Determining the Geographical Origin of Fuji Apple from China by Multivariate Analysis Based on Soluble Sugars, Organic Acids, and Stable Isotopes

Zhang, Hui; Nie, Jiyun; Tong, Zhaoguo; Li, Jing; Li, An; Shen, Youming; Zhang, Jia; Kuang, Lixue

doi:https://doi.org/10.1155/2022/5415257

Journal of Food Quality

On this page

Abstract Introduction Materials and Methods Results Discussion Conclusions Data Availability Conflicts of Interest Acknowledgments References Copyright Related Articles

Special Issue

Multiple Spectrum Analysis for Food Quality and Safety

View this Special Issue

Research Article | Open Access

Volume 2022 | Article ID 5415257 | https://doi.org/10.1155/2022/5415257

Determining the Geographical Origin of Fuji Apple from China by Multivariate Analysis Based on Soluble Sugars, Organic Acids, and Stable Isotopes

Hui Zhang,¹Jiyun Nie,²Zhaoguo Tong,³Jing Li,¹An Li,⁴Youming Shen,¹Jia Zhang,¹and Lixue Kuang¹

Academic Editor: Yong-Jie Yu

Received14 Jan 2022

Accepted23 Mar 2022

Published28 Apr 2022

Abstract

The aim of this study was to explore the regional characteristics of soluble sugars, organic acids, and stable isotopes (δ²H, δ¹⁸O, and δ¹³C) in Fuji apple and the viability of tracing the geographical origin. Totally, 181 Fuji apple samples from 2017 and 2018 from three main apple production regions in China, Bohai Bay (BHB), Loess Plateau (LP), and Northwest region (NW) were collected. The parameters of soluble sugars, organic acids, and stable isotopes in samples were analyzed with HPLC, IC, and IRMS, respectively. The results of regional difference analysis, multiway variance analysis, and correlation analysis indicated that sorbitol (Sor), glucose (Glu), fructose (Fru), sucrose (Sucr), δ²H, and δ¹³C can be used to distinguish the samples from the three regions. Stepwise linear discriminant analysis (SLDA) showed that the correct discriminant rate of samples from the advantageous production areas of apples in China (BHB and LP) was 82.2%, and the most effective indexes were Glu, Fru, Sucr, and δ²H. Moreover, satisfactory classification can be achieved in samples from BHB and NW, with a correct classification rate of 90.0%, and Sor, Glu, and Fru were included in the discrimination model. Furthermore, the validity of the discriminant model was verified by the prediction set. The study also found that organic acids were not suitable to distinguish the apple samples from the three regions. In addition, soluble sugars and stable isotopes could not effectively distinguish LP and NW samples, which was also the reason that the samples from the three main apple production regions could not be distinguished well.

1. Introduction

In recent years, due to the occurrence of food safety issues such as heavy metal pollution, pesticide residues, adulteration, and fake and inferior agricultural products [1], consumers want to know more relevant information to confirm the safety of the food they purchase [2]. In this situation, consumers are more inclined to buy agricultural products with a clear geographical origin, especially certified foods, such as agricultural products with geographical indications, which comply with relevant standards [3, 4]. However, some unauthorized operators may use false labels to impersonate agricultural products with clear geographical indications in order to obtain higher profits, as the agricultural products with guaranteed origin will charge a high premium, which will damage the interests of operators and consumers [5]. Therefore, the development of suitable analytical techniques for tracing the origin of agricultural products is highly desirable for consumers, production operators, and government regulators [6].

Due to the different principles and characteristics, especially the limitations of technology in the application, it is difficult to accurately trace the origin of agricultural products with only one traceability technology [7–9]. For example, stable isotope ratios analysis technology is an increasingly accepted tool for tracing the geographical origin of agricultural products, but the isotopic composition is mainly affected by geographic and climatic conditions, such as latitude, distance from the sea, altitude, temperature, and precipitation [10]. Therefore, the isotopic compositions of agricultural products from areas with similar climates and topography may not be effectively distinguished [5]. Using organic fingerprints to trace the origin of agricultural products, the discriminant rates were varied greatly due to the selected indexes. Therefore, screening out the organic ingredients which are unique or closely related to environmental conditions is the key to use organic compounds fingerprint analysis technology to trace the origin of agricultural products [11]. Therefore, it is the current development trend to use multiple technologies and multiple indexes to trace the origin of agricultural products. For instance, by combining the stable isotopes with volatile compounds, organic compounds, and multiple elements, the geographical origin of potatoes, kiwifruit, apple juices, and red wine can be well classified, and the total correct discrimination rate was varied from 83.9% to 100% [8, 12–14].

