An appropriate amount of shrimp sample in the refrigerator at –80°C was transferred into a grinding tube, adding some protein lysate with 8 M urea + 1% sodium dodecyl sulfate (SDS) (containing protease inhibitor). For surface bacteria, the 0.9% (w/v) sterile saline was mixed with shrimp and shaken for 5 min in a sterile plastic bag, and then the mixed liquids were transferred into a centrifugal tube and centrifuged at 16,000 × g and 4°C for 30 min. The supernatant was discarded, and then some protein lysate with 8 M urea + 1% SDS (containing protease inhibitor) was added. A high-throughput tissue grinder was used to vibrate the sample three times, for 40 s each time. Then, the sample was cracked on ice for 30 min, with vortex mixing 5–10 s for every 5 min. The mixture was centrifuged at 16,000 × g and 4°C for 30 min, the supernatant was collected, and the protein content was measured using the bicinchoninic acid assay (Smith et al., 1985).

The SDS-polyacrylamide gel electrophoresis protocol was modified according to Li et al. (2020c). Shrimp protein samples, which were extracted as mentioned above, were mixed with loading buffer in the ratio of 1:4 and then placed in a water bath for 10 min at 95°C. The concentration of spacer gel was 6%, and the concentration of separating gel was 10%. Each well of the gel was injected with 15 μl of the sample. The running buffer with a pH of 8.3 consisted of 28.8 g glycine, 6.0 g Tris, and 1.0 g SDS. The constant voltage power supply for cataphoresis was 100 V, running for around 2 h. The gel was then put into dyeing liquor for 1.5 h. After dyeing, the gel was rinsed three times and then placed in the detainer until the background became transparent.

A weight of 100 μg protein was measured and triethylammonium bicarbonate buffer (TEAB) was added to make the final concentration of 100 mM. Tris (2-carboxyethyl) phosphine was added to obtain a final concentration of 100 mM and allowed to react at 37°C for 60 min. The IAM (iodoacetamide) was added to make the final concentration of 40 mM, and the solution was held for 40 min at room temperature without light. For each tube, the precooled acetone was added in the ratio of 6:1 (acetone:sample, v/v) and then precipitated at –20°C for 4 h; centrifugation was done at 10,000 × g for 20 min, the precipitate was collected, and 100 μl of TEAB (100 mM) was used to treat the sample until fully dissolved. Trypsin was added according to the ratio enzyme:protein (M/M) = 1:50, and the enzymatic hydrolyses were conducted overnight at 37°C.

After performing the abovementioned procedures, the shrimp samples were processed as follows. The TMT reagent (Thermo Fisher Scientific No. a44522) was taken out at 20°C and restored to room temperature. The reagent was centrifuged, acetonitrile was added, and centrifuge was vortex mixed. A tube of TMT reagent was added for every 100 μg polypeptide and incubated at room temperature for 2 h. Then, hydroxylamine was added, and the samples were kept at room temperature for 30 min. Each group of moderately labeled products was mixed into a tube and drained using a vacuum concentrator.

The polypeptide samples were redissolved with UPLC loading buffer (2% acetonitrile with ammonia adjusted to pH 10) and separated by a reversed-phase C18 column. Column information: ACQUITY UPLC BEH C18 Column 1.7 μm, 2.1 mm × 150 mm (Waters, United States); Chromatographic instrument: Thermo Scientific Vanquish Flex Binary UHPLC system; phase A: 2% acetonitrile (ammonia adjusted to pH 10); phase B: 80% acetonitrile (adjust ammonia to pH 10); UV detection wavelength: 214 nm; flow rate: 200 μl/min; and gradient: 48 min. According to the peak type and time, 20 fractions were collected and combined into 10 fractions. After vacuum centrifugation and concentration, the fractions were dissolved in mass spectrometry loading buffer (2% acetonitrile and 0.1% formic acid) for the second-dimensional analysis.

After performing the abovementioned procedures, the surface bacteria samples were processed according to this section. After trypsin digestion, the peptide segments were dried with a vacuum pump. The digested peptides were redissolved with 0.1% trifluoroacetic acid (TFA). After desalting the peptide segments with hydrophilic and lipophilic balance and a mixed-mode cation exchanger (MCX) solid-phase extraction column, each sample was divided into two parts and dried with a vacuum concentrator. Peptide quantification was conducted using a Thermo Fisher Scientific peptide quantification kit (No. 23275).

