Effect of salt mixture on flavor of reduced‐sodium restructured bacon with ultrasound treatment

Abstract Flavor loss from sodium reduction is a large challenge faced in the meat industry. The effects of salt mixture (KCl: CaCl2 = 1:1, w/w) content (0%–1.0%) on flavor of reduced‐sodium (1.5% NaCl) restructured bacon with ultrasound treatment (UT, 600 w for 30 min) were investigated. The results showed that 0.5% salt mixture (0.25% KCl and 0.25% CaCl2) could significantly (p < .05) enhance the lipid oxidation, the protein oxidation, and the formation of free amino acids of reduced‐sodium UT‐restructured bacon and could also markedly (p < .05) improve its flavor and the overall quality of sensory evaluation via promoting the release of five kinds of volatile phenolic compounds (o‐cresol, m‐cresol, 2‐methoxy‐phenol, 2‐methoxy‐4‐methylphenol, and 2‐methoxy‐5‐methylphenol) and the formation of five kinds of volatile aldehyde compounds (hexanal, nonanal, decanal, furfural, and 5‐methyl furfural). It is interesting to understand the mechanism for the effect of salt mixture on flavor and to efficiently develop a technique for improving the flavor of reduced‐sodium products in the meat industry.

become an important part of meat products in the Chinese market.
Therefore, there is a lot of room to reduce sodium salt in restructured bacon in order to ensure the healthy diet of consumers.
The quality, such as flavor, color, and appearance, is a dominant factor to determine the meat character and purchasing decision of the consumers. However, directly decreasing the salt content in meat products can lead to flavor loss, texture deterioration, shelf life shortening, more rapid growth in natural flora (Desmond, 2006;Jeremiah, Ball, Uttaro, & Gibson, 1996), and reduced consumer purchasing intentions (Delgado-Pando et al., 2018). Salt reduction also affects the development of those volatile compounds (Corral, Salvador, & Flores, 2013). Accordingly, determining how to enhance the flavor of reduced-sodium restructured bacon is a significant challenge.
Ultrasound treatment can significantly increase the relative contents of volatile flavor compounds, especially aldehydes, ketones, and alcohols, and then produce a better flavor in spiced beef (Zou et al., 2018), especially improve the taste and purchase intention of ham with 0.75% of salt (Barretto, Pollonio, Telis-Romero, & da Silva Barretto, 2018). Moreover, the addition of KCl and CaCl 2 into low-sodium meat products can also improve the flavor and overall acceptance (de Almeida, Villanueva, da Silva Pinto, Saldaña, & Contreras-Castillo, 2016), increase the degree of lipid oxidation (dos Santos, Campagnol, Fagundes, Wagner, & Pollonio, 2017), and promote the content of amino acids (dos Santos et al., 2015).
And the best sensory acceptance can be yielded in mortadella when 50% NaCl was replaced by 25% CaCl 2 and 25% KCl (Horita, Morgano, Celeghini, & Pollonio, 2011), although the replacement of NaCl by 50% of KCl or CaCl 2 could lead to bitter taste of drycured ham (Armenteros, Aristoy, Barat, & Toldrá, 2012) or texture deterioration of salted meat (Vidal, Bernardinelli, Paglarini, Sabadini, & Pollonio, 2019). CaCl 2 has also a greater lipid oxidation capacity when compared to NaCl and KCl (Nachtigall et al., 2019) and is not well accepted. The mixture of KCl and CaCl 2 may be an interesting strategy depending on the meat matrices. To our knowledge, no literature has reported the effects of both KCl and CaCl 2 on oxidation and flavor compounds of reduced-sodium restructured bacon with ultrasound treatment.
The objective of this work was to investigate the effect of the salt mixture (KCl: CaCl 2 = 1:1, w/w) at different content levels (0%-1.0%) on the flavor compounds of reduced-sodium restructured bacon (1.5% NaCl) with ultrasound treatment (UT, 600 w; 30 min).