The cultivation of apples in China is mainly distributed in the advantageous production areas of apple, Bohai Bay (BHB), and Loess Plateau (LP), including the seven major apple-producing provinces of Shaanxi, Shandong, Henan, Shanxi, Hebei, Liaoning, and Gansu. In 2018, the apple cultivated area and the yield in BHB and LP accounted for 84.22% and 89.12% of the total cultivated area and yield in China [15]. Xinjiang and Ningxia are characteristic apple-producing areas in China and belong to the Northwest region (NW). In recent years, the cultivation area and yield of apples in the NW have continued to increase; moreover, with the changes in the layout of apple production, Xinjiang has gradually become an important apple-producing area in China. Furthermore, the Fuji apple is the main cultivated variety, and its cultivation proportion accounts for 72.7% of the total cultivated area [16]. As we all know, the quality and price of agricultural products will be different due to their geographic origin [17]; in addition, studies have found that the physiological quality of apples was related to the latitude, longitude, and altitude of the growing area [18]. For example, fructose has the highest proportion in the soluble sugars in Fuji apple from Shaanxi Province which belongs to the LP [19], with a higher price than in other provinces; moreover, apples in some specific areas of Shaanxi have been recognized as “protected designation of origin” (PDO) products [20]. Therefore, tracing the origin of the Fuji apple in the three main apple production regions in China is of great significance to achieving the traceability of apple and protecting the regional advantages of agricultural products. What’s more, there has not been literature that specifically studied the use of organic compounds and stable isotope fingerprint characteristics to trace the origin of the Fuji apple.

Since the isotope ratios of carbon, hydrogen, and oxygen were altitude-, latitude-, and climate-dependent, these three stable isotope ratios can provide information about the geographic origin and metabolism of plants, so the isotopic ratios of carbon, hydrogen, and oxygen are suitable for tracking the geographical origin of agricultural products related to regional climate conditions [17, 21–23]. In addition, soluble sugars and organic acids were important nutrients in apples, as well as the main flavor compounds in apples. Moreover, studies have found that soluble sugars and organic acid compounds in Fuji apple were different between regions [19]. Therefore, the soluble sugars, organic acids, and stable isotopes (δ²H, δ¹⁸O, and δ¹³C) will be used to study the effectiveness of multi-index in tracking the origin of Fuji apples from the three main apple producing areas (BHB, LP, and NW) in China. It is expected to provide references for tracing the origin of apples.

2. Materials and Methods

2.1. Sample Collection

A total of 181 Fuji apple samples were collected from the 2017 and 2018 harvests from three main apple production regions in China, BHB (Hebei, Shandong, and Liaoning provinces), LP (Shaanxi, Gansu, Shanxi, and Henan provinces), and NW (Ningxia and Xinjiang autonomous regions). For each sample, approximately 5 kg of fruit was collected at the harvest maturity stage from September to November. Samples were transported to the laboratory 3 days after collection. The information on samples was shown in Table 1.

2.2. Sample Preparation

About 2.0 kg of fruit was ground into a powder in liquid nitrogen by a SPEX Sample Prep system (New Jersey, USA) and stored at −20°C for the determination of soluble sugars and organic acids. For the rest of the samples, wipe the surface of the peel and core; then the sample was homogenized by a homogenizer (JT C, Jintian, China), which was used for the determination of stable isotope ratios. The extraction of soluble sugars and organic acids was carried out according to the previously reported method [24, 25] and the extracted solution was stored at −20°C and analyzed by the instrument later. The homogenized sample was placed in the sample chamber of a vacuum freeze dryer (VFD-1000, Bilon, China); then, the sample was freeze-dried for 72 hours under the instrument conditions of −40°C and under 15pa. The dried sample was passed through a sieve (100 mesh, Huafeng, China) immediately. Finally, a powdered sample was obtained to detect the isotope ratio ¹³C/¹²C in samples.

2.3. Instrument Conditions

2.3.1. High-Performance Liquid Chromatography (HPLC) Measurement Conditions [25,26]

After filtered through a 0.22 μm pore filter (Fine, Japan), the extracted solution was used to detect the organic acid compounds directly, using HPLC (LC-10A, Shimadzu, Japan), with SPD-10A UV-VIS detector, and a C18 column (Ultimate LP-C18, 4.6 mm × 300 mm, 5 μm, Ultimate, China). A total of five organic acid compounds, including oxalic acid (Oxa), tartaric acid (Tar), quinic acid (Qui), malic acid (Mal), and shikimic acid (Shi), were identified by comparison with the standards. The calibration curves were prepared by plotting different concentrations ranging from 0.05 to 50 mg/L (for Mal, the concentrations ranging from 25 to 400 mg/L) of standards versus the area measurements in UPLC (R2 ≥ 0.9992). Results were expressed as mg/g. The reproducibility of the chromatographic separation of the organic acid compounds was carried out according to the previously reported method (Wu et al., Wang). The results of the reproducibility of chromatographic separation for organic acids were expressed as relative standard deviation (RSD%) as follows: 0.63 for Oxa, 1.41 for Tar, 0.43 for Qui, 0.62 for Mal, and 0.81 for Shi. The content of total acid in the samples was the sum of the contents of the five organic acids.

2.3.2. Ion Chromatograph (IC) Measurement Conditions [24, 25]

The soluble sugars were detected by an Ion Chromatograph (ICS-5000, Dionex, USA) with a conductivity detector, an anion exchange analytical column (Dionex CarboPacTM PA10, 4 mm × 250 mm, Thermo Fisher Scientific, USA), and a guard column (IonPac AG23, 4 mm × 50 mm, Dionex, USA). The extracted solution after being diluted 50 times with deionized water can be used to measure sorbitol (Sor), glucose (Glu), fructose (Fru), and sucrose (Sucr) in the samples. The soluble sugars were identified by comparison with the standards. The calibration curves were prepared by plotting different concentrations ranging from 0.5 to 100 mg/L of standards versus the area measurements in IC (R² ≥ 0.9965). The result was expressed as mg/g. The results of the reproducibility of chromatographic separation for soluble sugars, expressed as RSD%, were as follows: 1.75 for Sucr, 1.52 for Glu, 1.69 for Fru, and 0.37 for Sor. The content of total sugar in the samples was the sum of the contents of the four soluble sugars.

