Quantitative Analysis of Fragrance and Odorants Released from Fresh and Decaying Strawberries

The classes and concentrations of volatile organic compounds (VOC) released from fresh and decaying strawberries were investigated and compared. In this study, a total of 147 strawberry volatiles were quantified before and after nine days of storage to explore differences in the aroma profile between fresh strawberries (storage days (SRD) of 0, 1, and 3) and those that had started to decay (SRD = 6 and 9). In terms of concentration, seven compounds dominated the aroma profile of fresh strawberries (relative composition (RC) up to 97.4% by mass, sum concentration): (1) ethyl acetate = 518 mg·m−3, (2) methyl acetate = 239 mg·m−3, (3) ethyl butyrate = 13.5 mg·m−3, (4) methyl butyrate = 11.1 mg·m−3, (5) acetaldehyde = 24.9 mg·m−3, (6) acetic acid = 15.2 mg·m−3, and (7) acetone = 13.9 mg·m−3. In contrast, two alcohols dominated the aroma profile of decayed samples (RC up to 98.6%): (1) ethyl alcohol = 94.2 mg·m−3 and (2) isobutyl alcohol = 289 mg·m−3. Alternatively; if the aroma profiles are re-evaluated by summing odor activity values (ΣOAV); four ester compounds ((1) ethyl butyrate (6,160); (2) ethyl hexanoate (3,608); (3) ethyl isovalerate (1,592); and (4) ethyl 2-methylbutyrate (942)) were identified as the key constituents of fresh strawberry aroma (SRD-0). As the strawberries began to decay; isobutyl alcohol recorded the maximum OAV of 114 (relative proportion (RP) (SRD = 6) = 58.3%). However, as the decay process continued, the total OAV dropped further by 3 to 4 orders of magnitude—decreasing to 196 on SRD = 6 to 7.37 on SRD = 9. The overall results of this study confirm dramatic changes in the aroma profile of strawberries over time, especially with the onset of decay.


Introduction
Strawberries are one of the most widely consumed fruits with a good flavor and high nutritional value [1,2]. The aroma and odor-quality of strawberries depend on the type and concentration of volatile hydrocarbons (HC) in the aroma profile [3][4][5]. In fact, it is estimated that more than 360 volatile organic compounds (VOC) are emitted from strawberries [6,7]. However, only a small number of these contribute significantly to the strawberry fragrance and impact its perceived quality [8]. The intensity of strawberry fragrance has also been found to vary with the degree of freshness of the fruit. If the characteristics of VOC emitted from strawberries are evaluated thoroughly over time, this information can be used to improve our understanding of the natural strawberry fragrance and allow growers and retailers to optimize their harvesting, packing, storage and retail display procedures.
In order to assess the VOCs released from strawberry, researchers have used several different analytical approaches. Gas chromatographs (GCs) equipped with either flame ionization (FID) or mass spectrometer (MS) detectors have been the most common choices [9][10][11]. Recently, strawberry fragrances were also evaluated by combining olfactometry and GC techniques-i.e., harnessing state-of-the-art analytical technology alongside the particular selectivity of the human nose [1,2,12].
In this study, the concentrations and chemical types of strawberry volatiles were analyzed to characterize the fragrance (aroma profile: freshness staging) and offensive odorants (due to decay). All volatile compounds released from strawberry samples were collected at five different intervals (up to 9 days of storage period) at 25 °C. For the quantification of volatile components, liquid-phase standard was prepared containing a total of 19 odorous compounds for external calibration (Table 1S) The numbering of all supplementary (S) Tables and Figures are made with an S symbol following the number and placed in the Appendix section at the end. These calibration results were then used to develop predictive equations based on effective carbon number (ECN) [13]. These equations were then used for an extensive list of 'compounds lacking authentic standards/surrogates (CLASS)' due to the absence of standard material (i.e., authentic compounds) or to the synthesis complexities or costs involved in standard preparation [14]. The use of the predictive equations based on response factor vs. effective carbon number (ECN) linear correlation allowed robust, statistical estimation of all CLASS. The results of this approximation method allowed us to characterize the emission pattern of most fragrance and odorous components released from strawberry samples in a quantitative manner. In this research, we undertook measurements of strawberry aromas and odorants to provide detailed descriptions on their emission patterns in relation to storage duration. The results of this study will thus help us understand the characteristics of the flavor changes in strawberries that occur during storage.

