3.1. Orthogonal Optimization of DPCD Parameters
TBC was significantly reduced by DPCD treatment, and the germicidal efficacy was closely related to the DPCD parameters shown in
Table 1. Temperature was the most important factor, followed by time and pressure, according to the range analysis results (
Table 1), since R
C > R
B > R
A. The variance analysis results (
Table 2) indicate that temperature and treatment time had a significant impact (
p < 0.05) on the bacteria inactivation in cheese samples. Lactic acid bacteria (LAB) acted as the major microorganisms in our cheese samples, which were sensitive to temperature rise and the prolonging of time when treated by DPCD [
13]. Although an increase in temperature reduced the solubility of CO
2, it also increased CO
2 diffusivity and cell membrane fluidity, resulting in cell component extraction, protein denaturation and cytoplasmic membrane collapse [
14]. The importance of temperature in DPCD treatment was also demonstrated by Ji et al. [
14] and Tang et al. [
6]. Pressure was found to play the least important role in microbial inactivation of quark cheese samples. This was consistent with the contribution of pressure to sterilization in paprika with DPCD [
15]. On the contrary, the significant effects of DPCD pressure on microbial reduction have been reported elsewhere, such as in coconut water [
16]. This varying effect of pressure could be due to material statuses, with solid foods being less sensitive to pressure variation in terms of microbial inactivation [
6].
The range analysis results exhibited that 30 MPa (K
1 > K
2 > K
3), 45 min (K
2 > K
1 > K
3) and 55 °C (K
1 > K
2 > K
3) were optimal parameters for DPCD treatment. However, according to variance analysis for the pressure term (data not shown), there was no significant difference between the 30 MPa and 20 MPa groups. Higher pressure would cause higher costs and severe protein degeneration [
11], so 20 MPa was the optimal choice. Thus 20 MPa, 45 min and 55 °C (A
2B
3C
3) were considered as the optimal DPCD treatment parameters for bacteria reduction.
3.4. Proteolytic Analysis
During storage, proteolysis is inevitable in cheese due to enzymatic decomposition during storage. In this study, the proteolytic activity of quark cheese ranged from 0.287% to 0.348%. (
Table 3). The minimal effect of DPCD processing on the proteolytic activity of cheese on day 0 suggested that the molecular effect of CO
2 under this pressure and temperature did not result in protein degradation. This was consistent with the results reported by Liao et al. [
9], who found that the peptide chain of lipoxygenase subunits could not be decomposed by DPCD treatment. For all cheese samples, there was a significant increase in proteolytic activity during storage. Additionally, we could see that the proteolytic activity index of the control increased from 0.0296 to 0.0348, which was more intense than the change in the DPCD-processed cheese. This demonstrated a slower proteolytic action in cheese with DPCD treatment during storage. On the one hand, proteases such as plasmin and rennet in cheese samples could possibly be inactivated by structural alterations [
11,
19], which consequently preserved milk proteins from vigorous hydrolysis. On the other hand, smaller amounts of microorganisms in DPCD-processed cheese, as mentioned previously, might lead to less involvement of proteins in the metabolism of LAB and yeast. Silva et al. [
10] also found that the presence of CO
2, effective in controlling spoilage, was associated with lower proteolysis in cheese. Similar inhibition of proteolysis occurred in cheese processed using high-pressure treatment [
21,
22]. Although the generation of polypeptides and amino acids caused by proteolysis contributes to the unique flavor of cheese, it is also responsible for undesirable flavors such as maltiness and bitterness [
3,
23].
