*Article* **Assessment of the Hypoglycemic and Hypolipidemic Activity of Flavonoid-Rich Extract from** *Angelica keiskei*

**Lanlan Tu <sup>1</sup> , Rui Wang 2, Zheng Fang 3, Mengge Sun 1, Xiaohui Sun 1, Jinhong Wu 1,\*, Yali Dang 2,\* and Jianhua Liu <sup>4</sup>**


**Abstract:** *Angelica keiskei* contains a variety of bioactive compounds including chalcone, coumarin, and phytochemicals, endowing it with pharmacological effects such as lipid-lowering activity, antitumor activity, liver protection, and nerve protection. This study aims to study the hypoglycemic and hypolipidemic effects of the flavonoid-rich extract from *Angelica keiskei* (FEAK) in an effort to exploit new applications of FEAK and increase its commercial value. In this paper, flavonoid compounds in *Angelica keiskei* were extracted using 50% ethanol, and the contents of the flavonoid compounds were analyzed by UPLC-MS/MS. Then, the hypoglycemic and hypolipidemic activities of the FEAK were investigated through in vitro enzyme activity and cell experiments as well as establishing in vivo zebrafish and Caenorhabditis elegans (*C. elegans*) models. The UPLC-MS/MS results show that the major flavonoid compounds in the FEAK were aureusidin, xanthoangelol, kaempferol, luteolin, and quercetin. The inhibitory rates of the FEAK on the activity of α-amylase and cholesterol esterase were 57.13% and 72.11%, respectively. In cell lipid-lowering experiments, the FEAK significantly reduced the total cholesterol (TC) and total triglyceride (TG) levels in a dose-dependent manner, with 150 μg/mL of FEAK decreasing the intracellular levels of TC and TG by 33.86% and 27.89%, respectively. The fluorescence intensity of the FEAK group was 68.12% higher than that of the control group, indicating that the FEAK exhibited hypoglycemic effects. When the concentration of the FEAK reached 500 μg/mL, the hypoglycemic effect on zebrafish reached up to 57.7%, and the average fluorescence intensity of *C. elegans* in the FEAK group was 17% lower than that of the control group. The results indicate that the FEAK had hypoglycemic and hypolipidemic activities. The findings of this study provide theoretical references for the high-value utilization of *Angelica keiskei* and the development of natural functional food with hypoglycemic and hypolipidemic activities.

**Keywords:** hypoglycemic activity; hypolipidemic activity; *Angelica keiskei*; flavonoid; function evaluation

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

#### **1. Introduction**

Diabetes is a metabolic disease characterized by abnormal glucose metabolism, and type 2 diabetes accounts for about 90% of all diabetes patients [1]. Insulin resistance, one of the most prominent features of type 2 diabetes, mainly results from disorders of lipid metabolism, which is, in turn, one of the frequent complications of diabetes. According to a China Health and Nutrition Survey, a total of 20~90% of diabetes patients suffer from hyperlipidemia [2]. Therefore, the prevention and treatment of type 2 diabetes as well as lipid metabolism disorders are of utmost importance. At present, metformin and acarbose tablets are proven to be relatively reliably safe long-term in prediabetic patients, while

**Citation:** Tu, L.; Wang, R.; Fang, Z.; Sun, M.; Sun, X.; Wu, J.; Dang, Y.; Liu, J. Assessment of the Hypoglycemic and Hypolipidemic Activity of Flavonoid-Rich Extract from *Angelica keiskei*. *Molecules* **2022**, *27*, 6625. https://doi.org/10.3390/molecules 27196625

Academic Editors: Jolanta Sereikaite,˙ Weiying Lu and Yanping Chen

Received: 12 September 2022 Accepted: 3 October 2022 Published: 6 October 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

145

there is insufficient evidence for the long-term safety of other drugs [3]. The drugs used to regulate blood lipids, including statins and bast, may cause adverse reactions such as liver injury and rhabdomyolysis, despite their high efficacy [4]. It has been reported that traditional Chinese medicine treatment is a useful therapeutic option for regulating and treating type 2 diabetes patients with hyperlipidemia because of its safety and efficacy [5]. In addition to conventional drug therapy, functional foods have become increasingly attractive options for helping lower blood sugar and fat.

*Angelica keiskei* is a perennial herb plant from the Umbelliferae family that can be used as both medicine and food [6]. Praised in "Foods for Health in the 21st century" [7], *Angelica keiskei* has pharmacological functions such as antitumor and anti-inflammatory activities, and it can lower hypertension, hyperlipidemia, and hyperglycemia, prevent cardiovascular and cerebrovascular diseases, regulate the intestinal tract, and improve human immunity and sleep quality.

*Angelica keiskei* is a rich source of flavonoid compounds. Studies have shown that flavonoid compounds possess a wide range of pharmacological activities, such as vasodilation, lipid-lowering activity, anticoagulation, anti-inflammation, antitumor activity, analgesia, and nonenzymatic glycosylation, and these properties have become one of the focuses of phytochemical studies. Riezki Amalia et al. showed that extract from *Angelica keiskei* stems protects HEK293 cells from N-acetyl-*p*-benzoquinone imine (NAPQI) damage and has nephroprotective properties [8]. Zhang studied the effect of chalcone from *Angelica keiskei* on the lipid metabolism in rats with type 2 diabetes. Compared with the diabetic model group, the serum levels of the TG, TC and FFA contents were significantly lower in the intervention groups [9]. Zhang found that a flavonoid-rich ethanol extract from *Angelica keiskei* leaves in a dose of 800 mg/kg could exhibit the same effect on TC with metformin [10]. Liu studied the effect of chalcones extracted from *Angelica keiskei* (AC) on the hepatocytes of rats with type 2 diabetes. Compared with rats in the diabetic control group, the levels of blood glucose and serum insulin in the 10 mg/kg AC group were decreased [11].

Most of the reports were about the single functional study of *Angelica keiskei*, and the simultaneous glucose-lowering and lipid-lowering effects of the *Angelica keiskei* extract were not investigated. However, further development of *Angelica keiskei* as a natural functional food requires systematic and comprehensive studies of its hypoglycemic and hypolipidemic activities, including its detailed flavonoid composition and the dose–effect relationship of the hypoglycemic and hypolipidemic activities of the flavonoid-rich extract. Therefore, this study aimed to assess the hypoglycemic and hypolipidemic effects of the flavonoid extract obtained from *Angelica keiskei* by in vitro and in vivo experiments so as to lay the foundation for further development of the functional food or drugs from *Angelica keiskei*.

#### **2. Results and Discussions**

#### *2.1. The Flavonoid Compositions of FEAK*

UPLC-MS/MS was used to analyze the components of flavonoids in the FEAK. As can be seen from Table 1, the most abundant flavonoids were aureusidin, xanthoangelol, kaempferol, luteolin, and quercetin. Studies have shown that aureusidin, xanthoangelol, and quercetin play a crucial role in regulating the cardiovascular system through the inhibition of blood lipids, prevention of the elevation of low-density lipoprotein in serum, and reduction in blood sugar and serum cholesterol. Wang Junbo [10] showed that 10 μg/mL luteolin and quercetin cholesterol significantly reduce the TG content in HepG2 cells. Busu et al. [12] showed that quercetin and kaempferol significantly inhibit the expression of C/EBPα, PPARγ, and SPEBP-1c at high concentrations, thus improving lipid metabolism and preventing lipid overaccumulation. The flavonoid composition of FEAK, therefore, supports its potential application in the development of functional material with hypoglycemic and lipid-lowering activities.


**Table 1.** Component analysis of flavonoids in the ethanol extract of *Angelica keiskei*.

#### *2.2. The Inhibition Activities on the α-Amylase and Cholesterol Esterase*

α-amylase inhibitors destroy the α-1,4-glycosidic bond of starch in the intestinal tract and are considered as therapeutic targets for type 2 diabetes [13]. Therefore, the inhibition of α-amylase activity is of great significance for the study of hypoglycemia. Figure 1 shows that, with increasing concentrations of the FEAK, the inhibitory rate of the FEAK on α-amylase significantly increased overall. When the concentration was 40 mg/mL, the inhibition rate of α-amylase reached 57.13 ± 1.88%.

Moreover, pancreatic cholesterol esterase contributes to the bile salt-dependent hydrolysis of dietary cholesterol esters as well as the hydrolysis of triglyceride-phospholipids which are potential targets for the prevention of dietary cholesterol absorption [14]. Figure 1 shows that the inhibitory rate of the FEAK on the activity of cholesterol esterase first increased and then decreased with the increasing concentration. When the concentration of the FEAK was increased to 15 mg/mL, the inhibition rate reached a maximum value of 72.11 ± 1.69%. These results demonstrate that FEAK acts as both an α-amylase inhibitor and a cholesterol esterase inhibitor.

**Figure 1.** The inhibitory rate of the FEAK on the activity of α-amylase and cholesterol esterase. Note: different lowercase letters in the same enzyme activity test group represent the existence of significant differences at *p* < 0.05 in the different concentration group of the FEAK.

#### *2.3. Effect of FEAK on the Intracellular Levels of TC and TG in HepG2 Cells*

Blood fat is the general name for various lipid substances in blood, among which TC and TG are the most important and abundant, so they are important routine clinical test indices of blood lipid levels [15]. There have been many studies on the hypolipidemic activity of flavonoids. The TG and FFA levels in cells and FFA levels in the cell supernatant in cells treated with low, medium, and high doses of total chamomile flavonoids are significantly reduced [16]. Kobayashi et al. found that persimmon flavone regulates the expression of cholesterol 7α strengthening the enzyme gene and the cholesterol regulatory primary binding protein (SREBP) gene to affect the lipoprotein receptor [17]. Liu Chang showed that 200 μg/mL of mangrove berry extract reduces lipid and TG levels in HepG2 cells by 25.93% and 37.23%, respectively [18]. Lv Yichun et al. showed that 100 μg/mL of blueberry polyphenols decrease TG levels in liver cells by 40%, demonstrating that these compounds can clear fat accumulation in liver cells [19].

Figure 2 shows the effect of the FEAK on the TC and TG levels in the HepG2 cells. Compared with the normal control groups, the TC and TG levels in the sodium oleate and sodium palmitate model groups were significantly increased (*p* < 0.05), while the levels in the groups treated with the FEAK were significantly decreased. Compared with the model control groups, the change in TC and TG levels showed a dose-dependent relationship. When the concentration of the FEAK increased to 150 μg/mL, the intracellular levels of TC in the HepG2 cells decreased by 33.86% (*p* < 0.001). When the concentration of the FEAK increased to 150 μg/mL, the intracellular levels of TG in the HepG2 cells decreased by 27.89% (*p* < 0.001).

**Figure 2.** Changes of TC (**a**) and TG (**b**) values in HepG2 cells after treatment of FEAK. Compared with the model control groups, \* *p* < 0.05, \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001.

#### *2.4. Effect of FEAK on Glucose Uptake in HepG2 Cells*

In a glucose uptake model using HepG2 cells, hypoglycemic ingredients were added, the cells in each group were treated with fluorescent-labeled glucose, and the fluorescence intensity of the cells was measured using flow cytometry. The fluorescence intensity of the treated cells was higher than that of the control group, indicating that the cells treated with hypoglycemic ingredients absorbed more glucose. These observations reflect the increase in the glucose utilization rate of cells in the human body and the subsequent decrease in blood glucose, demonstrating that the samples were exhibiting hypoglycemic effects [20].

According to Table 2, increasing the FEAK leads to stronger fluorescence intensity, and the change due to the FEAK treatment was higher than that of the contrast group. When the concentration of the FEAK reached 100 mg/mL, the contrast value reached the highest value of 68.12% (*p* < 0.01). This result indicates that the FEAK absorbs the glucose in cells and exerts a hypoglycemic effect.

**Table 2.** Results of cell fluorescence intensity.


The sample without FEAK was used as the control group used to calculate the δ value according to the Equation (2) shown in the Section of Materials and Methods. Note: the different lowercase letters in the fluorescence intensity test group represent the existence of significant differences at *p* < 0.05 in the different concentration group of the FEAK.

#### *2.5. Evaluation of Hypoglycemic Efficacy of FEAK In Vivo in a Zebrafish Model*

Zebrafish are often used as a hypoglycemic model for the research of human diabetes drugs because they employ similar mechanisms of blood glucose regulation to mammals [21]. Zebrafish are an ideal model with which to study diabetes because of their high carbohydrate diet. Sugar in food is metabolized under catalytic enzymes, such as hexokinase and glucose 6 phosphatases, via glycometabolism. Therefore, the lack of any intermediate steps in the carbohydrate metabolism will have irreversible effects on the nervous system and the development of zebrafish.

Several zebrafish models of diabetes have been established. Gleeson et al. [22] induced type 2 diabetes in zebrafish by soaking the fish in a 2% glucose solution, but the induction period of this method is long (28 days). Capiotti et al. [23] established a hyperglycemia model of zebrafish by soaking the fish in a 0.111 mol/L glucose solution for 14 days. This induction cycle is relatively short, but the model is unstable, and the glucose concentration is high, which can lead to the death of the zebrafish. In this study, a zebrafish hyperglycemia model was established by combining glucose solution immersion and egg yolk powder feeding. The hypoglycemic effect of 125~500 μg/mL of FEAK in vivo was evaluated with this model.

To determine the maximum tolerance concentration (MTC) of the FEAK in the zebrafish model experiment, the mortality rate of adult zebrafish was calculated 2 days after the FEAK administration (Table 3). No death was found in the concentration range from 31.2 μg/mL to 500 μg/mL, and the phenotypic appearance, feeding and swimming abilities of the treated fish were the same as those of the control group (Figure 3), indicating that the FEAK is not toxic to adult zebrafish. As 500 μg/mL is a large dose, the MTC of the FEAK was considered to be 500 μg/mL and was used to determine the experimental doses.

**Table 3.** Evaluation of MTC (*n* = 30).


**Figure 3.** *Cont*.

**Figure 3.** The phenotypic appearance of zebrafish fed with different concentration of FEAK. (**a**) Normal control group. (**b**) Model control group. (**c**) 31.2 μg/mL of FEAK group. (**d**) 62.5 μg/mL of FEAK group. (**e**) 125 μg/mL of FEAK group. (**f**) 250 μg/mL of FEAK group. (**g**) 500 μg/mL of FEAK group.

As shown in Table 4, the blood glucose level of zebrafish in the hyperglycemia group was 2.31 ± 0.129 mmol/L, which is about 2.5 times that of zebrafish in the normal control group (0.920 ± 0.044 mmol/L, *p* < 0.001), demonstrating that the hyperglycemia model was successful. As a short-term measurement of blood glucose, blood glucose levels can be directly used for blood glucose monitoring. Blood glucose levels were measured after zebrafish in the hyperglycemia group were treated with 125, 250, or 500 μg/mL of the FEAK for 2 days. Our results show that the hypoglycemic effect improved with increasing concentrations of the FEAK, indicating that the hypoglycemic effect of FEAK is dosedependent. The blood glucose levels of the zebrafish dropped by 57.7% upon treatment with 500 μg/mL of FEAK (*p* < 0.001), and this hypoglycemic effect was similar to that of the positive control drug pioglitazone hydrochloride.


**Table 4.** Experimental result of FEAK-assisted hypoglycemic effect (*n* = 10).

Note: the normal control group was prepared with pure water; the model control group was prepared with 0.15% yolk powder solution in the daytime, and 3% glucose solution was added in the evening after the removal of the yolk powder solution; the positive control group was prepared with pioglitazone hydrochloride tablets based on the model control group; and the sample groups were prepared with different concentrations of FEAK based on the model control group. Statistical differences between groups were obtained by comparing the data of each group with the model control group, \* *p* < 0.05, \*\*\* *p* < 0.001.

#### *2.6. Evaluation of Hypolipidemic Efficacy of FEAK In Vivo in a C. elegans Model*

*C. elegans* is a good model organism for studies of lipid storage because it possesses similar regulatory factors and metabolic pathways that regulate adipose deposition and related metabolic diseases to mammals. Many mutants of lipid deposition have been generated in *C. elegans.* The dhs-3::gfp mutant exhibits green fluorescence on the surface of lipid droplets, allowing changes in lipid deposition in *C. elegans* to be easily observed. Therefore, we used dhs-3::gfp mutants to observe the effect of the FEAK on lipid accumulation in *C. elegans*. The fluorescence images were processed using ImageJ. The results in Figure 4 show that the mean fluorescence intensity of the DMSO group was 15.10 ± 0.086, while the mean fluorescence intensity of the 500 μg/mL FEAK group was 12.56 ± 0.108 (*p* < 0.001), a reduction of almost 17%. This result indicates that the FEAK decreases lipid accumulation in *C. elegans* and suggests that the FEAK lowers adipose deposition in vivo.

#### **3. Materials and Methods**

#### *3.1. Materials and Chemicals*

*Angelica keiskei* (Xiancao Health Management Group Co., Ltd., Shandong, China); human hepatoma cells (HepG2) (Yongchuan Biotechnology, Shanghai, China); sodium oleate (Sigma, St. Louis, MO, USA); sodium palmitate (Sigma, St. Louis, MO, USA)); Dulbecco's modified Eagle medium (DMEM) (Hyclone, Beijing, China); TC and TG assay kits (Nanjing Jiancheng Biology Co., Ltd. Nanjing, China); α-Amylase (Yuanye Biology Co., Ltd. Shanghai, China); cholesterol esterase (Yuanye biology Co., Ltd. Shanghai, China); zebrafish (AB wild-type, Shanghai Feixi Biotechnology Co., Ltd. Shanghai, China); Caenorhabditis elegans (presented by the Wu Ziyun research group of Shanghai Jiaotong University, Shanghai, China); M9 buffer: Na2HPO4·12H2O 0.15 g, KH2PO4 0.03 g, NaCl 0.05 g, MgSO4 0.0025 g; and pure water to a final volume of 100 mL.

#### *3.2. Preparation of Flavonoid-Rich Extract from Angelica keiskei*

*Angelica keiskei* samples were shattered after vacuum freeze-drying for 24 h. The dried samples were milled to a powder using a 100-mesh sieve. The powder was mixed with 50% ethanol at a mass/volume ratio of 1:10 (g:mL) and was heated and stirred in a water bath at 37 ◦C for 3 h. After cooling to room temperature, the extraction was collected by centrifuging at 8000 rpm/min at 4 ◦C. The ethanol extract process was repeated, and all extracts were placed in a rotary evaporator to remove trace amounts of ethanol. After lyophilization, the flavonoid-rich extract from *Angelica keiskei* (FEAK) was obtained and stored at −20 ◦C [24].

#### *3.3. Analysis of the Flavonoid Composition of FEAK*

FEAK samples were ground to a powder in a laboratory mill (30 Hz, 1.5 min). Then, 100 mg of powder was weighed and dissolved in 1.2 mL 70% methanol, stirred every 30 min for 30 s 6 times. The extract sample was placed in a refrigerator at 4 ◦C overnight. After centrifugation (12,000 rpm, 10 min), the supernatant was obtained, and the sample was filtered through a 0.22 μm microporous membrane and stored in the sample bottle for UPLC-MS/MS [25].

The UPLC-MS/MS analytical conditions were as follows. The UPLC was equipped with an Agilent SB-C18 column (1.8 μm, 2.1 mm × 100 mm). The mobile phase consisted of solvent A, pure water with 0.1% formic acid, and solvent B, acetonitrile with 0.1% formic acid. Sample measurements were performed with a gradient program that employed the starting conditions of 95% A, 5% B. Within 9 min, a linear gradient to 5% A, 95% B was programmed, and a composition of 5% A, 95% B was kept for 1 min. Subsequently, 95% A, 5.0% B was reached within 1.1 min and kept for 2.9 min. The flow velocity was set to 0.35 mL per minute. The column oven was set to 40 ◦C. The injection volume was 2 μL. The effluent was alternatively connected to an ESI-triple quadrupole-linear ion trap (QTRAP)-MS.

The relative content of flavonoid compound above 1% in the extract was detected by a relative quantification method, namely, it was calculated by the ratio of the peak area of each flavonoid component divided by the peak areas of all flavonoid components.

#### *3.4. Evaluation of the Hypoglycemic and Hypolipidemic Activity of FEAK In Vitro* 3.4.1. Measurement of α-Amylase Inhibitory Activity

We used the method of Yuca et al. [26], with slight modifications. The α-amylase solution (1 U/mL) was mixed into samples at varying concentrations and preheated in 37 ◦C water bath for 5 min, followed by addition of 1% preheated starch solution. The solution was incubated in a 37 ◦C water bath for 5 min. Then, 0.25 mL DNS solution was immediately added for color responses, and the reaction mixtures were incubated in boiling water for 5 min. After cooling in an ice bath, the reaction mixtures were diluted to 5 mL with distilled water. The absorbance was recorded at 540 nm on a microplate reader. The enzyme solution and starch solution were prepared in PBS buffer at pH 6.8, 0.1 mol/L. The additive and concentration of each tube are shown in Table 5. Equation (1) was used to calculate percent inhibition.


**Table 5.** Inhibition system of α-amylase activity.

$$\text{Inhibtivity activity} = 1 - \frac{\text{A}\_3 - \text{A}\_4}{\text{A}\_1 - \text{A}\_2} \times 100\% \tag{1}$$

A1 is the absorbance of the blank group (equal volume buffer replaces the sample solution); A2 is the absorbance of blank control group (equal volume buffer replaces sample solution and enzyme solution); A3 is the absorbance of the sample group; and A4 is the absorbance of the sample control group (equal volume buffer instead of enzyme solution).

#### 3.4.2. Measurement of Cholesterol Esterase Inhibitory Activity in Porcine Pancreas

We used the method of Su et al. [27], with slight modifications. In brief, the measurement was performed in 0.1 mol/L sodium phosphate (pH 7.04) containing 0.1 mol/L NaCl, 0.2 mmol/L *p*-nitrophenyl butyrate (*p*-NPB) and 5.16 mmol/L sodium taurocholate buffer (STC). First, 10 U/mL of porcine pancreatic cholesterol esterase and PNPB were predissolved in acetonitrile and stored at −20 ◦C. Porcine pancreatic cholesterol esterase was added to the reaction tube, and samples were incubated at 25 ◦C for 5 min. Then, the absorbance of the solutions was measured at 450 nm using an ultraviolet-visible spectrophotometer. The additives and corresponding concentrations of each tube are shown in Table 6. Equation (1) was used to calculate percent inhibition.