2.3.3. Isotopic Ratio Mass Spectrometer (IRMS) Measurement Conditions

(1)Stable Hydrogen and Oxygen Isotope Determination Conditions [27,28]. About 2 g of the homogenized sample was weighed and placed in a glass sample bottle. The fully automatic vacuum condensed water extraction system (LI-2100, LICA, China) was used to extract the water in the samples; then, the extracted water was put into the automatic sampling tray of the element analyzer (Flash HT2000, Thermo Fisher Scientific, USA). The elements of H and O in the samples were converted into H₂ and CO by splitting at 1,400°C. After passing through the chromatographic column, H₂ and CO successively entered the stable isotope mass spectrometer (MAT 253, Thermo Fisher Scientific, USA) to detect the isotope ratios ²H/¹H and ¹⁸O/¹⁶O. The flow rates of helium carrier gas and helium purge gas were both 100 mL/min.(2)Stable Carbon Isotope Determination Conditions [27,28]. About 0.8 mg of the powder sample was weighed, which was wrapped in a tin cup, and placed in the element analyzer. The element of C in the sample was converted into CO₂ in a combustion furnace at 960°C. After passing through the chromatographic column, CO₂ entered the stable isotope mass spectrometer for the determination of the isotope ratio ¹³C/¹²C. The flow rates of helium carrier gas and helium purge gas were both 100 mL/min.(3)The Calculated Stable Isotope Ratios. The isotope values were calibrated against standard materials, which were Vienna Standard Mean Ocean Water (V-SMOW), USGS45 (δ²HVSMOW = −10.3‰, δ¹⁸OVSMOW = −2.238‰), USGS46 (δ²HVSMOW = −235.8‰, δ¹⁸OVSMOW = −29.8‰), and USGS40 (δ¹³CVPDB-LSVEC = −26.39‰, δ¹⁴NAIR = −4.52‰). According to the equation, R was the isotope ratio of heavy isotope to light isotope abundance in the samples (¹⁸O/¹⁶O, ²H/¹H, and ¹³O/¹²O). R_sample and R_standard were the isotope ratio of the samples and the international standards. Results were referenced to Vienna Pee Dee Belemnite (V-PDB) for δ¹³C and Vienna Standard Mean Ocean Water (V-SMOW) for δ²H and δ¹⁸O. The analytical precision was 2‰, 0.4‰, and 0.2‰ for H, O, and C, respectively. The stable isotope ratios were calculated as follows:

2.4. Statistical Analysis

When data conformed to both normality and homogeneity assumptions, a regional different analysis was performed by one-way analysis of variance (ANOVA) ( value < 0.05), and the Pearson correlation analysis was used to investigate the relationship between indexes and regional conditions. When data conformed to the normality assumption but failed on the homogeneity of variances, analysis of variance was performed using Welch’s test. For not normally distributed data, Kruskal-Wallis (K-W) was used for the analysis of regional differences, and Spearman correlation analysis was used to investigate the relationship between indexes and regional conditions. Multiway ANOVA was applied was to quantify the contributions of geographical origin, harvest year, and their interactions to the total variance in stable isotope ratio, soluble sugars, and organic acids levels. In addition, the sample set was randomly split up into two groups, with three-quarters of the total samples as a training set and the remaining one-quarter of samples as a prediction set. According to F-value, the training set was used to create the discriminant model by Linear Discriminant Analysis (LDA), and the predictive ability of the discriminant model was estimated by the prediction set. Stepwise linear discriminant analysis (SLDA) was used to select necessary and optimal indexes to incorporate into the discriminant model, and the coefficients of the indexes in the discriminant model were standardized coefficients. The statistical analysis of the data was performed with SPSS 26.0 software (SPSS, IBM Corp., USA).

3. Results

3.1. Analysis of Soluble Sugars in Apples

The proportion of each sugar component in samples from high to low was Fru, Sucr, Glu, and Sor. For BHB, LP, and NW samples, the proportion of Fru to total sugars was 45.93%, 49.65%, and 49.33%, respectively (Figure 1(a)), which was the main soluble sugar component, in addition, the proportion of Sucr (26.44%, 23.03%, and 28.36%, respectively) was higher than Glu (23.12%, 22.51%, and 16.19%, respectively), and the lowest content was Sor (4.51%, 4.81%, and 6.12%, respectively). Moreover, the radar chart showed that the soluble sugar content in samples from different regions had different patterns (Figure 2(a)).

(a)

(b)

(a)

(b)

(c)

Regional differences of soluble sugars were shown in Table 2, and the contents of soluble sugars were expressed as means ± standard deviation (SD). In LP samples, the average content of Fru was 59.5 ± 12.8 mg/g FW, which was higher than in BHB and NW samples, which are 54.3 ± 9.41 mg/g FW and 55.2 ± 13.3 mg/g FW, respectively. The average content of Sucr in LP samples was 27.4 ± 9.21 mg/g FW, which was lower than BHB and NW samples (31.6 ± 9.34 mg/g FW and 32.2 ± 10.0 mg/g FW, respectively). The contents of Fru and Sucr in the LP samples were significantly different from the BHB and NW samples (), and the content of Sucr was significantly different in LP and NW samples (), while the differences in Fru and Sucr contents in BHB and NW samples were not significantly (). The content of Glu in NW samples was significantly different from that in BHB and LP samples (). In addition, the content of Sor was significantly different between the BHB and NW samples ().