Materials and Methods
In this research, a total of 19 VOCs that had relatively strong odor intensities with a wide range of volatility and polarity were selected for external calibration (Table 1S). The calibration results obtained using this standard mixture was used to derive predictive equations based on 'effective carbon number (ECN)' theory [13]. These ECN-based predictive equations were then used to calculate the concentrations of 'CLASS' due to the absence of standard material (i.e., authentic compounds) or to the complexity involved in standard preparation [14].
The concentrations of CLASS were derived from the predictive equations based on linear regression equations between RF values of target standard compounds (Table 3S) 19 VOCs contained in the L-WS, they were quantified directly using the calibration data of the L-WS. However, we did not prepare standards for the remaining 137 volatiles for many different reasons. For simple quantitation of those strawberry aroma components, we treated them as CLASS to quantify their concentrations based on the ECN approach (Table 4S).

The Collection of Strawberry Volatiles
The sorbent tube sampling method was used for the collection of the VOCs released from the cut strawberry sample. In the case of ammonia and RSC, the bag sampling method (polyester aluminum-PEA bag) was used as discussed in the next subsection. The sorbent tube was prepared as a three-bed sorbent by packing with 100 mg of Tenax TA, Carbopack B, and Carboxen 1000 (Supelco, Bellefonte, PA, USA) in a SS tube holder (tube size: length: 9 cm, OD: 6 mm, and ID: 5 mm; Camsco, Houston, TX, USA). The strawberries for this study were grown in Jin Ju city, Gyeong Sang Nam Do Province, Korea. The strawberries (1 kg in a Styrofoam tray) were purchased from a local market within one day after harvesting. An approximately 50 gram of strawberry sample was sliced and placed inside a 750 mL capacity impinger with the gas inlet and outlet positioned as the side arm and nozzle cap, respectively (ID: 3 mm and length: 30 mm). The strawberry samples were cut into four pieces (3 cm × 3 cm × 5.5 cm) (The weight of sample was adjusted to ~50 g) ( Figure S2). The inlet and outlet of the impinger were connected to a 10 L PEA bag filled with back-up gas (ultra-pure nitrogen > 99.999%) and the inlet of the sorbent tube (ST), respectively. A Teflon tube was used to connect the impinger and the PEA bag at one end and the ST at the other. The outlet of the ST was connected to the mini vacuum pump interfaced with mass flow controller (MFC) (Shibata ΣMP-30, Saitama, Japan) using silicon tubing. The VOCs emitted from the strawberry were flushed onto the ST by pumping the nitrogen gas through the impinger containing the strawberry samples at a flow rate of 50 mL•min −1 for 1 min. The impinger was maintained at 25 °C using a temperature-controlling water bath ( Figure 1). VOCs emitted from the strawberry sample were collected five times during the entire storage period of 9 days set for this study at 0, 1, 3, 6 and 9 days (Table 5S). For each selected day, the collection of samples was made as replicate for each target group (VOC, RSC, ammonia, and olfactory analysis). To initiate each sampling at a given day, a pre-purge was conducted by supplying ultra-pure nitrogen into the impinger at a flow rate of 50 mL•min −1 for 20 min. Throughout the storage period, the strawberry sample in the impinger was maintained in an aerobic state as the inlet and outlet of the impinger were left open to the air in the laboratory.

Instrumental Setup for VOC Analysis
All the analyses in this study were carried out using a GC-2100 (Shimadzu, Kyoto, Japan) equipped with a QP2010 MS (Shimadzu) and a UNITY thermal desorber (TD: Markes International, Ltd, Llantrisant, UK). The TD focusing cold trap was packed with Tenax TA and Carbopack B in a 1:1 volume ratio (inner diameter = 2 mm and total sorbent bed length: 50 mm). The VOCs were separated on a CP Wax column (diameter = 0.25 mm, length = 60 m, and film thickness = 0.25 μm) using a 50 min GC analytical cycle. The separated VOCs were detected by MS system and identified through library searching (NIST mass spectral library, NIST, Gaithersburg, MD, USA). The detailed conditions are also presented in Table 5S.

