Exploring the Lipids Involved in the Formation of Characteristic Lactones in Japanese Black Cattle

The meat from Japanese Black cattle (Japanese Wagyu) is finely marbled and exhibits a rich and sweet aroma known as Wagyu beef aroma. To clarify the key metabolites involved in the aroma, we analyzed the correlation between lactone and lipid composition in Japanese Black cattle. Using gas chromatography-olfactometry, we identified 39 characteristic odorants of the intermuscular fat. Seven characteristic lactones considered to be involved in Wagyu beef aroma were quantified and compared in the marbled area and intermuscular fat using a stable isotope dilution assay. Among them, γ-hexalactone was the only lactone whose level was significantly higher in the marbled area. To explore the lipid species involved in lactone formation, we analyzed samples with different aroma characteristics. Liquid chromatography-mass spectrometry revealed eight lipid classes and showed significant differences in triacylglycerides (TAGs). To determine the molecular species of TAGs, we performed high-performance liquid chromatography analysis and identified 14 TAG species. However, these analyses showed that seven lactones had a low correlation with the TAGs. However, γ-hexalactone showed a positive correlation with linoleic acid. This study suggests that lipid composition affects the characteristic lactone profile involved in the Wagyu beef aroma.


Introduction
Japanese Black cattle, also known as Japanese Wagyu, are used to produce one of the world's most renowned types of beef [1]. Its defining characteristics are the excellent marbling of crossed fat in muscle tissues and its rich and sweet aroma (the so-called Wagyu beef aroma). Matsuishi et al. confirmed that lactones of cyclic esters produced during cooking contributed to the Wagyu beef aroma [2]. There are two types of lactones produced during beef cooking, namely, γ-lactone and δ-lactone, and they have different heterocyclic carbon atoms. The quality of the sweet odor changes with the length of the carbon chain bonded to the heterocycle [3].
Holstein cattle are bred as dairy cattle, but male cattle castrated before weaning are used for meat production. Holstein cattle, which are mostly characterized by lean meat, are often used experimentally in comparisons with Japanese Black cattle, as their proportions of intramuscular fat [4], which is responsible for marbling in the muscles, differ significantly. Japanese Black cattle have a considerably higher marbling rate than Holstein cattle and are rich in monounsaturated fatty acids (MUFAs) [5]. As the characteristic lipid metabolism of Japanese Black cattle has a significant influence on meat quality (tenderness, flavor, and juiciness) [6,7], there are several studies on genes which are crucial for the marbling trait [8][9][10][11]. In addition, studies have attempted to analyze the aroma of Wagyu beef [12][13][14].

GC-O Analysis of Intermuscular Fat Aroma
In the gas chromatography-mass spectrometry (GC-MS) analysis, the aroma components from Japanese Black cattle ribeye steak (musculus longissimus) formed more than 1000 peaks after cooking, making it difficult to identify the odorants that contributed to Wagyu beef aroma; see Figure S1. Consequently, the ribeye steak was divided into two parts, namely, intermuscular fat around the steak and marbled beef (called marbled area in this study).
In this study, we hypothesized that lipids are involved in the formation of Wagyu beef aroma. The aroma characteristics of the intermuscular fat were analyzed using GC-O analysis. The focus was on the 39 odorants previously detected in the marbled area using the GC-O analysis [11]. Similar to previous data, the odorants generated when the boiled intermuscular fat of Japanese Black cattle of different pedigrees (Type A and Type B) and Holstein cattle were compared.
The characteristic odorants of intermuscular fat and marbled area are displayed as visual plots using the multivariate analysis of their FD values in Figure 1. The left side of the S plot shows the prominent odorants (3, 9, 13, 14, 33, 35, and 38) of the intermuscular fat. Among them, 3-methyl-2-butene-1-thiol (3) is a sulfur compound that causes the aging flavor of beer [25], whereas 4-vinylphenol (33) and 9-decenoic acid (35) have unique flavors. These unique odorants (3, 33, and 35), which have high FD values in intermuscular fat, may contribute to the distinctive complex aroma generated from fat tissues [26]. In contrast, the right side shows the prominent odorants of marbled area (2, 4, 10, 18, 22, 25, 28, 30, 31, 34, and 39). Methional (10) is a degradation compound of methionine, generating a stewed potato odor [27]. Maltol (25) is a caramel-like odorant produced by the Maillard reaction; it is also known to have an e-cigarette fragrance [28]. Hexanoic acid (22) and decanoic acid (34) are medium-chain fatty acids with a dust cloth odor [29]. These odorants (10, 22, 25, and 34) are inferred from the odor quality associated with the beef flavor specific to the muscle tissue. Fat tissues and muscle tissues from the musculus longissimus of Japanese Black cattle Type A and Type B, and Holstein cattle were used for the analysis. The multivariate analysis showed characteristic odorants in the intermuscular fat and marbled area (R2X = 0.474, Scaling, Par). The OPLS-DA model was calculated based on the data presented in Table 1 and previous study data [11].
The score plots of the OPLS-DA model were R2 (cum) = 1.00 and Q2 (cum) = 0.997. The plot number indicates the number of compounds in the corresponding odorants Table 1.