According to the quantitative results of peptide segments, the peptide with an equal concentration of 0.25 μg/μl was dissolved in mass spectrometry loading buffer (2% acetonitrile and 0.1% formic acid) for mass spectrometry analysis.

After treating the shrimp samples and surface bacteria samples as stated above, liquid tandem mass spectrometry was performed according to the following parameters:

Data acquisition software: Thermo Xcalibur 4.0 (Thermo, United States); reversed-phase column information: C18 column (75 μm × 25 cm, Thermo, United States); chromatographic instrument: EASY-nLC 1200 (Thermo, United States); mass spectrometer: Q_Exactive HF-X (Thermo, United States); chromatographic separation time: 120 min for TMT proteomics, 90 min for label-free proteomics; phase A: 2% acetonitrile and 0.1% formic acid; phase B: 80% acetonitrile and 0.1% formic acid; and flow rate: 300 nl/min.

The EASY-nLC liquid-phase gradient for TMT proteomics was shown as follows: 0 min, 5%; 65 min, 23%; 81 min, 29%; 90 min, 38%; 92 min, 48%; 93 min, 100%; 120 min, stop. MS scanning range (m/z): 350–1,500; acquisition mode: DDA, top 15 (select the 15 with the strongest signal in the parent ion for secondary fragmentation); primary mass spectrometry resolution: 120,000; AGC target: 3e6; maximum injection time: 50 ms; fragmentation mode: HCD; secondary resolution: 45,000; AGC target: 2e5; maximum injection time: 120 ms; fixed first mass: 110 m/z; minimum AGC target: 1e4; intensity threshold: 8.3e4; and dynamic exclusion time: 30 s.

The EASY-nLC liquid-phase gradient for label-free proteomics was shown as follows: 0 min, 5%; 53 min, 23%; 65 min, 29%; 73 min, 38%; 74 min, 48%; 75 min, 100%; 90 min, stop. MS scanning range (m/z): 350–1,500; acquisition mode: DDA, top 20 (select the 20 with the strongest signal in the parent ion for secondary fragmentation); primary mass spectrometry resolution: 60,000; AGC target: 3e6; maximum injection time: 20 ms; fragmentation mode: HCD; secondary resolution: 15,000; AGC target: 1e5; maximum injection time: 50 ms; fixed first mass: 100 m/z; minimum AGC target: 8e3; intensity threshold: 1.6e5; and dynamic exclusion time: 18 s.

The software version used for library search is Proteome Discoverer™ Software 2.4. When searching the database, submit the raw file to the proteome discoverer server, select the established database, and then search the database.

For TMT proteomics, the parameters were set as follows: protein database: uniprot-taxonomy-6689.unique.fasta; cys alkylation: iodoacetamide; dynamic modification: oxidation (M), acetyl (protein N-terminus), met-loss (protein N-terminus), and met-loss + acetyl (protein N-terminus); static modification: carbamidomethyl (C), TMTpro (K), and TMTpro (N-terminus); enzyme name: trypsin (Full); maximum missed cleavage sites: 2; precursor mass tolerance: 20 ppm; fragment mass tolerance: 0.02 Da; and validation based on q-Value. The resultant filtering parameter is peptide FDR ≤ 0.01. For label-free proteomics, the parameters were set as follows: protein database: merge-uniprot-taxonomy-1590 + 6689 + 62322.unique.fasta; cys alkylation: iodoacetamide; dynamic modification: oxidation (M), acetyl (protein N-terminus), met-loss (protein N-terminus), and met-loss + Acetyl (protein N-terminus); static modification: carbamidomethyl (C); enzyme name: trypsin (full); max. missed cleavage sites: 2; precursor mass tolerance: 10 ppm. The resultant filtering parameter is peptide FDR ≤ 0.01.

All the proteins identified in the proteomics study of L. vannamei were classified by GO, KEGG, and COG pathways, and the results are presented in Figure 4. In the GO pathways, 1,754 proteins were identified, and the detailed GO classifications are shown in the pie charts in Figure 4A. The different colors in each pie chart represented different GO terms, the area represents the relative proportion of proteins, and the number corresponding to the color represents the number annotated to the GO term in the identified proteins. The three pie charts from left to right represented the three branches of GO, namely, biological process, cellular component, and molecular function (MF). In total, 38 kinds of proteins are classified as relating to the myosin complex. Myosin plays a vital role in cell movement and intracellular material transmission and plays a vital role in the process of muscle contraction and cell division (Scarff et al., 2020). The texture changes of L. vannamei could be highly related to the changes in those proteins. Except for texture-related differential proteins, there are many other differential proteins that participated in the protein activity of L. vannamei. Binding, catalytic activity, and cellular anatomical entity were the top 3 categories containing the largest number of differential proteins caused by the absence of the luxS gene.

