Longissimus dorsi muscles were collected from six Simmental cattle (24 ± 1 months old, 470 ± 30 Kg) in a commercial abattoir and transferred to the laboratory in an ice box within 2 h. All beef were rinsed with sterile water to remove stains and blood, and then dried with sterilized paper towels. The muscles were cut into 300 g samples (6 cm × 6 cm × 10 cm) and vacuum packaged at the O₂-transmission rate of 40 cm³/m²/day and 85% relative humidity. Samples were divided into three groups, and aged at respective temperature (4, 10, or 15°C). Samples were analyzed at different time points (0, 2, 4, 6, 8, 10, 14, and 18 day for 4°C, 1, 2, 3, 4, 5, 6, 8, 10, 12, and 14 day for 10°C; 0, 1, 2, 3, 4, 5, 6, and 7 days for 15°C). The temperature was monitored continuously by using remote temperature recorders with high precision control. Each analysis was performed using triplicate samples.

The beef samples (3 cm × 3 cm × 6 cm) were packaged and cooked in a water bath at 80°C to achieve a core temperature of approximately 70°C. After cooking, the samples were cooled at 4°C until the core temperature cooled down to 10°C, surface-blotted with paper towels, and reweighed for weight loss. Cooking loss was determined by calculating the weight difference of the samples before and after cooking and expressed as the percentage of initial weight. Shear force values of samples (1.0 cm × 1.0 cm × 3.0 cm) weredetermined across the longitudinal direction of muscle fibers by a texture analyzer (TA-XT plus) attached with a Warner-Bratzler blade (V-notch, HDP/BSW). The cutting line was positioned to avoid fatty and connective tissues and was perpendicular to the muscle fiber direction. The shear force value was calculated as the mean of the maximum force and expressed as in Newtons (N).

The surface color of the beef samples was measured on each analysis day after the samples were blooming for 30 min at room temperature. The lightness (L*), redness (a*), and yellowness (b*) of the beef samples were measured by a spectrophotometer (model CR-410, Minplta, Tokyo, Japan) with a diameter of 8 mm. The instrument was set for illuminant D65 and calibrated with a standard white plate before measurement. Measurements were taken in triplicate at different locations within each sample.

Briefly, 10 g of raw meat (free of subcutaneous fat) was weighed into a closed glass vessel with 75 ml of distilled water with intermittent shaking for 30 min (room temperature). Immediately prior to distillation, 1 g of magnesium oxide was added to the sample, and the sample was steam distilled for 5 min. The distillate was condensed into a receiving flask containing boric acid (20 g/L) with a mixed indicator solution of bromocresol green/methyl red. The solution was back-titrated with 0.01 M hydrochloric acid solution, and TVB-N was calculated as mg/100 g raw meat.

A total of 25 g of samples from the top surface (depth: 0.5 cm) were blended in 225 ml of 0.85% sterile saline solution for 90 s in a stomacher (Ningbo Jiangnan Instrument Factory) at room temperature. Samples for microbial testing were prepared in a series of decimal dilutions by using sterile saline. Plate count agar (PCA) was used for total viable counts (TVCs). Violet red bile glucose agar (VRBA) was used for Enterobacteriaceae counts. Rogosa and Sharpe agar (MRS) was utilized for LAB counts. Cephalothin–sodium fusidate–cetrimide (CFC) agar with CFC-selective supplement was applied to determine Pseudomonas spp. counts. Streptomycin thallous acetate agar (STAA) with STAA selective supplement was employed for Brochothrix thermosphacta counts. Each diluent (0.1 ml) was spread on the selective medium agar in triplicate. CFC and STAA plates were incubated at 25°C for 48 h. PCA and MRS plates were incubated at 30°C for 48 h, and VRBA plates were incubated for 24 h at 37°C. The number of colonies was counted and expressed as colony forming units per gram (log CFU/g).

The volatile compounds present in beef samples were identified and quantified by a GC–IMS flavor analyzer (FlavourSpec^®, Shandong, China) equipped with a syringe and an autosampler unit for headspace analysis. Briefly, 2 g of homogenized beef sample was transferred into a 20 ml headspace vial, incubated at 60°C for 20 min, the vial was put in an incubator and shook at 500 rpm to facilitate the emitting of volatiles into headspace, the vial was sealed by using a magnetic cap with a silicone septum. The temperature of head space injection needle was 60°C and the injection volume was 500 μl. The analytical conditions for this test are as follows: Chromatographic column type: MXT-5, 15 m, 0.53 mmID, 1.0 μm df (RESTEK, United States), to separate the volatile components and coupled to ion mobility spectrometry (IMS); carrier/drift gas: N₂, with the flow ramp starting at 2 ml/min for 2 min, then increasing to 20 ml/min in 8 min and increasing to 130 ml/min in10 min, finally 130 ml/min for 5 min, introducing the sample into the capillary; column syringe temperature: 85°C. The total GC runtime was 25 min, triplicate injections and analysis of samples were performed.