**Table 6.** The amount of additive agents for cholesterol esterase inhibitory activity measurement.

3.4.3. Glucose Consumption Assay in HepG-2 Cells

The effect of FEAK on glucose consumption was investigated in insulin-resistant HepG-2 cells. After thawing, subculturing, and plating, HepG-2 cells were processed as follows. First, 100 nmol/L of recombinant human insulin was added, and cells were incubated at 37 ◦C for 30 min. Then, 0, 10, 25, 50 to 100 mg/mL FEAK was added, and cells were incubated at 37 ◦C for 1 h. Finally, 50 μM 2-NbDG was added, and cells were incubated at 37 ◦C for 1 h. After the incubation, the culture medium was removed, and cells were washed with PBS buffer twice. After digestion with 1 mL trypsin-EDTA solution, culture medium was added, and the resulting single cell suspension was placed in a centrifugal tube and centrifuged at 1000 RPM for 5 min. Afterward, the supernatant was removed and discarded, and the solution was centrifuged at 1000 rpm for 5 min. This process was repeated three times. HepG-2 cells were collected into flow cytometry tubes and placed on ice to maintain a low temperature for further determination of cell fluorescence intensity [28].

Fluorescence intensity was measured by flow cytometry in specific channels (10,000 cells per tube). The contrast value of fluorescence intensity was calculated according to the following Equation (2):


The larger the value of δ, the stronger the ability of cells to absorb glucose and the stronger the hypoglycemic effect [20].

3.4.4. Total Cholesterol (TC) and Triglyceride (TG) Assay in HepG-2 Cells

The high lipid cell culture medium consisted of DMEM supplemented with 500 μmol/L sodium oleate and 250 μmol/L sodium palmitate [7]. HepG-2 cells were treated with 0.25% trypsin and transferred to petri dishes. The sample dose was set based on the cytotoxicity test. The blank control group and the model group were cultured in basal DMEM medium and high lipid medium at 37 ◦C and 5% CO2 for 24 h, respectively. The sample intervention groups were first cultured in high lipid medium at 37 ◦C and 5% CO2 for 24 h, then 50, 100, or 150 μg/mL FEAK was added for an additional 24 h [12].

After HepG-2 cells were cultured, the culture medium was discarded. The cells were washed with the PBS at 4 ◦C 3 times. Radio immunoprecipitation assay (RIPA)lysis buffer was added to each petri dish, and the cells were placed on ice for 30 min. The lysate was collected and centrifuged at 10,000 rpm for 10 min at 4 ◦C. The supernatant was obtained for the determination of TG and TC levels by using TC and TG assay kits.

*3.5. Experimental Analysis of In Vivo Hypoglycemic Effect in a Zebrafish Model*

#### 3.5.1. Sample Preparation

Sample group: 50.0 mg/mL mother liquor was prepared using ethanol extract of FEAK and DMSO and stored at −20 ◦C.

Positive control group: 10.0 mg/mL mother liquor was prepared using pioglitazone hydrochloride tablets and DMSO and stored at −20 ◦C.

#### 3.5.2. Evaluation of the Hypoglycemic Effect of FEAK in a Diabetic Zebrafish Model

Zebrafish of the AB wild-type strain were bred naturally in pairs. Zebrafish were all raised in fish culture water at 28 ◦C. Zebrafish at 5 days postfertilization (5 DPF) were used to evaluate the maximum tolerable concentration (MTC) and auxiliary hypoglycemic efficacy of FEAK.

The MTC of FEAK was determined. Zebrafish of the AB wild-type strain at 5 DPF were randomly selected and placed in 25 mL beakers, with 30 zebrafish in each beaker (experimental group). FEAK was dissolved in the water (31.2, 62.5, 125, 250, and 500 μg/mL). All experimental groups except for the normal control group were given 0.15% yolk powder solution in the daytime, and 3% glucose solution was given in the evening after the removal of the yolk powder solution to establish the zebrafish hyperglycemia model. After treatment at 28 ◦C for 48 h, the MTC of FEAK on zebrafish in the model control group was determined.

The auxiliary hypoglycemic efficacy of FEAK was further evaluated. Zebrafish of the AB wild-type strain at 5 DPF were randomly selected and placed in beakers, with 30 zebrafish in each beaker (experimental group). FEAK was given in aqueous solution (concentrations as shown in Tables 1 and 2), and 20.0 μg/mL of pioglitazone hydrochloride was used as a positive control. The normal control group and model control group were set up at the same time. After establishing the zebrafish hyperglycemia model and treating the zebrafish with FEAK at 28 ◦C for 48 h, the zebrafish were washed three times with fish culture water, and the glucose level of zebrafish was detected using a blood glucose meter. The auxiliary hypoglycemic efficacy was evaluated by statistical analysis of the results obtained using this indicator.

#### *3.6. Evaluation of Hypolipidemic Effect of FEAK in C. elegans Model*

First, *Escherichia coli* (*E. coli)* solution with FEAK was prepared as follows: 500.0 mg/mL mother liquor of FEAK was prepared with DMSO and stored at −20 ◦C; 10 μL of mother liquor was added to 10 mL concentrated *E. coli* solution and mixed, then the mixture was added to the nematode growth medium (NGM) agar plate and dried by airing for later use. The DMSO control group received 10 μL DMSO instead of FEAK.

The effect of FEAK on lipid deposition was observed by determining the green fluorescence of the dhs-3::gfp mutants. The animals were first synchronized at the L1 stage and then divided into groups. About 50 nematodes were cultured in each plate in a constant temperature incubator set to 20 ◦C for 6 days. The animals were then washed with M9 buffer, anesthetized with 40 mmol/L imidazole, observed, and photographed under fluorescence microscope. The average fluorescence intensity was calculated using ImageJ.

#### *3.7. Statistical Analysis*

Each experimental group was set up in three parallel groups. Statistical results were expressed as mean ± SE. SPSS 26.0 software (IBM, 2022, Shanghai, China). was used for statistical analysis, with *p* < 0.05 indicating a statistically significant difference. GraphPad Prism 9.0 software (GraphPad Software, 2021, Shanghai, China) was used for plotting.

#### **4. Conclusions**

This study aimed to investigate the effect of FEAK on regulating glucose and lipid metabolism. The hypoglycemic and hypolipidemic activities of FEAK were analyzed through a series of in vitro and in vivo experiments, including assay testing the inhibition of α-amylase and cholesterol esterase, the determination of the intracellular TC and TG levels and glucose uptake in HepG2 cells, and the evaluation of hypoglycemic and hypolipidemic efficacy in vivo using zebrafish and *C. elegans.* We found that FEAK is rich in flavonoids, including aureusidin, xanthoangelol, kaempferol, luteolin, and quercetin. We also found that 40 mg/mL of the FEAK showed an inhibition rate of 57.13 ± 1.88% against α-amylase and that 15 mg/mL of the FEAK showed a maximum inhibition rate of 72.11 ± 1.69% against cholesterol esterase. Further, 150 μg/mL of FEAK decreased intra-

cellular TC levels in the HepG2 cells by 33.86% (*p* < 0.001), and 150 μg/mL of the FEAK decreased the intracellular TG levels in the HepG2 cells by 27.89% (*p* < 0.001). When the concentration of the FEAK reached 100 mg/mL, the contrast value indicating the glucose uptake reached its highest value of 68.12% (*p* < 0.01). Moreover, 500 μg/mL of FEAK decreased the blood glucose levels of zebrafish by 57.7% (*p* < 0.001), similar to the positive control drug pioglitazone hydrochloride. Finally, 500 μg/mL of the FEAK decreased the fluorescent intensity of *C. elegans* by 17% (*p* < 0.001) compared to that of the DMSO group. These findings provide strong evidence that FEAK has hypoglycemic and hypolipidemic activity and could be a promising natural product with potential value for the development of functional foods or drugs to prevent or treat type 2 diabetic or hyperlipidemia. However, the mechanism of the hypoglycemic and hypolipidemic effects of FEAK is not very clear, and further research is needed to understand the detailed action mechanism.

**Author Contributions:** Conceptualization, J.W. and Y.D.; Methodology, L.T. and R.W.; Software, L.T. and M.S.; Validation, L.T. and X.S.; Formal analysis, Z.F.; Data curation, L.T.; Writing—original draft preparation, L.T. and R.W.; Writing—review and editing, L.T.; Visualization, L.T. and J.L.; Supervision, J.W.; and project administration, J.W. All authors have read and agreed to the published version of the manuscript.

**Funding:** The authors are grateful to the financial fund of the Bayannaoer City National Industrial high-tech Industrial Demonstration Zone key project of "Science and technology to promote Inner mongolia development" (NMKJXM202209-2) and the National Natural Science Foundation of China (Grant No. 31972017).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

**Sample Availability:** Samples of the plant material are available from the authors.

#### **References**


## *Article* **Comparison of Different Volatile Extraction Methods for the Identification of Fishy Off-Odor in Fish By-Products**

**Yuanyuan Zhang 1, Long Tang 1, Yu Zhang 1,\*,†, Huanlu Song 1,\*,† , Ali Raza 1, Wenqing Pan 2, Lin Gong <sup>2</sup> and Can Jiang <sup>3</sup>**

	- <sup>3</sup> Wuzhou Testing Co., Ltd, Jining 273200, China

† These authors contributed equally to this work.

**Abstract:** This study was conducted to analyze volatile odor compounds and key odor-active compounds in the fish soup using fish scarp and bone. Five extraction methods, including solid-phase microextraction (SPME), dynamic headspace sampling (DHS), solvent-assisted flavor evaporation (SAFE), stir bar sorptive extraction (SBSE), liquid-liquid extraction (LLE), were compared and SPME was finally selected as the best extraction method for further study. The volatile odor compounds were analyzed by gas chromatography-olfactometry-mass spectrometry (GC-O-MS) and comprehensive two-dimensional gas chromatography-olfactometry-mass spectrometry (GC × GC-O-MS) techniques, and the key odor-active compounds were identified via aroma extract dilution analysis (AEDA) and relative odor activity value (r-OAV) calculation. A total of 38 volatile compounds were identified by GC-O-MS, among which 10 were declared as odor-active compounds. Whereas 39 volatile compounds were identified by GC × GC-O-MS, among which 12 were declared as odor-active compounds. The study results revealed that 1-octen-3-one, 2-pentylfuran, (E)-2-octenal, 1-octen-3-one, hexanal, 1-octen-3-ol, 6-methylhept-5-en-2-one, (E,Z)-2,6-nondienal and 2-ethyl-3,5-dimethylpyrazine were the key odor-active compounds in the fish soup.

**Keywords:** fish scraps; fish soup; GC-O-MS; GC × GC-O-MS; AEDA; r-OAV

#### **1. Introduction**

China is a large producer of freshwater fish, and the production of fish products is steadily increasing year by year. China has been rich in fish products since ancient times, and Dongting Lake in Hunan Province is the main producer of freshwater fish, especially having large amounts of white chub [1]. During the processing of freshwater fish, especially in the processing of minced fish, large amounts of scraps such as fish heads, fish bones, fish tails, fish skins, and offal are produced [2]. Previously, it was found that the meat extraction rate of seven kinds of bulk freshwater fish, namely mackerel, grass carp, silver carp, bighead carp, crucian carp and bream, was the highest for mackerel (54.33%) and the lowest for bighead carp (32.80%), while the remaining 45% to 67% were fish head and fish bone scraps [3]. At present, these scraps are not utilized as value-added products, except a limited portion of them are used for processing feed fish meal, and the majority of them are directly discarded, which not only wastes a lot of resources but also pollutes the environment and increases the cost of environmental waste management [4]. An important issue faced by the aquatic food industry these days is the effective utilization of fish waste in order to generate economic benefits. The fish head is rich in nucleotides, amino acids and inorganic elements such as potassium, calcium, sodium and magnesium that contribute relatively to the flavor, and it is valuable to utilize it in seasoned soup, fish bone paste or aquatic seasoning base after proper treatment [5]. However, the removal of a fishy odor is a key step in the production of seasoning bases, which first requires the investigation of the

**Citation:** Zhang, Y.; Tang, L.; Zhang, Y.; Song, H.; Raza, A.; Pan, W.; Gong, L.; Jiang, C. Comparison of Different Volatile Extraction Methods for the Identification of Fishy Off-Odor in Fish By-Products. *Molecules* **2022**, *27*, 6177. https://doi.org/10.3390/ molecules27196177

Academic Editors: Weiying Lu and Yanping Chen

Received: 27 August 2022 Accepted: 16 September 2022 Published: 21 September 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

nature of a fishy taste. In this study, fish scraps (fish heads and bones) were used as raw materials for boiling fish soup and then analyzing the off-flavor compounds in fish soup in order to provide a research background for the removal of fishy odor compounds as well as a scientific basis for studies such as the production of seasoning bases using fish scraps.

Recently, the odor compounds in fish have emerged as one of the most important studies for the development of fish seasonings, which are directly related to the sensory quality of the product. The fishy flavor is one of the main indicators used to determine the quality of fish flavoring, and it is formed by the joint action of a variety of odor compounds. Liu et al. [6] used SPME combined with the GC-MS technique to analyze fishy odor substances in the muscle (35 kinds), head (35 kinds) and skin (38 kinds) of tilapia, and nonanal, octanal and (E)-octenal were found to be the key fishy substances which were identified via AEDA analysis. Similarly, in the study of Li et al. [7], the DHS-GC-MS technique was employed to analyze the odor compounds produced during hydrolysis of grass carp, and 37 odor compounds were detected. In previous studies, a series of investigations were conducted on the volatile compounds of different varieties of fish, although the site, type and source of the study subject were not specified. However, these volatile compounds do not necessarily determine the key aroma compounds in the products. Therefore, the study of key odor-active compounds is also necessary. Although there have been some advances in the study of aroma-active substances in fish and boiled fish soup, most of the studies have been conducted with fish raw materials; it is interesting to think about the Maillard reaction, caramelization reaction and oxidation of fats and oils that may occur after heat treatment of foods, and these reactions produce various volatile compounds [8,9]; these types of reactions are also present in the boiling process of fish scraps, which helps to further investigate the odor compounds in them and provides a research basis for the preparation of seasonings using fish scraps.

GC-O-MS is the key approach to analyze odor compounds in fish soup. However, during the boiling process, the odor compounds in the fish soup may react in various ways, causing changes in the volatile components [10]. In this study, a new gas chromatographic technique, namely two-dimensional integrated gas chromatography-olfactometry-mass spectrometry (GC × GC-O-MS), was compared with conventional one-dimensional gas chromatography in order to analyze the sample; this two-dimensional gas chromatography allows the conversion between two modes (one- and two-dimensional analysis modes). The two-dimensional mode allows volatile compounds in the sample to be detected more quickly and efficiently [11]. With the continuous promotion of GC × GC-O-MS, this technique has been applied to the analysis of volatile odor compounds in various samples. For example, Yang et al. [12] used GC × GC-O-MS to identify the differences in volatile odorants of cinnamon tea leaves at different roasting temperatures and detected 97 aroma substances. Very recently, SPME combined with GC-O-MS and GC × GC-O-MS techniques were employed by Zhao et al. [13] to identify volatile odor compounds of pepper via AEDA and OAV analysis.

Therefore, in order to make the useful recovery of volatile compounds from fish byproducts, the offal (fish head and bones) of silver carp from Dongting Lake, Hunan Province was utilized. The present study was mainly aimed to accomplish the following objectives: (1) to select the most suitable volatile extraction method from solid-phase microextraction (SPME), dynamic headspace sampling (DHS), solvent-assisted flavor evaporation (SAFE), stir bar sorptive extraction (SBSE) and liquid-liquid extraction (LLE); (2) the optimal extraction method from (1) was selected and combined with the aroma extraction dilution analysis to identify the key odor-active compounds of fish soup. The significance of this study includes the selection of an appropriate method to analyze the odor compounds in fish broth and to study the key odor compounds, in order to provide a research basis for futuristic studies dealing with the removal of fishy odor from fish soup, and to provide a theoretical basis for the production of flavors from fish broth boiled with fish scraps.

#### **2. Materials and Methods**

#### *2.1. Samples and Chemicals*

The fish heads and bones used in the study were supplied by Hunan Jiapinjiawei Biotechnology (Hunan, China). Firstly, 60 ◦C warm water was added to a heavy-bottomed pot having a fish head and bone; furthermore, the waste water was removed, and washing was performed. The samples were mixed with water in a ratio of 1:1 by volume, boiling was performed under high pressure at 121 ◦C for 2 h, and the dregs were filtered off to obtain fish soup. Finally, the samples were concentrated to about 25% of the total soluble solid content with a rotary evaporator at 55◦C.

Ethyl ether, n-hexane, and anhydrous sodium sulfate, all having purities >99%, were purchased from Lab Gou e-mall (Beijing, China). 2-Methyl-3-heptanone and n-alkanes (C7–C30) were provided by Sigma-Aldrich (St. Louis, MO, USA.). Nitrogen gas (99.9992% purity) was obtained from Beijing AP BAIF Gases Industry Co., Ltd. (Beijing, China) and the liquid nitrogen was obtained from Xian Heyu Trading Co., Ltd. (Beijing, China).

#### *2.2. Solid-phase Microextraction (SPME)*

The SPME method was applied to extract the flavor compounds from fish soup according to the protocols mentioned by Li et al. [14]. Five grams of the sample and 0.3 μ L of 2-methyl-3 heptanone (0.816 μg/mL internal standard) were placed in a headspace vial (20 mL, Beijing Banxia Technology Development Co., Ltd., Beijing, China). The stock was equilibrated in a constant temperature water bath at 60 ◦C for 20 min, and then the volatile compounds were extracted with SPME needles having divinylbenzene/carboxyl/polydimethylsiloxane fibers (50/30 μm, Supelco, Bellefonte, PA, USA) at 60 ◦C for 40 min. Immediately after the extraction was completed, the extraction needle was inserted into the GC-O-MS and GC × GC-O-MS instruments for analysis and thermal desorption at 250 ◦C for 5 min. For accuracy, each sample was analyzed three times.

#### *2.3. Solvent-Assisted Flavor Evaporation (SAFE)*

The SAFE system is a compact combination of a distillation unit and a high vacuum pump; 40 g of fish soup and 5 μL of 2-methyl-3-heptanone (0.816 μg/μL, internal standard) was added to a Teflon bottle with diethyl ether/pentane (V1/V2 = 2:1). The resulting preparation was extracted by stirring for 8 h on a shaker (4 ◦C), and the volatile odor compounds were extracted using a SAFE apparatus (Deutschen Forschungsanstalt für Lebensmittelchemie, Freising, Free State of Bavaria, Germany). The distillation process was carried out by a molecular turbine pump (Edwards, Munich, Germany) for 2 h at 10-4 torr. The resulting fractions were collected in a trap that was submerged under liquid nitrogen. After collection, water was removed from the collected fraction by adding anhydrous Na2SO4 and the collected fraction was concentrated to 10 mL through a Vigreux column (50 cm × 1 cm I.D.; Beijing Banxia Technology Development Co., Ltd., Beijing, China). The concentrate was further reduced to 200 μL by a nitrogen stream (purity ≥ 99.999%) purging. Finally, one microliter of the sample was injected into the GC-O-MS and GC × GC-O-MS instruments for detection using a 5 μL syringe. All the samples were analyzed in triplicates for statistical accuracy.

#### *2.4. Dynamic Headspace Sampling (DHS)*

Fish soup (20 g) and 2 μL of 2-methyl-3-heptanone (0.816 μg/μL, internal standard) were added to a dynamic headspace flask, which was purged with 99.9992% high purity nitrogen at a flow rate of 150 mL/min at one end and inserted into a Tenax TA tube at the other end for adsorption. The temperature of water bath was established at 60 ◦C, equilibrated for 20 min, and the adsorption was performed for 40 min. After that, the Tenax TA tube was removed and placed in a thermal desorption unit (TDU) (Gerstel, Germany) for GC-O-MS or GC × GC-O-MS analysis. For statistical accuracy, all the samples were analyzed in triplicates.

#### *2.5. Stir Bar Sorptive Extraction (SBSE)*

Fish soup (15 g) and 1.5 μL of 2-methyl-3-heptanone (0. 816 μg/μL, internal standard) were added to a 40 ml headspace vial, and a Twister®(Gerstel, Germany) (such as PDMS, EG (ethylene glycol) or silica gel) was immersed inside the headspace vial (carrying the sample) and stirred at 60 ◦C for 40 min. After that, the stirring rod was picked out, rinsed with deodorized water, and placed in a hollow glass tube, which was further transferred to a TDU for thermal desorption in order to perform GC-O-MS and GC × GC-O-MS analysis. For accuracy, all the samples were analyzed three times.

#### *2.6. Liquid-Liquid Extraction (LLE)*

The volatile substances in the samples were extracted according to the conventional liquid-liquid extraction (LLE) method [15]. Fifty grams of sample was transferred to a triangular flask, after that, 50 mL distilled water, 50 mL dichloromethane, 50 mL diethyl ether and 5 μL of internal standard 2-methylhept-3-one (0. 816 μg/μL, internal standard) were added together. The mixture was stirred at 800 rpm for 10 min. After centrifugation at 8,000g for 40 minutes, the extracts were separated using a partition funnel. The resulting extract was added with anhydrous sodium sulfate and left to dry at 4 ◦C for 12 h. The dried extract was then concentrated to 100 μL using a nitrogen stream, and finally the concentrated extract was analyzed by GC-O-MS and GC × GC-O-MS. The samples were analyzed in triplicates for accuracy and statistical analysis.