KEGG is a practical database resource for understanding advanced functions and biological systems (such as cells, organisms, and ecosystems), providing information at the molecular level, especially genome sequencing and other high-throughput experimental technologies generated from sizable molecular data sets. In the histogram of KEGG (Figure 4B), 119 proteins were identified to relate to infectious diseases caused by the bacteria. During the storage of L. vannamei, the infection of diverse microorganisms and the inoculated S. baltica could affect those proteins.

COG is a database for the homologous classification of gene products. It is obtained by comparing many protein sequences of various organisms. As shown in Figure 4C, 138 kinds of proteins are identified to be related to the cytoskeleton. Cytoskeleton refers to the protein fiber network structure in eukaryotic cells. It plays a considerate role in maintaining cell morphology, bearing external forces, maintaining the order of internal cell structure, and participating in many important life activities. The changes in cytoskeleton-related proteins could affect the texture of L. vannamei during storage.

The distribution of differential proteins in the group SS-128 and the control group on L. vannamei are presented in Figures 5A,B. In the volcano (Figure 5A), the abscissa represents the change value of the protein expression between samples. The ordinate was the protein expression change difference (p-value). The p-value was negatively correlated to the expression difference. The values of the abscissa and ordinate are logarithmized. Each point in the figure represents a specific protein. The point on the left is the protein with differentially downregulated expression, and the point on the right is the protein with differentially upregulated expression. In the MF of GO annotations analysis, nine kinds of differential proteins related to structural molecular activity were raised, whereas four kinds of differential proteins in this classification were reduced. Except for those proteins, some proteins classified into other classes could also have texture-maintaining functions.

For example, the degree of biological adhesion could be related to the texture (Behera et al., 2020). The top 10 differential proteins associated with the texture of L. vannamei are listed in Table 1A. All the quantities of those 10 proteins of L. vannamei were raised significantly more in the group SS-128 than in the control group, corresponding to the better texture data. The luxS affected the AI-2/LuxS QC system of SS-128 during the storage (Kaur et al., 2018). The SS-128 surpassed S. baltica in the growing competition, with the quantitative growth of SS-128, the AI-2 signals secreted more. By sensing the existence of AI-2 signals, the quantitative growth and the metabolic process of S. baltica could be inhibited. Although the growth and metabolism of SS-128 might use L. vannamei as an energy source, its damage to L. vannamei was far less than that from S. baltica, which is a recognized spoilage microorganism for aquatic products (Feng et al., 2021). The AI-2/LuxS QC system of SS-128 retarded the damage of texture proteins from microorganisms, especially from S. baltica. In contrast, most of the vital texture-related proteins in the control group were decomposed, resulting in the collapse of muscle structure, represented as the texture deterioration of L. vannamei.

Although the AI-2/LuxS QC system of SS-128 introduced mainly by luxS played a critical role in the texture maintenance of L. vannamei, the protection mechanisms from SS-128 could be multiplex. As shown in Figures 5C,D, numerous kinds of differential proteins were identified in L. vannamei between the luxS-mutant group and the control group, indicating that the protecting effect of the luxS-mutant group could be significant. These results support the phenomenon in texture analysis. Table 1B shows that in the top 10 texture-related proteins, nine kinds were raised in the luxS-mutant group, and eight kinds of those proteins were also identified in the differential proteins identified in SS-128 vs. CG, whereas the change folds were minor. The mutant of luxS weakened the texture-protecting effect of SS-128 by inhibiting the AI-2/LuxS QC system. However, the acidic environment formed by L. plantarum could effectively inhibit the growth of spoilage bacteria (Ruiz-Moyano et al., 2019). Since it has been reported that the optimal pH for S. baltica was under meta-alkalescence (Kim et al., 2017), the organic acids produced by luxS-mutant SS-128, including acetic acid and lactic acid could also contribute to the inhibition of spoilage microorganisms. Therefore, under an acidic environment formed by SS-128, the growth of S. baltica could be affected. Besides, the bacteriocin produced by luxS-mutant SS-128 was unfavorable for spoilage microorganisms (Bagde and Nadanathangam, 2019), because the production of some bacteriocin was shown to be unrelated to the absence of luxS (Wang et al., 2021).

