Abstract

   The freshness of Atlantic salmon is influenced mainly by tissue
   metabolism, which in turn is affected by storage time and conditions.
   The alterations in taste profiles and nutritional values of salmon when
   packaged using vacuum methods have not been fully understood, and the
   factors contributing to these changes require further research. In this
   work, the extraction method for flavor nutrients from salmon was
   optimized via the Plackett–Burman (PB) test. A sensitive and rapid
   targeted metabolomics method for the simultaneous determination of 34
   nutrients was successfully established via ultra-performance liquid
   chromatography–triple quadrupole/linear ion trap composite mass
   spectrometry (UHPLC-QTRAP/MS), and various nutritional compositions
   during storage at 0 °C under different vacuum conditions (0 kPa or −90
   kPa) for 4 and 8 days were analyzed. Results showed that storage time
   had a significant effect on salmon metabolism. The total amino acids
   decreased by 62.95% and 65.89% at 0 kPa and −90 kPa, respectively.
   Notably, a marked reduction in histidine after 8 days at −90 kPa may
   have diminished bitterness, while decreased levels of umami-tasting
   amino acids like glutamine and aspartic acid affected the overall
   flavor profile. Overall, the packaging conditions at 0 °C and 0 kPa
   were more suitable for the preservation of most nutrients in salmon.
   Pathway enrichment analysis revealed that the reduction in substances
   was mainly related to the alanine, aspartate, and glutamate metabolism
   pathways. Alanine, inosine, and histidine, whose levels changed
   significantly, can bind to the typical umami taste receptor TIR1/TIR3
   and can be biomarkers to monitor and determine the freshness or
   spoilage of salmon after 4–8 days of storage. This study revealed the
   changes in small-molecule nutrients in salmon during storage under
   different packaging conditions, which provides a reference for the
   packaging preservation technology of fresh salmon and new ideas for the
   evaluation of salmon quality and determination of freshness.

   Keywords: Atlantic salmon, vacuum packaging, flavor nutrient, storage
   time, TIR1/TIR3

1. Introduction

   Atlantic salmon is one of the most traded and consumed aquatic products
   worldwide because of its high nutritional value. Atlantic salmon is
   known for its bright red–orange flesh and deep, rich flavor; it
   contains a variety of high-quality animal proteins and essential amino
   acids [[36]1] and is easy to digest and absorb by the human body
   [[37]2]. It is also a good source of n-3 polyunsaturated fatty acids
   (n-3 PUFAs), fat-soluble vitamins, trace elements, and other important
   biologically active substances for humans [[38]3,[39]4,[40]5,[41]6].
   These nutrients can promote growth and development [[42]7], reduce
   cholesterol levels, and have important functions that prevent and treat
   cardiovascular and cerebrovascular diseases [[43]8,[44]9], diabetes,
   and cancer [[45]10]. However, salmon is perishable during storage due
   to its smaller muscle fiber dimensions but higher moisture content
   compared to other fish; salmon quality degradation during storage is
   caused by different damage mechanisms, such as microbial activities
   [[46]11], autolytic enzymatic reactions, lipid oxidation [[47]12], and
   pathogenic microorganisms [[48]13]. The taste, nutritional content, and
   safety of salmon are crucial for both public health and consumer
   preference, as well as for the global seafood industry [[49]14].
   Therefore, developing reliable methods for the evaluation of salmon
   quality is highly important.