#### *2.7. Gas Chromatography-Olfactometry-Mass Spectrometry (GC-O-MS) Analysis*

A GC-MS (7890A-7000, Agilent Technologies Inc., Santa Clara, CA, USA) instrument combined with an olfactory detection port (ODP4, Gerstel, Germany) was used to identify volatile odor compounds. Separation of odor-active substances in samples was performed on a polar DB-WAX capillary column (30 mm × 0.32 mm, 0.25 μm film thickness; J & W Scientific, Folsom, CA, USA). The gas chromatographic instrument condition includes an initial column temperature setting of 40 ◦C, holding for 3 min, followed by an increase in temperature up to 230 ◦C at 4 ◦C/min and holding for 3 min. Ultra-pure helium (99.999%, Beijing AP BAIF Gas Industry Co., Ltd., Beijing, China) was used as the carrier gas. The electron impact mass spectra were generated at an ionization energy of 70 eV with an *m*/*z* scan range of 25–370 amu. The temperatures of the mass spectrometer source and quadrupole were programed at 230 ◦C and 150 ◦C, respectively. Moisture gas was delivered to the olfactory detection port through a blank capillary column.

#### *2.8. GC* × *GC-O-MS Analysis*

A GC-MS instrument (8890A-5977B; Agilent Technologies Inc., Santa Clara, CA, USA) with an olfactory detector (OFD) Sniffer 9100 (Brechbühler, Schlieren, Switzerland) was used to identify volatile odor compounds in the samples. Two columns were used to separate the components, DB-WAX (polar, 30 m × 0.25 mm × 0.25 μm; Agilent Technologies) and DB-17 (mid polar, 2.22 m × 0.18 mm × 0.18 μm; Agilent Technologies). The initial column temperature was established at 40 ◦C and held for 3 min, followed by a ramp-up to 230 ◦C at 4 ◦C/min, held at 230 ◦C for 3 min. Ultra-pure helium (99.999%, Beijing APB Gas Industry Co., Ltd., Beijing, China) was used as the carrier gas. The electron impact mass spectra were generated at an ionization energy of 70 eV with an *m*/*z* scan range of 35–375 amu. The temperatures of the mass spectrometer source and quadrupole were programmed at 230 ◦C and 150 ◦C, respectively. A solid-state modulator SSM1800 (J&X Technologies, Shanghai, China) was placed between the two columns for the heating and cooling phases. The temperature of the cold zone was kept at −50 ◦C and the modulation cycle was set to 5 s.

The concentrated fractions in the instrument were analyzed at the sniffing port by three trained sensory panelists who were members of the Molecular Sensory Laboratory from Beijing University of Technology; they were trained for four weeks to analyze the effluent components at the sniffing port for 2 h per day. During the GC-O analysis, moisture was delivered to the sniffing port through a blank capillary column. The odor descriptors and label

odor intensity values, as well as retention times, were recorded [16]. The compounds that were perceived by two or more panelists were tentatively identified as odor compounds.

#### *2.9. Aroma Extraction Dilution Analysis (AEDA)*

AEDA was analyzed by varying the fractionation ratio of GC-O. The importance of each volatile component was ranked by the average flavor dilution (FD) factor, which was determined at the sniffing port according to the AEDA procedure. According to the ratios of 1:3n, which were 0, 1:3, 1:9, 1:27, 1:81, etc. The corresponding flavor dilution (FD) factors were defined as 1, 3, 9, 27, 81, etc. The result of the FD factor analysis refers to the maximum value for which the compound can be detected. Consequently, the FD factor of the odor compound that was analyzed at the sniffing port was 81, which indicates that the odor-active substance was not analyzed by the panelist at the sniffing port when the GC-O dilution ratio was 81:1. Usually, the higher the FD factor obtained for the odor-active compound, the more critical it is to be analyzed.

#### *2.10. Qualitative and Quantitative Analysis*

Volatile odor compounds were identified by mass spectrometry (MS) and comparison of identified peaks with the NIST library mounted on GC-MS. The linear retention index (RI) was calculated for dual identification according to the following equation:

$$\text{ARI} = 100\text{N} + 100\text{n} \langle \text{l} \,\text{Ra} - \text{t RN} \rangle / \,\left[ \text{t} \,\text{R} (\text{N} + \text{n}) - \text{t RN} \right] \tag{1}$$

where N represented the number of carbon atoms with peaks in front of the identified compound, n represented the numerical difference between the upper and lower n-alkanes in gas chromatography, and the variables t Ra, t RN, t R (N + n) denoted the retention time of the identified compounds and the upper alkane and lower alkanes, respectively. The unknown odor compounds were positively identified by the three methods: (1) comparison of RI and odor descriptions (O) with reference compounds; (2) comparison of MS data with NIST library; (3) finally by standard (STD) compounds verification [17].

#### *2.11. Relative Odor Activity Value (R-OAV) Calculation*

By adapting the method of Yang et al, the r-OAV was calculated using the following formula:

$$\text{Ir-OAV} = \text{Ci/OTi} \tag{2}$$

where Ci represented the relative concentration of a certain volatile odor compound in the fish soup and OTi denoted the odor threshold of the compound [18]. The threshold of the compound used in this study was the actual threshold in water [19].

#### *2.12. Omission Experiment*

An aroma model composed of the selected key odor-active compounds was prepared by reducing one of the 11 key aroma compounds at a time. A triangulation experiment assessed by a sensory panel was used to compare the differences between the blended quintile and the complete reconstituted fraction according to the PRC national standard. In other words, if 8, 9, and 10 or more of the 12 sensory members were able to identify differences between a compound when it was omitted and completed the reconstitution model, the results were considered significant (α ≤ 0.05), highly significant (α ≤ 0.01), or very highly significant (α ≤ 0.001), respectively. If fewer than 8 were identified, the result was considered "not significant". Sensory evaluation is a common method used in the food science area, which also does not involve informed consent and ethics.

#### *2.13. Statistical Analysis*

All experiments were performed in triplicate, and the data was expressed as mean ± standard deviation. Statistical analysis was performed using Microsoft Excel 2019 (Microsoft Corp., Redmond, WA, USA).

#### **3. Results and discussion**

#### *3.1. Comparison of Different Extraction Methods for Analysis of Odor Compounds*

Based on the efficiencies of the five extraction methods including SPME, SAFE, DHS, SBSE and LLE the best extraction method was selected and the odor substances in the fish soup were analyzed by GC-O-MS and GC × GC-O-MS. As shown in Table 1, in the 1D GC mode, a total number of 38 odor compounds were extracted by SPME, including 5 alcohols, 15 aldehydes, 7 ketones, 6 acids and 5 others. Ten odor compounds were detected from SAFE extract, including 3 alcohols, 2 aldehydes, 1 ketone, 1 acid and 3 others. Eighteen different types of odor compounds were extracted by DHS, including 8 aldehydes, 3 ketones, 2 esters, 2 acids and 3 others. Whereas, 11 odor compounds were extracted from SBSE, including 6 aldehydes, 2 ketones, 2 esters and 1 acid. While for LLE, only one aldehyde and one ketone were extracted. In the 2D GC mode. A total of 39 odor compounds were extracted by SPME, including 3 alcohols, 12 aldehydes, 9 ketones, 1 ester, 2 acids and 12 others; 33 odor compounds were extracted by SAFE, including alcohols (5), aldehydes (7), ketones (8), ester (1), acids (2) and others (10); by DHS, 21 volatiles were extracted including alcohols (2), aldehydes (8), ketones (4), esters (2), acids (3) and others (2). A total of 14 odor compounds were extracted from SBSE, including alcohol (1), aldehydes (4), ketones (2), ester (1), acids (3), and others (2); and only 10 volatile compounds were extracted through LLE, including alcohol (1), aldehydes (4), ketones (2) and esters (3). The results of the five extraction methods using 1D GC mode were compared, and the results are presented in Figures 1 and 2; it can be clearly concluded that SPME extracted a higher number of odor compounds.

To better determine the best method for the extraction of volatile compounds, OPLS-DA was chosen for further differentiation. As shown in Figure 3a,b, the OPLS-DA analysis showed that SPME was better differentiated from the other four extraction methods. In 1D mode, LLE could not be distinguished from SBSE. In 2D mode, three extraction methods cannot be distinguished, namely DHS, LLE and SBSE.

It has been shown that among the alcohols, 1-octen-3-ol is one of the potent compounds that contributed the most to the fishy odor [10,20], and it was predicted that lipoxygenases cause the breakdown of polyunsaturated fatty acids or the reduction of carbonyl compounds [21]. 1-Octen-3-ol was detected in SPME, SAFE and DHS, and the highest level (7.07 ± 0.98 μg/Kg) was detected when analyzed in the 2D mode in SPME; this was in agreement with the study of Xue et al. [10]. However, the aforementioned compound was not detected in both the SBSE and LLE extraction methods. Aldehydes are a relatively large group of compounds that contribute to fish odor [22]. Some of the aldehydes that can be easily sniffed were detected in SPME, but almost no aldehydes were extracted via LLE; this also confirms that LLE is not suitable for the analysis of odor compounds in fish. The extraction of odorous substances from fish by the LLE method has not been reported in recent studies. 1-Octen-3-one has the same mushroom flavor as 1-octen-3-ol, which has a low threshold and can be distinctly sniffed; it is worth noting that this substance is only detected in the 2D mode of SPME and is not detected by any other method. A few previous reports have shown 1-octen-3-one as a fishy substance, but to the best knowledge of the authors, this is the first time that 1-octen-3-one has been identified in fish. Pyridine was once considered as one of the main fishy substances in freshwater fish [23]. However, it has not been reported in recent studies [24–26]; this may be related to the detection method. In this study, only the 2D mode of SAFE was detectable at minimal levels and was not detected by olfaction. In addition, small amounts of pyrazines and furans were detected in SPME, DHS and SAFE, probably due to substances generated by prolonged high temperature heating during the boiling process [27], and the contents in SPME were higher than the other two methods. In summary, the best choice was concluded as SPME.

**Table 1.** Volatile compounds identified by five extraction methods in combination with GC-O-MS and GC×GC-O-MS.







trillion, ng/g).

**Figure 1.** The number of volatile compounds measured by GC-O-MS for five extraction methods.

**Figure 2.** Number of volatile compounds measured by GC × GC-O-MS for five extraction methods.

#### *3.2. Comparison of Aroma Compounds between GC-O-MS and GC* × *GC-O-MS Analysis*

For a clearer comparison of the methods of extracting volatile compounds and analysis of odor-active substances in fish soup, GC-O-MS (1D, one dimension) and GC × GC-O-MS (2D, two dimension) were employed. For this, the results of the two-dimensional mode (GC × GC-O-MS) were compared with the analysis of all volatile compounds detected through the one-dimensional column (first column). Compared with GC-O-MS, GC × GC-O-MS can analyze the odor compounds in the sample more accurately based on its high resolution and high sensitivity. In other words, the GC × GC-O-MS analysis can be used to confirm whether each odor region is shared by multiple odor compounds; and it can also be used to quickly distinguish the compounds in multiple odor regions. Therefore, the GC × GC-O-MS method can be used to analyze volatile odor compounds more efficiently than GC-O-MS, although the analytical results of both instruments show a high degree of agreement.

**Figure 3.** OPLS-DA analysis chart of five extraction methods. (**a**) OPLS-DA analysis of five extraction methods in 1D mode. (**b**) OPLS-DA analysis of five extraction methods in 2D mode.

As shown in Figure 4a–e, among the five extraction methods, more compounds were detected in 2D mode than in 1D mode. As shown in Table 1, the relative content of the compounds analyzed by the 2D model was also higher. The 3D view of the 2D mode was presented in Figure 5a,b. Specifically, 9 compounds, such as decanal, dimethyl trisulfide, and 2,5-dimethylpyrazine, were not analyzed in the 1D mode, but several of these substances could be identified in the 2D mode and could be analyzed in the sniffing port. Interestingly, it was observed that 1-octen-3-one can be smelled but not identified in the 1D mode, which does not reach the detection limit of the instrument, while it can be both analyzed and smelled in the 2D mode, which further confirms the high resolution of the 2D mode. In addition, other substances, such as 1-octen-3-ol, acetic acid, and (E)-2-octenal, can be analyzed in the 1D mode but cannot be identified in the sniffing port, while these compounds can be easily analyzed and sniffed in the 2D mode. Some compounds such as furfural, (E,Z)-2,6-nonadienal, and 4-hydroxy-2,5-dimethyl-3(2H) furanone, which were not identified in 2D mode, could be analyzed or smelled in the 1D mode. Clearly, the above results revealed that the use of both analytical methods can analyze the odor compounds in the samples more comprehensively.

**Figure 4.** Five extraction methods combined with GC and GC × GC to measure the amount of volatile compounds. (**a**) SPME combined with GC and GC × GC to measure the amount of volatile compounds; (**b**) SAFE combined with GC and GC × GC to measure the amount of volatile compounds; (**c**) DHS combined with GC and GC × GC to measure the amount of volatile compounds; (**d**) SBSE combined with GC and GC × GC to measure the amount of volatile compounds; (**e**) LLE combined with GC and GC × GC to measure the amount of volatile compounds.

**Figure 5.** (**a**) The 3D analytical images of GC × GC-O-MS in sample; (**b**) The 2D analytical images of GC × GC-O-MS in sample.

#### *3.3. Key Aroma-Active Compounds Identified by AEDA and r-OAV*

It was found that all the odor compounds in the sample contributed to the overall odor profile. To analyze the compounds that contributed more, a dilution analysis was performed in order to identify the key odor compounds in the sample. The critical level of the odor-active compounds was determined by the FD factor obtained from the dilution analysis. In other words, the FD factor is positively correlated with the contribution of the odor compounds to the overall odor profile [28].

By using SPME combined with GC-O-MS and GC × GC-O-MS, 55 odor compounds were identified, including 5 alcohols, 19 aldehydes, 12 ketones, 1 ester, 6 acids, 5 pyrazines, 3 furans, and 6 others (Table 1), and their FD factors were obtained by AEDA, as shown in Table 2. Among these, 2-pentylfuran, and 1-octen-3-one showed the highest FD factor (FD = 729), followed by 4-hydroxy-2,5-dimethylfuran-3(2H)-one, (E)-2-octenal, 1-octen-3-ol and hexanal with FD factors of 243. However, hexanal, 2-pentylfuran, 1-octen-3-one, (E)- 2-octenal, and 1-octen-3-ol contributed significantly to the odor characteristics of the fish soup; these results were consistent with previous studies [10]; these compounds delivered the odor of mushroom, beany, grass and cucumber, creating a characteristic odor in fish soup. The FD factors of these odor-active compounds were then obtained, including 6 methylhept-5-en-2-one (81), (E,Z)-2,6-nonadienal (27), 2-ethyl-3,5-dimethylpyrazine (27) and 3-methylpentanoic acid (27). Since these compounds showed a high FD factor, they were considered to be the key odorants in fish soup. However, some typical substances that usually affect the fish odor, such as heptanal, nonanal and trimethylamine, did not contribute to the overall odor profile and were not even detected in this study, probably due

to their changes in content after boiling; this was for the first time that 3-methylpentanoic acid was detected as a compound with sweaty odor with an FD factor of 9, and also as a contributing compound to the fish odor. Importantly, the advantages of GC × GC-O-MS in the detection of odor compounds were also confirmed. In addition, the FD factor for acetic acid, 2,6-dimethylpyrazine and 5-methyl-2-thiophenecarboxaldehyde was 9, while furfural and dimethyl trisulfide had an FD factor of 3. Although the FD factors of these compounds are relatively small compared to other substances, since they were also smelled at the sniffing port, so it was predicted that they also made contributions to the overall odor profile of the fish soup.


**Table 2.** FD factors and relative odor activity values (r-OAV) of key aroma compounds.

a: RI, The Retention index on capillaries DB-WAX and DB-5. b: The FD factors of all aroma compounds in three samples were calculated by varying the split ratio of the GC inlet. c: Odor thresholds were referenced from a book named Odor thresholds compilations of odor threshold values in air, water and other media (second enlarged and revised edition).

The concentrations of the compounds after AEDA analysis were calculated by a semiquantitative method. Table 2 summarizes the concentrations as well as the results of r-OAV calculations. Among the odor compounds, 1-octen-3-one had the highest r-OAV of 7500, attributed to the fishy odor; this result was also consistent with the previous results [29]. Secondly, 2-pentylfuran had a higher r-OAV of 3436, and its odor attribute was beany; this compound has not been previously reported in the literature as a fishy odorant and may be generated from fatty acids during the boiling of fish scraps. (E)-2-Octenal, hexanal, 4-hydroxy-2,5-dimethylfuran-3(2H)-one, 1-octen-3-ol, and 6-methylhept-5-en-2-one had the r-OAVs of 1452, 1212, 904, 869 and 602, respectively. The r-OAVs of these compounds were consistent with the AEDA results. Additionally, these compounds were previously reported as fishy substances [30]. However, there was one exception, such as 6-methylhept-5-en-2-one with an FD factor of 81, but it showed a lower FD factor relative to other substances with an FD factor of 27 (2-ethyl-3,5-dimethylpyrazine and (E,Z)-2,6-nonadienal). The results of the above analysis indirectly suggested the existence of some synergistic effect between odor compounds, rather than a single relationship, which could potentially enhance or diminish the overall aroma; these results better illustrate the importance of olfactory detection when analyzing the odor compounds.

#### *3.4. Omission Experiment*

Aroma model was constructed by selecting the top 11 key odor compounds in order of r-OAV value. Triangle tests were performed for the deletion of the 11 key odor compounds one by one and the results are presented in Table 3. The absence of 2-pentylfuran, (E)-2 octenal, 1-octen-3-one and hexanal was detected by all panelists, which indicates the importance of these compounds in contributing to the overall odor profile. While in the absence of acetic acid, 4-hydroxy-2,5-dimethylfuran-3(2H)-one, and 3-methylpentanoic acid, the results showed insignificant, i.e., less impact on the overall aroma. Thus, the omission experiments further confirmed that 2-pentylfuran, (E)-2-octenal, 1-octen-3-one, hexanal, 1-octen-3-ol, 6-methylhept-5-en-2-one, (E,Z)-2,6-nonadienal and 2-ethyl-3,5-dimethylpyrazine were the key odor compounds in the fish soup.



a: The compounds which r-OAV in Table 3 showing a significant difference among OF, PC and AO. b: Number of correct judgments from 12 assessors by the triangle test. c: Significance: "\*", significant (α ≤ 0.05); "\*\*", highly significant (α ≤ 0.01); "\*\*\*", very highly significant (α ≤ 0.001); "-", non-significant.

#### **4. Conclusions**

In this study, fish scraps were used as raw materials for the preparation of fish soup. Five volatile extraction methods were compared in order to select the best method for extracting odor compounds from fish soup. Finally, SPME combined with GC-O-MS and GC × GC-O-MS were selected to analyze odor-active compounds together. A total of 57 volatile compounds were identified. Among these, 38 volatile compounds were detected by GC-O-MS, of which 10 were declared as odor-active compounds, i.e., substances that could be smelled at the sniffing port; while 39 volatile compounds were detected by GC × GC-O-MS, of which 12 were odor-active. The above results revealed that the combination of these two analytical methods can provide a more comprehensive analysis of the volatile aroma compounds in fish soup. The results of AEDA analysis showed that 1-octen-3-one and 2-pentylfuran contributed most significantly to the odor of fish soup, followed by (E)-2-octenal, hexanal, 4-hydroxy-2,5-dimethylfuran-3(2H)-one, 1-octen-3-ol, 2-ethyl-3,5 dimethylpyrazine, (2E,6Z)-nona-2,6-dienal, 6-methylhept-5-en-2-one and 3-methylvaleric acid. From the r-OAV results, it was concluded that the compound with the highest r-OAV was 1-octen-3-one, followed by 2-pentylfuran. In summary, 2-pentylfuran, (E)-2-octenal, 1-octen-3-one, hexanal, 1-octen-3-ol, 6-methylhept-5-en-2-one, (E,Z)-2,6-nonadienal and 2-ethyl-3,5-dimethylpyrazine were the most important odor-active compounds in the fish soup as confirmed by omission experiment and triangle test. In the subsequent studies, based on the fishy compounds in this study, enzymatic digestion of fish soup using protease was performed to remove the fishy compounds and enhance the umami taste of the fish soup, contributing to the production of a full-flavored and distinctive seasoning; this approach not only improves the comprehensive utilization value of fish scraps, but also avoids the environmental pollution caused by discarding fish scraps.

**Author Contributions:** Writing—original draft, formal analysis, Y.Z. (Yuanyuan Zhang); conceptualization, investigation, validation, L.T.; conceptualization, methodology, Y.Z. (Yu Zhang); writing—review & editing, H.S., Y.Z. (Yu Zhang) and A.R.; supervision, resources, H.S., W.P., L.G. and C.J.; methodology, A.R. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was financially supported by Hunan Province Jiapinjiawei Biotechnology Co., Ltd. Funder: Huanlu Song. No funding number is available.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data will be available at request.

**Acknowledgments:** Authors gratefully acknowledge the panelists of Laboratory of Molecular Sensory Science, Beijing Technology and Business University.

**Conflicts of Interest:** The authors declare no conflict of interest. The methods used in the manuscript are those commonly used in the field of food science and do not require ethical proof.

**Sample Availability:** Samples of the compounds are not available from the authors.

#### **References**


## *Review* **Recent Progress on Techniques in the Detection of Aflatoxin B1 in Edible Oil: A Mini Review**

**Shipeng Yin 1, Liqiong Niu <sup>2</sup> and Yuanfa Liu 1,\***


**Abstract:** Contamination of agricultural products and foods by aflatoxin B1 (AFB1) is becoming a serious global problem, and the presence of AFB1 in edible oil is frequent and has become inevitable, especially in underdeveloped countries and regions. As AFB1 results from a possible degradation of aflatoxins and the interaction of the resulting toxic compound with food components, it could cause chronic disease or severe cancers, increasing morbidity and mortality. Therefore, rapid and reliable detection methods are essential for checking AFB1 occurrence in foodstuffs to ensure food safety. Recently, new biosensor technologies have become a research hotspot due to their characteristics of speed and accuracy. This review describes various technologies such as chromatographic and spectroscopic techniques, ELISA techniques, and biosensing techniques, along with their advantages and weaknesses, for AFB1 control in edible oil and provides new insight into AFB1 detection for future work. Although compared with other technologies, biosensor technology involves the cross integration of multiple technologies, such as spectral technology and new nano materials, and has great potential, some challenges regarding their stability, cost, etc., need further studies.