| Materials
Fresh pork back fat and lean pork of Landrace pig were purchased from the Hejiafu Supermarket (Hefei City, Anhui, China) and stored at 4°C until use (within 4 hr). The pigs were slaughtered in compliance with ethical guidelines for animal care published by the Ministry of Science and Technology of the People's Republic of China in 1988. The composition of the lean pork was 71.8% water, 18.2% crude protein, and 5.6% crude fat. The composition of the back fat was 6.1% water, 2.1% crude protein, and 91.3% crude fat (AOAC, 1996). Food-grade NaCl, KCl, and CaCl 2 were purchased from Yinuo Biotechnology Co., Ltd. Food-grade sodium nitrite, soy protein isolate, sodium tripolyphosphate, and sugar were purchased from Jinshan Pharmaceutical Co., Ltd. Liquid smoke was purchased from Jinan Hualu Food Co., Ltd. Commercial bacon (Smithfield, the smoked wood used in the bacon was the same as the raw material of the liquid smoke) was purchased from Shineway Group. Other reagents and chemicals in the research were of analytical grate or better.

| Preparation of the reduced-sodium restructured bacon
The ratio of fat to muscle (3:7, w/w) of the restructured bacon was determined via investigating the fat ratio of bacon in the Chinese market. The salt mixture contained KCl and CaCl 2 in a ratio of 1:1 (w/w), according to the result of Horita et al. (2011). The pork back fat and lean pork were minced through a meat grinder (MG-1220, Shunde Jinyimei Electrical Appliances Co., Ltd.). The salts, liquid smoke, and the other ingredients were sequentially added to the minced meat as depicted in Table 1 and were thoroughly mixed in a mixer machine (CL-877, Chulaikesi Electrical Equipment Co., Ltd.) at 60 rpm for 5 min. The mixed meat was put into a polyethylene bag and sealed (approximately 200 g) by a vacuum-packing machine (DZ-260, Rayleigh Packaging Machinery Co., Ltd).
The sealed bag of mixed meat and a Ф15 probe were fully immersed in a 3,000-mL beaker filled with ice water and were subjected to UT for 30 min under 600 w at a frequency of 20 kHz (Zou et al., 2018) using a SCIENTZ-ⅡD ultrasound processor (Ningbo Xinzhi Biotechnology Co., Ltd.). The constant distance between the bag and probe head was 20 mm on the basis of the result of Kang et al. (2016).
The restructured bacon was prepared on the basis of the procedures described by Aaslyng, Vestergaard, and Koch (2014) and preliminary tests. The meat batter from cured bags, after UT and being stored at 0-4°C for 12 hr, was cooked at 72°C for 60 min in a stainless-steel mold of bacon, followed by drying at 65°C for 60 min using an electric blast drying box (DGF30/7-2, Nanjing experimental instrument factory), and then cooled. The restructured bacon was vacuum-packed before slicing (2.8 mm thick) and frozen to −18°C.
The bacon was thawed approximately 4 hr at 4°C before the values of thiobarbituric acid reactive substances (TBARS), protein carbonyl content, free fatty acids, and free amino acids were measured.

| Determination of TBARS
Thiobarbituric acid reactive substances values of the restructured bacons were determined as described by Li et al. (2019). Five grams of the minced restructured bacon was mixed with 50 ml of a 7.5% aqueous solution of trichloroacetic acid (TCA) containing 0.01% ethylene diamine tetraacetic acid (EDTA) in a 150-mL conical flask.
These conical flasks were incubated at 50°C for 30 min in a thermostatic water bath oscillator (WE-3, Ouyi Co., Ltd.) at 150 rpm. After filtering through double-layer filter paper, 5.0 ml of filtrated volume was mixed with 5.0 ml of 0.02 M aqueous thiobarbituric acid (TBA) in a 25-mL stoppered colorimetric cylinder. These cylinders were incubated at 90°C for 30 min in a thermostat water bath (HH-56, Jintan Guosheng Experimental Instrument Factory) and were then cooled for 5 min in running water. Absorbance was measured at 532 nm using a spectrophotometer (UV-6000PC, Yuanxi Instrument Co., Ltd.) against a blank containing 5 ml of 7.5% TCA and 5 ml of TBA reagent. The results, expressed as milligrams malondialdehyde (MDA)/ kg meat, were calculated from the standard curve prepared by TEP (1,1,3,3-tetraethyoxypropane). Each treatment was performed for three times, and the determinations were replicated three times.