   Atlantic salmon are highly perishable, and temperature is highly
   important in determining their shelf life. Freshness is one of the most
   important attributes when Atlantic salmon quality is assessed. A wide
   variety of chemical and physical methods have been used to assess the
   freshness of fish during storage, including sensory evaluation
   [[50]15], microbial inspection [[51]16], and physical and chemical
   detection [[52]17]. Jääskeläinen et al. [[53]18] evaluated the
   microbial changes in raw salmon and tuna fillets stored under vacuum at
   3 °C over a period of 12 days and reported that trimethylamine was the
   main spoilage product, and that the bacterial colony count reached an
   average of 7.3 lg cfu/g in salmon. Moreover, the concentration of
   glucose decreased, and the concentration of organic acids increased
   during storage. Physical and chemical detection methods evaluate the
   quality of fish meat by analyzing the stiffness index, total volatile
   base nitrogen (TVB-N) value, K value, drip loss, water holding
   capacity, protein content, fatty acid content, and other indicators
   [[54]16,[55]19,[56]20,[57]21]. Pan and Lin [[58]21] analyzed the TVB-N
   value, color difference value, K value, and other indicators of
   Atlantic salmon at -1.8 °C and 4 °C and reported that the ice
   temperature (−1.8 °C) preservation method can extend the shelf life by
   3 days. However, few studies have evaluated the fresh quality of salmon
   on the basis of changes in small-molecule nutrients. Fidalgo et al.
   [[59]22] analyzed the fatty acid and polyethylene indices of
   vacuum-packaged fresh Atlantic salmon loins after hyperbaric storage at
   low temperature via GC–MS and reported that hyperbaric storage samples
   retained fresh-like alcohols and aldehydes. The nutrient components of
   salmon are complex and diverse, except for fatty acids, alcohols, and
   aldehydes. The spoilage process caused by microbial activity produces
   many low molecular weight metabolites [[60]23]. This study aimed to
   identify comprehensive changes in small-molecule metabolites that
   quantitatively indicate the degree of freshness of Atlantic salmon.

   Metabolomics encompasses the qualitative, quantitative, and dynamic
   study of all endogenous small molecules within organs and tissues
   [[61]24]. Metabolomics has been widely used in screening biomarkers
   related to the freshness of pork patties [[62]25], chicken [[63]23],
   eggs [[64]26], yellowfin tuna [[65]18], shucked mussels [[66]27], and
   soybean sprouts [[67]28]. Kritikos, Aska, Ekonomou, Mallouchos,
   Parlapani, Haroutounian, and Boziaris [[68]19] researched the evolution
   of volatile organic molecules during the storage of European seabass
   (Dikentrarchus labrax) fillets and Atlantic salmon (Salmo salar) slices
   under modified atmosphere packaging at 2 °C and monitored total
   volatile basic nitrogen (TVB-N), microbiological, and sensory changes.
   The results revealed that the volatile organic molecules
   3-hydroxy-2-butanone, 2,3-butanediol, 2,3-butanedione, and acetic acid
   could be potential spoilage markers. Free amino acids can be produced
   as a result of fish muscle protein autolysis, leading to fish muscle
   deterioration, which is a result of microbial growth and biogenic amine
   production [[69]29]. The levels of adenine nucleotides inosine and
   hypoxanthine are also commonly used as indicators of muscle freshness
   in fish [[70]30].

   However, little research has been conducted on evaluating the freshness
   of salmon by small-molecule metabolites (amino acids, nucleosides, and
   vitamins) via targeted metabolomics. The aim of this study was to
   investigate the metabolite changes in Atlantic salmon stored at 0 °C
   under different vacuum packaging (0 kPa or −90 kPa) conditions and to
   screen for biomarkers of freshness via targeted metabolomics and
   molecular docking. This study provides more comprehensive information
   on salmon storage and freshness packaging by measuring the low
   molecular weight nutrient (amino acid, vitamin, and nucleoside) content
   of salmon, which provides a reference for packaging preservation
   technology for fresh salmon and new ideas for the evaluation of salmon
   quality and determination of freshness.