**Keywords:** aflatoxin B1; edible oil; chromatographic technology; spectroscopic technology; biosensor technology; recognition elements

#### **1. Introduction**

Food security has always been an issue of concern in the international community, and, in recent years, food contamination has become a major factor affecting food security. Contaminated food can not only adversely influence human health (poisoning events, chronic diseases, etc.) but also affect and slow down the economy. When people consume contaminated food, they need to spend a lot of money and time on treatment. There are many factors causing food contamination, such as biological, chemical, and physical factors. Among these, microbial contamination is common and mainly includes contamination by bacteria, fungi, molds, viruses, or their toxins and by-products [1,2]. Mycotoxins are common food contaminants, which can cause changes in the appearance, flavor, smell, and other characteristics of food [3–7]. Mycotoxins are secondary metabolites produced by fungi (e.g., *Fusarium*, *Aspergillus*, and *Penicillium*) that have multiple toxic effects on organisms and contaminate agricultural products (cereals, milk, etc.). More than 400 kinds of mycotoxins have been identified. Among them, aflatoxins (AFs) have become one of the major concerns due to their high toxicity and carcinogenicity, causing approximately 25% of animal deaths [8–12].

Edible vegetable oil plays an irreplaceable role in the human diet. The world oil crop output has increased year by year and had reached 635.5 million tons by 2021 [13]. From the growth of oil crops to the final product, i.e., oil, each link may be affected by external factors (such as mycotoxins), which may affect the quality and safety of edible vegetable oil [14]. This is because most oil crops, such as corn, peanut, soybean, rapeseed, sunflower seeds, olives, and nuts, are seasonal. During the growth process, they will be

**Citation:** Yin, S.; Niu, L.; Liu, Y. Recent Progress on Techniques in the Detection of Aflatoxin B1 in Edible Oil: A Mini Review. *Molecules* **2022**, *27*, 6141. https://doi.org/10.3390/ molecules27196141

Academic Editors: Yanping Chen and Weiying Lu

Received: 10 August 2022 Accepted: 15 September 2022 Published: 20 September 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

affected by climate, pests, and other factors and can be easily be infected by *Aspergillus* flavus. After harvest, the oil may deteriorate or be affected by mildew due to storage conditions (such as temperature and humidity, etc.) and storage methods [15]. At the same time, during the production of edible oil, fresh-pressed edible oil is vulnerable to contamination of raw materials infected with *Aspergillus* by aflatoxin B1 (AFB1) [16–22]. Therefore, contamination of edible vegetable oil products by AFB1 is a serious food safety problem (Figure 1) [20,23–25].

**Figure 1.** Harmful effects of different types aflatoxins contaminated edible oil.

The presence of aflatoxin is usually detected by using precision instruments, such as high-performance liquid chromatography–mass spectrometry (HPLC–MS), high-performance liquid chromatography–fluorescence detection (HPLC–FD), or other molecular techniques, while rapid detection is mainly realized by enzymatic immunoassay ELISA [26,27]. Although different methods are available for the detection of AFB1 toxicity, these methods require expensive equipment and complex sample pretreatment or can only be performed at relatively high concentrations [28]. Therefore, simple, sensitive, efficient, economical, rapid, and stable AFB1 detection methods are required. Recently, new technologies, such as biosensors, have been applied in many fields, such as health care and food detection. Because of their key advantages, such as convenient operation, rapid response, and excellent portability, these technologies can detect harmful substances in food sensitively and accurately, helping effectively avoid their harmful effects. They have attracted increasing attention of researchers and also promoted the rapid development of biosensors. With progress in nanotechnology, scientists are paying special attention to biosensors based on nanomaterials. These new biosensors or detection systems are sensitive, rapid, consistent, and cost-effective and can be used to detect AFB1 in food [29–33].

Regarding the increased importance of biosensors for accurate detection of AFB1 in edible oil, we have summarized the recent advances in biosensors for AFB1 analysis, specifically from the points of view of the development of novel bioinspired recognition elements and nanomaterials-based electrochemical biosensors.

Therefore, we searched PubMed and web of science for publications describing the detection technology of aflatoxin B1 in edible oil. Search terms were as follows: aflatoxin B1 OR AFB1 OR *Aspergillus* OR mycotoxins OR AFB2 OR AFG1 OR AFG2 OR AFM1 OR AFM2 OR AFs OR AFBO OR CYP450 OR edible oil OR vegetable oil OR corn oil OR peanut oil OR soybean oil OR sesame oil OR rapeseed oil OR sunflower seeds oil OR olives oil OR nuts oil OR maize oil OR canola oil OR blend oil OR coconut oil OR almond oil OR rice oil OR palm oil OR tea oil OR chromatographic technology OR spectroscopic technology OR immunological technology OR biosensor technology OR QuEChERS OR Fluorescence spectrophotometry OR Infrared spectroscopy OR Terahertz spectroscopy OR surface-enhanced raman spectroscopy (SERS) OR enzyme-linked immunosorbent assay (ELISA) OR amperometric OR impedometric OR electrochemical impedance spectroscopy (EIS) OR voltammetry (potentiometric) OR Conductometric OR LOD OR chromogenic OR Luminogenic OR Chemiluminescence OR Gravimetric OR Piezoelectric OR Magnetoelastic OR Acoustic OR electrodes (SPEs)OR SRP OR biosensors OR Nanomaterial-based biosensors OR electrochemical biosensors OR bioinspired recognition elements OR antibodies OR aptamers OR molecularly imprinted polymers OR Phylogenetic Evolution of Ligands for Exponential Enrichment (SELEX) OR fluorescence resonance energy transfer (FRET).

Publications until 29 August 2022 were included. This review only had the detection technology targeted at aflatoxin B1 in edible oil, and that had not included other types of toxins or other food carriers. After 4692 publications were searched, 596 full-text articles were reviewed and 132 articles were finally identified to meet our requirements.

#### **2. Importance of Aflatoxins**

Aflatoxins are a type of mycotoxins. They are highly toxic metabolites of fungi, produced in food and agricultural products. They have severe toxic effects, such as immunosuppressive, nephrotoxic, teratogenic, carcinogenic, and mutagenic, on human and animal health [34–38].

Aflatoxins can be divided into aflatoxin B1 (AFB1), aflatoxin B2 (AFB2), aflatoxin G1 (AFG1), and aflatoxin G2 (AFG2) according to their fluorescence properties and chromatographic mobility (Figure 1) [39–41]. Aflatoxin M1 (AFM1) and aflatoxin M2 (AFM2) are hydroxylated metabolites of AFB1 and AFB2, respectively. AFB1 is the most toxic among all AF species, with a high incidence rate and the most complex detection mechanism (Figure 2) [42].

**Figure 2.** Main mechanisms of toxicity of aflatoxin B1 for humans.

AFB1 is a powerful carcinogenic, teratogenic, mutagenic, immunotoxic, hepatotoxic, and reproducible poison. Previous studies have shown that the toxicity of AFB1 is 10, 68, and 416 times that of KCN, arsenic and melamine, respectively [43,44] (Figure 2). Therefore, AFB1 has been classified as a class 1 carcinogen by many international authoritative organizations or institutions [45,46]. Due to the structural double bonds in the furan ring, AFB1 has high carcinogenicity and toxicity [17,47]. The lipophilic structure of atrial fibrillation promotes its entry into the blood through gastrointestinal and respiratory tracts [48,49]. Once AFB1 enters blood, it is distributed in various tissues and accumulates in the liver or other organs, resulting in liver cancer (Figure 3). In the liver, AFB1 produces a variety of metabolites through the hydroxylation and demethylation of the first-stage drug metabolism enzymes (for example, cytochrome P450 oxidase and CYP450 superfamily members, such as CYP1A2, CYP3A4, and CYP2A6) [50]. Metabolic reaction (internal and external) activates the final carcinogen AFB1 -8,9-epoxy metabolite, which covalently binds to cellular macromolecules (DNA, RNA, or protein) and plays a key role in acute and chronic poisoning. AFB1 residues also destroy the function of tumor suppressor genes (p53 and Rb) in the liver, which affects normal cells and leads to liver injury, increasing the probability of tumor and liver cirrhosis [51–55]. It is estimated that about 30% of liver cancers in the world are caused by AFB1. Its toxicity increases the infection rate of hepatitis B virus (HBV) and the risk of liver cancer [56]. A recent study found that the synergistic effect of AFB1 and HBV leads to liver cancer [50]. The reason is that HBV infection directly or indirectly exposes hepatocytes to AFB1 sensitive to tumors. The toxic effect of AFB1 is also related to dose, age, sex, nutrition, exposure time, and type [57]. In addition, AFB1 can be transmitted to the fetus through the placenta and affect the health of infants [58]. AFB1 exposure also inhibits immunity, thereby increasing the susceptibility to immunodeficiency virus attack and the probability of infection with other infectious diseases [59–63].

**Figure 3.** Illustration of the mechanism of hepatocellular carcinoma caused by ingestion of AFB1 contaminated foods.

#### **3. AFB1 Regulations on Edible Oil**

Because AFB1 poses many hazards to the human body, many governments and international research institutions have made many efforts to control AFB1 pollution in different foods. For example, the FAO and the European Commission and Codex Alimentarius Commission have formulated regulations regarding the content of AFB1 in various foods to ensure consumer safety [64–69].

As for edible oil, most countries have no legislative restrictions and only a few countries, such as China, have effective regulations, laws, and standards for the highest level of

AFB1 in different edible oils (Table 1). Due to some adverse conditions in the traditional oil processing process, AFB1 is usually degraded to the normal level in the extraction and refining process [17,70]. The EU has strict regulatory norms. The total amount of AFB1 and AF allowed in oilseeds is restricted to 2 and 4 μg kg−1, respectively. However, the maximum limit of AFs in oils has not been determined. The corresponding regulations in China, the United States, Kenya, and Thailand clearly stipulate the maximum level of total AFs in all edible oils, but the maximum level required is different. It is worth mentioning that in China, the AFB1 limit in corn and peanut oil is stipulated to be 20 μg kg−1, which may be because corn and peanut are most vulnerable to aflatoxin pollution [71,72]. See Table 1 for specific differences.


**Table 1.** The maximum limits (μg kg−1) established for major AFB1 in some countries/regions for edible oils.

#### **4. Methods for Detecting AFB1 in Edible Oil**

The matrix is too complex for edible oil, and the mycotoxin content is relatively low, making it difficult to detect AFB1. Therefore, researchers have developed various traditional and modern methods to detect AFB1 in oil. AFB1 detection technology is mainly divided into chromatographic technology, spectroscopic technology, immunological technology, and biosensor technology [16,77].

Figure 4 briefly summarizes the LOD timelines for AFB1 detection in edible oils published from 2007 to 2022 mentioned in this review. It can be seen from the figure that with the advancement of time, no matter what type of detection technology or which specific detection method is used, the LOD of AFB1 in edible oil tends to be lower. This shows that people have a great interest in the detection of AFB1 in edible oil. At the same time, the wide use of new materials represented by nanomaterials highlights the interdisciplinary characteristics of new sensors. Next is a brief introduction of the identification method of AFB1, including its advantages and disadvantages, combined with actual cases.

**Figure 4.** Timeline of the limit of detection on AFB1 in edible oil by different (**a**) detection technology and (**b**) detection method. CI: Chemiluminescence immunoassay.

#### *4.1. Chromatographic Technology*

#### 4.1.1. High Performance Liquid Chromatography (HPLC)

High-performance liquid chromatography is a common official detection method. Many countries and institutions have used it, such as China's national standard, the European Committee for Standardization (CEN), and the association of analytical organizations (AOAC). One characteristic of the HPLC method is that it can measure multiple targets with high sensitivity [78]. In recent years, researchers have developed new detection strategies combining HPLC with other sensors, such as fluorescence detection (FLD), ultraviolet (UV) detection, diode array detection, and mass spectrometry (MS) [79,80]. Compared to traditional HPLC, this further improves the reliability, sensitivity, and accuracy of target analytes and is widely used to detect harmful substances in food. For example, HPLC combined with FLD is the standard method for detecting AFB1 in edible vegetable oil [81–86]. HPLC–FLD was able to detect AFB1 levels as low as 0.01–0.04 μg kg−<sup>1</sup> [81] and 0.005–0.03 μg L−<sup>1</sup> [82].

Recently, liquid chromatography–tandem mass spectrometry (LC–MS–MS) methods are being increasingly used for the analysis of mycotoxins [85]. They have the advantages of not having a sample purification limitation during extraction, high resolution, high sensitivity, and suitability for various edible vegetable oils [19,87–99]. GC analysis is mostly used for volatile substances, and most mycotoxins are non-volatile, further limiting

the application of GC in mycotoxin detection. A similar procedure to HPLC, UHPLC or UPLC is also used on the column to improve the resolution of AFB1. Hidalgo et al. [100] developed a new analytical method by coupling UHPLC to a triple quadrupole analyzer (UHPLC–QqQ–MS/MS), which was well validated and applied to monitor mycotoxins, including AFB1, in 194 samples of edible vegetable oil.

Many commonly used methods require sample preparation due to the different matrices of edible oil. Currently, a variety of methods are available for the extraction and isolation of mycotoxins from oil, such as liquid–liquid extraction or partitioning (LLE), frequently reported in the literature [101–105]; solid–phase extraction (SPE) [105–109]; immune affinity columns (IACs) [81,94]; IAC combined with dispersive liquid–liquid microextraction (DLLME) [91]; multifunctional cleanup columns [110]; the QuEChERS system [90]; gel permeation chromatography (GPC) [111]; immune assay extraction; and low-temperature cleanup (LTC) [112–115]. However, each method has its advantages and limitations. Thus, which method to choose still depends on the type of food matrix, mycotoxin characterization, and detection techniques [116].

#### 4.1.2. Thin-Layer Chromatography (TLC)

Thin-layer chromatography (TLC) is an adsorption thin-layer chromatographic separation method suitable for complex mixed samples [117,118]. Since its development in the 1950s, thin-layer chromatography has been widely used in, for example, biology, medicine, and the chemical industry. It has recently been used in food analysis and quality control and has become a conventional technology in laboratories. Many reports have shown that TLC can be applied to all stages of the food industry, such as the stage of traditional substances, represented by food raw materials, ingredients, and additives, and the stage of unconventional substances, represented by harmful substances and pollutants. The detection and determination of compounds cover almost all substance categories [119–122].

Thin-layer chromatography uses the different adsorption capacities of each component to the same adsorbent so that when the mobile phase (the solvent) is flowing through the stationary phase (the adsorbent), there is continuous adsorption, desorption, readsorption, and redesorption to achieve the mutual separation of each component [123].

Although the TLC method has matured, it still has shortcomings, such as a low detection accuracy, volatility during the experiment being harmful to the experimental operators and the environment, and complex sample pretreatment [124,125]. In recent years, an interdisciplinary approach, such as the combination of TLC with image analysis and with new technologies, such as surface-enhanced Raman spectroscopy, mass spectrometry, and nuclear magnetic resonance, has further promoted the development of thin-layer chromatography and enhanced the practicability of this method in food analysis [126–129]. TLC is used to detect harmful substances in various foods, such as AF in edible oil, making it an effective analytical tool in food science methods [124,130].

#### *4.2. Spectroscopic Technology*

#### 4.2.1. Fluorescence Spectrophotometry

Spectrum-based sensing technology has been developed and used to assess AFs contamination in food [131]. Among many spectral techniques, fluorescence spectrometry shows certain potential in determining AFs in a variety of agricultural products and foods [125,132,133]. Fluorescence spectrometry uses the target molecules in the sample to absorb ultraviolet or visible light to produce fluorescence and determine its molecular structure. It has excellent detection sensitivity and specificity in the study of AFs and other chemical components [134,135]. The study found that the fluorescence phenomenon is conducive to the characterization and monitoring of target detection objects. For example, AFB1 can emit a specific range of fluorescence (425–500 nm) under the excitation of UV light source (340–400 nm), which provides the possibility of using fluorescence spectroscopy to analyze AFB1 in different foods [135,136]. In recent years, laser-induced fluorescence (LIF) technology has developed rapidly and attracted more attention because it uses a certain

wavelength of laser light source and has better specificity and detection sensitivity. The advantage of LIF is that it can realize online, rapid and nondestructive direct detection according to the characteristic fluorescence peak of AFB1. Researchers have developed a detection model based on LIF, which can quickly and accurately screen AFB1 in different edible oils. The information and conclusions obtained in the study further show that LIF technology can be used for rapid and nondestructive detection of AFB1 in different edible oils [19,137]. However, LIF is also vulnerable to the interference of external factors, such as the power and accuracy of the instrument, the environmental factors of temperature and humidity, and the physical and chemical index factors of the detected object. Although this limits the wide application of LIF technology, researchers are still trying and exploring.

#### 4.2.2. Infrared (IR) Spectroscopy

Infrared spectroscopy (IRs) has the characteristics of rapid detection, simple sample preparation process and strong adaptability. It has been widely proven to be an effective food safety detection and control technology. Because IR covers a wide range of electromagnetic spectra (780 to 2500 nm), IR can be applied to the detection of a variety of foods including edible oil, meat, aquatic products, fruits and vegetables [138–144]. When IRs radiation penetrates the sample, the radiation is reflected, absorbed or transmitted by molecular bonds, resulting in the energy change of light, which can reflect some characteristic chemical bonds, thus reflecting the characteristics of the tested product [145,146]. In the application of edible oil, IR shows many abilities, such as distinguishing different kinds of oil, grading the quality of oil, detecting harmful substances in oil, etc. [138,143,147–150]. Using near infrared (NIR) technology to detect mold in edible oil has also been a research hotspot in recent years. Researchers have promoted the further application and development of IR technology by establishing qualitative and quantitative analysis models for AFB1 pollution in edible oil [151–153].

#### 4.2.3. Terahertz (THz) Spectroscopy

With the development of optical and electronic technology, terahertz spectroscopy (THz) has been a revolutionary development, and shows great potential as a new technology tool for nondestructive food testing [154–157]. As a technical information link between microwave spectroscopy and infrared spectroscopy, THz has the characteristics of both, making it widely used in basic research and industrial practice [158,159]. Like other spectral technologies, thanks to the development of chemometrics methods, THz has become a powerful technical tool in the food industry, due to its strong detection and quantification capabilities [156,157,160]. Through the combination of THz and chemometrics methods, researchers have constructed a rapid nondestructive detection model for AFB1 in edible oil. Although the accuracy is slightly lower than other conventional analysis methods, it provides a possibility for THz in food safety detection [161]. In a recent study, researchers further improved the accuracy of THz in detecting AFB1 in edible oil by adding pretreatment and other methods on the basis of predecessors, and reduced the LOD of AFB1 to 1 μg kg−1, and the accuracy is improved to more than 90% [161,162]. The cross integration of THz and chemometrics and other disciplines is conducive to promoting its application and development in the detection of AFS in the edible oil industry. At the same time, the limitations of THz should also be clear, such as the low detection limit and sensitivity advantage are not obvious, the penetration of the detected object is limited, there is scattering effect, the technology is expensive, the database is lack, etc. [163].

#### 4.2.4. Surface-Enhanced Raman Spectroscopy (SERS)

As a complementary analysis technology of IR, the Raman spectroscopy (RS) technology is sensitive to the symmetrical vibration of covalent bonds of non-polar groups (such as C=O, C-C and S-S) [164–166]. Therefore, RS has the advantages of being fast, sensitive and simple in the detection and evaluation system of food [165,167,168]. However, traditional RS has some limitations, such as Raman scattering. Therefore, researchers

have developed SERS signal enhancement technology represented by electromagnetic field enhancement and chemical enhancement [165]. At present, the application of SERS technology in the detection of AFs is still challenging, and the intersection of technology development and multidisciplinarity (such as materials science, stoichiometry, etc.) is the focus of researchers [165]. In recent years, researchers have reported a variety of SERS schemes for AFB1 detection in edible oil, such as SERS tag detection using antibodies and aptamers, sandwich immunoassay based on SERS, etc. [169–174]. The growing research results show that SERS technology is becoming a powerful tool to ensure the safety development of the food industry, especially in the safety supervision of AFs. However, it cannot be denied that challenges still exist, such as the development of targeted new materials, the optimization of key core technologies, and the practical application of research results [175–177].

#### *4.3. Immunological Technology*

#### Enzyme-Linked Immunosorbent Assay (ELISA)

In recent years, researchers have often used immunochemical methods to determine mycotoxins in food, in addition to traditional chromatographic techniques. The core of immunochemistry is the specific interaction between immunoglobulin (Igs) and antigen (Ag). Several immunochemical methods have been applied to detect mycotoxins in edible vegetable oils, such as enzyme-linked immunosorbent assay (ELISA) and biosensors based on immunoassay.

ELISA is one of the most commonly used methods for detecting mycotoxins [24]. It has been designed and developed on the basis of the principle of specific immune responses between Igs and Ags. The specificity of this immunoassay is due to the use of enzyme-labeled Igs or Ags and solid-matrix-restricted immunoglobulins to capture unlabeled silver in the analyte and detect it with labeled immunoglobulins. Although ELISA is well developed and widely used in food analysis, clinical practice, biotechnology, environmental, chemical, and other industries, it still has several deficiencies, such as excessive dependence on the matrix caused by the interaction between the target antigen and matrix components. The standard ELISA is composed of four main parts (immunerecognition element, sorbent substrate, enzyme label, and chromogenic reagent), and the deficiency of the central part is the root cause of the limitation of ELISA. In recent years, researchers have used the cross-fusion of multiple technologies to drive the performance of one of the components or the whole ELISA, especially in terms of sensitivity, accuracy, and stability [27].

For mycotoxins, due to the high singularity of ELISA, the developed kit has specific recognition ability and has been widely used in the detection of mycotoxins [70]. For example, Qi et al. [20] used ELISA and UPLC–MS/MS to detect AFB1 in peanut oil, although the LOD was only 1.08 μg kg−1, much higher than the LOD of UPLC–MS/MS (the LOD is 0.01 μg kg−1) [20]. It has been affirmed because of its accuracy, rapidity, and other advantages. For the actual detection of other harmful substances, such as AFB1, AFB2, AFG1 and AFG2, in different edible oils (oils of soybean, coconut, peanut, fennel, melon, and palm kernel), ELISA showed satisfactory results and the concentration was lower than the legislative limit [178–180]. On this basis, the researchers developed a commercial ELISA kit that can detect AFB1, which can be applied to a variety of samples including edible oil, and the detection limit can be as low as 3 ppb. Although the current ELISA technology or kit still has problems such being as time-consuming, high cost, and cumbersome operation, with the advancement of technology, ELISA technology shows strong application potential [27].