| Determination of protein carbonyl content
The protein carbonyl content was determined in accordance with the method described by Berardo et al. (2016). Carbonyl was reacted with 2,4-dinitrophenyl hydrazine (DNPH) to produce the hydrazones of protein, which were detected by measuring the absorbance at 370 nm. The protein carbonyl content was calculated as nmol carbonyl/mg protein based on an absorption coefficient of 22,000 M −1 cm −1 . Each treatment was performed for three times, and the determinations were replicated three times.

| Free fatty acid analysis
The free fatty acid (FFA) composition of the restructured bacon was analyzed as described by Huang, Li, Huang, Li, and Sun (2014). FFAs were converted to fatty acid methyl esters (FAMEs) and analyzed using a gas chromatograph (Agilent 7,693, Agilent Technologies Inc. US) equipped with a capillary column (DB-WAX, 30 m length, 0.25 mm diameter, 0.25 μm film thickness; Agilent Technologies Inc. US). The injector and detector temperatures were both set at 250°C. The oven temperature program was 60°C for 1 min and then increased at 50°C/min up to 200°C, where it was maintained for 1 min, then increased to 250°C at 3°C/min and held at 250°C for 3 min. Helium was used as the carrier gas at a flow rate of 3 ml/min, and the split ratio was 10:1. Identification of the FAMEs was carried out by comparing their retention times with those obtained for standards. The results are expressed as the FFA percentages of the total FFAs. Each treatment was performed for three times, and the measurements were performed in triplicate.

| Free amino acid analysis
The free amino acid (FAA) content in the restructured bacon was analyzed utilizing an Amino Acid Automatic Analyzer (L-8900, Hitachi) and was determined according to the method described by Zou et al. (2018). One gram of the minced restructured bacon was placed in a centrifuge tube and brought to 10 ml with 5-sulfosalicylic acid (40 g/L), after which it was shaken and mixed by vortex mixer (SA8, Mercy Scientific Instruments Co., Ltd.). After incubating it in a dark place for 1 hr, the tube was centrifuged at 10,000 g at 4°C for 30 min in a high-speed freezing centrifuge (CT14RD, Tianmei Biochemical Instrument Engineering Limited Company). Finally, the supernatant was filtered through a 0.22-μm membrane, and the resulting filtrate was analyzed by the automatic analyzer. Each treatment was performed for three times, and the measurements were performed in triplicate.

| Analysis of volatile compounds
Volatile compounds of the restructured bacon were analyzed as described by Wen et al. (2019). The restructured bacon was baked in an oven at 210°C for 10 min and immediately minced, and five grams of the minced restructured bacon was packed into 20-ml headspace sample vials (Supelco). The volatile compounds in the restructured bacon were collected in a solid phase microextraction fiber (50/30 μm DVB/CAR/PDMS, Supelco) and exposed to TA B L E 1 Content of NaCl and salt mixture in reduced-sodium restructured bacon under ultrasound treatments (U0, U1, U2, and U3) a Note: The content of other additives in U0-U3 were the same and were 0.5% sugar, 0.01% sodium nitrite, 2% soy protein isolate, 0.4% sodium tripolyphosphate, and 0.5% liquid smoke, 11.1% water, respectively.

Treatments of four
Abbreviation: UT, ultrasound treatment. a Deviation of weight of all salts was not in excess of ± 0.5%. the headspace sample vials for absorption at 60°C for 40 min.
After absorption, the fiber was immediately inserted into the GC syringe port and was then thermally desorbed at 250°C for 3 min.
Volatile compounds were analyzed utilizing a GC-MS system (QP2010 Shimadzu) with a DB-5 ms capillary column (60 m length, 0.25 mm diameter, 0.25 μm film thickness; Agilent Technologies Inc.). Helium was used as the carrier gas at a flow rate of 1 ml/ min, and the split ratio was splitless. After desorbing, the oven temperature was held at 40°C for 2 min, followed by an increase at 5°C/min to 60°C, with a subsequent increase to 100°C at 10°C/ min and a final increase to 240°C at 18°C/min, where it was maintained for 6 min. The mass-selective detector was applied 250°C as the ion source temperature. The ionization potential of the detector was 70 eV, and the mass spectrometer scan range was 33-350 m/z. The identification of volatile compounds was achieved by comparing their mass spectra with those in the NIST 11 library.
The area of each peak was integrated and reported as an indicator of volatile compounds from the restructured bacon. Each treatment was performed for three times, and the measurements were performed in triplicate.