Quantification of Odorants Using the Stable Isotope Dilution Assay (SIDA)
In a previous study, the lactones related to beef aroma were studied using solventextracted lipids [14]. In this study, we compared intermuscular fat and marbled area to examine the lactone characteristic of Wagyu beef aroma generated from the edible part. The box plots in Figure 2 show the seven lactones measured using the SIDA. Although γ-hexalactone (16), γ-heptalactone (20), γ-decalactone (29), and γ-undecalactone (ND) were not or were slightly detected in the intermuscular fat using the GC-O analysis Table. 1, these lactones could be quantified using the SIDA. γ-Heptalactone (20), γ-decalactone (29), γ-nonalactone, δ-decalactone, and γ-undecalactone were detected at higher levels in the intermuscular fat than in the marbled area. Whereas, γ-hexalactone (16) and γ-octalactone (23) presented higher levels in the marbled area than in the intermuscular fat. The level of γhexalactone was the highest in the edible marbled area (p < 0.001).
Lactones are sweet odorants that are produced from hydroxy fatty acids [30]. Their volatility depends on the number of carbon atoms. γ-Hexalactone is the most volatile lactone (boiling point 220 °C ) and exhibits a sweet fruity and nutty aroma [31]. Previous studies have shown that γ-hexalactone correlates with the strength of Wagyu beef aroma and proposed that γ-hexalactone can be an indicator of Wagyu beef aroma [11]. The difference in lactone profiles determined using the SIDA is considered to reflect the quality of Wagyu beef aroma between the fat and muscle tissue. As the marbled area is the edible part, we focused on that in the next comprehensive analysis of lipid composition.  Table 1 and previous study data [11]. The score plots of the OPLS-DA model were R2 (cum) = 1.00 and Q2 (cum) = 0.997. The plot number indicates the number of compounds in the corresponding odorants Table 1.

Quantification of Odorants Using the Stable Isotope Dilution Assay (SIDA)
In a previous study, the lactones related to beef aroma were studied using solventextracted lipids [14]. In this study, we compared intermuscular fat and marbled area to examine the lactone characteristic of Wagyu beef aroma generated from the edible part. The box plots in Figure 2 show the seven lactones measured using the SIDA. Although γ-hexalactone (16), γ-heptalactone (20), γ-decalactone (29), and γ-undecalactone (ND) were not or were slightly detected in the intermuscular fat using the GC-O analysis Table 1, these lactones could be quantified using the SIDA. γ-Heptalactone (20), γ-decalactone (29), γ-nonalactone, δ-decalactone, and γ-undecalactone were detected at higher levels in the intermuscular fat than in the marbled area. Whereas, γ-hexalactone (16) and γ-octalactone (23) presented higher levels in the marbled area than in the intermuscular fat. The level of γ-hexalactone was the highest in the edible marbled area (p < 0.001). Metabolites 2021, 11, x FOR PEER REVIEW 6 of 16 Figure 2. Comparative analysis of the characteristic odorants identified in the intermuscular fat and marbled area of Japanese Black cattle. The lactones were measured after boiling using the stable isotope dilution assay (SIDA). The box plot is an exclusive median and shows all plots, including outliers. The cross marks indicate the mean values. The samples analyzed were the intermuscular fat and marbled area from Japanese Black cattle (17 cattle for each tissue, 10 Type A and 7 Type B). Type A Japanese Black cattle are a typical pedigree (non-Tajima), and Type B is a closed breeding pedigree (Tajima). The total lactone is the sum of the γ-hexalactone, γ-heptalactone, γ-octalactone, γ-nonalactone, γ-decalactone, δ-decalactone, and γ-undecalactone. Significant differences are indicated as follows: ** p < 0.01, * p < 0.05.