Calibration of the Liquid Working Standards of VOCs
Five-point calibration curves were prepared by analyzing sorbent tubes loaded with 1 μL of L-WS at these different concentration levels: (1) 1.30, (2) 6.52, (3) 13.0, (4) 26.1, and (5) 65.2 ng•µL −1 ) ( Table 2S). Details of the approach used to introduce the liquid standards to the sorbent tubes in the vapor phase have been described in detail elsewhere [15]. In short, a micro-syringe was used to introduce the liquid standard into the sampling end of the ST in a 50 mL•min −1 flow of nitrogen for 10 minutes. Each loaded sorbent tube was then analyzed by the TD-GC-MS system described below.
The sensitivity of the instrumental system remained fairly constant leading to stable response factors throughout the 9-day study period (Table 3S). All coefficients of variation (CV (%): SD/mean × 100) for the RF values were fairly stable (CV = 1.46 ± 1.29% (<4%)) allowing mean RF values to be applied to all the data. The correlation coefficients (R 2 ) of nearly all VOCs were above 0.99 (mean = 0.9954 ± 0.0075%), although there was a slight anomaly in the case of AA (0.9619 (SRD = 0) and 0.9698 (SRD-9)). In addition, to assess reproducibility, the L-WS with a mean of 26.1 ng•µL −1 was analyzed repeatedly prior to analysis of each batch of samples (SRD -0, 1, 3, 6, and 9). If the RF values of the 19 VOCs determined using this consecutive series of analyses, the RSE values generally fell below 4% (mean RSE = 1.39 ± 0.82%).

The Analysis of Ammonia and Reduced Sulfur Species
Although strawberries are noted for their attractive fragrance, the fruit can release unpleasant odors if stored for too long. Important offensive odorants such as ammonia and reduced sulfur compounds (RSC) [16,17] were therefore also quantified in this study. As expected, these compounds were found to be difficult to detect in the fresh fruit but they became increasingly abundant as the decay progressed. The TD-GC-MS setup optimized for the VOC analysis in this study is not optimal choice for ammonia or RSCs, thus these compounds were determined using alternative analytical approaches.
For the collection of samples to analyze for ammonia and RSC, the bag sampling method (PEA bag) was employed. The inlet and outlet of the impinger filled with the strawberry samples were connected to a gas cylinder filled with ultra-pure air (>99.999%) and a 10 L empty PEA bag, respectively. Teflon tubing was used to connect the impinger and the gas cylinder at one end and the empty PEA bag at the other. Air for the cylinder was plowed through the impinger containing the strawberry sample and into the 10 L empty PEA bag at a flow rate of 100 mL•min −1 for 100 min aftera pre-purge was conducted by supplying ultra-pure air into the impinger at a flow rate of 100 mL•min −1 for 10 min. Ammonia was analyzed using absorption photometry (Genysys 10 series, Thermo Scientific, Waltham, MA, USA) based on the indophenol method, a well-known approach for amino compounds [18,19]. The reduced sulfur compounds were analyzed using an on-line thermal desorption system (UNITY-Air Server, Markes International, Ltd.) coupled with a GC (CP-3800, Varian, Palo Alto, CA, USA) and pulsed flame photometric detector (PFPD: Varian). The RSCs in the PEA bag were transferred to the TD system using pump and collected into the focusing trap (cold trap) in TD system. The RSCs loaded on the focusing trap were then thermally desorbed and transferred to the GC column for separation and detection. Finally, the dilution-to-threshold (D/T) ratios of the strawberry samples were also determined using an air dilution sensory (ADS) test, as a direct means to assess odor intensity [20].