| Determination of the key volatile compounds
The relative odor activity value (ROAV) was calculated to determine the key volatile compounds of the restructured bacon as described by Zhuang et al. (2016)

| Sensory analysis
Sensory evaluation of the restructured bacon was implemented by a nine-member trained team according to the method described by Soladoye et al. (2017). The restructured bacon was baked in an oven at 210°C for 10 min and was then served on a white plate with twodigit codes in a randomized order. A nine-point scale was utilized, and the intensities of all sensations (aroma, taste, color, texture and aftertaste) were expressed from 1 (none of the sensation) to 9 (maximum of the sensation). A commercial bacon was used as an evaluation reference (7 points). And the overall quality was calculated using a weighted rubric (30% aroma, 30% taste, 10% color, 10% texture, and 20% aftertaste). During sensory evaluation, water at room temperature was provided to panelists for palate cleansing and to remove residual flavors between each sample.

| Statistical analysis
Statistical analysis of the data was performed using IBN SPSS Statistics software (SPSS Inc.25) and was expressed as the mean ± standard deviation (SD). A one-way analysis of variance (ANOVA) with Duncan's multiple range tests was used to determine the significance of the treatment effects (p < .05).

| Lipid oxidation and protein oxidation
Thiobarbituric acid reactive substances is a major index to evaluate the extent of lipid oxidation, which is primarily used to determine the secondary products of lipid oxidation (Wen et al., 2019). As shown in Figure (Rhee, Smith, & Terrell, 1983). The TBARS contents usually increase with increasing ionic strength of the processed meats in a certain range (Hernández, Park, & Rhee, 2002). Additionally, the UT could generate free hydroxyl radical by cavitation and promoted the effective diffusivity of salt mixture and NaCl (Kang et al., 2016;Zou et al., 2018) and further promoted the oxidation of the restructured bacon by the salt mixture. The increasing trend of TBARS content from U0 to U2 could be attributed to the enhanced ionic strength (Table S3) that was induced by the addition of salt mixture and the diffusion of the salt mixture promoted by UT in the restructured bacon. However, high ionic strength can decrease the solubility of oxygen, exhibiting an inhibitory effect on lipid oxidation (Sharedeh, Gatellier, Astruc, & Daudin, 2015), thus resulting in the decreased TBARS value (p < .05) of the U3 group compared with the U2 group ( Figure 1a). This result is similar to that described in some reports (Nachtigall et al., 2019;dos Santos et al., 2017). In addition, it should be noted that the lipid oxidation could not be the primary cause of the change in the bacon quality because of its low TBARS values.
Formation of carbonyl compounds is a major indicator of protein oxidation in meat products (Estévez, 2011). Carbonyl groups can be formed in proteins by the direct oxidation of the side chains from amino acids (threonine, lysine, proline, and arginine), oxidative cleavage of the peptide backbone via the α-amidation pathway or via oxidation of glutamyl side chains; covalent binding to nonprotein compounds such as MDA from lipid oxidation could also generate carbonyl (Estévez, 2011). The effects of salt mixture on the carbonyl content of the restructured bacon are illustrated in Figure 1b.
The carbonyl contents in the treatment groups (U1, U2, and U3) also increased significantly compared with the control group (U0) (p < .05), but the difference between 0.5% (0.25%KCl + 0.25%CaCl 2 ) and 1.0% (0.5%KCl + 0.5%CaCl 2 ) of salt mixture was insignificant and induce a destabilized structure of the protein, resulting in an enhanced exposure to pro-oxidants and, thus, promoting carbonyl formation (Liu, Xiong, & Chen, 2011). The free radicals produced by the UT can be contributed to the increased protein carbonyl contents in beef (Kang et al., 2016 Wen et al., 2019). It has been implied that the accelerated protein oxidation in the restructured bacon formulated with salt mixtures could also be related to the enhanced lipid oxidation in these groups ( Figure 1a). The present study showed that the salt mixture could promote the protein oxidation of the restructured bacon, but the oxidation degree of the bacon could reach saturation with the addition of 1.0% salt mixture. It implied that the protein oxidation contribution to the quality of the bacon was greater.