3.5. Color Changes
Color is known as an important factor for cheese quality, and the effect of DPCD treatment on cheese color is shown in
Table 4. After DPCD treatment at day 0, the L* value of cheese hardly changed, while the a* value decreased from −0.47 to −1.04 and the b* value increased from 10.32 to 11.05 significantly (
p < 0.05), indicating greener and yellower cheeses. This was consistent with the color change of cheese after high-pressure treatment reported by Evert-Arriagada et al. [
3]. It is important to highlight that DPCD treatment exhibited greater advantages in preserving cheese color than traditional heat sterilization according to the fact that treatment at 117 °C for 20 min would inevitably cause the cheese color to degrade visually [
5], while changes in cheese color were indistinguishable following DPCD treatment in this work referring to sensory descriptions. During the storage, a significant increase in L* value could be observed in the control ranging from 94.13 to 94.84 (
p < 0.05), which was greater than the DPCD-processed cheese. This was because protein hydration decreased in quark cheese samples and the increased number of free moisture droplets consequently increased the degree of light scattering [
18]. Additionally, we found an obvious increase in redness and decrease in yellowness in our cheese samples, which are possibly related to changes in the cheese structure such as the protein network during storage [
24].
3.6. Microstructure
The microstructure of cheese samples was monitored using CSLM to determine the changes in the protein matrix (stained in green) and fat distributions (stained in red). It could be observed that a relatively continuous protein matrix existed in the control initially, accompanied, however, by slight heterogeneity and a dispersed fat phase in the form of discrete and large globules (
Figure 1). This could be explained by the milk used for the cheese preparation not being homogenized, which probably resulted in an increase in particles and a lack of active fillers (homogenized milk fat globules) in the acidified casein network [
25]. At day 0, an ununiform and loose microstructure of cheese resulted from DPCD treatment distinguishably, since many black cracks appeared in the protein matrix. This was due to the fact that DPCD treatment could lead to aggregation of milk proteins in cheese by enhancing the protein–protein interactions [
11,
19]. The transition from its protein hydration to protein–protein interactions then resulted in a more discontinuous protein matrix [
26]. Additionally, this supported our sensory descriptions stating that DPCD-processed cheese had a slightly gritty flavor. After the storage for 14 days, we found an obvious extension of the serum phase and a concomitant reduction in the size of, and an increase in the number of, fat globules in the control. Similar microstructural changes have also been described in cheese subjected to high-pressure treatment [
27]. However, there was no distinguishable change in the microstructure of DPCD-processed cheese. Combined with proteolytic analysis, the intense proteolysis might lead to a weakening of the connection between casein as well as a loosening of the protein skeleton structure, which could explain the greater extension of the serum phase in the control.
3.7. Rheological Properties
Processing with DPCD and the storage times affected the rheological performance of cheese samples, as displayed in
Figure 2. The storage modulus (
G′) reflects the ability of cheese to store energy while maintaining its complete structure and loss modulus (
G″) indicating the ability to dissipate mechanical energy by converting mechanical energy into heat through molecular motion [
28]. For all cheeses, we found that
G′ >
G″, demonstrating the elastic structure of our quark cheeses (
Figure 2A,B). This was consistent with the strain sweep test results for Akawi cheese reported by Abdalla et al. [
29]. The viscoelastic properties (
G′ and
G″) of cheese were obviously enhanced by DPCD treatment. However, these values were diminished during storage for both cheeses. Power law rheological parameters (
k′,
n′,
k′,
n″) were calculated and are shown in
Table S1 to further characterize the viscoelastic properties of cheese. The values of
n′ and
n″ reflected the frequency dependence of rheological parameters and were used to describe the type of bonding of the structural elements present in the matrix [
30]. These values for all cheese samples were always lower than 0.20, which is considered a relatively low level, indicating the presence of strong and cross-linked gels with permanent covalent bonds [
30]. We found a significant increase in
n′ value following DPCD treatment on day 0, indicating an increase in the frequency dependence of rheological moduli. This suggested a less structured cheese matrix resulted from DPCD processing. As we previously discussed, this was consistent with the more discontinuous protein phase and extended serum phase of cheese after DPCD treatment.