When group SS-128 was compared to the luxS-mutant group, 347 types of proteins were identified as differentially expressed proteins (Figures 5E,F), illustrating that the luxS played a crucial role during the storage of L. vannamei. Myosin heavy chain is the basic unit of myosin and plays a vital role in ensuring the daily activity of muscle cells. Its function is to lengthen muscle fibers to produce movement and strength (Luo et al., 2019). Myosin heavy chain showed a significant increase in fold-change (Table 1C), indicating that the luxS in SS-128 had a substantial effect on the protection of myosin heavy chain. Beta-actin was also a structure-related protein that was protected in the group SS-128. Actin is a globular multifunctional protein family that forms microfilaments in the cytoskeleton and filaments in muscle fibers. Beta-actin coexists in most cell types as a component of the cytoskeleton and a mediator of internal cell movement, contributing to the texture performance of muscle (Wiame et al., 2018).

The endogenous enzymes of L. vannamei participate in every aspect of changes in the internal environment, and they have an important role in the regulation of spoilage. The top 10 differential endogenous enzymes in Litopenaeus vannmei of SS-128 vs. ΔluxS are shown in Table 2. The content of three types of endogenous enzymes increased with the presence of the luxS gene, while most of the endogenous enzymes exhibited downregulation trends. The top 2 upregulated endogenous enzymes are kinase, which participates in the phosphorylation to transfer phosphate groups from high-energy donor molecules ATP to specific target molecules, leading to changes in the state of ion channel proteins and channel gates (Steinberg and Carling, 2019). Except for the two kinases, two kinds of ATP synthase were identified as the differential endogenous enzymes, which was evidenced by the difference in energy metabolism between the L. vannmei in SS-128 and the luxS-mutant group.

Of the 7 kinds of downregulated endogenous enzymes, three are hydrolase (alpha-L-fucosidase, epoxide hydrolase, and alpha, alpha-trehalase), two are dehydrogenase (putative retinol dehydrogenase 12-like and 3-hydroxybutyrate dehydrogenase), and one is a hydrolase. The spoilage of muscle food could be highly related to the effect of the endogenous hydrolase, and the downregulation of the hydrolase indicated the slower structure deterioration of L. vannmei in group SS-128. The downregulation of dehydrogenase could be an indicator of low-level redox reactions for Litopenaeus vannmei in group SS-128 (Kuk et al., 2019). The results of differential endogenous enzymes showed that the existence of luxS in SS-128 could maintain the texture and quality of Litopenaeus vannmei by inhibiting the hydrolysis and energy metabolism.

The metabolism of endogenous enzymes and enzymes in spoilage microorganisms are considered the two primary reasons related to the texture change of muscle foods (Dehghani et al., 2018). Thus, the differences in crucial enzymes’ expression between the group SS-128 and the luxS-mutant group could be related to the texture. This section collected the supernatant fluid of L. vannamei at the endpoint of storage in the group SS-128 for label-free proteomics analysis. According to the origination, the supernatant fluid mainly contained four types of proteins: proteins from L. plantarum (wild strain or luxS-mutant strain), proteins from spoilage bacteria (in this study, only analyzed those from S. baltica), endogenous enzymes from L. vannamei, and other soluble proteins from L. vannamei.

Figure 6 provides an excellent illustration of all the identified differential proteins. Figures 6A,B shows that 284 kinds of upregulated proteins and 168 kinds of downregulated proteins were identified. Figure 6C shows the heatmap of differential proteins. In the tree view of sample clustering, the closer branches of the two samples were positively related to the closer expression patterns. The GO annotation summary in Figure 6D presents that 89 kinds of differential proteins participating in the catalytic activity were upregulated, whereas 135 were downregulated. The expression changes of those proteins might influence the deterioration degree of texture-related proteins in L. vannamei. COG classification in Figure 6E indicates that four kinds of proteins regarding coenzyme transportation and metabolism were upregulated and three kinds were downregulated; 15 kinds of differential proteins in signal transduction were identified. KEGG enrichment chord in Figure 6F and Table 3 presents the corresponding relationship between the target protein set and the annotation and enrichment of the KEGG pathway, and the top 50 target proteins with the most annotated pathways and the pathways with the enrichment significant p-value of the pathways containing these target proteins in the enrichment results were selected for display. The pentose phosphate and oxidative phosphorylation pathways in some proteins are upregulated. The histograms of KEGG are displayed in Figures 6G,H.

The overview of differential proteins proved that the enzyme expression caused by luxS had significant differences, which could directly influence the texture of L. vannamei by enzymatic kinetics.

