening limits the success of the fluorescence method.Another limitation of the fluorescence technique is that only a small number of representative apples in storage are used to measure chlorophyll fluorescence.Our results support the definition of what a representative sample might be.We propose not only to store apples with the same degree of ripeness under DCA storage but also to use the fluorescence parameter Fv/Fm as a standard value for selecting a representative sample.Apples with Fv/Fm < 0.7 should not be used for the fluorescence measurement.

Fig. 1
Fig. 1 The setup of the experiments with 96 fluorescence measuring points on four layers (a); Mini-Apple-PAM measurement device (b); Mini-Apple-PAM individual fruit measurement (c)

Fig. 2
Fig. 2 Firmness (a to d) and starch index (e to h) of 'Elstar' and 'Braeburn' at the time of lower oxygen limit test runs in the season 2020/21 and 2021/22; Standard error (p < 0.05; n = 96).ULO Ultra Low Oxygen

Fig. 3
Fig. 3 TSS (a to d) and TA (e to h) of 'Elstar' and 'Braeburn' at the time of LOL test runs in the season 2020/21 and 2021/22; Standard error (p < 0.05; n = 96 (a-d), n = 4 (e to h)); ULO Ultra Low Oxygen

Fig. 4
Fig. 4 The Lower Oxygen Limit (LOL) (a to and LOL error cases (e to h) of 'Elstar' and 'Braeburn' at three harvest dates and after 4 months of ULO storage in the season 2020/21 and 2021/22; Standard error (p < 0.05; n = 96 minus LOL error cases).ULO Ultra Low Oxygen

Fig. 5
Fig. 5 Chlorophyll content as Normalized Difference Vegetation Index (NDVI) (a to d) and Fv/Fm in the absence of lowoxygen stress (≥ 1.5 kPa) (e to h) of 'Elstar' and 'Braeburn' at the time of LOL test runs in the season 2020/21 and 2021/22; Standard error (p < 0.05; n = 96).ULO Ultra Low Oxygen

Figure 6
Figure 6 shows the relationship between Fv/Fm and the ripeness parameters firmness (a), starch index (b), TTS (c) and TA (d) of 'Elstar' and 'Braeburn' (data from both

Table 1
Overview of the different harvest dates

Table 3
The test variants of experiment B and the time of the lower oxygen limit (LOL) test runs different dates to start cold storage were examined.(n = 48).