Lipidomic Analysis of Japanese Black Cattle Meat by Liquid Chromatography-Mass Spectrometry (LC-MS)
To comprehensively detect multiple lipid classes in the marbled area of Japanese Black cattle meat, we used LC-MS analysis. For this analysis, we selected loin (musculus longissimus) and adductor magnus (round) as the negative control from the same Japanese Black cattle to match the genetic and physiological conditions. The musculus longissimus has more marbling than the adductor magnus, and it is highly valued for its flavor [32]. We also added samples of different meat quality grades (A3 and A5 rank) to the comparison. A5 is the highest quality grade, and A3 is the middle grade with a moderate marbling rate [33].
After normalization and removal of the nonspecific peaks, our LC-MS analytical conditions revealed 108 lipids in the chloroform/methanol fraction. The mean value of the coefficient of variation for the internal standards was 8.1%, and good reproducibility was confirmed. The lipid species identified by LC-MS were classified into eight classes; see Figure  S2. The molecular species in the lipid class included the following: acylcarnitine (AcCa), . Comparative analysis of the characteristic odorants identified in the intermuscular fat and marbled area of Japanese Black cattle. The lactones were measured after boiling using the stable isotope dilution assay (SIDA). The box plot is an exclusive median and shows all plots, including outliers. The cross marks indicate the mean values. The samples analyzed were the intermuscular fat and marbled area from Japanese Black cattle (17 cattle for each tissue, 10 Type A and 7 Type B). Type A Japanese Black cattle are a typical pedigree (non-Tajima), and Type B is a closed breeding pedigree (Tajima). The total lactone is the sum of the γ-hexalactone, γ-heptalactone, γ-octalactone, γnonalactone, γ-decalactone, δ-decalactone, and γ-undecalactone. Significant differences are indicated as follows: ** p < 0.01, * p < 0.05.
Lactones are sweet odorants that are produced from hydroxy fatty acids [30]. Their volatility depends on the number of carbon atoms. γ-Hexalactone is the most volatile lactone (boiling point 220 • C) and exhibits a sweet fruity and nutty aroma [31]. Previous studies have shown that γ-hexalactone correlates with the strength of Wagyu beef aroma and proposed that γ-hexalactone can be an indicator of Wagyu beef aroma [11]. The difference in lactone profiles determined using the SIDA is considered to reflect the quality of Wagyu beef aroma between the fat and muscle tissue. As the marbled area is the edible part, we focused on that in the next comprehensive analysis of lipid composition.