#### *4.4. Electrochemical Biosensing Technology*

Due to rapidity, small footprint, economy, sensitivity, and unique capabilities, electrochemical biosensing devices have received particular attention in assessing food quality, mainly reflecting AFB1 levels in food samples [181]. The AFB1 electrochemical biosensor can produce various types of analytical signals, such as voltage, current, and impedance [182,183]. The standard transduction methods are amperometric, electrochemical impedance spectroscopy (EIS), and voltammetry (potentiometry).

#### 4.4.1. Amperometric Biosensors

The amperometric biosensor is an electrochemical device with high selectivity and sensitivity that takes the change in the measuring current as the analysis signal. Because the change in the current is closely related to the concentration of AFB1 in food samples and the change can be achieved by maintaining a stable potential, an amperometric biosensor is relatively perfect. A typical amperometric biosensor consists of two- or three-electrode systems (containing a functional electrode, a reference electrode, and an auxiliary electrode), and the analytical performance of the latter is significantly higher than that of the former (Figure 5) [181]. This is because the additional auxiliary electrode not only increases the area of the detection surface but also increases the current between it and the functional electrode, as well as the operating potential between the functional electrode and the reference electrode, thereby enhancing the changes in the detection process of AFB1 in food in electronic dynamics. On the contrary, the dual-electrode system does not include auxiliary electrodes, which may lose their function at high temperatures. Therefore, amperometric biosensors with dual-electrode systems are not used to analyze the quality of food samples [181].

**Figure 5.** Scheme of the two or three-electrode setup used in electrochemical methods.

Even functional electrodes are usually made of inert metal materials (such as platinum, gold) or carbon (graphite, glassy carbon). The main drawback is reproducibility of measurements. Currently, printed electrodes have become a good substitute because their cost and mass production can be controlled [70,78].

Researchers used two kinds of nanomaterials with different charges to deposit on the electrode alternately, obtaining a multilayer electrode with a sandwich structure with excellent conductivity and rich electrochemical active sites [184]. Such a biosensor has good selectivity, reproducibility, and stability. In the subsequent optimization test, the optimized electrochemical biosensor was found to have significant stability and even after being placed for a period of time, it showed good LOD (0.002 ng mL−1). This sensor is believed to be one of the best biosensors for detecting mycotoxins. Researchers applied the electrochemical biosensor to detect AFB1 in real oil samples and found that it has a good recovery rate (98.11–103.36%).

Xuan et al. [185] developed an integrated AFB1 detection platform that uses disposable screen-printed electrodes (SPEs), allowing routine detection without electrode modification. According to the SPE used, the platform can simplify the tedious sample processing process through high-throughput processing, reduce operating errors, and improve experimental reproducibility, which can benefit large-scale sample processing. The detectable concentration range of AFB1 was 0.08–800 μg kg−<sup>1</sup> with a LOD of 0.05 μg kg−1. Analysis of real samples and verification of the method showed the results of the new sensor to be consistent with those of the classical method (LC–MS/MS), indicating that the developed method has the potential to monitor AFB1 in peanut oil.

Another study reported a new aflatoxin biosensor based on the AFB1 inhibition of acetylcholinesterase (AChE) [186]. The core of this method is to immobilize choline oxidase on a screen-printed electrode modified with Prussian blue (PB). The electrode used in the biosensor can detect H2O2 at low potential. As per the results, the linear operating range of the biosensor is estimated to be 10–60 ppb and the LOD is 2 ppb. On using real olive oil samples to evaluate the sensor, the recovery rate was found to be 78 ± 9% at 10 ppb.

#### 4.4.2. Electrochemical Impedance Spectroscopy (EIS)

Electrochemical impedance spectroscopy (EIS) technology is an effective monitoring tool for identifying and monitoring changes in mycotoxins at the interface between electrode surface modifications. When the target analyte is combined with a biometric element, it generates an electrochemical response by changing conductivity and capacitance through an impedance biosensor [187]. These biosensors monitor the impedance changes caused by the interaction between the target detection object, such as AFB1, and the biometric element fixed on the working electrode, and display the detection results in the form changed electron flow on the working electrode [188,189]. Typical potentiometric sensors are also suitable for the three-electrode system.

The main parameter of the EIS is the charge transfer resistance value (RCT), which is closely related to the reaction between immobilized mycotoxins and the antibody antigen and is also proportional to the target detector/concentration of the target [65,66]. For determining AFB1, Yu et al. [190] reported a sensitive and convenient EIS method involving MWCNT/RTIL/Ab-modified electrodes coated in bare GCE. The experimental results show that the resistance of the MWCNT/RTIL/Ab-modified electrode (605.6 Ω) is higher than that of bare GCE (151.9 Ω). When AFB1 was immobilized on MWCNT/RTIL/Abmodified electrodes, the increase in the electron transfer resistance (Ret) value was found to be directly related to the AFB1 amount. The specific interaction between AFB1 and Ab causes an increase in the Ret value, which leads to the production of electrically insulating biological conjugates, which will prevent the electron transfer process of redox probes. Therefore, the EIS measurement results are consistent with the above cyclic voltammetry results. Because of its simple characteristics, this method can be widely used to detect various agricultural products and edible oils.

For many researchers exploring mycotoxin detection methods, aptamer-based EIS has become a hot research topic. Aptamer-based impedance biosensors have achieved satisfactory results in detecting mycotoxins in food and have great potential for practical application in edible oils.

#### 4.4.3. Voltammetry Biosensors

Voltammetric biosensors solve the problem of obtaining analytical data using ionselective electrodes. Similar to amperometric biosensors, voltammetry also requires a twoor three-electrode system. When the current is constant, it can detect target analytes, such as AFB1, in food samples by evaluating the change in circuit potential between the functional electrode and the reference electrode [60,191].

Biosensors have also shown promising results in detecting the AFB1 content in edible oils. For example, Wang et al. [192] developed a new disposable electrochemical biosensor based on stripping voltammetry to detect copper ions released from copper apatite. The biosensor uses copper ions as a signal label to immobilize AFB1 antibody on a screen-printed carbon electrode (SPCE) modified by gold nanoparticles. The detection is performed by the voltammetric signal of the dissolution of copper ions released from acid hydrolysis of copper apatite, and copper apatite increases the number of loaded copper ions. The electrochemical signal is further amplified. Peanut oil was used to evaluate the reliability and application potential of biosensors. Researchers believe that this new method will be applied to many fields in the near future because of its many excellent characteristics (low cost, rapidity, accuracy, and high sensitivity).

#### 4.4.4. Nanomaterial-Based Biosensors

Recently, different nanomaterials, such as carbon and metal, have been used to modify the active surfaces of macroelectrodes and microelectrodes to design electrochemical biosensors for the detection of AFB1 [193–195]. This is because new biosensors directly use nanomaterials or other materials containing nanoparticles that show significant characteristics, such as high sensitivity and specificity for detecting targets, reliability, and consistency of products [181,196,197]. Nanomaterials significantly increase the effective surface area of biosensors and further improve the analytical performance [60,194]. Nanomaterials also enhance some characteristics of biometric elements in biosensor devices in terms of electrical, catalytic, optical, and thermal properties [198]. According to previous studies, some of the key functional enhancements are the enhanced immobilization of biomolecules, generation and expansion of analytical signals, and enhanced usability of fluorescent labels.

#### Characteristic of Nanomaterials Based Electrochemical Biosensors

The role of nanomaterials in biosensors is mainly reflected in the immobilization of biomolecules, signal generator, fluorescent labeling, and signal amplification.

Nanomaterials not only immobilize biomolecules but also increase the interaction between different molecular materials. In addition, nanomaterials enhance the stability of biomolecular immobilization, thereby increasing the signal strength of the immunoassay [31]. Metal nanomaterial particles, such as AgNPs and MOFs, can increase the surface area and biocompatibility of biomolecules bound to the detection target. However, nonmetallic nanomaterials show negatively charged functional groups, which can be used as an effective carrier to bind and fix with positively charged targets.

#### Signal Generator

Xue et al. [31] reported that, when the photoelectric signal changes, nanoparticles such as gold and silver can act as a signal generator. By adjusting the fluorescence signal generated by nanomaterials, a new AFB1 nanoprobe can be constructed. In addition, because these nanoparticles can be prepared in different sizes according to need, they have good functionality, stability, and scalability [133,199].

#### Fluorescent Label

Nanomaterials have unique optical properties that enable them to be widely used in a variety of disciplines, especially in the detection of hazardous substances in food. Nanomaterials can detect AFB1 by sensing optical signals (absorbance, chemiluminescence, fluorescence, etc.) [31]. Some nanomaterials, such as metal nano-ions and quantum dots, have been used as fluorescence quenching agents because of the ability of AFB1 to directly quench or reduce the fluorescence intensity. In addition, quantum dots have transformed fluorescein into a marker element that binds to aptamers or antibodies.

#### Signal Amplification

Nanomaterials can also be used as functional materials for various electrodes, signal components, etc., to amplify signals in various ways. For example, on the electrode surface of electrochemical sensors, nanomaterials such as gold and silver can amplify the analytical signal by enhancing the redox reaction. Some metal nanoparticles, such as gold, can amplify signals related to their characteristics, such as unique catalytic activity, biocompatibility, and multiple absorption sites. Carbon, graphene, and other non-metallic nanomaterials improve the analytical performance by increasing the surface area.

#### *4.5. Bioinspired Recognition Elements for Biosensors*

A biosensor is independent quantitative analysis equipment used to study the analytes required in different types of food samples. A biosensor consists of many parts [29,30] (Figure 6). Biometric elements are the core components of biosensors and can detect specific target analytes. The quality of biometric elements usually determines the specificity and sensitivity of analysis [200,201]. Biorecognition elements, including antibodies, aptamers, molecularly imprinted polymers, and enzymes, have been used to manufacture biosensors [34,64,202]. These elements show increased sensitivity and selectivity for target analytes. Critical biometric elements for developing biosensors to detect AFB1 in edible vegetable oil are elaborated below.

**Figure 6.** Schematic diagram of typical biosensor. Including analyzer, bioreceptor, transducer, electronic system (amplifier and processor), detector (for data processing).

#### 4.5.1. Antibody

Antibodies have been used as recognition elements for developing biosensors because of their specificity and sensitivity [200]. Biosensors that use antibodies as recognition elements are called immunosensors, and their mechanism relies on the specific recognition of aflatoxin epitopes by antibodies.

The first batch of polyclonal antibodies, developed in 1976, became the basis for most mycotoxin detection methods. In the following decades, polyclonal and monoclonal antibodies were the basis for most mycotoxin detection methods [192,200,203]. Today, in addition to monoclonal and polyclonal antibodies, various other types of antibodies are being used to detect target analytes. Researchers have developed an antibody-based immunosensor that can directly recognize AFB1 and is used in peanut oil with a concentration range of 0.001 to 100 ng mL<sup>−</sup>1, with a detection limit of 0.2 pg mL−<sup>1</sup> [192].

However, the production of monoclonal antibodies and polyclonal antibodies is complex and the antibodies degrade, denature, and aggregate easily [204,205]. In recent years, with the development of protein- and DNA-based new engineering technology, it has become possible to develop modified and recombinant antibodies (RAbs). RAbs integrate many advantages of biosensors, such as simple operation and a high degree of automation, high throughput screening, low requirements for configuration attributes, and the trend of

more miniaturization [200]. Zhao et al. [206] developed a novel method of MB-dcELISA for AFB1 based on the mimotope of an RAb and nanobody. This study effectively proved that compared with monoclonal antibodies, an RAb is more economical and easier to prepare. Compared with chemically synthesized toxic antigens, immunoassay is safer and performs better in validation studies. In real samples (corn germ oil and peanut oil), the LOD of AFB1 is as low as 0.13 ng mL−1. Other researchers also designed an RAb with increased sensitivity to low-molecular-weight haptens, and this RAb was validated in olive oil with a lower LOD (0.03 ng mL<sup>−</sup>1) for AFB1 [190].

Researchers recently found that, by increasing the immobilization of antibodies and giving full play to the characteristics of specific antibodies, the performance of the sensor can be effectively improved, and on this basis, some immunosensors have been developed for detecting AFB1 in edible oil [133,196]. For example, to determine AFB1, Shi et al. [207] proposed a novel immobilized immunosensor based on graphene supported with hybrid gold nanoparti-cles-poly4-aminobenzoic acid. In the study, after the reduction in graphene oxide by PABA via an epoxy ring opening reaction, the nanocomposite PABA-r-GO was obtained. Then, gold nanoparticles (AuNPs) were prepared on this basis to form a Au-PPABA-r-GO nanohybrid. The final sensor was obtained by the covalent binding of the COOH group of functional nanocomposites with an AFB1-specific antibody. The sensor has good performance (linear range 0.01–25 ng mL−<sup>1</sup> and LOD 0.001 ng mL−1) and has been successfully applied to detect real vegetable oil. This sensor also has good reproducibility and selectivity, especially stability, and can be stored at a low temperature for a long time.

#### 4.5.2. Aptamers

Aptamers are single-stranded RNA or DNA (20–90 oligonucleotide sequences with specific sequences) that can bind to various targets, such as ions, antibodies, proteins, cells, and organic molecules [208]. The particular recognition ability of aptamers relies on the three-dimensional structure of a high-affinity target-induced DNA three-dimensional structure. The researcher procured specific targets for aptamers by screening oligonucleotides using the Phylogenetic Evolution of Ligands for Exponential Enrichment (SELEX) program. Aptamer sensors are biosensors integrated with aptamers developed in the 1990s [200,208–211].

Recently, aptamers have attracted significant attention in food contamination analysis and are used for various sensing applications due to their inherent benefits: (1) aptamers are obtained from in vitro synthesis, so animals are not necessary; (2) aptamers have lower toxicity, immuno-genicity, and production cost; (3) aptamers have enhanced chemical and thermal stability; (4) aptamers have excellent batch-to-batch reproducibility; (5) aptamers have a smaller size and show a remarkable ability to penetrate the tissue and adhere to target molecules; and (6) it is possible to change their structure [193,212,213].

Notably, the immobilization of aptamers is a critical step in biosensor design because it can affect the affinity of aptamers for their targets and their long-term stability in fundamental sample analysis.

Therefore, researchers have developed many strategies to immobilize aptamers: (1) adsorption or π–π stacking interactions between DNA bases and modified graphene oxide (GO) interface [214], (2) aptamers with carboxylic acids on surfaces or nanomaterial covalent bonding of groups [184,215], (3) binding of sulfide aptamer with CdTe quantum dots (QDs) or gold-based materials [216], (4) binding with avidin or other affinity interactions based on biotin streptomycin affinity [200,217,218], and (5) hybridization with partially complementary single-stranded DNA previously fixed on the surface of nanoparticles [219–222].

Table 2 describes some examples of aptasensors recently reported for the detection of AFB1 in edible oils. About half of the previous reports have been based on fluorescent mycotoxin aptamer sensors. Some of them use metal or nanostructured materials, such as gold nanoparticles (AuNPs), GO, single-walled carbon nanotubes, or TiO2 tubes, and are used to prepare aptamer sensors.

Nanometer material has always been the focus of research, and its applications in biosensors are also diverse. Black phosphorus nanosheets (BPNSs) have great application prospects in biosensors due to their unique characteristics [223]. Wu et al. [224] developed a highly specific and sensitive aptamer sensor (UCNPs-BPNSs) based on the team's research on upconversion nanoparticles (UCNPs) [196]. The research team attached UCNPs to the surface of BPNSs at a very small space distance (less than 10 nm) through glutaraldehyde crosslinking method and π-π stacking effect method, and then constructed the fluorescence resonance energy transfer (FRET) system. This aptamer sensor can effectively detect AFB1 in peanut oil and other foods quantitatively with good linear range (0.2–500 ng mL<sup>−</sup>1) and LOD (0.028 ng mL<sup>−</sup>1).

Xia et al. [225] proposed a label-free, single-tube, homogeneous, and inexpensive assay for AFB1 based on fine-tunable double-ended stem aptamer beacons (DS) and the effect of aggregation-induced emission (AIE). The structure of the DS aptamer beacon can provide end protection against exonuclease I (EXO I) to the aptamer probe and endow it with specificity and a rapid response to the target AFB1. Compared with the traditional molecular beacon structure, the stability of the DS aptamer beacon can be adjusted by adjusting its two terminal stems so that the affinity and selectivity of the probe can be precisely optimized. Using an AIE-active fluorophore, which is illuminated by the aggregation of negatively charged DNA, AFB1 can be measured label-free. The method has been successfully applied to the analysis of AFB1 in peanut oil, with a total recovery of 93.59–109.30%. Therefore, beacon-based DS assays may help in real-time monitoring and control of AFB1 contamination.

Yang et al. [226] first devised a selection method based on rational truncation and postsplicing and developed a bivalent anti-AFB1 chimeric aptamer (B72) that was measured by micro-thermophoresis (MST) compared to the initial selection. The affinity of the anti-AFB1 aptamer (B50) increased by 188-fold, and the study also found that B72 has a dual binding site for AFB1, which is consistent with the experimental results obtained by isothermal titration calorimetry (ITC) and molecular docking simulations. Therefore, on the basis of the peroxidase-like activity of gold nanoparticles catalyzing 3,3,5,5-tetramethylbenzidine (TMB), an aptamer sensor of gold nanoparticles (AuNPs) was developed by the colorimetric detection of AFB1. The assay further validates the practical applicability of the chimeric aptamers. The aptasensor could identify AFB1 with an excellent linear range (5–5120 nM) and detection limit (1.88 nM) in the corn oil environmental test of H2O2. Therefore, this study can be called a general selection method for designing high-affinity aptamers and constructing novel aptamer-based biosensing platforms for high-sensitivity and specificity analysis of other targets.

Zhong et al. [227] manufactured an electrochemical aptamer sensor in a similar way for the sensitive detection of AFB1. The researchers used electrodeposited AuNPs to prepare AuNPs/ZIF-8 nanocomposites on glassy carbon electrodes (GCEs) decorated with the eight zeolite imidazolate framework (ZIF-8), which increased the surface area of the electro desorption molecular load. Compared with other previously reported sensors, the aptasensor developed under optimized conditions shows a more comprehensive linear range (10.0–1.0 × <sup>10</sup><sup>5</sup> pg mL−1) and a lower detection limit (1.82 pg mL−1). In addition, the constructed aptasensor possesses excellent selectivity, reproducibility, and stability. Moreover, the aptamer sensor has been successfully used to detect AFB1 in corn oil and peanut oil samples, and the recovery was between 93.49% and 106.9%, which proves the potential application value of this method. Researchers are very interested in this kind of electrochemical aptamer sensor. Wang et al. [228] also developed an AFB1 electrochemical aptamer sensor for detecting peanut oil in a similar way. The difference lies in the use of different composite materials (zinc and nickel bimetallic organic skeleton materials).

The hybridization chain reaction (HCR) is a commonly used isothermal nucleic acid amplification technique, and due to the characteristics such as no enzyme, high amplification efficiency etc., HCR is usually used as a new synthetic material technology and is widely used in various sensors. Researchers have fully combined the characteristics of HCR

to build a signal amplification strategy, which has been successfully applied to the sensitive detection of AFB1 [229–235]. Wang et al. [236] proposed a fluorescent aptamer sensor based on DNA walker, DNA tetrahedral nanostructures (DTNs) and network HCR. Among them, DNA walker was used as the signal amplifier induced by AFB1 target, and combined with self-assembled DTNs. Finally, based on network HCR, signal amplification is realized and sensitive detection of AFB1 in peanut oil was realized with with LOD of 0.492 pg mL−<sup>1</sup> and the linear range of 1–1000 pg mL−1. In the other report, Zuo et al. [237] combined DNAzyme with substrate chain (Zn-Sub) and enzyme chain (Zn-Enz) with HCR products to form a Y-shaped structure, which can significantly enhance the fluorescence intensity of the detection target. The fluorescent aptamer sensor proposed by researchers shows excellent performance with LOD of 0.22 nmol L−<sup>1</sup> and the linear range of 0.4–16 nmol L<sup>−</sup>1.

The emerging quantum dots (QDs), represented by carbon quantum dots (CQDS), graphene quantum dots (GODS) etc., have attracted great attention and are widely used in various sensor fields since their discovery because of their excellent optical properties, low toxicity, stability and low cost etc. [238–240]. QDs-based sensors can adopt different working mechanisms and be applied to detect different substances, including AFS in edible oil [51,173,238]. According to the characteristics of QDs, Xuan et al. [185] and Ye et al. [241] developed and constructed different magnetic control pretreatment platforms, which were actually applied to the detection of AFB1 in peanut oil and agricultural products, and both showed good detection characteristics. Other researchers used a quencher system composed of quantum dots and graphene oxide to detect AFB1 in peanut oil, which also showed good detection characteristics [242].

#### SERS

As mentioned above, SERS is a promising analytical tool with many advantages over traditional AFB1 detection methods, including high sensitivity, easy sample preprocessing, and non-destructive testing [166,174,243]. Compared with antibodies, aptamers have the advantages of low cost, easy synthesis, good stability and strong specificity to target molecules. With the mature development of aptamer manufacturing technology, they have gradually become one of the most potential recognition elements in SERS labeling detection.

Recently, several authors have combined advanced composite materials with SERS aptamer sensors to develop new procedures for AFB1 detection. For example, on the basis of the combination of a multifunctional capture probe (Fe3O4@Au report the strong Raman signal of probe 1 (AU)-4MBA@AgNSs-Apt), an ultrasensitive assay was successfully developed for a high-performance SERS aptamer sensor of AFB1. He et al. [174] reported that, in the presence of AFB1, the probe was released from the capture probe, resulting in a decrease in SERS intensity, possibly due to the specific binding affinity between the aptamer and AFB1. For AFB1 detection, a wide linear range, from 0.0001 to 100 ng mL<sup>−</sup>1, was obtained, with an R<sup>2</sup> of 0.9911, and the LOD was calculated as 0.40 pg mL<sup>−</sup>1. Finally, after extracting AFB1 from peanut oil samples, the SERS aptamer sensor was successfully applied to the analysis of AFB1, and the recovery was between 96.6% and 115%. Therefore, the novel SERS aptamer sensor is a promising analytical tool for detecting AFB1 in actual samples.