3.2. Analysis of Organic Acids in Apples

Different from soluble sugars, the content characteristics of organic acids had similar patterns among regions, and the content of organic acids in descending order was Mal, Qui, Oxa, Tar, and Shi (Figure 2(b)). In the samples from BHB, LP, and NW, Mal accounted for 85.18%, 85.66%, and 88.73% of the total acids, respectively (Figure 1(b)), indicating that Mal occupies an absolute advantage in the total acid content.

Regional differences of organic acids are shown in Table 2, and the contents of organic acids were expressed as means ± standard deviation (SD). Mal was the main organic acid component in Fuji apple; the average content of Mal in BHB, LP, and NW samples was 4.11 ± 1.31 mg/g FW, 3.70 ± 0.828 mg/g FW, and 3.85 ± 1.20 mg/g FW, respectively; however, the differences in the three regions were not significant (). The average content of Qui in the BHB samples was significantly different from NW samples (). The content of Tar in NW samples was significantly different from that in BHB and LP samples (). The average content of Oxa and Shi in samples from the three main production regions was not significantly different ().

3.3. Analysis of Stable Isotopes in Apples

The characteristics of the δ¹⁸O and δ¹³C values in the samples from the three main apple-producing areas showed similar patterns, while the δ²H value in the BHB samples was higher than that in NW and LP samples (Figure 2(c)). The mean δ²H value of the BHB samples was −39.6‰, which was higher than that of the LP and NW samples (−44.6‰ and −44.4‰, respectively), and the results of ANOVA showed that the δ²H value of BHB samples was different from LP and NW samples, and the difference was significant (). In addition, there was no significant difference in the mean δ²H values between LP and the NW samples (). The distribution of δ¹⁸O values in the samples was similar to the δ²H value; the mean δ¹⁸O value of BHB samples was higher than that of LP and NW samples, which were −4.67‰, −5.26‰, and −5.22‰, respectively. The results of ANOVA showed that the mean δ¹⁸O value in the BHB samples was significantly different from the LP sample (). In addition, the mean δ¹⁸O value of NW samples was not significantly different from BHB and LP samples (). The distribution of the δ¹³C value of samples roughly showed that the δ¹³C value of the NW samples was higher than that of LP samples, while the δ¹³C value of the BHB samples was the lowest, and the mean δ¹³C values of NW, LP, and BHB samples were −25.5‰, −25.7‰, and −26.3‰, respectively. The results of ANOVA analysis showed that the mean δ¹³C value of BHB samples was different from the LP and NW samples, and the difference was significant (). In addition, the mean δ¹³C value of LP and NW samples was not significantly different () (Table 2).

3.4. Multiway ANOVA

A combined analysis of variance of two years and three geographical origins was performed using the general linear model (GLM). Harvest year and geographical origin were considered as the fixed factors. The effects were portioned into different sources: year, region, and region × year (Table 3). Oxa was not statistically significant in this analytical model (). Sor, Glu, Fru, Sucr, Tar, Qui, Mal, Shi, δ²H value, and δ¹⁸O value were significantly influenced by year (). Sor, Glu, Fru, Sucr, δ²H value, and δ¹³C value were significantly affected by region (). Qui, Mal, and δ¹⁸O values were significantly affected by region (). Region × year had a significant effect on Sucr () and δ¹³C (). The results showed that geographical origin was the most important source of the concentration variations of δ¹³C value, and the contribution percentage was 44.76%. The harvest year was the most source of variation for soluble sugars, organic acids, δ²H, and δ¹⁸O.

3.5. Correlation between Indexes and Regional Conditions

To identify the relationship between indexes and regional conditions (longitude, latitude, altitude, mean annual temperature, and annual precipitation), correlation analysis was employed to assess whether or not positive/negative correlation patterns existed. The results showed that longitude was significantly negatively correlated with Sor (r = −0.150, ) and δ¹³C value (r = −0.244, ) had a significant positive correlation with Glu (r = 0.320, ) and δ²H value (r = 0.228, ) (Table 4). Latitude was negatively correlated with Glu (r = −0.328, ) and Fru (r = −0.205, ) and significantly positively correlated with Sucr (r = 0.189, ). Altitude was significantly positively correlated with Sor (r = 0.290, ) and δ¹³C value (r = 0.336, ) and was correlated with Glu (r = −0.326, ) and Tar (r = −0.170, ) showing a significant negative correlation. The temperature was significantly positively correlated with Glu (r = 0.199, ) and was significantly negatively correlated with Sucr (r = −0.222, ) and Mal (r = −0.162, ). There was a significant negative correlation between precipitation and Sor (r = −0.196, ) and δ¹³C value (r = −0.178, ), and there was a significant positive correlation with Glu (r = 0.398, ), Tar (r = 0.160, ), and δ²H value (r = 0.181, ). In addition, in the sample collection area, Oxa, Qui, Shi, and δ¹⁸O values had no correlation with the regional conditions studied. Therefore, geographical location has a significant impact on most soluble sugars, organic acids, and stable isotopes studied (), which provides a basis for distinguishing the samples from different geographical regions by using soluble sugars, organic acids, and stable isotopes.