Lipidomic Analysis of Japanese Black Cattle Meat by Liquid Chromatography-Mass Spectrometry (LC-MS)
To comprehensively detect multiple lipid classes in the marbled area of Japanese Black cattle meat, we used LC-MS analysis. For this analysis, we selected loin (musculus longissimus) and adductor magnus (round) as the negative control from the same Japanese Black cattle to match the genetic and physiological conditions. The musculus longissimus has more marbling than the adductor magnus, and it is highly valued for its flavor [32]. We also added samples of different meat quality grades (A3 and A5 rank) to the comparison. A5 is the highest quality grade, and A3 is the middle grade with a moderate marbling rate [33].
After normalization and removal of the nonspecific peaks, our LC-MS analytical conditions revealed 108 lipids in the chloroform/methanol fraction. The mean value of the coefficient of variation for the internal standards was 8.1%, and good reproducibility was confirmed. The lipid species identified by LC-MS were classified into eight classes; see Figure S2. The molecular species in the lipid class included the following: acylcarnitine (AcCa), 10 species; lysophosphatidylcholine (LPC), 9 species; lysophosphatidylethanolamine (LPE), 4 species; phosphatidylcholine (PC), 44 species; phosphatidylethanolamine (PE), 12 species; sphingomyelin (SM), 5 species; diacylglyceride (DAG), 1 species; and triacylglyceride (TAGs), 23 species; see Figure 3. However, it was difficult to accurately separate the TAG molecular species because their three fatty acids were bound to the glycerol skeleton in various combinations.  Figure 3. However, it was difficult to accurately separate the TAG molecular species because their three fatty acids were bound to the glycerol skeleton in various combinations. AcCa and PC were significantly higher in the adductor magnus with less marbling than in the samples. The tendency for the PC values to be higher in the adductor magnus was consistent with that observed in the analysis of phospholipids in New Zealand beef [34]. AcCa consists of carnitine and fatty acids and is involved in the β-oxidation of lipids. AcCa is abundant in the mitochondria, particularly in lean muscle [35]. In contrast, TAG was significantly higher in the musculus longissimus, depending on the meat quality grade (level of marbling). TAG is the major lipid stored in the lipid droplets of adipocytes. Lipid droplets have a single membrane structure of phospholipids surrounding the hydrophobic TAG and cholesterol esters [36]. This difference in lipid composition, consisting of lipid droplets, may be reflected in the LC-MS results. It is known that the A5 musculus longissimus has a strong Wagyu beef aroma [7,37]. Next, we focused on the TAG, which is the most abundant in A5 musculus longissimus, and measured its molecular species composition. AcCa and PC were significantly higher in the adductor magnus with less marbling than in the samples. The tendency for the PC values to be higher in the adductor magnus was consistent with that observed in the analysis of phospholipids in New Zealand beef [34]. AcCa consists of carnitine and fatty acids and is involved in the β-oxidation of lipids. AcCa is abundant in the mitochondria, particularly in lean muscle [35]. In contrast, TAG was significantly higher in the musculus longissimus, depending on the meat quality grade (level of marbling). TAG is the major lipid stored in the lipid droplets of adipocytes. Lipid droplets have a single membrane structure of phospholipids surrounding the hydrophobic TAG and cholesterol esters [36]. This difference in lipid composition, consisting of lipid droplets, may be reflected in the LC-MS results. It is known that the A5 musculus longissimus has a strong Wagyu beef aroma [7,37]. Next, we focused on the TAG, which is the most abundant in A5 musculus longissimus, and measured its molecular species composition.
Next, we used high-performance liquid chromatography (HPLC) to analyze TAG molecular species; this approach is conventionally used to separate TAG molecules [38]. After normalization, the HPLC analytical conditions revealed 14 peaks in the TAG fraction; see Figure S2. The details of the TAGs are shown in Table 2. The major TAGs were POO, 29.9%; POP, 9.8%; PPoO, 8.2%; POS, 7.9%; SOO, 7.3%; OOO, 7.2%; MOP, 4.9% (O, oleic acid; P, palmitic acid; S, stearic acid; Po, palmitoleic acid; and L, linoleic acid). Seven TAGs accounted for 75.2% of the total TAGs. The total peak area for the unknown TAGs was 10.2%. To clarify the TAG contributing to Wagyu beef aroma, we quantified the above seven lactones using the SIDA. The lactone levels produced from musculus longissimus during boiling are shown in Table 2.