| Free fatty acids
Free fatty acid formation is due to lipolysis; the composition of FFAs is one of the most important components that can influence the nutritional value and sensory properties, especially flavor, of meat products (Wood et al., 2008). As shown in Table 2 It could be seen that the percentages of SFA and MUFA increased significantly (p < .05), while the percentage of PUFA decreased significantly (p < .05) in the all treatment groups (U1, U2, and U3) compared with the control (U0). With respect to the PUFA/SFA ratios, the nutritional recommendations are that this ratio should be approximately 0.45 (Wood et al., 2008). In this study, the ratios of PUFA/SFA approached the values recommended for the human diet ( Table 2). The PUFA/SFA ratios were significantly decreased in the U1, U2, and U3 groups compared with the U0 group (p < .05).
The composition of FFA is directly affected by lipid oxidation, and PUFA are more subject to oxidation than MUFA and SFA (Coutron-Gambotti & Gandemer, 1999). According to present results of the TBARS values, it was speculated that salt mixture accelerated the oxidation of PUFA in the UT condition, thus inducing the decreased PUFA/SFA ratio. Moreover, the FFAs predominately found in the restructured bacon were C16:0 (palmitic), C18:0 (stearic), C18:1 (oleic), and C18:2 (n-6) (linoleic), as found in dry-cured loin (

| Free amino acids
Free fatty acids, the final products of proteolysis, have been considered to be of significance not only for their contribution to specific taste but also for their participation in the generation of volatile compounds that provide the flavor in meat products (Toldrá & Flores, 1998). As presented in Table 3, the major FAA in each restructured bacon group was His, followed by Ala and Thr.

| Volatile flavor compounds
As shown in Note: Different superscript letters ( a-d ) in the same row indicate statistically significant differences at p < .05 (n = 3). AU: area units resulting of counting the total ion chromatogram (TIC) for each compound. -: not queried; n.d.: not detected.
Aldehydes were the second most abundant group of compounds in the restructured bacon (Table S1) Alcohols mainly arise from the reaction products of lipid oxidation (Petričević et al., 2018). Similar to hydrocarbons, alcohols are generally not thought of as important to contributing flavor to meat products because of their high odor thresholds (Domínguez et al., 2014). 2-Furanmethanol was found in U0, while it was not present in the U1, U2, and U3. Additionally, levels of linalool were significantly decreased with the addition of salt mixture (p < .05).
Heterocycle compounds are formed by the Maillard reaction and provide important contributions to sweet, caramel, and barbecue-like notes, leading to the improved flavor of meat products (Romero, Raghavan, & Ho, 2013;Zhang et al., 2019). Of the heterocycle compounds, furans are usually generated through the heating process. The levels of most of the furans (such as 2-acetylfuran and 2-propionylfuran) were higher in the treated groups (U2 and U3) compared with the control group (U0). The heterocycle and carbonyl contents showed similar variation tendencies, and it could be considered that the change in heterocycle content was due to the promotion of protein oxidation by the salt mixture.

| Relative odor activity value
Due to the variety of volatile flavor compounds and their different threshold values, the content of each compound cannot reflect the true contribution to the whole aroma profile. Therefore, ROAV was used to evaluate the contribution of volatile flavor compounds in the whole aroma profile. The compounds with ROAV ≥ 1 were considered as the key volatile flavor compounds in the restructured bacon.
The higher ROAV was, the higher the degree of compound contribution was to the flavor. The compounds with 0.1 ≤ ROAV < 1 were considered to be potential flavor compounds (Zhuang et al., 2016).
As shown in Table 5, the types of key volatile compounds in all restructured bacon groups were the same, including four phenols (ocresol, m-cresol, 2-methoxy-phenol, 2-methoxy-4-methylphenol) and three aldehydes (hexanal, nonanal, and decanal). The addition of salt mixture could significantly increase the release of the four phenols, especially 2-methoxy-phenol. The relative contents of 2-methoxy-phenol were the highest of the total volatile compounds (Table 4). In agreement with the present study, 2-methoxy-phenol had the highest content in bacon and was essential for the sensory characteristics of smoked meat products (Saldaña et al., 2019). The