2. Materials and Methods

2.1. Chemicals and Reagents

   The standards of 34 metabolites, including arginine (Arg), methionine
   (Met), glycine (Gly), valine (Val), tryptophan (Try), histidine (His),
   cysteine (Cys), serine (Ser), leucine (Leu), isoleucine (Iso)
   phenylalanine (Phe), threonine (Thr), alanine (Ala), proline (Pro),
   lysine (Lys), glutamine (Glu), glutamate (Gln), tyrosine (Tyr),
   aspartic acid (Asp), asparagine (Asn), γ-aminobutyric acid (GABA),
   pyroglutamic acid (Pyr), taurine (Tau), thiamine (VB[1]), riboflavin
   (VB[2]), nicotinic acid (VB[3]), folic acid, lumichrome, cytosine,
   2-hydroxyadenine, guanosine hydrate, inosine, cytidine,
   1-methyladenosine, and 2′-deoxyuridine, were purchased from Alta
   Scientific Co., Ltd. (Tianjin, China) and J&K Scientific Co., Ltd.
   (Beijing, China). MS-grade methanol and acetonitrile were purchased
   from Merck (Darmstadt, Germany). Water was purified with a Milli-Q
   system (Millipore, France).

2.2. Salmon Samples and Storage Conditions

   Fresh salmon samples were collected from local markets in Beijing. The
   salmon weighed between 3 and 4 kg, and were slaughtered before sexual
   maturity. After slaughtering, the salmon were gutted and rinsed in
   running water for 30 min, followed by chilling to 0 °C in slush ice.
   The entire process adheres to the cold chain standards, ensuring the
   product is delivered fresh to the supermarket. Once there, the salmon
   is cut and labeled as ready-to-eat. The salmon purchased for this study
   were all fresh, from the same batch, and brought back to the laboratory
   on ice. They were then cleaned with ice water and low-temperature
   sterile water. Approximately 15 g (±0.5) of a salmon slice was vacuum
   packaged in sterile plastic nylon packaging bags at 0 kPa and −90 kPa.
   All salmon samples were stored at 0 °C for 8 days, and analyses were
   performed every 4 days. The experimental conditions are shown in
   [71]Table S1. Three replications were set for all batches.

2.3. Extraction Method Optimized by the PB Experimental Design

   The extraction solvent was optimized via single-factor experiments, and
   the extraction results were obtained from three typical extraction
   solvents: acetonitrile, water (containing 1% formic acid), and
   methanol/water (1:1). A Plackett–Burman factorial design experiment
   with 12 runs was designed to identify the significant parameters
   affecting the extraction of 36 nutrients. Parameters such as extraction
   solvent volume (V), vortex time (t[1]), ultrasonic time (t[2]),
   ultrasonic temperature (T), and ultrasonic intensity (E) were
   investigated via variance analysis to analyze the systematic error
   affecting extraction efficiency. The PB design experiment is shown in
   [72]Table S2.

   A 0.2 g homogenized salmon sample was weighed and transferred to a 5 mL
   centrifuge tube. After 2 mL of extraction solvent (methanol/water =
   1:1) was added, the sample was mixed for 3 min via a vortex mixer to
   disperse the sample. Afterward, the salmon samples were sonicated for
   10 min at 20 °C and 60 W and centrifuged at 10,000 rpm for 10 min at 4
   °C. Subsequently, 1 mL of the supernatant was filtered through a 0.22
   μm membrane filter and placed in vials at 4 °C until analysis. Quality
   control (QC) samples were prepared by mixing equal aliquots of the
   supernatants from all of the samples.

2.4. Metabolite Analysis via UHPLC-QTRAP/MS

   The analyses of nutrients and metabolites were performed via an
   ultra-performance liquid chromatography–triple quadrupole/linear ion
   trap composite mass spectrometer (UPLC-QTrap 6500, AB SCIEX, Redwood
   City, CA, USA) with an Xbridge C18 column (4.6 mm × 150 mm, 3.5 µm,
   Waters, Milford, MA, USA). Mobile phase A consisted of formic
   acid/water (%), and mobile phase B consisted of formic
   acid/acetonitrile (%). Separation was carried out with an elution
   gradient. The injection volume was 2 μL, and the flow rate was set
   constant at 500 μL/min. The mass spectrometry conditions were set as
   follows: MRM-ESI scanning mode, atomizing gas pressure (Gas 1), 55 psi;
   heating auxiliary gas pressure (Gas 2), 55 psi; curtain gas, 30 psi;
   ion spray voltage floating (ISVF), 4.5 kV; and ion source temperature,
   600 °C. Mass spectrometry was performed via electrospray ionization in
   positive ion mode, and the corresponding parameters are presented in
   [73]Table S3.