Correlation between the Composition of Lipids and Odorants Related to Wagyu Beef Aroma
Next, we examined the correlations among TAGs, fatty acid composition, and the seven lactones. As expected, the TAG molecular species and fatty acid compositions showed a high correlation; see Table 3. Of the fatty acids, myristic acid, stearic acid, linoleic acid, palmitoleic acid, and margaric acid presented low TAG composition ratios. These fatty acids showed a high correlation with specific TAGs as follows: myristic acid (MOP:

LACTONES (ng/g Beef) Muscle
Tissue In the correlation between fatty acid composition and lactones, several fatty acids were correlated. γ-Hexalactone was positively correlated with linoleic acid (r = 0.54) and negatively correlated with myristic acid (r = −0.53). Among other lactones, γ-decalactone was correlated with myristoleic acid (r = 0.54) and linolenic acid (r = −0.49). The heatmap overview showed a similar tendency for low molecular weight γ-hexalactone (six carbon atoms; C-6) and γ-heptalactone (C-7). PUFAs linoleic acid and linolenic acid were correlated with γ-hexalactone and γ-heptalactone. The cyclization of hydroxy fatty acids generally causes the formation of γ-lactones [39]. γ-Nonalactone (C-8) and γ-decalactone (C-10) are formed via hydroxy fatty acids during fermentation from MUFAs, such as oleic acid and palmitoleic acid [30]. Unlike other γlactones, the details of the molecular mechanism of γhexalactone formation while cooking beef are unknown. From the results of the correlations between fatty acids and γ-hexalactone, it is speculated that PUFAs are involved in the formation of low molecular weight γ-lactones. In addition, as γ-hexalactone is produced in a marbled area, it is suggested that metabolites (organic acids, amino acids, and sugars) derived from muscle tissue are involved [17]. Metabolites (glutamine, decanoic acid, sedoheptulose 7-phosphate, creatinine, and xanthine) that are positively correlated with γhexalactone in muscle tissue may be candidates involved in the formation of low molecular weight γ-lactones [11]. The production of hydroxy fatty acids may also contribute to lipase activity, resulting in the production of free fatty acids in the muscle tissue of the marble area.
In this study, we did not find a strong correlation between the TAG molecular species and the seven lactones. As more combinations can be inferred from the fatty acid compositions, the presence of many unidentified TAG molecular species is expected. However, γ-hexalactone showed a positive correlation with linoleic acid. This study suggests that lipid composition affects the characteristic lactone profile involved in Wagyu beef aroma.

Sample Collection
We purchased the muscle blocks commercially from meat wholesalers. The blocks were then sliced into steaks and individually packaged. These samples were then stored in a freezer at −30 • C (Table 4 and Figure 4). f beef samples used for testing. Beef samples were obtained from Type A and Type B Japanese Black the aim of the study [11]. Type A Japanese Black cattle are a typical pedigree (non-Tajima) that exhibit ight growth. Type B is a closed breeding pedigree (Tajima) that is highly traded, and it has an excellent is, Kobe beef grade [40]. Meat quality grade is according to the carcass trading standards (The Japan iation, Tokyo, Japan). The grades range from 1 to 5, depending on marbling, meat color and brightness, texture, fat color, luster, and quality (higher values indicate high quality) [1].

Position
Meat Quality Grade

GC-O Analysis of the Odorant Concentrations in Boiled Beef
As in a previous report [11], the aroma characteristics of Wagyu beef were compared with those of the intermuscular fat from Type A and Type B Japanese Black cattle and Holstein cattle. A total of 50 g of fat tissue (intermuscular fat around the steak) was boiled in 500 mL of distilled water for 30 s. The fat tissues were cooled with ice, ground with   [11]. Type A Japanese Black cattle are a typical pedigree (non-Tajima) that exhibit excellent body weight growth. Type B is a closed breeding pedigree (Tajima) that is highly traded, and it has an excellent meat quality, that is, Kobe beef grade [40]. Meat quality grade is according to the carcass trading standards (The Japan Meat Rating Association, Tokyo, Japan). The grades range from 1 to 5, depending on marbling, meat color and brightness, meat hardness and texture, fat color, luster, and quality (higher values indicate high quality) [1].