3.6. Linear Discriminant Analysis (LDA)

In order to explore the effectiveness of soluble sugars, organic acids, and stable isotopes in tracing the origin of the Fuji apple in China, the indexes had significant differences between regions () and were significantly affected by regional factors (); besides, the selected indexes were significantly correlated with the regional conditions. Finally, Sor, Glu, Fru, Sucr, δ²H, and δ¹³C were selected for stepwise discriminant analysis (SLDA) of BHB, LP, and NW samples. About three-quarters of the samples (136 samples) were used as training set and the remaining 45 samples, BHB (11 samples), LP (19 samples), and NW (15 samples), were used as prediction set to verify the stability of the discriminant models.

SLDA was used to classify Fuji apple samples from BHB, LP, and NW. The classification of the training set by the canonical discriminant functions was shown in Figure 3; it can be seen that BHB, NW, and LP samples were overlapped, mainly because the LP samples were misjudged as BHB and NW samples, and the correct classification rate and cross-validation rate were only 69.1% and 67.6% (Table 5); in addition, Sor, Glu, Fru, Sucr, and δ²H, which had a significant impact on sample discrimination (), were included in the discriminant model, and the correct discrimination rate of the discriminant model for the prediction set was only 75.6% (for BHB, NW, and LP samples was 81.8%, 80.0%, and 68.4%, respectively) (Table 6). In addition, it was found that the classification effect of the samples from two producing areas was better (Figure 3). So, SLDA was performed on samples from the two production regions to explore the effectiveness of the combination of soluble sugars and stable isotopes in distinguishing the regions of samples.

(a)

(b)

(c)

(d)

Glu, Fru, Sucr, and δ²H were selected by SLDA to classify Fuji apple samples from the advantageous production areas of apple in China (BHB and LP). The correct classification rate and cross-validation rate of the training set of BHB and LP samples were both 82.2% (Table 5), and the correct discrimination rate of the prediction set was 80.0% (Table 6). When SLDA was used to classify Fuji apple samples from BHB and NW, a satisfactory classification was obtained, and the correct classification rate of the training set and cross-validation rate were 90.0% and 86.3%, respectively (Table 5); moreover, the discriminant model, including Sor, Glu, and Fru, had a correct discrimination rate of 84.6% for prediction set (Table 6). However, it was found that the samples from the LP and NW were not easy to be distinguished, and the correct classification rate of the training set was only 71.6% (Table 5), which was the reason that the classification of samples from the three production regions was not satisfactory.

4. Discussion

In this study, Fru was the main soluble sugar in samples, accounting for 45.93%–49.65% of the total sugar content, and Mal was the main organic acid, accounting for 85.18%–88.73% of the total acid, which was consistent with previous studies [24,29]. For apple fruits, the formation and accumulation of the flavor substances (soluble sugars and organic acids) were closely related to ecological and climatic conditions such as temperature, light intensity, and altitude. In this study, LP (Shaanxi, Gansu, Shanxi, and Henan) is next to BHB (Shandong, Hebei, and Liaoning) and NW (Ningxia and Xinjiang). The altitude and annual average temperature of provinces on the LP are similar to those of BHB and NW, such as Gansu, Ningxia, and Xinjiang. This may cause the content of soluble sugar in the LP sample to be similar to the NW and BHB samples, making it difficult to be classified from BHB and NW samples. In fact, acid metabolism and regulation in apples are very complex, and the acid content is jointly regulated by the relevant genes and enzymes [30]. In addition, it is also affected by many factors such as environment and cultivation conditions. For example, a certain degree of deficit irrigation will affect the content of total acid in fruit [31,32]. Moreover, fruit acidity rather than sweetness was likely to have undergone selection during apple domestication [26]. In this study, it was found that the organic acids were not suitable for distinguishing the Fuji apple samples from these three producing regions. Therefore, it was speculated that organic acids are less affected by temperature, light, and altitude than soluble sugars, which was consistent with the results of correlation analysis of soluble sugars, organic acids, and regional conditions (Table 4).