2.5. Method Performance

   The analytical methods were validated following the SANTE guidance
   document (European Commission, 2019) [[74]31] and NY/T 788–2018
   [[75]32], using linearity, the limit of detection (LOD), the limit of
   quantification (LOQ), accuracy, and precision. The LODs and LOQs for
   each metabolite were calculated by the minimum concentration with
   signal-to-noise (S/N) ratios of 3 and 10, respectively.

2.6. Homology Modeling of the Umami Receptor T1R1/T1R3 and Molecular Docking
with Low Molecular Weight Compounds

   Homology modeling of the umami receptor T1R1/T1R3 was conducted via
   Discovery Studio (DS, version 2.1; Neo Trident Technology Ltd.)
   according to previous reports [[76]33]. The metabotropic glutamate
   receptor (PDB ID: 1EWK) was used as a template for homology modeling.
   Moreover, to verify the reliability of the selected homology model, a
   Ramachandran plot was used for evaluation. In addition to applying
   AlphaFold 2.1 for modeling ([77]Figure S2), the calculated Ramachandran
   plot demonstrated that close to 99.3% of the residues were located in
   the allowed regions (89.3% of the residues in the most favored regions
   and 10.0% of the residues in the allowed regions), and only 0.7% of the
   residues were located in the disallowed regions. According to the
   critical evaluation principle of 90%, the conformation of homologous
   models with the metabotropic glutamate receptor as the template was
   reasonable.

   The structure of the model needs to be optimized via energy
   minimization. The energy of the peptides was minimized via minimization
   and the CHARMM force field; that is, the energy minimization calculated
   via molecular mechanics was used to correct the unfavorable noncovalent
   bond contact, expansion bond geometry, and energy minimum state. To
   obtain a model closer to the real structure, the minimization module of
   Discovery Studio was used to minimize energy. In this study, the CHARMM
   force field was selected as the force field of the initial structure
   energy. The smart minimizer algorithm, with an energy minimization of
   5000 steps, was revised and repeatedly optimized according to the
   molecular structure in operation and combined with the evaluation of
   the evaluation software; finally, the most reasonable and realistic
   structure was formed. Semiflexible docking of the receptor–ligand
   complex (CDOCKER) was conducted via Discovery Studio. The compound
   conformation of the receptor and ligand was assessed by the docking
   interaction energy, which was also used as the basis for ranking low
   molecular weight compounds.

2.7. Data Processing, Statistical Analysis, and Metabolic Pathway Analysis

   The biomarkers between different groups were recognized by FC < 0.5
   (downregulation) or FC > 2 (upregulation) with p < 0.05. Multivariate
   data analysis, including principal component analysis (PCA) and
   orthogonal partial least squares discriminant analysis (OPLS-DA), was
   performed via SIMCA-P 14.1 software (Umetrics, Umea, Sweden).
   Hierarchical clustering analysis (HCA) and heatmaps were constructed
   via Multi Experiment Viewer (MeV) software. One-way ANOVA was conducted
   via IBM SPSS Statistics 23 according to Duncan’s test (p < 0.05).
   Metabolomic pathway analysis of biomarkers with significant differences
   was performed via MetaboAnalyst 5.0 ([78]https://www.metaboanalyst.ca)
   (accessed on 26 November 2022) on the basis of the KEGG metabolic
   pathway.

3. Results and Discussion

3.1. Optimization of the Extraction Step

   The extraction solvent was optimized, and the extraction results were
   obtained from three typical extraction solvents, acetonitrile, 0.1%
   formic acid aqueous solution, and methanol/water = 1:1. The results
   revealed that methanol/water = 1:1 resulted in the optimal extraction
   efficiency. [79]Figure 1 shows the effects of three typical extraction
   solvents on the extraction of metabolites.