Shear properties of cheese should be estimated in the complete characterization of rheological profiles. The change in apparent viscosity with shear rate is exhibited in
Figure 2C. It could be observed that the apparent viscosity of cheese decreased continuously with increasing shear rate, indicating the shear thinning flow behavior. The flow behavior index (
n) and the consistency index (
k) were also calculated through the power law model and displayed in
Table S1. Similarly to the viscoelastic results, the cheese processed with DPCD showed lower
n values than the control at day 0. This demonstrated the weakened ability to resist shearing led by DPCD because of the discontinuous protein network structure. However, DPCD significantly increased the apparent viscosity of cheese at day 0 according to
Figure 2C and
k value changes in
Table S1. Changes in moisture distribution, which has been confirmed to be closely related to the rheological properties of cheese, were most likely to blame [
30]. The prolonging of storage time caused an obvious decrease in apparent viscosity, which corresponded to alterations in
k′ and
k″ values. The proteolysis might be responsible for this phenomenon. Surprisingly, DPCD-processed cheese after 14 days of storage displayed similar rheological properties to the control at day 0. So, DPCD treatment could somewhat compensate for the degradation of cheese’s rheological properties during storage.
3.8. Moisture Distribution
The proton relaxation signal of the cheeses was investigated, which was affected by the cheese matrix due to interactions between water and macromolecules [
31]. Due to the heterogeneity of our cheese samples, five different proton populations were identified (
Table 5). The water population peaked at around 0.15 ms, which could be attributed to the
1H of water being strongly bound to the casein structure [
30]. DPCD treatment resulted in a slight increase in T
21 at day 0 according to
Table 5. This might be due to the enhancement of casein surface hydrophobicity caused by DPCD, which consequently weakened the protein hydration [
26]. This was consistent with the slight extension of the serum phase in cheese microstructure resulting from DPCD. The water in T
23 and T
24 was attributed to protons of water trapped in the protein meshes and considered as immobilized-state water [
30,
31]. We also found that the DPCD-processed cheese initially displayed significantly lower T
23 and T
24 on day 0, which was consistent with the rheological properties. Tidona et al. [
30] also illustrated a similar negative correlation between the relaxation times of these two populations and viscoelasticity. After 14 days of storage, it could be observed that T
21 decreased significantly and the population in T
22 migrated and became undetectable in the control (
p < 0.05). This was probably due to the intense proteolysis leading to an alteration in protein structure as well as a destruction of the protein network [
32].
3.9. Volatile Profile
Flavor is recognized as a critical factor in determining consumer acceptance of cheese. The flavor of cheese is produced by a combination of a large number of volatile substances, which are mostly derived from complex biochemical reactions during the preparation and deterioration of cheese [
33]. A total of 40 volatile compounds were detected in our cheese samples including sulphocompounds and aliphatic hydrocarbons, heterocyclic compounds, acids, alcohols, aldehydes, ketones, ethers, and aromatic hydrocarbons, as shown in
Table 6. There was significant reduction (
p < 0.05) in the contents of acids including acetic acid, butyric acid, caproic acid, caprylic acid, and
n-decanoic acid in cheese after treatment with DPCD, among which the acetic acid content decreased most obviously (
Table S2). Similar results have also been carried out by Evert-Arriagada et al. [
34], who found that the content of acids in cheese volatile compounds was significantly reduced after high-pressure treatment. There was a relatively high content of alcohols in cheeses, and this could be possibly explained by the use of a high incubation temperature, which might not correspond to the mesophilic starters used in the study. Evert-Arriagada et al. [
34] also reported such initial alcohol content in alcohols. Significant alterations in the volatiles of cheese were determined after 14 days of storage. Due to its lower odor threshold, sulphocompounds with a strong odor of garlic, onion, cabbage and mature cheese had an important contribution to the flavor of cheese, though the detected content is very low [
35,
36]. However, dimethyl disulfide is the only sulphocompound detected in cheese, and it became undetectable after 14 days of storage for both cheese samples.