For the differential proteins between SS-128 and its luxS-mutant strain, six were upregulated, and four were downregulated (Figure 7A). Those changes resulted from the luxS and the competitive inhibition with another microorganism, which was mainly S. baltica in this study. IS3 family transposase and elongation factor TU were significantly more highly expressed in SS-128 than in the luxS-mutant strain, whereas the DNA-binding protein II and ornithine carbamoyltransferase were concentrated in the luxS-mutant strain. Furthermore, it was reported that the expression of DNA-binding protein was related to the AI-2/LuxS QS system (Niazy, 2021). The 6-phospho-beta-glucosidase was upregulated in SS-128, which participates in the energy metabolism (Acin-Albiac et al., 2021).

In the supernatant, the differential proteins identified in L. vannamei could be divided into endogenous enzymes and other soluble proteins from shrimp muscle. The enzymes could affect the spoilage degree of L. vannamei, and the differential expressed enzymes in L. vannamei are listed in Figure 7B. Compared to the luxS-mutant strain, some proteases (proteasome assembly chaperone 2-like, zinc protease Mpc1, and 26S proteasome non-ATPase regulatory subunit 6-like) were downregulated in group SS-128, which retarded the degradation rate of proteins and maintained the texture of L. vannamei. However, some energy-consumption-related enzymes were upregulated in the luxS-mutant strain: NADH dehydrogenase is an enzyme located in the inner membrane of mitochondria that catalyzes the transfer of electrons from NADH to coenzyme Q (Ishikawa et al., 2020). This enzyme is an “entry enzyme” for oxidative phosphorylation in mitochondria (Guo et al., 2018). ATP synthase subunit g catalyzes the synthesis of the energy substance ATP in cells. These enzyme activities indicate the higher metabolic rate of L. vannamei in this group. That might lead to a faster spoilage rate, worsening the texture. Putative protein kinase C and casein kinase substrate in neurons protein 1, arginine kinase Lit v 2, putative phosphorylase b kinase regulatory subunit beta isoform X1, and putative protein kinase C and casein kinase substrate in neurons protein 1 were four downregulated kinases in the group SS-128. Protein kinase catalyzes the transfer of ATP phosphate group to substrate protein amino acid residues, called protein phosphorylation, and plays a vital role in cell signal transduction. Protein phosphorylation is the most basic, universal, and essential mechanism regulating and controlling protein activity and function (Batista et al., 2020). After phosphorylation, the protein has an electric charge, which changes the structure and further causes the change of protein activity. The lower content of those kinases in group SS-128 was consistent with the slower spoilage rate in this group.

Although endogenous enzymes in L. vannamei play essential roles in the texture changes during storage, the damage from exogenous spoilage microorganisms could be fiercer. As previously illustrated, the AI-2/LuxS QC system of SS-128 contributed to the texture maintenance through competitive inhibition toward S. baltica, and the proteins affected by luxS in S. baltica are listed in Figures 7C,D.

In the group SS-128, the content of thiamine-phosphate kinase was significantly decreased. Phosphokinase is a class of the enzyme that catalyzes the transfer of phosphate groups from ATP to other compounds. The purpose of phosphorylation is to “activate” or “enable” substrate molecules and increase their energy to participate in the reaction of subsequent negative changes in free energy. Protein kinases act on specific proteins and alter their activities (Leopold et al., 2018). The decreasing content of thiamine-phosphate kinase corresponded to the slower growth rate and metabolic strength of S. baltica (Rodionov et al., 2017). Phosphorylation could activate or inactivate an enzyme. The AI-2/LuxS QC system of SS-128 could affect the critical phosphorylated proteins and phosphorylation sites in S. baltica, control its metabolic pathways, and change the ability to deteriorate the shrimp muscle.

The ATP-dependent protease subunit HslV and zinc metalloprotease FtsH were two major downregulated differential proteins. The decreased expression of protease leads to more negligible protein hydrolysis and destruction in L. vannamei. Shewanella baltica preferentially produces elastase, collagenase, trypsin, and other proteases with vigorous enzyme activity. Those enzymes promoted protein decomposition in muscle, caused mucous membranes, and increased TVB-N release to accelerate the corruption and deterioration of L. vannamei (Odeyemi et al., 2020). Rather than directly affecting the secretion of protease, the AI-2/LuxS QC system of SS-128 was more likely to adjust its content by regulating the growth and metabolism of S. baltica. In the essential proteins that might be directly regulated, kinases play several roles in cell signal transduction and complex life activities (Jiao et al., 2018), and many phosphorylases were identified as differential proteins.