Figure 1.

   [80]Figure 1
   [81]Open in a new tab

   Effects of three typical extraction solvents on the extraction of
   metabolites (a); the Plackett–Burman design and the concentrations of
   typical metabolites (b).

   The PB factorial design experiment was used to further optimize the
   pretreatment conditions. The chosen low-level (−1) values and
   high-level (1) values of extraction solvent volume (V), vortex time
   (t[1]), ultrasonic time (t[2]), ultrasonic temperature (T), and
   ultrasonic intensity (E) were 2 and 4 mL, 1 and 3 min, 10 and 20 min,
   20 and 40 °C, and 60 and 100 W, respectively. The results of the PB
   design experiment indicated that the extraction solvent volume (V) and
   vortex time (t[1]) had significant effects on the extraction
   efficiency, and 2 mL and 3 min were determined to be the optimum
   extraction solvent volume and vortex time, respectively, through
   single-factor experiments. Finally, the ultrasonic temperature, time
   and power were determined to be 20 °C, 3 min and 60 W, respectively. PB
   design has been used to optimize multiple parameters that affect
   extraction efficiency. Li et al. [[82]34] used aqueous alcohol as the
   extraction solvent to obtain high yields of flavonols and anthocyanins
   from blackcurrant via a PB design. Li et al. [[83]35] investigated the
   factors affecting the high extraction yield of purines via PB design
   for screening.

3.2. Method Optimization and Quality Control

   The total ion chromatograms of the 39 mixed metabolomics solutions are
   shown in [84]Figure S1. The linear regression equation, correlation
   coefficient, linear range, limit of detection, limit of quantitation,
   and retention time of 34 nutrients and metabolites were obtained by
   repeated injection of a matrix-matched standard solution that included
   12 gradients within 0.1–500 μg/L. The results are shown in [85]Table
   S4. The LODs and LOQs of the 34 nutrients ranged from 0.01 to 5 μg/kg
   and 0.05 to 10 μg/kg, respectively.

3.3. Variations in Total Nutrient Metabolism during Storage

   Initially, an analysis was conducted based on changes in odor and color
   at different sampling times. The fresh salmon purchased had a bright
   color and was accompanied by a fresh, sweet fragrance. During the
   experiment, it was observed that the muscle color of the salmon became
   increasingly pale over time. After 4 days of storage, the salmon stored
   at 0 °C still retained a faint pleasant aroma. By the 8th day of
   storage, the salmon stored at 0 °C had developed a slight fishy odor,
   with the pleasant fragrance had disappeared. On the basis of the
   established metabolomics method, the nutritional components of small
   molecules in salmon vacuum packaged at 0 kPa and −90 kPa and stored for
   8 days were analyzed. The nutrients and metabolites present during
   storage can generally be divided into three groups according to the
   results of targeted metabolomics analysis. The changes in 34
   metabolites, amino acids, vitamins, and nucleosides are shown in
   [86]Figure 2.

Figure 2.

   [87]Figure 2
   [88]Open in a new tab

   Changes in the concentrations of amino acids (a), vitamins (b), and
   nucleosides (c) after storage at 0 kPa and −90 kPa for 4 and 8 days.
   Bars with different letters (a, b, c) indicate significant differences
   between groups.

   The decreasing trends were most notable for vitamins and nucleosides.
   Overall, the impact of vacuum pressure on the concentrations of these
   three constituents were not significantly. The total amino acids in the
   group stored under −90 Kpa were observed to be lower in comparison to
   those in the 0 Kpa group. The total vitamin and nucleoside contents
   were 5.13 mg/100 g and 5.30 mg/100 g at 0 kPa and −90 kPa,
   respectively, after 8 days, which decreased by 43.33% and 41.42%,
   respectively. The total amino acid content was 173.27 mg/100 g and
   159.55 mg/100 g under storage conditions of 0 kPa and −90 kPa after 8
   days, which decreased by 62.95% and 65.89%, respectively.