Major Volatile Components Emitted from Strawberry Samples
The concentration and occurrence frequency of these species is classified in terms of functional groups in Table 1. The types and concentration levels of 147 VOCs detected from all strawberry samples are also summarized in Table 6S. If they are arranged by the chemical grouping and occurrence frequency, they can be classified as follows: (1) ester = 61, (2) alcohol = 21, (3) aldehyde = 11, (4) ketone = 9, (5) fatty acid = 5 and (6) miscellaneous (etc.) = 40.
Esters (n = 47), with total concentration of 52,648 μg•m −3 (76.8% by mass), represented more than half of the 81 VOCs detected at the start of the study (SRD = 0). Similarly, approximately 130 different types of esters have been reported from strawberry fragrances [21], where they were found to represent 25 to 90% of strawberry volatiles [22][23][24]. At SRD-1, esters (n = 38) still recorded the highest concentration (42,713 μg•m −3 ). By day 3, while the total concentration of esters had continued to increase (sum concentration = 705,447 μg•m −3 ) their numbers had fallen down to 31. However, once decay had started, their concentration dropped dramatically down to 533 (SRD = 6) and 45.7 μg•m −3 (SRD = 9). In contrast, alcohols exhibited a reversed trend. Although alcohols were much less abundant in fresh strawberries (sum concentration of SRD-0, 1, and 3 = 1,582 μg•m −3 ), they tended to peak noticeably at 380 mg•m −3 on SRD-6. Thus, the best indicators of the fresh and decayed stages of strawberries are identified as esters and alcohols, respectively. If the other classes of chemicals are considered, aldehydes were detected in all samples and recorded the highest concentration (19,054 μg•m −3 ) on day 3. Ketones and fatty acids were also relatively abundant in fresh strawberries, although they faded away during decay ( Figure 2).
In order to evaluate the indicative fragrance of strawberry, the relative composition (RC) of the strawberry volatiles was assessed initially by normalizing the concentration of an individual compound against the total concentration of all species at each sampling day ( Table 2). If any compound with more than 0.05% of RC (total mass) on one or more sampling day was selected, 53 were observed. The compound contribution pattern of these 53 VOCs was then analyzed both in terms of concentration and odor intensity. The sum concentration for these major VOCs (RC > 0.05%) generally exceeded 99.9% of the total mass of VOCs from each individual measurement (99.4% (SRD-0) to 99.99% (SRD-6).      Although our analysis focused mainly on volatile organics by GC-MS, we also analyzed some offensive odorants like reduced sulfur compounds (RSC) and NH 3 (Table 7S). It can be seen that three RSCs and ammonia were detected from the strawberry sample. Especially, methane thiol and dimethyl disulfide were seen fairly consistently and recorded fairly high concentrations of 267 μg•m −3 and 196 μg m −3 in SRD-1, respectively. In contrast, ammonia was detected apparently only under the decaying conditions (concentration (μg•m −3 ) = 169 (SRD-6) and 445 (SRD-9)) relative to the fresh period below 81.3 (SRD-0, 1, and 3).

The Variety of VOC Threshold Values for Strawberry Volatiles and Their Relationship with Molecular Weights
The odor threshold of a compound is defined as the lowest concentration that can be detected by human olfaction [25]. The lower the odor threshold, the stronger the odorant will be. However, many authors have investigated the threshold values of various volatiles and results for individual compounds can be very variable. In this study, a literature survey was conducted for the odor strengths (thresholds) of the strawberry volatiles. Although we measured a total of 147 VOCs during this study period, we were only able to obtain threshold values for up to of 79 species (Table 6S). The results of this survey are also summarized in Table 8S.
As reported previously, the odor strengths of VOCs tend to exhibit strong relationships with their physicochemical properties, e.g., the number of carbons and molecular weight [26]. Hence, a number of combinations between such variables (e.g., log thresholds vs. molecular weights) were tested to seek for such linear relationship. For this comparative analysis, fatty acids and some miscellaneous groups were however excluded due to the lack of threshold data. As shown in Table 9S, an inverse correlation was seen consistently between the log (thresholds) and molecular weights of VOCs without a single exception. However, the magnitude of slope values differed greatly between the VOC groups, while the strongest correlation with molecular weights was seen from the maximum (out of all available) threshold values.

Figure 3.
Plots of correlation between molecular weight and log (odor thresholds (ppbv)) for alldata (n = 62) and optimal fit (n = 54) of the four major VOC groups (Ester, Alcohol, Aldehyde, and Ketone) emitted from strawberry samples.