GC-O Analysis of the Odorant Concentrations in Boiled Beef
As in a previous report [11], the aroma characteristics of Wagyu beef were compared with those of the intermuscular fat from Type A and Type B Japanese Black cattle and Holstein cattle. A total of 50 g of fat tissue (intermuscular fat around the steak) was boiled in 500 mL of distilled water for 30 s. The fat tissues were cooled with ice, ground with a mixer with 500 mL water containing the fat eluted by boiling, added to 500 mL dichloromethane, and extracted for 16 h at 25 • C with stirring. Many nonvolatile compounds, such as the fats and oils derived from the fat tissues, were removed using an SAFE apparatus at an ultralow temperature (−196 • C). After drying with anhydrous sodium sulfate, the sample was concentrated using a Kuderna-Danish evaporative concentrator [11].
The GC-O analysis was performed under the same conditions to compare the data with those of a previous analysis that used the marbled area [11]. For the GC-O analysis, we used CharmAnalysis (DATU, Geneva, NY, USA) with an Agilent 6890 gas chromatograph (Agilent Technologies, Santa Clara, CA, USA) equipped with a DB-WAX capillary column (length, 15 m; inner diameter, 0.32 mm; film thickness, 0.25 mm; Agilent Technologies) [15]. The odor extract was diluted stepwise (four-fold) with dichloromethane. Compounds were identified by comparing their odor quality, RI, and mass spectrum with those of authentic compounds on a DB-WAX column.

Quantification of Lactone
The concentration of the odorants was measured using the SIDA with a liquid extraction method in which stable isotopes were added to dichloromethane as an internal standard. Stable isotopes of γ-nonalactone, γ-octalactone, γ-decalactone, γ-undecalactone, γ-hexalactone, γ-heptalactone, and δ-decalactone were purchased from AromaLAB (Martinsried, Germany). Quantitation was performed using a GC-tandem quadrupole mass spectrometer (Agilent 7000C TripleQuad GC-MS system; Agilent Technologies) equipped with a DB-WAX capillary column (length, 30 m; inner diameter. 0.25 mm; film thickness, 0.25 mm; Agilent Technologies). The oven temperature was programmed to hold at 35 • C for 5 min, and then increased from 35 • C to 217 • C at a rate of 4 • C/min. The system was operated in the multiple reaction monitoring (MRM) mode. Two microliters of the concentrate were then injected into the instrument with the inlet temperature set at 250 • C in the spitless mode.

LC-MS Analysis
One gram of adductor magnus or musculus longissimus was frozen in liquid nitrogen, and then milled with a multibead shocker µT-48 (Token, Chiba, Japan). The powdered samples (20 mg) were homogenized with 500 µL methanol using an ultrasonic cleaner for 10 min and stirred at 2500 rpm for 5 min in a shaking incubator. Chloroform (500 µL) was then added, and they were shaken at 2500 rpm for 5 min in a shaking incubator and centrifuged at 9100× g for 5 min to obtain a chloroform/methanol fraction. A sample was prepared by adding an internal standard of phosphatidic acid (16:0 D30/18:1) to the supernatant to a final concentration of 0.2 µg/mL. The internal standard was purchased from Avnati Polar Lipids (Alabaster, AL, USA). High-resolution Fourier transform mass spectrometry (LC-MS) analysis consisted of ultrafast liquid chromatography (UFLC XR; Shimadzu, Kyoto, Japan) with an L-column2 ODS metal-free column (inner diameter, 2 mm; length, 50 mm; particle size, 3 µm; Chemicals Evaluation and Research Institute, Tokyo, Japan) and LTQ-Orbitrap XL (Thermo Fisher Scientific K. K., Tokyo, Japan). The separation buffer for the UFLC consisted of solvent A (1 mM ammonium formate solution), solvent B (ammonium formate/isopropanol), and solvent C (acetonitrile). For the UFLC, the injection volume was 3 µL, sampler temperature was 4 • C, column temperature was 40 • C, and flow rate was 0.3 mL/min. Mass spectrometry was performed using the thermal electrospray ionization method in the data-dependent Top N3 scan mode at a resolution of 30,000 from 200 to 1600 m/z.
For peak selection, the estimations for the lipid species and alignments between the samples were analyzed using the lipid identification software Lipid Search (Mitsui Information, Tokyo, Japan). The search options were set as follows: precursor tolerance, 5.0 ppm; product tolerance, 0.5 Da; merge range, 2.0; and min peak width, 0.0. The quantitation option was set as follows: MZ tolerance, −5.0 to +5.0 ppm; RT range, −0.5 to +0.5 min. The estimated lipids were confirmed based on the retention time and product ions. In this study, only lipids satisfying the conditions (peak intensity ≥5000, coefficient of variation ≤0.15, and mean signal-to-noise ratio ≥3) were selected. The area of the detected peak in each sample was corrected using the peak area value of the internal standard [41].