In the report by Yang et al. [169], with the help of the specific interaction between AFB1 and aptamer, a novel SERS-based universal aptamer sensor platform was constructed to detect AFB1. First, gold nanotriangle (GNT)-DTNB@Ag-DTNB nanotriangles (GDADNTs) were synthesized and used as SERS active substrates. These magnetic beads and aminoterminal-aptamer-conjugated magnetic beads (CS-Fe3O4) were then used as capturer and reporter of AFB1, respectively. Finally, the platform showed excellent sensitivity under optimized assay conditions, with a lower LOD (0.54 pg mL<sup>−</sup>1) and a more comprehensive linear range (0.001–10 ng mL−1). In addition, the high stability of SERS substrate activity was maintained for at least three months, with an RSD of ~5%, which has good selectivity for general coexistence interference. The excellent sensitivity and selectivity of micro-AFB1 detection are mainly due to the substantial Raman-enhancing effect of GNTs as the core of GDADNTs, which results from the bilayer of reporter molecules, aptamer specificity, and the super-paramagnetic CS-Fe3O4, respectively. The researchers also evaluated and confirmed that the established SERS aptamer sensor can be used to detect AFB1 in peanut oil samples.

In a subsequent study, on the basis of previous research, another simple and sensitive SERS aptamer sensor was developed for detecting AFB1 in peanut oil [170]. In this study, the researchers used an aminoterminal AFB1 aptamer (NH2-DNA1) as a SERS aptamer sensor, magnetic beads conjugated to a thiol-terminal-complementary AFB1 aptamer (SH-DNA2) (CS-Fe3O4) as enrichment nanoparticle probes, and AuNR@DNTB@Ag nanorods (ADANR) as reporter nanoprobes. 5,5 -Dithiobis (2-nitrobenzoic acid) (DNTB) is embedded in gold and silver core/shell nanorods as a Raman reporter molecule, which has a large Raman scattering cross section and no fluorescence interference. Furthermore, CS-Fe3O4 has good biocompatibility and superparamagnetism, which can quickly enrich signals. Therefore, NH2-DNA1-CS-Fe3O4 and SH-DNA2-ADANRs were prepared by a mixed reaction between aptamers and complementary aptamers. When present, AFB1 will compete with NH2-DNA1-CS-Fe3O4 to induce SH-DNA2-ADANRs to dissociate from CS-Fe3O4, further reducing SERS signals. According to the SERS aptamer sensor, the lower detection limit of AFB1 is 0.0036 ng mL−<sup>1</sup> and the correlation coefficient is as high as 0.986. The effective linear detection range is 0.01–100 ng mL<sup>−</sup>1, obtained with a correlation coefficient as high as 0.986. Finally, the specificity and accuracy of the SERS aptasensor were proved by detecting AFB1 in natural peanut oil.

Similar research strategies are reflected in other reports. Jiao et al. [244] developed a gold-silver core-shell nanoparticles (Au@Ag CSNPs) SERS sensor decorated with 5 aminotetramethylrhodamine (NH2-Rh). Based on the optimization of experimental conditions, the sensor can be combined with solid phase extracts of peanut oil, hazelnut, and other samples to achieve a quantitative analysis of AFB1 with detection range and LOD 0.1–5.0 ng·mL−<sup>1</sup> and 0.03 ng·mL<sup>−</sup>1, respectively.

Various AFB1 sensors are also identified in edible vegetable oil by electrochemical detection, which has some unique advantages, such as low cost, high sensitivity, and the possibility of micromachining. For example, Xiong et al. [245] revealed a highly innovative method based on dual-DNA-tweezer nanomachines to detect AFB1 in olive and peanut oils. Wu et al. [246] presented a method based on ferrocene and β-cyclodextrin (β-simple electrochemical aptamer sensor for host–guest recognition between CD) to detect AFB1 in peanut oil, with a low LOD (0.049 pg mL<sup>−</sup>1).

#### 4.5.3. Molecularly Imprinted Polymers (MIPs)

MIPs have been used as recognition elements to develop biosensors, and synthetic polymers have displayed precise target recognition [133,202,247]. These artificial materials can recognize specific targets in complex mixtures because of specific recognition sites for binding or catalysis and functional groups with shapes and geometries complementary to those of the template molecule. These polymers self-assemble with template molecules and active/functional monomers through the polymerization of crosslinking agents. Therefore, when the template molecule is removed, pores with multiple active sites appear in the polymer, which match the spatial configuration of the template molecule [248,249]. In recent years, traditional MIPs have been applied in many cross fields, such as chromatography, drug delivery, solid-phase extraction, controlled release, bioremediation, and sensors [200,250–256]. In AFB1 detection studies, MIP-based biosensors have shown many advantages, such as unique selectivity, sensitivity, user-friendliness, and cost-effectiveness [32,42,200,202]. For instance, Li et al. [173] exploited MIPs by preparing an electrochemiluminescence (ECL) platform for AFB1 detection with an ultra-low LOD, of 8.5 fg mL−1, and a wide linear range (10−<sup>5</sup> to 10 ng mL−1). While the MIP–ECL platform was used, the recovery rate of corn oil samples was close to that obtained by HPLC, indicating the reliability of the sensor and its potential in food safety evaluation. It is worth mentioning that, as of the publication of this review, this is the lowest LOD of AFB1 in edible oil.

However, MIP-based biosensors also have some disadvantages, such as generally poorer affinity and specificity than antibodies, slower binding kinetics than biological receptors, incomplete template elimination, and lower utilization of binding sites [133,200,202]. Therefore, there is increasing interest in developing improved MIPs [257–259]. The key to the success of the sensor of an MIP is whether the MIP is effectively attached to the transducer. Three commonly used immobilization methods are in situ polymerization, electropolymerization, and physical coating. Additionally, the number of applications of MIP sensors for detecting AFB1 in edible vegetable oil is limited [200,202,260].


**Table 2.** Techniques used for the detection of AFB1 in different types of edible oil.




#### **5. Conclusions and Perspectives**

Mycotoxin contamination, especially AFB1 contamination in edible oil, is usually unavoidable. A more sensitive and rapid sensor-based early warning tool for AFB1 detection would help to reduce risk. Various traditional, modern, and biosensing technologies have been used to detect toxins in contaminated food. Spectroscopic techniques, chromatographic techniques are general methods for the detection of AFB1 in edible oils. In recent years, based on the cross-integration of multiple disciplines, the innovation, progress and development of general methods have also been promoted. Although traditional chromatographic techniques can effectively detect mycotoxins, their performance in all

aspects cannot achieve satisfactory results. Combined use with other sensor equipment can effectively improve reliability, sensitivity and accuracy. However, due to the high cost of equipment, on-site inspection cannot be performed, and sample pretreatment is required, which limits the use of chromatography technology in the detection of AFB1 in edible oil. The development of spectroscopic techniques has become increasingly diverse and can effectively detect mycotoxins, especially AFB1 in edible oils. However, these methods are not suitable for on-site detection, because they still have many shortcomings, such as low sensitivity and reliability, and the need for professional personnel to operate.

Unlike conventional detection techniques, novel biosensors show high accuracy, sensitivity, and specificity; better cost controllability and portability; and reliability and simplicity in operation.

This review also discusses the development of important recognition elements in sensors. The recognition element of the sensor should have sensitivity and specificity sufficient enough to detect small amounts of target toxins, even in samples with complex matrix systems. The development and use of nanomaterials further improve the efficiency of biosensor conversion systems, but these require further improvements in their sensitivity, selectivity, and reproducibility. Of course, the stability and cost will also affect the selection of identification elements, which can improve the practicability.

Despite significant progress in biosensors for the detection of AFB1, there are some problems and challenges in the future. (1) The recognition elements of biosensors (such as metal nanoparticles, quantum dots, and graphene) improve the efficiency of sensing systems, but all these require further improvements in terms of sensitivity, selectivity, and reproducibility. (2) Future studies can perform AFB1 toxicity measurements and develop advanced nanomaterial-integrated biosensors to improve the overall detection of harmful substances, such as AFB1, in contaminated food samples. (3) When detecting AFB1 in contaminated food samples, researchers can focus on combining biosensing systems with microarray technology to fabricate more portable devices. (4) Reagent-free, clean-free, calibration-free, or nonbiological contamination biosensors for aflatoxin analysis require more effort and will reduce the possible future hazards.

**Author Contributions:** S.Y. Conceptualization, investigation, writing, review and editing of the manuscript; L.N. conceptualization and investigation of the manuscript; Y.L. review of the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was financially supported by Key R&D project of Jiangsu Province (BE2021306); and Development Project (BE2021306) and the Shandong Province Key Research and Development Program (No. 2021CXGC010808).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Abbreviations**




#### **References**


## *Article* **Analysis and Comparison of Aroma Compounds of Brown Sugar in Guangdong, Guangxi and Yunnan Using GC-O-MS**

**Erbao Chen <sup>1</sup> , Shuna Zhao 1,2,\*,†, Huanlu Song 1,\*,†, Yu Zhang <sup>1</sup> and Wanyao Lu 2,3**


**Abstract:** Guangdong, Guangxi and Yunnan are the three provinces in China that yield the most brown sugar, a brown-red colored solid or powdered sugar product made from sugar cane. In the present study, the differences between odor compounds of brown sugar from Guangdong, Guangxi, and Yunnan provinces in China were compared and analyzed by gas chromatography-olfactometrymass spectrometry (GC-O-MS). A total of 80 odor compounds, including 5 alcohols, 9 aldehydes, 8 phenols, 21 acids, 14 ketones, 5 esters, 12 pyrazines, and 6 other compounds, were detected. The fingerprint analysis of the brown sugar odor compounds showed 90% similarity, indicating a close relationship among the odor properties of brown sugar in each province. Moreover, the orthogonal partial least squares discriminant analysis (OPLS-DA) was performed to identify the compounds contributing to the volatile classification of the brown sugar from three provinces, which confirmed that OPLS-DA could be a potential tool to distinguish the brown sugar of three origins.

**Keywords:** non-centrifugal cane sugar (NCS); GC-O-MS; fingerprint; orthogonal partial least squares discriminant analysis (OPLS-DA)

#### **1. Introduction**

Brown sugar, a traditional sweetener with a distinctive flavor, is mainly made from sugarcane through extraction, clarification, and boiling [1]. It is also called non-centrifugal cane sugar (NCS), which does not separate molasses, so it retains the original flavor and nutrients of sugarcane. Brown sugar is rich in flavonoids and phenols that may act as antioxidants and, therefore, exert benefits on organisms [2–4]. Furthermore, it exerts immunomodulatory, cytoprotective, anti-carcinogenic, and anti-cancer properties [5].

A study on the physicochemical properties and storage stability of brown sugar revealed darker color, increased water content and water activity, but decreased glucose and fructose contents due to the Maillard reaction [6]. Similarly, a study on the odor components of brown sugar revealed that acetaldehyde, 2-methylbutyraldehyde, 3-methylbutyraldehyde, 2,6-dimethylpyrazine, nonanal, 2,6-diethylpyrazine, 2,3,5-trimethy lpyrazine, furfural, 2,3-dimethylpyrazine, decanal, and 2-acetylpyrrole were the primary components based on their relative concentration [7]. Juliana et al. [8] extracted a total of six odor compounds from brown sugar beverages through simultaneous steam distillationsolvent extraction using a mixture of diethyl ether-pentane (1:1, *w/w*) as the solvent. Of the six components, 2-methylpyrazine was the key aroma compound in this beverage. Our previous research has proved that heating of syrup was the primary production step affecting the brown sugar flavor because of the production of a large number of pyrazine compounds [9].

**Citation:** Chen, E.; Zhao, S.; Song, H.; Zhang, Y.; Lu, W. Analysis and Comparison of Aroma Compounds of Brown Sugar in Guangdong, Guangxi and Yunnan Using GC-O-MS. *Molecules* **2022**, *27*, 5878. https:// doi.org/10.3390/molecules27185878

Academic Editors: Weiying Lu and Yanping Chen

Received: 15 August 2022 Accepted: 1 September 2022 Published: 10 September 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Brown sugar has a green and a strong caramel aroma. Some aroma compounds are inherent in sugarcane, while others are produced by microbial metabolism and Maillard reaction. Sugarcane varieties, growing regions, processing methods, storage conditions and other factors will affect the flavor of brown sugar [10]. The composition and concentration of odor compounds and nutrients in sugarcane from different producing areas are different, which leads to great differences in the flavor composition of brown sugar. However, it is difficult to distinguish the origin of brown sugars only by sensory evaluation. As an intuitive and reproducible method, GC-MS analysis has been effectively applied in origin differentiation studies [11]. Li et al. [12] and Zhao et al. [13] used GC-MS to analyze the volatile odor compounds of ham and rice, respectively, and the results proved that GC-MS played an important role in food odor analysis and origin identification.

Previous studies on brown sugar mostly focused on the identification of key aroma, and there is no study on the flavor differences of brown sugar in different regions. Guangdong, Guangxi and Yunnan are the three major producing areas of brown sugar in China. To the best of our knowledge, the discrimination of brown sugar according to origin has not been reported previously. Therefore, the purpose of this study is to (1) identify the odor compounds of the 18 brown sugar samples from Guangdong, Guangxi, and Yunnan using LLE/GC-O-MS; (2) determine the key odor compounds in brown sugar by calculating OAV; (3) establish the fingerprints of brown sugar from three different origins and (4) find out the compounds that cause the difference using OPLS, so as to provide the basis for selecting brown sugar from different regions when producing foods with different flavor characteristics.

#### **2. Results and Discussion**

#### *2.1. Volatile Aroma Components Analysis*

A total of 80 odor compounds, including 5 alcohols, 9 aldehydes, 8 phenols, 21 acids, 14 ketones, 5 esters, 12 pyrazines, and 6 other compounds, were detected in 18 samples from three different regions (Table 1). The brown sugar samples from Guangdong, Guangxi and Yunnan contained 72, 60 and 75 odor compounds, respectively. There are four kinds of alcohols in all three regions, but the types of acid compounds are quite different, with Guangdong and Yunnan containing 20 and 19 acid compounds, respectively, while Guangxi contained only 12 acid compounds. The types of pyrazines, aldehydes, ketones and phenols in the three regions are very close. By comparing the odor compounds in the three regions, it was found that the unique odor compounds of the brown sugar samples in Guangdong were 2-acetyl-5-methylpyrazine, 2-methylbutanoic acid and 3-phenylpropionic acid; the unique odor compound in Guangxi was propylene glycol; and the unique odor compounds in Yunnan were 1,3-dimethoxy-2-hydroxybenzene, 3-hydroxyl-2-methyl-4*H*-pyran-4-one, 3-methyl-1,2-cyclopentanedione, 4-methylpentanoic acid and γ-butyrolactone. These unique odor compounds are expected to be important indicators to distinguish the origin of brown sugar samples.

The average contents of odor compounds in brown sugar samples from the three regions are shown in Figure 1. It can be seen that the highest contents of acid compounds were found in all three regions with 25,595.06, 21,632.44 and 25,187.12 ng/g, followed by phenolic compounds with average contents of 111,69.29, 12,115.37 and 11,744.16 ng/g. In contrast, alcohols, esters, pyrazines and ketones had lower contents.


**Table 1.** Volatile aroma components of brown sugars from different producing areas.

















**Figure 1.** The average content of different kinds of compounds in three regions.

#### *2.2. Analysis of Key Aroma Compounds in Brown Sugar Samples*

A total of 46 aroma-active compounds were identified in 18 brown sugar samples by olfactometry, including 4 alcohols, 4 aldehydes, 3 phenols, 15 acids, 11 ketones, 7 pyrazines, and 2 other compounds. According to the odor properties of the aroma active compounds, these compounds can be classified into nine types: sweet/caramel, fruity, green/grassy, sour, sweaty/cheese, nutty, roasted, fatty and potato, which indicated that the aroma profile of brown sugar was the result of the synergistic effect of various odors.

In fact, it is the OAV of the aroma compound, and not its amount, that determines the contribution of the aroma compound. Aroma activity is generally defined as compounds with OAVs greater than 1 [14]. Therefore, the calculation of OAV was carried out for aroma compounds that can be sniffed (Table 2). Among the 18 brown sugar samples, 26 compounds with OAV >1 were considered as the key aroma active compounds of the brown sugar samples in this study and contributed to the overall flavor.

Alcohols: Among the four alcohols that can be sniffed, only furfuryl alcohol had OAV >1 and was only found in Guangxi and Yunnan. The content of furfuryl alcohol in Guangxi and Yunnan was 971.50 and 392.70 ng/g, respectively, and it contributed sweet, toast and caramel aroma to brown sugar. Sugar and amino acids react readily at elevated temperatures to form this compound [15]. The furfuryl alcohol contained in soy sauce has been considered to be one of the main components responsible for its odor, exhibiting a caramel scent, which contributes to the overall flavor of the sample [16].



**Table 2.** OAV of key odor compounds in brown sugar.

a Volatile

compounds

 that can be smelled at sniffer port. b Odor thresholds

 were referenced

 in a book, named: *odor thresholds compilations*

 *of odor threshold values in air, water and other media*.

Aldehydes: Among the aldehydes, there are four aldehydes with OAV >1, namely hexanal, (*E*)-2-nonenal, 3,5-dimethoxy-4-hydroxybenzaldehyde and benzaldehyde. (*E*)-2 nonenal and hexanal are probably oxidation products of polyunsaturated fatty acids [17], with high OAV due to their higher concentration and lower odor threshold, and are key aroma compounds among aldehydes, contributing to the green odor of brown sugar. The average content of benzaldehyde in Guangdong was higher than that in Guangxi and Yunnan, and it may be the degradation product of phenylalanine [14], contributing nutty and caramel aromas to the brown sugar. 3,5-dimethoxy-4-hydroxybenzaldehyde showed close OAV in Guangdong and Yunnan, and was higher than that in Guangxi, contributing sweet and nutty aroma to brown sugars. According to Chen, Song, Li, Chen, Wang, Che, Zhang and Zhao [9], 3,5-dimethoxy-4-hydroxybenzaldehyde is formed during brown sugar production, and the difference in content might be related to the raw materials and processing technology.

Ketones: Four ketones with OAV >1 were found in brown sugar samples, including 3-methyl-1,2-cyclopentanedione, 2-hydroxy-3-methyl-2-cyclopenten-1-one, 2,5-dimethyl-4 hydroxy-3(2*H*)-furanone, and 4-hydroxy-5-methyl-3(2*H*)-furanone. 2,5-Dimethyl-4-hydroxy-3(2*H*)-furanone has the highest OAV and contributes a strong caramel flavor to brown sugar, which is most likely formed by the Maillard reaction through deoxy sugars and is most abundant in strawberries [18,19]. 2-Hydroxy-3-methyl-2-cyclopenten-1-one has a strong caramel aroma and is one of the key odor compounds that contribute to the caramel odor in black tea, soy sauce and molasses [20–22]. 3-Methyl-1,2-cyclopentanedione was detected only in Yunnan brown sugar with OAV=14, which contributed sweet and bready aroma to Yunnan brown sugar. 2,5-Dimethyl-4-hydroxy-3(2*H*)-furanone was detected in all the three regions' samples, but the OAV was greater than 1 only in Guangxi brown sugar, which was caused by its high concentration in Guangxi brown sugar.

Pyrazines: Many products possess a distinctive aroma resulting from pyrazines, which are special Maillard reaction compounds [23,24]. Pyrazine is formed by condensing two α-aminocarbonyl compounds and forming a dihydropyrazine, which oxidizes spontaneously to form the pyrazine [23,25,26]. Among the twelve pyrazines detected in the eighteen samples, there are five kinds of pyrazines with OAV greater than 1, namely 2,3,5 trimethylpyrazine, 2,5-dimethylpyrazine, 2,6-dimethyyl-3-ethylpyrazine, 2,6-dimethyl-3-ethylpyrazine and 2-acetyl-6-methylpyrazine. 2,6-Dimethyl-3-ethylpyrazine exhibited the highest OVA due to its low threshold (OT=0.04 ng/g), contributing a strong roasted potato flavor to brown sugar. 2,5-Dimethylpyrazine and 2,6-dimethylpyrazine were previously reported to be key odor compounds in coffee, exhibiting strong roasted and nutty aroma [27].

Acids: A total of 21 kinds of acid compounds were detected in 18 brown sugars, among which the OAV of 11 kinds of acid compounds was greater than 1. Acetic acid, one of the most abundant compounds in brown sugar, had the highest OAV and contributed sour aroma to the samples. 2-Methylbutanoic acid and 3-methylbutanoic acid exhibited a sour aroma and had been reported to be the key aroma components in Japanese sweet rice wine, which played an important role in the overall flavor of sweet rice wine [28]. Benzoic acid, however, has an unpleasant urine-like odor, which may be caused by phenylalanine under the action of phenylalanine ammonia-lyase in plants [29].

#### *2.3. Fingerprint Analysis of Sugar Products from Three Different Regions*

A food fingerprint can be defined as molecular markers that indicate a characteristic state or condition of food, thus enabling more accurate product identification [30]. Each sample is regarded as a multidimensional space vector. If two samples are more similar, their space will be closer, and the angle between the two samples' space vectors will be smaller, which leads the cosine of the angle between the two vectors to move closer to 1. Therefore, the similarity of samples can be expressed by the cosine of the included angle. On the contrary, if the difference between the two samples is greater, the cosine of the

included angle becomes smaller. In this study, the samples were determined by GC-O-MS, and the odor-active compounds were selected for fingerprint and similarity evaluation.

It is worth mentioning that the similarity of samples becomes higher when the similarity or the cosine of the angle is above 90%. As depicted in Table 3 and Figure 2, of the six samples in Guangdong, except for Guangdong3, the similarity and cosine of the included angle of the other five samples were above 90%. This indicated that the odor properties of Guangdong3 were quite different than the other five samples, which might have happened due to different processing technology.