The Evaluation of the Odor Strengths with Changes in Freshness Status
In previous sections, the changes of VOC quantities and their threshold values were evaluated from fresh to decayed stages of strawberry. To learn more about strawberry fragrance, our results were examined further with respect to type and strength of strawberry odors. The selection of reasonable threshold value is important to help understand the contribution of a compound at its given concentration level to the overall perception of odor. It is however difficult to assign a single meaningful figure because the threshold of a given compound is often available as multiple reported values.
In this study, the odor strengths of strawberry were calculated in terms of odor activity values (OAV) by dividing the concentrations of the VOCs with the corresponding threshold in the same concentration unit: OAV = concentration (ppbv)/threshold value (ppbv) [27]. For the 53 selected major VOCs, multiple threshold values have been reported for many (21 (one value), 21 (two values), 10 (three values), and 1 (four values)). In case of two or more reported value thresholds, the one with maximum value was used to calculate the OAVs in light of consistency as seen in correlation analysis. Table 3 presents the specific description of odor types for each of the major VOCs with their OAV (n = 53). Information of the odor types was obtained by surveying the GC-olfactometry analysis of VOC [refer to a list of references (n = 24) in Table 3]. The ΣOAV values of the strawberries tended to decrease abruptly with storage time from their maximum at SRD-0 (OAV from SRD-0 to SRD-9 were 12,972, 6,992, 2,524, 196 and 7.37, respectively). The OAV values at SRD-0 decreased in order of ethyl butyrate (6,160), ethyl hexanoate (3,608), ethyl isovalerate (1,592), and ethyl 2-methylbutyrate (942). Table 3. Relationship between relative composition (RP) and odor activity values (OAV) of the major VOCs (n = 53) emitted from strawberry.  To assess the relative contribution of a given compound in terms of OAV, its relative proportion (RP) was also calculated by dividing OAV (a given compound) with ΣOAV (all) ( Table 3). This RP term for OAV is distinguished from the RC term used for relative mass concentration. If OAVs of these four esters are summed, their RP represents 94.8% of total OAV at SRD-0. It thus suggests that the fragrance of fresh strawberries is governed predominantly by these four esters. The scent of these esters is characterized as fruity, apple, and sweet (Table 3). Although their OAVs decreased from SRD-0 to SRD-1, they still recorded the highest OAVs among all the VOCs evaluated at SRD-1 with ΣRP = 91.1%. The fragrance pattern of SRD-0 and -1 is not likely to have changed because the key volatiles (the four esters) remained constant. In case of SRD-3, esters had high OAV along with significantly large ΣRP (97.2%), although their ΣOAV decreased considerably to 2,453 in SRD-3 relative to the earlier period (SRD-0 (12,917) and SRD-1 (6,942)). As a result, we were able to confirm that the esters should dominate the quality of fresh strawberry fragrance (SRD-0, 1, and 3).
To evaluate the occurrence patterns of fresh strawberry volatiles, our results were compared to a number of previous studies. Du et al. [1] analyzed the volatiles emitted from fresh strawberries using solid-phase micro-extraction (SPME)-GC-MS analysis. A total of 54 volatiles from two cultivar samples (Strawberry Festival and Florida Radiance) were selected as the main target compounds (with their respective standards). It is interesting to note that 52 target volatiles were found in both samples, while 23 of them were seen consistently in all fresh stage samples in this study. They further calculated the OAVs of detected volatiles using thresholds of 44 compounds. Accordingly, OAVs were seen to be  (SRD-3)). Likewise, methyl butyrate also showed high OAV with the RP values of 1.30% (SRD-0), 2.79% (SRD-1), and 0.50% (SRD-3). In contrast, the RP values of DMHF and linalool were relatively insignificant, although they were selected as the major VOCs in this study.
During decay, ΣOAVs of all detected esters (except ethyl 2-methylbutyrate with OAV = 46.9) were below 10, while accounting for 12.9% by ΣRP (ester (n = 12)). In contrast, isobutyl alcohol exhibited the highest OAV of 114 in SRD-6 (RP = 58.3%). As the odor of isobutyl alcohol is characterized as plastic and bad, is is distinguishable from pleasant fresh scents. As a result, odor intensity decreased with the progress of strawberry decay with the emergence of some offensive odors (e.g., ΣOAV of 7.37 in SRD-9). In SRD-9, only two compounds (acetaldehyde and isovaleraldehyde) showed OAVs above 1. For the reader's reference, volatiles that showed up at least once in terms of either OAV (above 100) or RP (above 5%) during the whole study period are illustrated in Figure 4.