Analysis of Fatty Acids and TAG Composition
Total lipids were extracted from ground beef (10 g) with t-butyl methylether/methanol (2:1) according to a previously described method [42]. TAG fractions were collected using solid phase extraction with an InertSep SI column (GL Sciences, Tokyo, Japan) according to the manufacturer's instructions. To analyze the fatty acid composition of the TAG samples, they were first dissolved in n-hexane and methyl esterified with potassium hydroxidemethanol solution [43]. The methylated sample (1 µL) was used for gas chromatography (GC-2010 Plus; Shimadzu, Kyoto, Japan) with a TC-70 capillary column (length, 60 m; inner diameter, 0.25 mm; film thickness, 0.25 µm; GL Sciences). The initial temperature was set at 150 • C, increased at 5 • C/min to 235 • C, and then held at that temperature for 8 min. The oven temperature was maintained at 150 • C for 30 min and was then increased to 250 • C at a rate of 10 • C/min and held isothermally for 13 min.
For the TAG analysis, the dissolved samples in isopropyl alcohol were subjected to HPLC using Agilent Technologies 1260 Infinity equipped with a refractive index detector and a Poroshell 120 EC-C18 LC column (three columns in a series, 3.0 mm × 50 mm, 3.0 mm × 50 mm, 3.0 mm × 100 mm; 2.7-Micron; Agilent Technologies). A mixture of acetonitrile and 2-propanol (4:6, v/v) was used. The flow rate was 0.2 mL/min, and the column temperature was maintained at 20 • C. The quantification of individual TAGs was performed by evaluating the corresponding relative percentage according to the normalization area procedure.

Statistical Analysis
Statistical significance was determined using a p-value (Student's t-test and Bonferroni test) with Excel 2019 software (Microsoft Japan, Tokyo, Japan). The correlation coefficient and uncorrelated test variables between lactones and lipids were recalculated using JMP12 (SAS Institute Japan, Tokyo, Japan). Multivariate data analysis of OPLS-DA, and PCA was performed using SIMCA14 software (Inforcom, Tokyo, Japan).

Conclusions
In this study, we compared and characterized 39 odorants that were detected using the GC-O analysis from the intermuscular fat and marbled area from Type A and Type B Japanese Black cattle and Holstein cattle. The GC-O analysis revealed various odorants that contribute to the flavor of cooked beef generated from the fat and muscle tissue. The qualitative analysis using SIDA revealed seven odorants that contributed to Wagyu beef aroma, and γ-hexalactone was identified as a characteristic odorant generated from the marbled area. The lipid composition was investigated, as it was expected to be involved in the formation of γ-hexalactone. The LC-MS analysis revealed 108 lipids belonging to eight lipid classes in adductor magnus and musculus longissimus from Japanese Black cattle. The semiquantitative analysis using HPLC revealed 14 TAGs contained in the marbled area. Contrary to our hypothesis, the correlation of the γ-hexalactone content with the major TAG species was small but positive (r = 0.55) for linoleic acid and negative (r = −0.53) for myristic acid.