| Sensory analysis
As shown in Table 6, for the attribute aroma, the U2 group showed a significantly higher score than the other groups (U0 and U1) (p < .05).
This was probably because the 0.5% salt mixture promoted lipid and protein oxidation and increased the contents of the key flavor compounds in the restructured bacon, generating better aroma in the final product. This finding was similar to the results of low-salt salami containing 0.25% KCl and 0.25% CaCl 2 (de Almeida et al., 2016). The taste of the restructured bacon exhibited a similar changing trend to the aroma when salt mixture was added. The increase in the taste score in the U2 group could be attributed to the increased generation of umami amino acids (Asp, Glu, and Gly) (Dashdorj, Amna, & Hwang, 2015) with the addition of the 0.5% salt mixture (Table 3).
The texture of the U3 restructured bacon had a significantly lower score than the other groups (U0, U1, and U2) (p < .05

| Principal component analysis (PCA)
Principal component analysis was used to explore the impacts of volatile flavor compounds of the restructured bacon with different treatments and the relationships among the key volatile flavor compounds, lipid oxidation, protein oxidation, FFA, FAA, and sensory evaluation. As shown in Figure 2a, the first principal com- Note: Different superscript letters ( a-b ) in the same row indicate statistically significant differences at p < .05 (n = 9).

TA B L E 6
The sensory scores of the reduced-sodium restructured bacon under ultrasound treatment with different salt mixture content (U0, U1, U2, and U3)* addition of salt mixture had a major influence on the volatile flavor compounds of the restructured bacon, but the effect was weaker between groups U2 and U3. This finding could be related to the fact that there were no significant differences between the contents of total volatile flavor compounds in the U2 and U3 groups

| Comprehensive discussion
As indicated above, the salt mixture does not affect the types of key volatile flavor compounds in the restructured bacon (Table 5).
The mixture of 0.25% KCl and 0.25% CaCl 2 could significantly promote the release of key phenol volatile flavor compounds and the formation of key aldehyde volatile flavor compounds (p < .05) ( Table 4, Table S2). Therefore, the aroma and overall quality were significantly enhanced in the U2 group (p < .05) ( Table 6). The key phenol volatile flavor compounds may mainly come from liquid smoke, which produces a smoky flavor and can greatly contribute to bacon flavor due to its low odor threshold (Petričević et al., 2018).
The three key aldehyde volatile flavor compounds belong to linear aldehydes and mainly result from unsaturated fatty acid oxidation (Wen et al., 2019). For example, hexanal is a product of linoleic acid oxidation, and decanal is produced in the oxidation of oleic acid; these aldehydes play an important role in the overall flavor of meat products due to their low odor thresholds (Zhuang et al., 2016). The salt mixture could promote the formation of linear aldehydes from the oxidation of PUFA, which was beneficial for improving the flavor of the restructured bacon (Table 3). The potential volatile flavor compounds (furfural and 5-methyl furfural) are generated by the Maillard reaction (Wen et al., 2019). The salt mixture could significantly improve the formation of these compounds (p < .05), and the change was generally consistent with the results of carbonyl content ( Figure 1b). It was suggested that the increases in two potential volatile aldehydes could be related to the promotion of protein oxidation in the restructured bacon by the salt mixture. Therefore, these results showed that the mixture of 0.25% KCl and 0.25% CaCl 2 could significantly enhance the lipid and protein oxidation of the restructured bacon (p < .05) with UT and could improve flavor of the restructured bacon (p < .05) by promoting the release of major volatile phenolic compounds and the formation of major volatile aldehyde compounds.

| CON CLUS ION
The mixture of 0.25% KCl and 0.25% CaCl 2 could significantly improve the flavor of reduced-sodium restructured bacon with ultrasound treatment (600 w) (p < .05) by promoting the release of five major volatile phenolic compounds (o-cresol, m-cresol, 2-methoxyphenol, 2-methoxy-4-methylphenol, and 2-methoxy-5-methylphenol) and the formation of five major volatile aldehyde compounds (hexanal, nonanal, decanal, furfural, and 5-methyl furfural). The increases in major volatile aldehyde compounds of the UT-restructured bacon were due to the promotion of lipid oxidation (p < .05) by the addition of the salt mixture. These findings implied that the combination of salt mixtures and UT could be used as a promising technique for the flavor improvement of reduced-sodium products in the meat industry.

ACK N OWLED G EM ENTS
The author thanks Guan-hua Yang, Sai Wang and Wei Wang, postgraduates of the Hefei University of Technology, Wan-guang Li and Bin Hu, senior engineers of the Grain & Oil Product Quality