The fractionation of hydrogen isotopes was affected by distance from the sea, temperature, rainfall, and atmospheric humidity. In this study, BHB (including Hebei Province) is located on the east coast of China and will be affected by coastal climate, so the δ²H value will be different from the samples in inland areas (LP and NW), which can be used to distinguish samples from coastal and inland areas, even from different coasts [10,28]. In addition, the altitude of BHB was lower than LP and NW; these factors will lead to the enrichment of heavy isotopes of hydrogen in BHB samples [33]. Therefore, the decreasing order of δ²H value in samples from the three production regions was BHB (−39.6‰), NW (−44.4‰), and LP (−44.6‰), and the δ²H value in BHB samples was significantly different from LP and NW samples (). Oxygen isotopes in plants were affected by latitude, distance from the sea, altitude, and degree of evaporation (mainly affected by humidity and temperature) [34–36]. In addition, in samples in coastal areas, the δ¹⁸O value was high [37], so it was speculated that the content characteristics of the δ¹⁸O value in the three production regions were consistent with δ²H. The study found that the decreasing order of δ¹⁸O value was BHB (−4.67‰), NW (−5.22‰), and LP (−5.26‰), and there was a significant difference between BHB and LP samples (), but NW samples had no significant difference with BHB and LP samples (), possibly because the increase of solar radiation and the decrease of precipitation in NW lead to the enrichment of oxygen heavy isotopes in samples [36]; furthermore, as the hydrogen isotopes in plants only come from water, the coastal climate had a greater impact on δ²H than δ¹⁸O [10]. Therefore, the variation in δ¹⁸O values within NW samples was observed. The δ¹³C value in plants is related to atmospheric photosynthetic carbon sequestration during growth and some environmental factors [38], such as water availability, relative humidity, light intensity, and temperature in the plant environment (atmospheric and soil water) [36,39,40]. In this study, the altitude of NW was higher than LP, and the altitude of BHB was the lowest, and the average δ¹³C value roughly conforms to the characteristics that it increases with the increase of altitude. The average δ¹³C value of samples from high to low was NW (−25.5‰), LP (−25.7‰), and BHB (−26.3‰). This phenomenon was considered to be related to the physiological changes of plants adapting to their growth conditions and the changes of carbon dioxide content with the rise of sea level [41–43].

The fingerprint characteristics of soluble sugars, organic acids, and stable isotopes formed under the influence of geographic and climatic conditions tend to distinguish samples without LP samples, and the correct discrimination rates of the training set were above 82.2%; furthermore, the correct discrimination rates of prediction set were above 80%. The storage time and the sample preparation for detecting sugars and acids in 2017 samples were different from 2018 samples, which may increase the differences in indexes between years (except δ¹³C value); however, the correct discrimination rates of the prediction set were similar to the training set, indicating that although the year may be the main source of variance, there is a variation of soluble sugar, organic acid, and δ²H value, and the discriminant model still has a certain stability. In addition, organic acids were not suitable for distinguishing the samples from the three main apple production regions, while soluble sugars and stable isotopes were not satisfactorily distinguishing LP and NW samples; therefore, in the following research, it is necessary to add indexes, suitable for distinguishing LP and NW samples, to improve the correct discrimination rate of samples from the three production regions.

5. Conclusions

In this study, as the soluble sugar, organic acid, and stable isotope studied were closely related to the regional conditions, which lead to the poor discrimination rate of LP and NW samples, the correct discrimination rate of the samples from the three main apple production regions was only 69.1%. Nevertheless, it can still be used for many specific applications, such as Glu, Fru, Sucr, and δ²H which can be used to distinguish the samples from the advantageous production areas of apple in China (BHB and LP), and the correct discrimination rates of the training set and prediction set were 82.2% and 80.0%, respectively. In addition, BHB and NW samples can be satisfactorily distinguished by Sor, Glu, and Fru, and the correct discrimination rates of the training set and prediction set were 90.0% and 84.6%, respectively.

Data Availability

Data are included within this article.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Acknowledgments

This research was supported by the Earmarked Fund for China Agriculture Research System (No. CARS-27), the Agricultural Science and Technology Innovation Program of Chinese Academy of Agricultural Sciences (No. CAAS-ASTIP), and the National Program for Quality and Safety Risk Assessment of Agricultural Products of China (No. GJFP2018003).