**Figure 2.** Fingerprint of brown sugar from Guangdong, Guangxi and Yunnan.

The cosine of the included angle of six samples in Guangxi was above 90%, and the similarity of Guangxi3 was just less than 90% (89.80%). This result indicated that the odor properties of these six samples in Guangxi were similar, without much difference

Of the six samples in Yunnan, only Yunnan2 had similarity and cosine of included angle lower than 90%, while the other five samples had similarity and cosine of included angle higher than 90%. This result indicated that the odor attributes of the other five samples were similar, but Yunnan2 had significant differences with them.


**Table 3.**Fingerprint results of brown sugar from each producing

 area.



#### *2.4. Verification of Fingerprint*

In order to verify whether the fingerprint method is suitable for the analysis of brown sugar, the verification was carried out. Fingerprint verification includes three parts: stability experiment, precision experiment, and repeatability experiment. Following the sample preparation described in Section 2.4, a brown sugar sample was selected and analyzed by GC-MS after 0, 2, 4, 8, 16, and 24 h. Furthermore, the relative standard deviations (RSD) of the relative retention times (RT) and relative peak areas of the odor-active compounds were calculated. The results showed that the RSD of the relative RT of the odor-active compounds was less than 0.3%, and the RSD of the relative peak areas was less than 5%, indicating that the samples were stable within 24 h and met the requirements of the fingerprint method.

A brown sugar sample was extracted and concentrated with the organic solvent, and then the concentration was injected six times consecutively to calculate the RSD of relative RT and relative peak area of the odor-active compounds. These results showed that the RSD of the relative RT of the odor active compounds was less than 0.5%, and the RSD of the relative peak area was less than 6%, indicating that the precision of the instrument was good and met the requirements of the fingerprint method.

Five brown sugar samples were extracted and analyzed for their odor compounds, followed by the RSD of relative RT and relative peak area of the odor active compounds analysis. The results showed that the RSD of relative RT was less than 0.3%, and the RSD of the relative peak area was less than 7%, indicating that they had good repeatability and met the requirements of the fingerprint method.

#### *2.5. Orthogonal Partial Least Squares Discriminant Analysis (OPLS-DA)*

The fingerprinting analysis of samples from the three origins of Guangdong, Guangxi, and Yunnan revealed that the majority of samples within each province had similar odor types. In addition, a supervised OPLS-DA multivariate statistical analysis method was used to establish a statistical model in order to distinguish odor compounds between Guangdong and Guangxi, Guangdong and Yunnan, and Guangxi and Yunnan.

By conducting OPLS-DA analysis on the brown sugar, a variable importance of projection diagram (VIP) of the model was obtained. A VIP is a vector that summarizes the contribution of a variable to the explanation of the model. Variables with a VIP >1 are generally considered to contribute to the explanation of the model [31,32]. The samples were assessed as independent variables, and the OPLS-DA model was fitted automatically.

The OPLS-DA and VIP results (Figure 3) indicate that the brown sugars from Guangdong and Guangxi were well separated. The brown sugar from Guangdong and Guangxi showed the greatest degree of separation and low intra-group differences, facilitating an accurate exploration of the differences in composition. VIP diagram elucidated that 4-hydroxybenzaldehyde, 3,5-dimethoxy-4-hydroxybenzaldehyde, n-hexadecanoic acid, butanoic acid, acetic acid, 2-methoxy-4-acetylphenol, 2-acetylpyrrole, pentadecanoic acid, furfuryl alcohol, 4-hydroxyacetophenone, etc., were the main contributors to the distinction between Guangdong and Guangxi samples. These compounds were basically aldehydes, acids, ketones, and phenols. Among these, 3,5-dimethoxy-4-hydroxybenzaldehyde and 4-hydroxybenzaldehyde played an important role in classifying Guangdong and Guangxi. 4-Hydroxybenzaldehyde and 3,5-dimethoxy-4-hydroxybenzaldehyde presented a pleasant nutty and creamy odor. Previously, 4-hydroxybenzaldehyde and 3,5-dimethoxy-4-hydroxybenzaldehyde were identified as the major volatile constituents in brown sugars [33]. Acetic acid is also one of the key compounds that can distinguish brown sugar from two provinces. Acetate is a well-known product of the thermal degradation of saccharides, and it is primarily formed during the early stage of the Maillard reaction, under neutral and alkaline conditions. Acetic acid is formed exclusively by hydrolytic cleavage of β-dicarbonyl in hexose-based systems [34].

**Figure 3.** OPLS-DA analysis and VIP diagram of brown sugar in Guangdong and Guangxi.

As shown in Figure 4, OPLS-DA analysis and VIP results indicate that the brown sugars from Guangdong and Yunnan are distinguishable. The principal compounds contributing to this distinction include n-hexadecanoic acid, acetic acid, dibutylphthalate, 2-acetylpyrrole, 2,5-dimethylpyrazine, and 2-methylpyrazine. Of the compounds with VIP greater than 1, pyrazine compounds appeared, which indicated that pyrazine compounds played a significant role in distinguishing brown sugar between Guangdong and Yunnan. The average content of pyrazines in Guangdong and Yunnan was 2897.28 ng/g and 1441.20 ng/g, respectively, and the pyrazine contents in Guangdong samples were higher than in Yunnan. These compounds could impart a popcorn, nutty, and roasted aroma to brown sugar.

**Figure 4.** OPLS-DA analysis and VIP diagram of brown sugar in Guangdong and Yunnan.

Based on the VIP diagram and OPLS-DA analysis of brown sugar between Guangxi and Yunnan (Figure 5), they were well separated. A number of compounds contributed to the differentiation between the two provinces, including 4-hydroxybenzaldehyde, 3,5dimethoxy-4-hydroxybenzaldehyde, n-hexadecanoic acid, acetic acid, butanoic acid, and 4 hydroxyacetophenone. Of these volatile compounds, the contribution of 4-hydroxybenzalde hyde was the greatest. The average content of 4-hydroxybenzaldehyde in Guangxi was 2728.55 ng/g, while the samples from Guangxi had no odor compounds. The average contents of 3,5-dimethoxy-4-hydroxybenzaldehyde in Guangxi and Yunnan were 926.34 ng/g and 2967.95 ng/g and the contents in Yunnan were significantly higher than in Guangxi. Perhaps these compounds play an important role in distinguishing the sugars from Guangxi and Yunnan.

**Figure 5.** OPLS-DA analysis and VIP diagram of brown sugar in Guangxi and Yunnan.

#### **3. Materials and Methods**

#### *3.1. Materials*

Eighteen brown sugar samples from Guangdong, Guangxi and Yunnan were provided by COFCO. These samples were stored in a refrigerator at −80 ◦C before analysis.

#### *3.2. Standards and Reagents*

Ether (purity > 99%), dichloromethane (purity > 99%), anhydrous sodium sulfate, 2-methyl-3-heptanone (purity > 99%) and n-alkane (C7-C30) were purchased from Sigma-Aldrich (St. Louis, MO, USA), and carrier gas (helium) was purchased from Beijing AP Baif Gases Industry Co., Ltd. (Beijing, China).

#### *3.3. Extraction of Odor Compounds from Sugars*

The odor compounds in brown sugar were extracted by a liquid–liquid extraction (LLE) method according to Chen et al. [33]. In brief, 50.00 g of brown sugar was placed in a triangular flask, 50 mL of distilled water was added to dissolve the brown sugar, then, 50 mL of ether, 50 mL of dichloromethane and 5 μL of internal standard 2-methyl-3-heptanone (81.6 mg/mL) were added, and the mixture was magnetically stirred at 1000 rpm for 10 min. After centrifugation (Hitachi, Japan) for 30 min at 10,000 rpm, the extract containing the volatile aroma compounds was separated by a funnel. Subsequently, 150.0 g anhydrous sodium sulfate was added to the extract and put into a refrigerator at 4 ◦C to remove water for 12 h, and filtered with a filter paper. A gentle nitrogen stream was used to concentrate the volume into 100 μL, and the odor compounds were extracted and stored at −80 ◦C for further analysis.

#### *3.4. GC-O-MS*

Three well-trained panelists conducted a GC-O analysis of the concentrated distillate. The panelists were recruited from Beijing Technology and Business University's Molecular Sensory Laboratory. To identify and describe the aroma characteristics of the reference compounds, they smelled several concentrations of reference compounds in model solutions 2 h per day before analysis. The training lasted for one month. For the GC-O analysis, wet gas was delivered to the nose using a blank capillary column to improve the sensitivity of the panelists. The aroma perceptions, intensity, and RT were recorded by the panelists. If two or more panelists detected the aroma, an aroma-active compound was identified [35].

To determine the volatile aroma profile of sugars, an Agilent 7890A gas chromatograph (GC) coupled with an Agilent 5977B mass spectrometer (MS) and a sniffing port (Gerstel, Germany) was used. The aroma extract (1 μL) was injected into a DB-Wax column (60 m × 0.25 mm i.d., film thickness 0.25 μm, Agilent J&W) through splitless mode, and the flow rate of the helium carrier gas was maintained at 1.7 mL/min. The oven temperature was initially programmed at 40 ◦C, further raised to 100 ◦C at a rate of 4 ◦C/min, following a gradual increase up to 200 ◦C at a rate of 3 ◦C/min for 5 min, and after achieving an ultimate temperature of 230 ◦C at a rate of 3 ◦C/min, it was maintained for 10 min. The interface and ion source were set at 250 ◦C and 230 ◦C, respectively, while the electron-impact ionization was set at 70 eV, the acquisition range (*m*/z) at 35–350 amu, and the scan rate at 1.77 scans/s. The transmission line temperature of the olfactory detection port (ODP) was maintained at 235 ◦C.

#### *3.5. Qualitative Analysis*

The ionization of a molecule in a vacuum produces a characteristic group of ions of different masses. The plot of relative abundance versus mass of these ions constitutes a mass spectrum. The spectrum can be used to identify the molecule. The unknowns were identified by comparing the fragments with the National Institute of Standards and Technology (NIST) MS Spectral Library (Version 2020), by comparing the odor percepts with the database (http://www.thegoodscentscompany.com) and by calculating the linear retention indices (LRIs) using a homologous series of n-alkanes (C7-C30). The use of multiple methods can increase the accuracy of qualitative results. Using the internal standard area, the resulting peaks were calibrated, and the aroma compound contents were expressed as nanograms per gram of sample [10].

#### *3.6. Odor Activity Value (OAV)*

In order to evaluate the contribution of each odorant to the overall aroma of brown sugar, the OAV (ratio of concentration to its odor threshold) was calculated [36]. These threshold values were derived from the literature in water [37].

#### *3.7. Statistical Analysis*

All experiments in this study were conducted in triplicates, and the data were expressed as mean ± standard deviation. The bar graph was drawn by OriginPro 2022 (OriginLab Corp., Northampton, MA, USA), the OPLS-DA analysis was conducted by SIMCA 14.1 (MKS Instruments, Andover, MA, USA), and the tables were organized by Microsoft Excel 2021 (Microsoft Corp., Redmond, WA, USA).

#### **4. Conclusions**

In summary, a total of 80 odor compounds, including 5 alcohols, 9 aldehydes, 8 phenols, 21 acids, 14 ketones, 5 esters, 12 pyrazines, and 6 other compounds, were detected in 18 brown sugar samples from three different provinces. The fingerprint analysis showed 90% similarity, indicating a close relationship among the odor components of brown sugars from each province without much difference. Further, the stability, accuracy, and repeatability of the fingerprint method were verified, and speculated that the method could meet the requirements of the fingerprint. In the future, fingerprint might have wider applica-

tions due to its characteristic of distinguishing geographical origin and food adulteration. Additionally, the OPLS-DA was employed to identify the tracing of brown sugar and to identify the compounds contributing to brown sugars' volatile classification. The results demonstrated that 4-hydroxybenzaldehyde, 3,5-dimethoxy-4-hydroxybenzaldehyde, nhexadecanoic acid, and acetic acid were the essential components in distinguishing the sugars from Guangdong, Guangxi, and Yunnan, validating the efficiency of OPLS-DA.

**Author Contributions:** Methodology, H.S.; Software, Y.Z. and W.L.; Writing—original draft preparation, E.C.; Writing—review and editing, H.S. and S.Z.; Supervision, H.S. and Y.Z. All authors have read and agreed to the published version of the manuscript.

**Funding:** COFCO Nutrition and Health Research Institute Co., Ltd.: No number.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** The present work was supported by China Oil and Food Import and Export Corporation (COFCO). No project number.

**Conflicts of Interest:** The authors declare no conflict of interest.

**Sample Availability:** Not available.

#### **Abbreviations**


#### **References**


## *Review* **Raman Method in Identification of Species and Varieties, Assessment of Plant Maturity and Crop Quality—A Review**

**Aneta Saletnik, Bogdan Saletnik \* and Czesław Puchalski**

Department of Bioenergetics, Food Analysis and Microbiology, Institute of Food Technology and Nutrition, College of Natural Science, Rzeszow University, Cwikli ´ ´ nskiej 2D, 35-601 Rzeszow, Poland; asaletnik@ur.edu.pl (A.S.); cpuchalski@ur.edu.pl (C.P.)

**\*** Correspondence: bsaletnik@ur.edu.pl

**Abstract:** The present review covers reports discussing potential applications of the specificity of Raman techniques in the advancement of digital farming, in line with an assumption of yield maximisation with minimum environmental impact of agriculture. Raman is an optical spectroscopy method which can be used to perform immediate, label-free detection and quantification of key compounds without destroying the sample. The authors particularly focused on the reports discussing the use of Raman spectroscopy in monitoring the physiological status of plants, assessing crop maturity and quality, plant pathology and ripening, and identifying plant species and their varieties. In recent years, research reports have presented evidence confirming the effectiveness of Raman spectroscopy in identifying biotic and abiotic stresses in plants as well as in phenotyping and digital selection of plants in farming. Raman techniques used in precision agriculture can significantly improve capacities for farming management, crop quality assessment, as well as biological and chemical contaminant detection, thereby contributing to food safety as well as the productivity and profitability of agriculture. This review aims to increase the awareness of the growing potential of Raman spectroscopy in agriculture among plant breeders, geneticists, farmers and engineers.

**Keywords:** Raman spectroscopy; digital farming; harvest maturity assessment; fruit and seeds quality diagnostics; non-invasive phenotyping

#### **1. Introduction**

Population growth and consequently the increasing demand for food as well as the decreasing availability of fertile land have led to reduced agricultural expansion and higher costs of agriculture and food. All these factors, combined with the long history of conventional agriculture, have forced food producers to seek new solutions. Owing to advancements in technology, these problems can be solved by introducing innovative methods in agriculture. The transformation of farming processes with the use of smart technologies, referred to as digital agriculture, aims to develop innovative technological methods to be used to maximise yield and minimise the environmental impact of agriculture [1–4]. Digital agriculture requires advanced methodologies for cultivation and selection of plants [5,6], detection and identification of biotic and abiotic stresses in plants, and for the acquisition of information about the health of plants and growth stages directly on the plantation. These methods are highly needed by crop producers [7].

In the related literature, there is more and more evidence showing the effectiveness of Raman spectroscopy in identifying the physiological condition of plants, crop quality, plant species or varieties, in assessing the maturity of plants and in pre-symptomatic diagnostics, or in detecting abiotic and biotic stresses in plants [1,8–17]. Raman spectroscopy also enables phenotyping and digital selection of plants in breeding. Owing to the immediate access to information about plant health, it is possible to detect and identify bacterial infections, secondary diseases, insect infestations, fungal infections, and other pathogens

**Citation:** Saletnik, A.; Saletnik, B.; Puchalski, C. Raman Method in Identification of Species and Varieties, Assessment of Plant Maturity and Crop Quality—A Review. *Molecules* **2022**, *27*, 4454. https://doi.org/ 10.3390/molecules27144454

Academic Editors: Weiying Lu and Yanping Chen

Received: 21 June 2022 Accepted: 11 July 2022 Published: 12 July 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

and the diseases they transmit in greenhouses and fields [8,18,19]. The information acquired this way is used to perform precise and site-specific chemical treatments which may prevent the spread of biotic stresses and save up to 30% of the crop. Rapid detection of physiological drought or nutrient deficiencies allows for supply of the nutrients promptly and accurately. By applying fertilisers to a strictly defined area, it is possible to reduce contamination of the crops and the soil [20,21].

Moreover, a tool effectively assessing the ripeness of the fruit and the quality of the crop helps the farmer to make sure that the crop is harvested at the right time. Researchers have demonstrated in a number of studies that Raman spectroscopy is a perfect method for enabling monitoring of field and greenhouse crops because of its high selectivity and specificity [1]. Recent research findings show that Raman spectroscopy (RS) can effectively be used in diagnostics of biotic and abiotic stresses [22–25]. RS is a label-free laser technique that does not require chemicals to perform analysis of plant material, so the farmer does not incur any costs by purchasing reagents. Furthermore, it takes only one second to perform analysis of a plant in order to detect pathogens or to identify the origin of abiotic stresses [26].

Agriculture and crop production are relatively new areas in research involving Raman spectroscopy [8]. This review focuses particularly on RS application for evaluating the physiological status of plants as well as the stage in the ripening process, for assessing quality of crops (fruit and seeds), and for identifying plant species, varieties, and their origin. This review aims to increase the awareness of the potential of Raman spectroscopy in agriculture among plant breeders, geneticists, farmers, and engineers.

#### **2. Principle, Instrumentation**

Raman spectroscopy is used to examine the structure, dynamics of changes, and function of biomolecules. It is a vibrational technique providing insight into the structure of tissues and their components at the molecular level [27–30].

In principle, Raman spectroscopy measures the frequency shift in the inelastic scattering of light when a photon of incident light strikes a particle and produces a scattered photon [31–36]. In the light scattered by the test medium there is, for the most part, a component of the same frequency as in the incident light (Rayleigh scattering, elastic scattering) [36], while in a minority of cases there are variable frequency components associated with the change in photon energy (inelastic scattering, Raman scattering). The outgoing scattered light can be a photon with a frequency lower than the incident photon and in such cases we call it Stokes Raman scattering, or it is of a frequency that is higher, and then it is known as anti-Stokes Raman scattering [29]. The Stokes band forms when the molecule, after interacting with the excitation radiation, shifts to a higher oscillatory level and the scattered photon has energy lower than the energy difference between the levels of vibrational energy. On the other hand, the anti-Stokes band may appear if the molecule was at the excited oscillatory level before the impact of the excitation radiation – that way there is a high probability that it returns to the basic oscillatory level. The scattered photon has an energy greater than the difference in energy of the oscillating energy levels [36]. Since the Stokes bands are of higher intensity than the anti-Stokes bands, in Raman spectroscopy, the measurement most often concerns only the Stokes part of the Raman spectrum, presented in the range from 4000 to 0 Δ cm−<sup>1</sup> (the so-called Raman shift) [29].

The phenomenon of Raman scattering of light by particles was predicted by Smekal in 1923 and observed experimentally for the first time by Sir Chandrasekhar Venkat Raman and his student Kariamanickam Srinivas Krishnan in 1928 [37,38].

In honour of the discoverer, the phenomenon of inelastic scattering of light is called Raman scattering [39]. An advantage presented by Raman spectroscopy lies in the fact that when it is used to examine biological materials, the spectra containing large amounts of information can be acquired from intact tissues [40,41], without interfering in their structure. This way it is possible to perform detailed chemical analysis of the biological material despite its high complexity. Organic compounds and functional groups have

characteristic spectral patterns, the so-called "fingerprints", enabling their identification, while the intensity of the bands can be used to calculate the concentration in the sample analysed [42]. The Raman spectrum can be used as a fingerprinting tool for various compounds [43]. Table 1 shows the assignment of bands in the Raman spectra of cell wall polysaccharides based on the literature, and Figure 1 shows the Raman spectra of the pure cell wall components pectin (A), xyloglucan (B), cellulose (C), and the Raman spectrum of the tomato cell wall (D) [44].

**Table 1.** Assignment of bands in the Raman spectra of cell wall polysaccharides based on the literature [8]. Copyright Front. Plant Sci. 2021.


The obtained analyte spectrum can be treated as a qualitative analysis of unknown samples or mixtures of components [45]. Importantly, Raman spectroscopy is sensitive even to small structural changes, which is useful in comparative studies [41]. Raman scattering in tissues provides a wealth of information about the vibrational structure of their constituent proteins, GAGs, lipids, and DNA. Another important advantage of Raman spectroscopy is the low intensity of water bands, which in other spectroscopic methods makes the analysis of biological materials difficult [39]. Raman spectroscopy extracts spatial information from complex biological samples, and as a result, it is an extremely accurate tool for examining a variety of plant materials [46], such as pollen [47,48], fruits [49,50], roots [51], and wood of various origins [52–57]. Its mechanism makes it possible to perform analysis of both the molecular composition and molecular structure of cell walls [58–60].

**Figure 1.** Raman spectra of the pure cell wall components (laser with green light at λ = 532 nm, with a power of 10 mW, the spectra were recorded within the range of 3500–150 cm−1): pectin (**A**), xyloglucan (**B**), cellulose (**C**), and the Raman spectrum of the tomato cell wall (**D**). Copyright Plant Methods 2014.

A Raman spectrometer consists of a laser source which generates a stream of photons. The light is then directed through a beam splitter and focused through a lens onto the test sample. This leads to scattering of light, which is then usually collected using the same optics and directed into the spectrometer. Elastically scattered photons are cut off by long-pass filters before entering the spectrometer. After the inelastically scattered photons (Raman photons) are scattered on the spectrometer gratings according to their energies, they are captured by the CCD [1,61]. An exemplary diagram of the Raman spectrometer operation is shown in Figure 2 (other settings are possible).