3.4. Changes of Free Amino Acids during Storage

   Salmon contains high amounts of free amino acids, which decrease
   significantly with increasing storage time. From 0 to 4 days, the total
   content decreased from 467.72 mg/100 g to 249.17 mg/100 g and 210.63
   mg/100 g at 0 kPa and −90 kPa, representing decreases of 46.73% and
   54.97%, respectively. From 4 to 8 days, the total content decreased
   from 249.17 mg/100 g to 173.27 mg/100 g at 0 kPa or from 210.63 mg/100
   g to 159.55 mg/100 g at −90 kPa, corresponding to decreases of 30.46%
   and 24.25%, respectively. The relevant mechanisms might include the
   degeneration of fish myofibril protein during storage, such as changes
   in spatial structure, solubility, Ca^2+-ATPase activity, sulfhydryl and
   disulfide bond content, and surface hydrophobicity, which lead to
   changes in amino acid content [[89]36,[90]37]. The activities of
   microorganisms and enzymes during storage [[91]38].

   However, the concentrations of some amino acids, including glycine,
   serine, GABA, taurine, and isoleucine, tended to differ; these
   compounds initially decreased on Day 4 but then increased on Day 8.
   This change probably resulted from their release from protein
   degradation by proteases [[92]39]. As described above, the storage of
   salmon causes changes in amino acids, and these changes differ
   depending on the degree of vacuum and storage time. After 4 days of
   storage at 0 °C, the amino acid content in salmon decreased rapidly,
   possibly because of the degeneration of fish myofibril protein during
   storage, such as changes in spatial structure, solubility, Ca^2+-ATPase
   activity, sulfhydryl and disulfide bond content, and surface
   hydrophobicity, which led to changes in amino acid content.

   To evaluate the taste contribution of the individual taste compounds,
   19 flavorful amino acids quantitatively determined in the salmon were
   grouped into five classes according to their taste qualities ([93]Table
   1). On the one hand, the umami-like taste sensation includes glutamic
   and aspartic acid. The glutamic acid exhibited an umami taste,
   differing in concentration from 7.90 mg/100 g to 4.56 mg/100 g after
   storage for 8 days at 0 °C/0 kPa. The aspartic acid concentration
   decreased from 3.08 mg/100 g to 0.82 mg/100 g after storage for 8 days
   at 0 °C/0 kPa. The total umami amino acid content decreased by 51.1%
   after storage for 8 days at 0 °C/0 kPa. On the other hand, alanine
   decreased significantly among the six kinds of sweet amino acids and
   was undetectable after 4 days of storage. Among the bitter-tasting
   compounds, the concentration of histidine was the highest, decreasing
   from 93.28 mg/100 g to 40.14 mg/100 g after storage for 8 days at 0
   °C/0 kPa. It is noteworthy that bitter-tasting amino acids such as
   phenylalanine, tyrosine, isoleucine, and histidine exhibit significant
   variation under different storage conditions. Particularly, the content
   of histidine notably decreased to 35.07 mg/100 g after 8 days of
   storage at −90 kPa, which could substantially affect the bitterness
   level of the salmon. Furthermore, glutamine and aspartic acid, known as
   umami-tasting amino acids, also show a decline in their content after 8
   days of storage at −90 kPa, potentially reducing the umami flavor of
   the salmon. Compared with the total content of sweet-tasting compounds
   and total content of bitter-tasting compounds, the total amount of
   sweet amino acids in fresh salmon is greater than that of bitter amino
   acids, but with increasing storage time, the content of bitter amino
   acids is greater than that of sweet amino acids, which may affect the
   taste of salmon. The compounds impart sour and/or mouth-drying
   sensations to the oral cavity, containing GABA and pyroglutamic acid
   [[94]40]. Our results revealed that GABA tended to increase with
   increasing storage time. The increasing trend of GABA may be due to the
   conversion of glutamate to GABA by glutamate decarboxylase. The content
   of pyroglutamic acid slightly decreased with increasing storage time.
   Overall, the total content of sour-tasting and mouth-drying compounds
   was reduced by 34.1%. The total content of salt-tasting compounds
   decreased by 24.2% during the storage process.