References

Y. m. Wei, B. l. Guo, S. Wei, S. Sun, and H. y. Zhao, “The principle of food geographical origin traceability and authenticity technique,” Scientia Agricultura Sinica, vol. 45, no. 24, pp. 5073–5081, 2012, in Chinese.
View at: Google Scholar
S. Hunt and L. J. Frewer, “Trust in sources of information about genetically modified food risks in the UK,” British Food Journal, vol. 103, no. 1, pp. 46–62, 2001.
View at: Publisher Site | Google Scholar
C. f. Zeng, L. f. Zhang, J. d. Xu, J. Shu, and D. h. Liu, “Research progress in traceability technology about geographical origin of agricultural products,” Science and Technology of Food Industry, vol. 34, no. 06, pp. 367–371, 2013, in Chinese.
View at: Google Scholar
X. Zhao, M. Kneafsey, and D. Finlay, “Food safety and Chinese geographical indications,” British Food Journal, vol. 118, no. 1, pp. 217–230, 2016.
View at: Publisher Site | Google Scholar
S. A. Wadood, G. Boli, Z. Xiaowen, I. Hussain, and W. Yimin, “Recent development in the application of analytical techniques for the traceability and authenticity of food of plant origin,” Microchemical Journal, vol. 152, Article ID 104295, 2020.
View at: Publisher Site | Google Scholar
D. M. A. M. Luykx and S. M. Van Ruth, “An overview of analytical methods for determining the geographical origin of food products,” Food Chemistry, vol. 107, no. 2, pp. 897–911, 2008.
View at: Publisher Site | Google Scholar
F. Longobardi, A. Ventrella, G. Casiello et al., “Instrumental and multivariate statistical analyses for the characterisation of the geographical origin of Apulian virgin olive oils,” Food Chemistry, vol. 133, no. 2, pp. 579–584, 2012.
View at: Publisher Site | Google Scholar
Y. y. Ma, B. l. Guo, Y. m. Wei, and S. Wei, “Identification of Kiwi fruit geographical origin using stable isotopes and organic compounds,” Journal of Chinese Institute of Food Science and Technology, vol. 15, no. 08, pp. 238–244, 2015, in Chinese.
View at: Google Scholar
A. Mahne Opatić, M. Nečemer, S. Lojen, and J. Masten, “Determination of geographical origin of commercial tomato through analysis of stable isotopes, elemental composition and chemical markers,” Food Control, vol. 89, pp. 133–141, 2018.
View at: Google Scholar
L. Bontempo, F. Camin, R. Larcher, G. Nicolini, M. Perini, and A. Rossmann, “Coast and year effect on H, O and C stable isotope ratios of Tyrrhenian and Adriatic Italian olive oils,” Rapid Communications in Mass Spectrometry, vol. 23, no. 7, pp. 1043–1048, 2009.
View at: Publisher Site | Google Scholar
H. Shi, S. Liu, and X. g. Zhao, “Application of stable hydrogen and oxygen Isotope in water circulation,” Journal of Soil and Water Conservation, vol. 17, no. 02, pp. 163–166, 2003, in Chinese.
View at: Google Scholar
D. McCarroll and N. J. Loader, “Stable isotopes in tree rings,” Quaternary Science Reviews, vol. 23, no. 7-8, pp. 771–801, 2004.
View at: Publisher Site | Google Scholar
F. Longobardi, G. Casiello, D. Sacco, L. Tedone, and A. Sacco, “Characterisation of the geographical origin of Italian potatoes, based on stable isotope and volatile compound analyses,” Food Chemistry, vol. 124, no. 4, pp. 1708–1713, 2011.
View at: Publisher Site | Google Scholar
H. Wu, T. Yue, and Y. Yuan, “Authenticity tracing of apples according to variety and geographical origin based on electronic nose and electronic tongue,” Food Analytical Methods, vol. 11, no. 2, pp. 522–532, 2018.
View at: Publisher Site | Google Scholar
J. t. Zhou, D. y. Zhao, Y. h. Chen, G. d. Kang, and C. g. Cheng, “Analysis of apple producing area changes in China,” Journal of Fruit Science, vol. 38, pp. 1–15, 2021, in Chinese.
View at: Google Scholar
L. X. Kuang, J. Y. Nie, Y. P. Li, Y. Cheng, and Y. M. Shen, “Quality evaluation of Fuji apples cultivated in different regions of China scientia,” Agricultura Sinica, vol. 53, no. 11, pp. 2253–2263, 2020, in Chinese.
View at: Google Scholar
M. T. Osorio, A. P. Moloney, O. Schmidt, and F. J. Monahan, “Multielement isotope analysis of bovine muscle for determination of international geographical origin of meat,” Journal of Agricultural and Food Chemistry, vol. 59, no. 7, pp. 3285–3294, 2011.
View at: Publisher Site | Google Scholar
Z. Li, Y. R. Guo, L. J. Sun, Q. I. Liu, J. J. Liu, and C. C. Fu, “Quality differences of Nagafu 2 apple from different habitats and its correlation with geographical coordinates,” Journal of Shaanxi Normal University (Natural Science Edition), vol. 40, no. 04, pp. 98–103, 2012, in Chinese.
View at: Google Scholar
X. Wang, Research on Quality Evaluation and Processing Suitability of Fuji Apple from Different Chinese Origins, Chinese Academy of Agricultural Sciences, Beijing, China, 2013, in Chinese.
J. Guo, T. Yue, Y. Yuan, and Y. Wang, “Chemometric classification of apple juices according to variety and geographical origin based on polyphenolic profiles,” Journal of Agricultural and Food Chemistry, vol. 61, no. 28, pp. 6949–6963, 2013.
View at: Publisher Site | Google Scholar
S. Kelly, K. Heaton, and J. Hoogewerff, “Tracing the geographical origin of food: the application of multi-element and multi-isotope analysis,” Trends in Food Science & Technology, vol. 16, no. 12, pp. 555–567, 2005.
View at: Publisher Site | Google Scholar
B. l. Guo, Y. m. Wei, and J. r. Pan, “Progress in the application of isotopic fingerprint analysis to food origin traceability,” Transactions of the CSAE, vol. 23, no. 3, pp. 284–289, 2007.
View at: Google Scholar