Raman spectroscopy was for a long time used exclusively as a laboratory technique, but there are now several commercially available handheld spectrometers. These instruments typically have laser excitation in the green (λ = 532 nm), red (λ = 785 or 830 nm), or infrared (λ = 1064 nm) parts of the electromagnetic spectrum [22–24,62–64]. The beam diameter or laser spot size on portable Raman spectrometers ranges from a few dozen microns to a few millimetres. Excitation wavelength is one of the most important parameters to pay attention to in spectroscopic examinations of plants. Results acquired by researchers show that by using radiation in the blue and green parts of the electromagnetic spectrum it is possible, in particular, to visualise carotenoid signals. This results from the strong absorption of carotenoids in this part of the electromagnetic spectrum [65]. Lasers with wavelengths above 561 nm and below 700 nm are not suitable for structural analysis of living plants due to the extremely strong fluorescence of chlorophyll. Chlorophyll fluorescence decreases exponentially at wavelengths above 700 nm. Therefore, laser excitations of 785–830 nm provide sufficient signal spectra for plant leaf noise [66]. Handheld Raman spectrometers provide accurate access to physiological responses in plants under field conditions [66], both for plant producers and researchers [67].

**Figure 2.** Schematic representation of a Raman spectrometer. Copyright Plant Methods 2021.

The Raman method is an optical spectroscopic technique which can effectively be used to detect and quantify key compounds without destroying the sample. As a result, it is a powerful tool for monitoring plant physiological status and assessing crop quality, pathology, and plant maturation [67–72].

Raman spectroscopy is not yet used widely in agriculture and in response to food analysis but has been used recently [73]. Ramaskie techniques do not require a laborintensive access service, and they can be tested directly in the environment with pits [74]. The Raman method is fast and non-invasive; there is a text, a spectrometric manual, that prevents spectral registration and fruit testing on the plant or during sorting. An additional advantage of Raman spectroscopy is a small queue for water or its steps, thanks to which you can analyse the attendance and dried fruit. The Raman methodology also enables food safety to be controlled by glass or polymer [75,76]. Raman techniques used in precision agriculture can significantly improve capacities for farming management, crop quality assessment, as well as biological and chemical contaminant detection, thereby contributing to food safety as well as the productivity and profitability of agriculture [77].

#### **3. Raman Spectroscopy in Assessment of Changes during Growth as Well as Harvest Maturity of Plants**

Analysis of the spectra collected from various crops at specific vegetation stages can be successfully used to determine the optimal harvest time [78,79]. As an example, Piot and colleagues applied confocal Raman microscopy to examine wheat grain and follow the evolution of protein content and structure during the growth of wheat grains of different cultivars. The cultivars differed in the level of grain hardness and the aptitude to separation of peripheral layers during milling process [80]. The study showed that RS is a suitable tool for acquiring information on grain structure and composition and, most importantly, can detect molecules present at low concentrations, such as α-helical protein. Researchers have noted that concentration of this protein increases as the grains harden during the maturation process [1].

Chyli ´nska and her team applied confocal spectroscopy to assess the ripening stage of tomatoes. Fruit was harvested at the mature green and red ripe stages. The samples were

examined using a Raman spectrometer equipped with a green light laser (λ = 532 nm). The researchers investigated the effect of biochemical parameters (such as cell wall polysaccharide content, phenolic compounds, ascorbic acid, and pectinolytic enzyme activity) on cell wall microstructure and changes in polysaccharide distribution during the process of physiological development of tomato fruit (Solanum lycopersicum cv Cerise). The study showed that confocal Raman spectroscopy makes it possible to visualise changes in the spatial distribution of polysaccharides in the plant cell wall (including the central lamella region). In mature green tomato, pectin concentrations were visible particularly in cell corners, whereas ripe red tomato were found with a homogeneous distribution of pectin in the cell wall (Figure 3). The study by Chyli ´nska et al. demonstrates that the Raman technique can visualise the changes occurring in the cell wall (mainly degradation of pectin polysaccharides) during tomato ripening [81].

**Figure 3.** Raman images of cross sections of tomato cell wall from mesocarp at laser with green light (λ = 532 nm); a power of 25 mW and integration time at 0.1 s was chosen. Raman images of all primary cell wall polysaccharides at 2936 cm<sup>−</sup>1, γ(CH) (**E**–**H**). Raman images of pectin at 854 cm<sup>−</sup>1, the (COC) skeletal (**I**–**L**). Raman images of cellulose at 1096 cm−<sup>1</sup> and 1115 cm−<sup>1</sup> (glycosidic bond) (**M**–**P**). Copyright Plant Physiology and Biochemistry 2017.

Lopez-Sanchez et al. [82] used the potential of Raman spectroscopy to assess changes taking place in the process of olive fruit growth and ripening. For this purpose, the researchers measured the spectra of different parts of olive fruit (skin, flesh, and stone) at different stages of development. They observed an increase in carotenoids and phenolic compounds during olive growth and a decrease during the ripening phase. The evolution of the different spectral bands was linked to the content of the olive constituents, such as triglycerides, water, carotenoids, and phenolic compounds. The study demonstrated the ability of RS to track oil accumulation in olive fruit. Increased intensity of the peaks at 1440 cm−<sup>1</sup> correlates well with oil content in the fruit measured using the standard Soxhlet extraction method. The researchers showed that increase in carotenoid and phenolic contents during olive growth and decrease in these values during the ripening stage can effectively be monitored using vibration bands at 1525 and 1605 cm<sup>−</sup>1.

Szyma ´nska-Chargot et al. [83] applied confocal Raman microscopy to assess changes in the distribution of polysaccharides in the cell wall of apple flesh in ripening fruit and during storage after harvest. The apples were collected on three dates: at 1 month and

at 2 weeks before the optimum harvest date and on the optimum harvest date. Apples collected on the optimum harvest date were kept in storage for 3 months. The researchers acquired Raman maps for each harvest date and after 1, 2, and 3 months of storage. Raman images of apple cell walls show significant changes in the quantity and distribution of the main cell wall polysaccharides. Analysis of the Raman maps showed degradation of pectins distributed in the middle lamella and primary cell wall. These findings were consistent with the results of chemical assays. During the process of apple ripening and ageing, the analyses based on Raman spectroscopy showed changes in the distribution of pectins, which in young fruit were dispersed along the cell walls, whereas in ripe fruit and those kept in storage they were concentrated in the cell wall corners. Analysis of apples after 3 months in storage showed a significant decrease in pectin content. The findings reported by this research team show that Raman imaging can be a very useful tool for early identification of changes in plant tissue composition during development.

Qin et al. [84] report that it is possible to use Raman spectroscopy to visualise the content of lycopene, the main carotenoid in tomatoes. According to those researchers, imaging of changes in lycopene content during fruit ripening is a good method for monitoring the stage of tomato maturity. During the study, the research team developed a laboratory-based point-source Raman chemical imaging system to detect and visualise the internal distribution of lycopene during the process of fruit ripening and post-harvest. The researchers analysed lycopene content in tomato fruit samples representing different stages of ripeness (i.e., green, breaker, turning, pink, light red, and red). Qin et al. applied the spatially offset Raman spectroscopy (SORS) technique for subsurface detection of a Teflon slab placed under samples of the outer pericarp from green and pink tomatoes. The findings showed that the Teflon spectrum acquired this way can be extracted from SORS measurements of tomato pericarp placed above Teflon. These results suggest a potential for the development of the SORS method in imaging lycopene concentration as a marker of tomato fruit ripeness.

Martin et al. [85] developed a tomato ripening model based on vibrational bands of carotenoids in Raman spectra. Tomato fruit during the growth phase and during the postharvest ripening stage were analysed using a laboratory Raman spectrometer equipped with a 532 nm laser. The researchers observed increase in the carotenoid signal at the start of the turning stage of the fruit ripening. The acquired data were used by the team to build a model describing the stages of the tomato ripening process and helping to accurately assess post-harvest fruit quality.

A similar study using handheld Raman spectrometers for hot pepper was first carried out by Langer et al. [86] Scientists in their reports described the Raman signals of carotenoids typical of hot pepper fruits and followed their evolution during maturation (Figure 4). Researchers proposed and compared a multivariate chemometric model and a simple one-dimensional model, resulting in a four-point scale for grading the ripeness of hot pepper fruit. The authors suggest that the model proposed is appropriate for assessing the ripening stage of fruit containing carotenoid, and thus for determining the maturity on site or during the sorting process in an automated manner.

Bands in the range of approximately 800 to 1600 cm−<sup>1</sup> have been assigned to carotenoids. The band at about 860 cm <sup>−</sup><sup>1</sup> can be attributed to the asymmetric stretching of the C–O–C glycosidic bond in acid pectins. A faint band at 1327 cm−<sup>1</sup> is attributed to chlorophyll a, which is known to have the highest intensities in the pure chlorophyll spectrum. The most intense bands in the spectrum were assigned to the carotenoids, observed at 1150–1170 and 1500–1550 cm−1. The band is formed as a result of vibrations in the C = C phase and C–C stretching of the polyene chain. In the range of 1000–1020 cm<sup>−</sup>1, methyl groups attached to the polyene chain were recognized, showing moderate intensity [86].

**Figure 4.** Raman spectra of hot peppers during the maturation process obtained with handheld Raman spectrometer at 785 nm laser wavelength with a power of 100 mW. Unripe pepper (green), ripening pepper (yellow), and fully ripe fruit (red) are each represented by four deliberately selected spectra. Copyright MDPI 2021.

Cabrales and team applied CRM to examine cross-sections of growing cotton fibres during the main five developmental stages. In course of the study, the researchers analysed vibrational bands at 383 cm−1, attributed to cellulose. The intensity of the Raman spectra for cellulose was significantly lower for fibres harvested at 21 days post-anthesis (dpa) compared to 56 dpa. The sub-micron resolution of the CRM provides insight into the deposition of cellulose in the secondary cell wall. The findings showed that it is possible to obtain information on the chemical composition and structure of deposited cellulose in developing cotton fibres. CRM can be used as a tool to differentiate between cotton fibres at different stages of growth [87].

#### **4. Raman Spectroscopy for Fruit and Seeds Quality Control**

RS can also be applied to perform non-invasive assessment of the nutritional value and quality of plants, fruits, and seeds, and as a result, it is a perfect tool to be used in digital agronomy [88,89]. Nekvapil et al. [41] investigated the relationship between the freshness of selected citrus fruit varieties and their Raman spectra. Citrus fruit freshness is associated with the appearance and colour and consequently the net carotenoid content of the peel. In that study the research team assessed freshness of commercially available citrus fruits (clementines and different varieties of mandarins) using a handheld Raman spectrometer. Evaluation of fruit freshness can be performed using Raman instruments quickly, objectively, and without destroying the material. The samples were excited with 532 nm and 785 nm waves. Nekvapil et al. evaluated the fruit for carotenoid contents in the peel during the time span between fresh fruit delivery and their physical degradation. The analyses found a strong correlation between carotenoid content of the peel and the intensity of the Raman signal. The findings showed that the intensity of the Raman signal for carotenoids is a good indicator of fruit freshness. Based on this, the researchers introduced a Raman coefficient of freshness (CFresh), which decreases linearly in time with a different slope for different citrus groups. After 7 days of the experiment, a decrease in signal intensity was observed in the case of all the samples, but this change was varied. It was observed that the decrease in Raman signal intensity was smaller in the case of fruit stored in daylight compared to fruit kept in dark storage. The Raman coefficient of freshness (CFresh) calculated for specific fruit appears to be a user-friendly, fast, and sensitive method for assessing citrus fruit freshness with a portable Raman spectrometer. The portable system shows great potential for extensive use in the evaluation of citrus freshness, both in ripening crops and in fruit supplied to the market or to the food industry.

Similarly, Nikbakht and colleagues [90] applied RS to determine tomato fruit quality. The researchers demonstrated that RS can be used to measure important tomato quality parameters such as soluble solids content (SSC), acidity (pH), and colour. They showed that RS can be very effective in assessing the quality of both the external and internal properties of tomatoes.

Zhu et al. [89] applied Raman spectroscopy to examine the mechanisms underlying fruit lignification. Knowledge of the fruit lignification process would facilitate operations aimed to optimise storage and preservation strategies and to reduce post-harvest deterioration. By investigating lignin deposition in fruit at cellular level it may be possible to work out the mechanisms underlying fruit lignification. The study mainly aimed to establish a procedure for applying Raman microspectroscopy to visualise cellular-level lignification in loquat fruit. Fruit lignification leads to increased fruit firmness and is important from the viewpoint of optimum post-harvest handling of fruit to minimise deterioration. The findings of the study showed that Raman spectroscopy can effectively be used to assess fruit lignification in order to determine fruit maturity.

The group of Morey and Kurouski et al. [91] used RS to assess nutrient content of potato tubers. The researchers showed that the intensity of the 479 cm−<sup>1</sup> band (correlating with starch) increases linearly with growing starch content in the test samples. These findings suggest that RS can effectively be used to measure starch content in intact potato tubers. Using such calibration curves, the researchers were able to accurately determine the absolute starch concentration in potatoes. The findings showed differences in the spectra collected from samples with different starch contents; the Raman spectra acquired from a sample containing 12% starch (6 g starch) were statistically different from the spectra collected from samples with starch at a content of 9% (4.5 g starch) and 15% (7.5 g starch). Similarly, the spectra acquired from samples with 15% starch content differed significantly from the spectra obtained from samples with 12 and 18% starch content. The results of the analyses and the standard deviations obtained suggest that this method can be used to identify starch content within 3% accuracy. It can be expected that more careful standardisation may change the prediction accuracy to 1% and below.

Krimmer et al. [92] used a Raman spectrometer to assess the nutrient content of maize grains. Maize is popular worldwide as an ingredient in human food and livestock feed as well as a raw material in industry and a biofuel. The researchers found that Raman spectroscopy can identify carbohydrates, fibre, carotenoids, and proteins in maize kernels [63]. Abreu and colleagues applied RS to monitor coffee quality. They collected spectra from coffee beans stored under different conditions for 0, 3, 6, 9, 12, and 18 months. The researchers observed that changes in kahweol visualised by means of vibrational bands can be helpful in predicting coffee quality and changes taking place in beans kept in storage.

#### **5. Raman Spectroscopy for Determination of Species and Origin of Plants, Fruit, and Seeds**

Related publications have reported successful attempts to build RS-based models enabling identification of fruit varieties. Feng and colleagues applied RS to study eight different citrus fruits. The team succeeded in building a model to distinguish between citrus varieties. This work shows that RS can be used to accurately, quickly, and efficiently identify varieties and assess citrus fruit quality [93].

Kurouski et al. [91] focused on using SR to distinguish potato varieties. Owing to their high starch content, simple cultivation process, and high yield, potatoes are one of the basic ingredients in the diet of people across the world. Potato tubers consist of approximately 83% water and 12% carbohydrates, whereas proteins, vitamins, and other trace elements account for the remaining 4%. The precise composition of potatoes varies relative to the type of potato and the place where they are grown. The researchers successfully used Raman spectroscopy to identify nine different potato varieties and to determine the origin of the crop. Using spatially offset Raman spectroscopy (SORS), the researchers observed that the peak intensity varied between potato varieties at 479 and 1125 cm−<sup>1</sup> for starch, 1600 and 1630 cm−<sup>1</sup> for phenylpropanoid, 1527 cm−<sup>1</sup> for carotenoid content, and 1660 cm−<sup>1</sup> for protein content. Based on the data obtained, Kurouski's team was able to identify the potato variety and determine the location of the potato crop with an accuracy of 81 to 100%.

Farber and colleagues [18] showed that by acquiring Raman spectroscopy spectra of peanut leaves it is possible to identify different peanut varieties and genotypes. This method can also be applied to determine plant resistance to nematodes and to measure the oleic/linoleic oil (O/L) ratio. Furthermore, analysis of peanut seeds based on Raman spectroscopy can be applied to accurately identify the genotype as well as carbohydrates, proteins, fibre, and other nutrients. In course of their experiment, the researchers grew plants of 10 different peanut genotypes and analysed their spectra. All the genotypes were found with similar profiles, with vibrational bands characteristic of carbohydrates, cellulose, pectins, carotenoids, phenylpropanoids, protein, and carboxylic acid. The findings acquired using the specially developed PLS-DA model showed that the Raman method makes it possible to identify peanut varieties with 80% accuracy. The results of the study suggest that the resistance of peanut plants to nematodes is related to changes in carotenoid- and phenylpropanoid-specific peaks. In the food industry, peanuts with a high oleic ratio are preferred because of their longer shelf life and consequently reduced rancidity. It has also been shown that high-oleic peanuts beneficially affect (i.e., decrease) serum cholesterol levels and reduce the risk of cardiovascular disease. RS showed that plants of this specific genotype had lower phenylpropanoid contents, while all other peaks remained almost identical. Farber and the team reported 82% accuracy of the Raman method in identifying peanuts with high versus normal oleic ratios. To compare the accuracy of the method, the researchers performed Raman analyses of peanut seeds. The findings show 95% accuracy of Raman spectroscopy in identification of peanut seeds, compared to 82% in the case of leaves.

Krimmer et al. [63] investigated the feasibility of Raman spectroscopy in identification of maize varieties. Using PLS-DA, Krimmer and colleagues showed that RS can be used to identify six different maize varieties based on their unique spectra. In the course of the experiment, Krimmer and colleagues collected over 600 spectra from six different maize varieties. All six varieties had similar spectral profiles, except for the scanned darker kernels, which had a lower intensity. This is due to the different light absorption and scattering properties of these maize kernels, which affect the scanning. This problem can be solved by normalisation, especially at the 1458 cm−<sup>1</sup> peak displayed by all spectra display. The authors analysed the intensity of the bands at 479 cm−<sup>1</sup> (starch), 1530 cm−<sup>1</sup> (carotenoids), 1600/1632 cm−<sup>1</sup> (both fibres), and 1640–1670 cm−<sup>1</sup> (protein region) to quantify the carbohydrate, carotenoid, fibre, and protein content of maize. Krimmer and colleagues showed that RS in combination with chemometric methods can be used for highly accurate typing of maize varieties.

Abreu and colleagues in their study showed that RS can be used to discriminate Arabica coffee genotypes very accurately. [92]. Using Raman spectroscopy and principal component analysis, Figueiredo and co-workers were able to identify four Arabica coffee genotypes: one Mundo Novo line and three Bourbon lines, with an accuracy of ~ 80% [94]. Keidel and co-workers [95] showed that spectroscopic measurements of both ground and whole beans can be used to predict the geographical origin of coffee beans.

#### **6. Conclusions**

Raman spectroscopy has been used in analysis of plant biomass for nearly 30 years [96,97]. In plant science, the Raman spectroscope was first used by Atella and Agarwal in the 1980s [52,96,97]. Owing to recent technological advancements, today Raman spectroscopy and the related instruments are recognised as precise tools for studying plant tissues [97,98]. Currently, there are over 25 different types of Raman spectroscopy techniques. For example, a Fourier transformation (FT) Raman spectrometer using a near-infrared (NIR) laser solves the problem of fluorescence interference [99]. The surface-enhanced Raman spectroscopy technique enhances the Raman scattering signal [100]. Raman confocal microscopy provides

three-dimensional images of the structure and composition of the material with micrometric resolution and clear image quality [101]. Coherent anti-Stokes Raman scattering–(CARS) provides spectral information with excellent sensitivity and low laser power [102]. Raman resonance scattering (RRS) allows the study of a spectrum of materials in the range of the photon energy itself [103]. Raman optical activity (ROA) monitors a small difference in the scattering of right- and left-circularly polarized light. ROA spectra are sensitive to chirality (e.g., to the enantiomeric excess) and fine variations in molecular geometry including conformational states [104].

Raman microscopy is widely used to examine the composition of the plant cell wall [105]. Numerous studies have reported that Raman imaging techniques can successfully be applied to investigate differences between tissues and to monitor changes in plant cells [35,80,82,106–112]. Raman spectroscopy can be used to acquire marker-free spectral maps of biological tissues and cells for the needs of chemical, structural, and environmental analyses. Raman images present information about molecular structure, composition, and interactions in micrometres or even nanometres [54,55]. An important advantage presented by the Raman techniques is the fact that Raman signal processing can be used to obtain analysis of several samples or plant fragments of one type [60].

Agriculture and plant growing, as well as plant pathology, are relatively new areas in RS-based research. The related literature published in the recent years shows that Raman spectroscopy can successfully be used in diagnostics of biotic as well as abiotic stresses in plants. Rapid assessment of plant phenotype allows farmers to intervene immediately and precisely to mitigate biotic and abiotic stresses [11,17,22–24,26]. Since they are highly sensitive to small changes in plant biochemistry, Raman spectrometers make it possible to identify plant species and plant varieties and allow farmers to select and grow plants with advantageous genotypes [18,63].

This review shows the potential of RS for the digital agriculture of the future. Portable and rapid Raman spectrometry analysis allows for quick detection of biotic and abiotic stresses in plants. Moreover, Raman techniques can be used as an advanced method for plant breeding and selection as they are both non-invasive and non-destructive. RS can also be used for plant phenotyping and nutrient analysis. Advantages presented by RS will certainly become more obvious to others, and the use of Raman spectrometry in digital agriculture will become more widespread. The relatively high cost of the related equipment is a significant drawback for farmers, adversely affecting widespread use of Raman spectroscopy in crop monitoring. It can be expected that continued advancements in the technology will bring the cost of these devices down in the near future. Furthermore, application of Raman spectrometry in agriculture is likely to be implemented in the near future as a service providing farmers with information on the condition of the field along with GPS coordinates of the assessed locations.

The above review takes a closer look at the potential of Raman spectrometers to be used for the advancement of digital agriculture. The related gains outlined here are gradually becoming more obvious for farmers and investors, and the use of Raman spectroscopy in digital agriculture will become more widespread [8].

**Author Contributions:** Conceptualization, A.S.; formal analysis, A.S.; investigation, A.S.; resources, A.S.; data curation, B.S.; writing—original draft preparation, A.S.; writing—review and editing, A.S. and B.S.; visualization, B.S.; supervision, B.S and C.P.; project administration, B.S.; funding acquisition, C.P. All authors have read and agreed to the published version of the manuscript.

**Funding:** The project is financed by the program of the Ministry of Science and Higher Education "Regional Initiative of Excellence" in the years 2019–2022, Project No.026/RID/2018/19, the amount of financing totaling PLN 9 542 500.00.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Abbreviations**


#### **References**


MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Molecules* Editorial Office E-mail: molecules@mdpi.com www.mdpi.com/journal/molecules

Academic Open Access Publishing

www.mdpi.com ISBN 978-3-0365-8231-3