Table 1.

   Analysis and comparison of 19 kinds of flavor amino acids in salmon.
   Tastant Compound Content (mg/100 g)
   Control 0 kPa-4d 0 kPa-8d −90 kPa-4d −90 kPa-8d
   Umami-tasting Glutamine 7.90 ± 0.14 4.86 ± 0.41 4.56 ± 0.11 3.78 ± 0.43
   4.45 ± 0.5
   Aspartic acid 3.08 ± 0.48 1.35 ± 0.18 0.82 ± 0.13 1.74 ± 0.11 2.13 ±
   0.17
   Umami-tasting compounds total content 10.99 ± 0.62 6.21 ± 0.59 5.37 ±
   0.24 5.52 ± 0.54 6.58 ± 0.67
   Sweet-tasting Glycine 19.45 ± 1.25 27.68 ± 3.57 26.35 ± 1.34 22.75 ±
   1.29 18.85 ± 1.49
   Serine 5.52 ± 0.31 21.35 ± 2.14 11.22 ± 2 5.25 ± 0.36 4.04 ± 0.11
   Threonine 13.03 ± 0.81 14.24 ± 0.64 12.45 ± 0.48 13.14 ± 1.04 12.44 ±
   0.16
   Proline 4.37 ± 0.22 5.28 ± 0.28 4.51 ± 0.14 5.67 ± 0.33 5.58 ± 0.48
   Lysine 3.38 ± 0.25 5.25 ± 0.57 4.14 ± 0.22 2.06 ± 0.13 3.06 ± 0.66
   Alanine 110.44 ± 7.02 NG NG NG NG
   Sweet-tasting compounds total content 156.19 ± 9.86 73.79 ± 7.2 58.67 ±
   4.18 48.88 ± 3.13 43.97 ± 2.91
   Bitter-tasting Arginine 2.31 ± 0.34 1.66 ± 0.11 1.26 ± 0.14 0.64 ± 0.05
   0.61 ± 0.11
   Valine 5.75 ± 0.14 3.7 ± 0.66 6.04 ± 0.55 9.15 ± 0.43 9.06 ± 0.36
   Tryptophan 1.31 ± 0.04 1.75 ± 0.2 1.62 ± 0.04 1.56 ± 0.09 1.62 ± 0.11
   Leucine 4.26 ± 0.13 3.58 ± 0.22 3.27 ± 0.23 3.94 ± 0.05 3.20 ± 0.14
   Phenylalanine 6.31 ± 0.18 7.54 ± 0.39 6.39 ± 0.27 7.13 ± 0.18 6.55 ±
   0.32
   Tyrosine 9.05 ± 0.37 3.23 ± 0.98 1.13 ± 0.15 0.81 ± 0.13 0.96 ± 0.02
   Isoleucine 2.53 ± 0.13 4.70 ± 0.04 4.27 ± 0.18 4.60 ± 0.31 3.53 ± 0.66
   Histidine 93.28 ± 8.13 90.52 ± 5.23 40.14 ± 2.03 75.06 ± 1.63 35.07 ±
   1.92
   Bitter-tasting compounds total content 124.8 ± 9.46 116.67 ± 7.83
   64.11 ± 3.59 102.88 ± 2.87 60.61 ± 3.66
   Salty-tasting Methionine 2.64 ± 0.14 2.86 ± 0.34 2.00 ± 0.07 2.96 ±
   0.28 2.21 ± 0.11
   Salty-tasting compounds total content 2.64 ± 0.14 2.86 ± 0.34 2.00 ±
   0.07 2.96 ± 0.28 2.21 ± 0.11
   Sour-tasting and mouth-drying GABA 0.27 ± 0.001 0.69 ± 0.1 0.77 ± 0.03
   1.37 ± 0.13 1.99 ± 0.19
   Pyroglutamic acid 12.19 ± 0.37 9.68 ± 0.32 7.44 ± 1.11 8.61 ± 0.39 8.82
   ± 0.5
   Sour-tasting and mouth-drying compounds total content 12.46 ± 0.37
   10.37 ± 0.41 8.21 ± 1.13 9.98 ± 0.51 10.81 ± 0.69
   [95]Open in a new tab

   NG: not detected.