Figure4.
Comparison between (odor activity value) OAV and relative proportion (RP) of the major VOCs which had OAV and RP of above 100 and 5%, respectively at storagetime of 0, 1, 3, 6, and 9 days.  If the OAVs of the RSCs and ammonia detected in strawberry aroma profiles are examined, their values from decayed strawberry samples (SRD-6 and 9) generally had very low OAV (below 1). Only methanethiol had high OAVs above 10 in SRD-0 and SRD-1 (OAV = 14.9 (SRD-0) and 64.6 (SRD-1)] compared with other RSCs or ammonia. However, if these results are compared with ΣOAV (all) of all hydrocarbons in fresh strawberry samples (SRD-0 and 1), the RP value of methane thiol were as low as 0.13% (SRD-0) and 1.04% (SRD-1). As a result, it is reasonable to infer that the volatile hydrocarbons should represent the odor/fragrance characteristics of strawberry most effectively.

Comparison between Odor Activity Value (OAV) and Dilution-to-Threshold Ratio
In this section, the TD-GC-MS odor profiles were compared between fresh and decaying strawberries. The results were then evaluated to assess the relationship between the classes of volatile components and their odor intensity. To estimate the key volatiles which dominate the strawberry scent, the odor strengths (OAV) were examined against the relative (mass) composition. Evaluation of our data indicated that the use of ΣOAV was useful to assess the actual occurrence of fragrance/odor from strawberry samples. As another means to explore the odor intensities of strawberry samples, we estimated the dilution-to-threshold (D/T) ratio derived experimentally based on air dilution sensory (ADS) test [20]. The D/T ratio is commonly used as a tool to measure the level of dilution by which the odor threshold is recognized [20]. As the D/T ratio of strawberry samples was measured concurrently with the analysis of their chemical composition, the ΣOAV values at each SRD interval can also be evaluated in relation to the D/T ratio.
The results of correlation analyses between storage day and log odor intensities (log ΣOAV and log D/T ratio) are plotted in Figure S3(A). In compliance with general expectation, the log D/T ratio decreased with the progress of decay to show the slope values of −0.2185 (with R 2 = 0.8646 and p-value = 2.21E−2) (log D/T ratio). A similar trend is also observable from ΣOAV. As such, the results of this correlation analysis between the log ΣOAV and log D/T ratio show a strong correlation with R 2 = 0.9338 (p-value = 7.33E−3). As a result, the computation of ΣOAV values can be used as sensitively as that of the D/T ratios to assess the fragrance occurrence pattern. In other words, the ΣOAV should be considered a good criterion to assess the fragrance or odor intensity of strawberry samples.

Conclusions
The volatiles emitted from strawberry are important components to accurately assess its fragrance. Although the fragrance of the strawberry can influence its flavor and taste, its smell types can vary greatly with aging conditions. If the volatiles from strawberry were analyzed at the fresh stage, the results should be useful enough to understand the natural flavor of strawberries. In this study, the mass concentrations and odor strength (odor activity values: OAV) from strawberries were analyzed by the sorbent tube method at storage times of 0, 1, 3, 6 and 9 days at 25 °C.
In this study, we were able to quantify a total of 147 strawberry volatiles (with 79 corresponding odor threshold values from literature survey). In order to assess the fragrance/odorant characteristics of strawberries, the OAV values were calculated for a total of 53 major volatiles that comprised more than 0.05% in mass concentration of all strawberry volatiles. The OAV values decreased abruptly with storage time (ΣOAV: SRD-0 (12,972) to SRD-9 (7.37)). If the ΣOAV values of a given strawberry volatiles are computed for the whole fresh period, their magnitude was dominated by four esters with fruity and sweet scents [(1) ethyl butyrate (11,422), (2) ethyl hexanoate (4,395), (3) ethyl isovalerate (2,751), and (4) ethyl 2-methylbutyrate (2,290)] [ΣRP (four esters) of 92.8%]. However, relative dominance of esters as the key strawberry fragrance changed abruptly with the onset of decay. Hence, as the transition proceeds, relationships between key parameters tended to vary widely. In case of SRD-6, the ΣOAV of the four esters decreased to 60.9 with ΣRC of below 1%, while isobutyl alcohol exhibited the highest OAV of 114 with RC of 75.3%. The ΣOAV was reduced further to reach the minimum value of 7.37 in SRD-9.
In this study, the strawberry volatiles were analyzed from fresh stage to 9 day storage at 25 °C, and the concentration of the strawberry volatiles were evaluated in relation to their occurrence patterns and olfaction data derived as D/T ratio. The overall results of our study suggest that strawberry volatiles are useful indicators to characterize the flavor changes of strawberry at the latter stages of its storage period.