Before the storage, three aliphatic hydrocarbons were detected in our cheese samples, among which 2,4-dimethylheptane was the main component. After 14 days of storage, there was no significant change in the content of aliphatic hydrocarbons, but a wider variety of aliphatic hydrocarbons became detectable, including decane, undecane, and d-limonene, and octane occupies the greatest content in aliphatic hydrocarbons. The formation of new aliphatic hydrocarbons during storage might be attributed to the oxidation of fats [
37]. However, aliphatic hydrocarbons were not the critical component for cheese flavor due to their high odor threshold [
38]. In terms of acids, their content decreased after storage in the control, which was mainly attributed to the decrease in acetic acid content. This could be related to the role of acetic acid as an intermediate in biochemical pathways [
39]. Such transformation was inhibited by DPCD treatment, as the acetic acid content remained constant in DPCD-processed cheese, which could possibly be explained by the weak metabolism. Similar results were obtained by Lues and Bekker [
40] and Upreti et al. [
41], who found that the content of acetic acid dropped rapidly during cheese storage. Additionally, butanoic acid and caproic acid increased obviously in both cheeses after storage, which played dominant role in total acid content alterations of DPCD processed cheese. Butanoic acid was derived from the fermentation of lactose and lactic acid, which changed greatly in the control with the existence of a large amount of yeast [
41]. In terms of aldehydes, the quantity and content of aldehydes detected were low. Additionally, these compounds were mainly derived from amino acids either by transamination, leading to an intermediate imide that can be decarboxylated, or by Strecker degradation [
42]. Aldehydes are short-lived compounds in cheese, which can be rapidly reduced to primary alcohols and even oxidized to corresponding acids. This is the reason why there was only a small number of aldehydes with low content detected. However, it has been confirmed that aldehydes, especially linear aldehydes, have an important contribution to the freshness and floral aroma of cheese [
42]. The contents of hexanal and nonanal increased significantly in DPCD-processed cheese after storage, while they remained unchanged in the control. Therefore, DPCD treatment might have the potential to maintain the freshness of cheese flavor. Additionally, acetoin is a critical compound of cheese that contributes to the desired creamy aroma [
42]. Due to the proteolysis during storage, the acetoin content in both cheeses significantly increased [
23].
Lactococcus lactis ssp.
cremoris, more vulnerable to heat and high pressure than
Lactococcus lactis ssp.
lactis, was considered more suitable for flavor development and contributed to acetoin generation. This could explain the limited formation of acetoin in DPCD processed cheese after storage [
43,
44,
45].
Our GC-MS results are equivocal in demonstrating the clear differences between the volatile profiles of cheese with or without DPCD treatment. Therefore, principal component analysis (PCA) was used to reduce the dimensions of our GC-MS data. Four principal components were extracted with a characteristic value > 1. PC1 and PC2 possessed variances of 43.9% and 22.9% (
Figure S1A), respectively, totaling 66.8%. According to the PCA score plots, the control day 0 group showed high similarity with the DPCD treatment group, indicating that the flavor property of cheese was hardly affected by DPCD treatment before storage overall. The control day 14 group and the DPCD treatment day 14 group were obviously isolated from the groups at day 0, and they could be clearly distinguished as they were located in the second and third quadrants, respectively. Therefore, DPCD treatment resulted in different changes in the cheese volatile profile during storage. The changes in
Lactococcus lactis ssp.
lactis and
Lactococcus lactis ssp.
cremoris by DPCD that led to changes in casein breakdown and flavor production from amino acids were considered one of the reasons for this difference [
45].
The OPLS-DA model was also used to establish the volatile components that led to the PCA score difference between the control day 14 group and the DPCD treatment day 14 group. The R
2Y and Q
2 values for the model were 0.978 and 0.933, respectively, and indicated a good predictive ability. The VIP values for these two classes are shown in
Figure S1C. Components with VIP values ≥1 are marked in red, and are considered as the differential metabolites of the two groups [
46]. Ethanol, acetoin, butyrate, octoic acid, acetic acid, octane, ether, 1,3-bis (1,1-dimethyl ethyl) benzene, n-hexanol, 2-heptanone, and hexanal were selected as the reasons for the differences in the volatile profile between these two groups.