P. Zhou, Z. Li, L. Ouyang et al., “A multi-element stable isotope approach coupled with chemometrics for the determination of Tieguanyin tea geographical origin and harvest season,” Analytical Methods, vol. 11, no. 3, pp. 346–352, 2019.
View at: Publisher Site | Google Scholar
L. j. Zheng, J. Y. Nie, Z. Yan et al., “Studies on the characteristics of the composition and content of soluble sugars in apple fruit,” Acta Horticulturae Sinica, vol. 42, no. 05, pp. 950–960, 2015, in Chinese.
View at: Google Scholar
J. Zhang, J.-y. Nie, J. Li et al., “Evaluation of sugar and organic acid composition and their levels in highbush blueberries from two regions of China,” Journal of Integrative Agriculture, vol. 19, no. 9, pp. 2352–2361, 2020.
View at: Publisher Site | Google Scholar
B. Ma, J. Chen, H. Zheng et al., “Comparative assessment of sugar and malic acid composition in cultivated and wild apples,” Food Chemistry, vol. 172, pp. 86–91, 2015.
View at: Publisher Site | Google Scholar
A. Li, Q. S. Chen, L. G. Pan et al., “Discriminations of the geographical origin of peach based on stable isotope and rare earth element fingerprint characteristics,” Food Science, vol. 41, no. 06, pp. 322–328, 2020, in Chinese.
View at: Google Scholar
A. Li, J. Zhao, J. Xi et al., “Geographical authentication of peach in China based on stable isotope combined with multielement analysis of peach juice,” Food Control, vol. 127, Article ID 108126, 2021.
View at: Publisher Site | Google Scholar
J. Wu, H. Gao, L. Zhao et al., “Chemical compositional characterization of some apple cultivars,” Food Chemistry, vol. 103, no. 1, pp. 88–93, 2007.
View at: Publisher Site | Google Scholar
J. Shi, F. F. Li, H. Ma, Z. Y. Li, and J. Z. Xu, “Effects of different dwarfing interstocks on key enzyme activities and the expression of genes related to malic acid metabolism in Red Fuji apples,” Genetics and Molecular Research, vol. 14, no. 4, pp. 17673–17683, 2015.
View at: Publisher Site | Google Scholar
A. Thakur and Z. Singh, “Responses of “Spring Bright” and “Summer Bright” nectarines to deficit irrigation: fruit growth and concentration of sugars and organic acids,” Scientia Horticulturae, vol. 135, pp. 112–119, 2012.
View at: Publisher Site | Google Scholar
X.-J. Xu, Q.-Y. Li, X.-H. Song, S. Qi-Rong, and D. Cai-Xia, “Dynamic regulation of nitrogen and organic acid metabolism of cherry tomato fruit as affected by different nitrogen forms,” Pedosphere, vol. 22, no. 1, pp. 67–78, 2012.
View at: Publisher Site | Google Scholar
M. Brencic and P. Vreca, “Identification of sources and production processes of bottled waters by stable hydrogen and oxygen isotope ratios,” Rapid Communications in Mass Spectrometry, vol. 20, no. 21, pp. 3205–3212, 2006.
View at: Publisher Site | Google Scholar
L. Bontempo, F. Camin, L. Manzocco et al., “Traceability along the production chain of Italian tomato products on the basis of stable isotopes and mineral composition,” Rapid Communications in Mass Spectrometry, vol. 25, no. 7, pp. 899–909, 2011.
View at: Publisher Site | Google Scholar
H. Jeon, S.-C. Lee, Y.-J. Cho, J.-H. Oh, K. Kwon, and B. H. Kim, “A triple-isotope approach for discriminating the geographic origin of Asian sesame oils,” Food Chemistry, vol. 167, pp. 363–369, 2015.
View at: Publisher Site | Google Scholar
K. Bizjak Bat, K. Eler, D. Mazej et al., “Isotopic and elemental characterisation of Slovenian apple juice according to geographical origin: preliminary results,” Food Chemistry, vol. 203, pp. 86–94, 2016.
View at: Publisher Site | Google Scholar
H. Wu, L. Tian, B. Chen et al., “Verification of imported red wine origin into China using multi isotope and elemental analyses,” Food Chemistry, vol. 301, Article ID 125137, 2019.
View at: Publisher Site | Google Scholar
M. F. Cengiz, O. Turan, D. Ozdemir, Y. Albayrak, F. Perincek, and H. Kocabas, “Geographical origin of imported and domestic teas (Camellia sinensis) from Turkey as determined by stable isotope signatures,” International Journal of Food Properties, vol. 20, no. 12, pp. 3234–3243, 2017.
View at: Publisher Site | Google Scholar
F. Camin, R. Larcher, M. Perini et al., “Characterisation of authentic Italian extra-virgin olive oils by stable isotope ratios of C, O and H and mineral composition,” Food Chemistry, vol. 118, no. 4, pp. 901–909, 2010.
View at: Publisher Site | Google Scholar
D. Potočnik, M. Nečemer, I. Perišić et al., “Geographical verification of Slovenian milk using stable isotope ratio, multi-element and multivariate modelling approaches,” Food Chemistry, vol. 326, Article ID 126958, 2020.
View at: Publisher Site | Google Scholar
K. A. Hobson, L. I. Wassenaar, B. Mila, I. Lovette, C. Dingle, and T. B. Smith, “Stable isotopes as indicators of altitudinal distributions and movements in an Ecuadorean hummingbird community,” Oecologia, vol. 136, no. 2, pp. 302–308, 2003.
View at: Publisher Site | Google Scholar
L. Bontempo, F. Camin, M. Paolini, C. Micheloni, and K. H. Laursen, “Multi-isotopic signatures of organic and conventional Italian pasta along the production chain,” Journal of Mass Spectrometry, vol. 51, no. 9, pp. 675–683, 2016.
View at: Publisher Site | Google Scholar
S. A. Wadood, G. Boli, and W. Yimin, “Geographical traceability of wheat and its products using multielement light stable isotopes coupled with chemometrics,” Journal of Mass Spectrometry, vol. 54, no. 2, pp. 178–188, 2019.
View at: Publisher Site | Google Scholar

Copyright

Copyright © 2022 Hui Zhang et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

PDF Download Citation

Download other formats

Order printed copies

Views

377

Downloads

604

Citations