3.5. Changes in Vitamins and Nucleosides during Storage

   The total vitamin content in salmon clearly decreased during storage,
   ranging from 9.06 mg/100 g to 5.13 mg/100 g during storage at 0 kPa.
   High vitamin B1, folic acid, and vitamin B2 contents were found
   ([96]Figure 3a). The vitamin B2 content varied from 3.93 mg/100 g to
   0.45 mg/100 g after 8 days at −90 kPa. However, the vitamin B2 content
   was close to zero (0.033 mg/100 g) after 8 days at 0 kPa. A high vacuum
   degree is relatively beneficial to the maintenance of vitamin B2.
   Interestingly, the contents of vitamin B3 and folic acid increased at 8
   days, whereas those of vitamin B3 and folic acid decreased first and
   then increased with increasing storage time, possibly because
   tryptophan can be converted into niacin [[97]41]; folic acid is a
   combination of p-aminobenzoic acid, tretin, and glutamic acid [[98]42];
   and the consumption of amino acids (such as tryptophan and glutamic
   acid) during storage may promote the production of folic acid. Vitamin
   B3 and folic acid are also involved in the synthesis of purines and
   pyrimidines and the mutual conversion of amino acids.

Figure 3.

   [99]Figure 3
   [100]Open in a new tab

   Changes in vitamin (a) and nucleoside (b) contents at different storage
   times and in the amount of vacuum used.

   Nucleosides are vital for the regulation of substance metabolism and
   biological functions in cells and are the second messengers of life
   information transmission. Purine metabolism is closely related to the
   metabolic response to temperature stress [[101]43]. The concentrations
   of 2′-deoxyuridine, cytidine, 1-methyladenosine and guanosine hydrate
   decreased below the limits of detection after 4 days of storage.
   2-Hydroxyadenine was also undetectable after 8 days of storage. There
   are more intense changes in nucleoside content within the first 4 days
   ([102]Figure 3b). This could indicate that the metabolic activities of
   the salmon’s cells are more active during this period.

3.6. Unique Biomarkers for Salmon during Storage

   In the present study, multivariate data analysis was performed to
   profile and visualize the variation in salmon under different storage
   conditions, and the correlations between the metabolites and storage
   conditions were analyzed. First, five groups of samples are shown with
   heatmaps generated via hierarchical Pearson clustering on the basis of
   the average contents of 34 components ([103]Figure 4a). The heatmaps
   show the contents of 34 nutritional compounds in salmon stored for
   different durations and under different degrees of vacuum packing. The
   cluster analysis on the basis of mean values revealed two major
   clusters on the basis of storing salmon for 4 days and 8 days. This
   means that the changes in 34 small molecule nutrients can distinguish
   between salmon stored for different periods of time under vacuum
   packaging conditions.

Figure 4.

   [104]Figure 4
   [105]Open in a new tab

   Heatmaps generated by hierarchical Pearson clustering for 34 nutrients
   in two different packing vacuums and storage times (a). S-plot of 34
   nutrients obtained via PCA after storage for 4 days (b) and 8 days (c).
   KEGG pathway analysis showing changes in metabolism after storage for 4
   days (d) and 8 days (e). Note: the size and color of the circle
   represent the pathway effect and pathway significance, respectively.
   The larger and more deeply colored circles in the upper right corner
   represent the pathways that had a strong effect with greater
   significance on salmon metabolism. (For interpretation of the
   references to color in this figure legend, the reader is referred to