All experiments were performed in triplicate. The mean values and standard errors of the means were recorded, outlying observations were identified, and implausible values were verified with the original source or coded as missing. Differences between treatments were analyzed through Tukey’s test. Statistical analyses were conducted at a 95% confidence level. Bioinformatics analysis was carried out on the microbiota by using the Majorbio Cloud platform on the basis of OTU information; rarefaction curves; and alpha diversity indices, including observed OTUs. The analytical software Laboratory Analytical Viewer and the built-in NIST and IMS databases of the GC–IMS Library Search software were used for the qualitative analysis of characteristic volatile compounds. One-way ANOVA was used to estimate the difference between means (p < 0.05).

Cooking loss is the water loss from the meat due to protein denaturation during cooking (Kim et al., 2019). Figure 1A illustrates the effect of different aging temperatures on cooking loss, which tended to increase as the beef aged: the higher the aging temperature, the higher the cooking loss. In particular, the cooking loss aged at 15°C always higher than aged at 4°C (p < 0.05). Similarly, the interaction of aging temperature had an effect on the Warner-Bratzler shear force (WBSF) of the beef (p < 0.05). Figure 1B shows that with the prolongation of aging time, the shear force values show a decreasing trend whether it is decreased at 4, 10, or 15°C; specifically, aging affected the shear force values of all the beef samples with the meat becoming increasingly tender at each time point (p < 0.05). The beef aged at elevated temperatures (10 or 15°C) for 5 days had equivalent shear force values as the beef aged at 4°C for 10 days (p > 0.05). The beef aged at 15°C for 3 days had equivalent shear force values to beef aged at 4°C for 8 days (p > 0.05). This indicated that the aging period could be shortened by placing the meat at a slightly higher temperature than the typical meat aging temperature (4°C). However, the final tenderness is not affected by the elevated temperature. Aging at 4°C resulted in no significant reduction in shear force values before 6 days due to lower temperatures, however, the shear force values of the beef aged at 10 and 15°C decreased rapidly to 47.9 and 47.1 N, respectively, at the same stage of aging. At 14 days after aging at 4°C, the beef samples had an average WBSF value of less than 50 N. Aging at 10°C and 15°C required about 6 days, to achieve the same results, likely due to the enhancement in proteolytic enzyme activity, proteolysis via calpains, and collagen fiber breakdown from lysosomal enzyme activity at high temperatures (Kim et al., 2016). Furthermore, when the beef shear force under 50 N, the change is no longer significant. The time point could be considered as the end of beef aging, to save aging time and energy consumption.

Meat color is a direct estimate of meat freshness and wholesomeness for consumers, who often associate discoloration with spoilage (Li et al., 2015). The L*, a*, and b* color attributes of beef samples are shown in Figures 2A–C. The L* value decreased with the extension of aging time. Groups aged at higher temperatures presented significantly decreased L* values compared with those aged at 4°C. a* is the most important color parameter for fresh meat (Yim et al., 2016). Generally, the higher a* value, the fresher the meat. Figure 2B shows that a* decreased with the extension of aging time. At the same aging time, the a* values of samples aged at 10°C and 15°C were lower than those of samples aged at 4°C, indicating that the higher temperature, the more unfavorable a* value of beef. With the progression of aging, the b* value increased, and the b* value of samples aged at 15 and 10°C was significantly higher than that of the samples aged at 4°C (p < 0.05). Furthermore, aged at 4°C for 14 days had equivalent b* values as the beef aged at elevated temperatures (10 and 15°C) for 5 days, indicating that the higher temperature, the more unfavorable the effect on meat color.

TVB-N refers to the combined action of endogenous enzymes and bacteria in the muscle during the storage of animal food (Holman et al., 2021). It is highly temperature-sensitive and increases rapidly with small increases in storage temperature (Frank et al., 2019). The trend of TVB-N during aging at 4, 10, and 15°C was shown in Figure 2D. During aging at 4°C, the TVB-N content increased from the initial 4.6 to 12.5 mg/100 g after 14 days (p < 0.05). In vacuum-packaged beef aged at 4°C, it increased to 15 mg/100 g in approximately 16 days. During aging at 10°C, the TVB-N content increased to 13.4 mg/100 g after 10 days and continued to increase to 15 mg/100 g on the next day. The TVB-N content exceeded 15 mg/100 g after 5 days of aging at 15°C, indicating that the TVB-N values of beef increased rapidly at high temperatures, rose slowly in the early stages of storage, then rapidly increased.

The changes in the microbial counts of beef samples aged at different temperatures are shown in Table 1. Temperature appears to be the most important factor that influences the spoilage and safety of meat (Kennedy et al., 2005). Significant aging time and temperature interactions affected the TVC and growth of Pseudomonas, Brochothrix thermosphacta, LAB, and Enterobacteriaceae (p < 0.05). Initially, microorganisms were present at low levels. The TVC in unaged beef samples was 3.1 ± 0.01 log CFU/g, indicating that the sample is of good quality (Yang et al., 2018). In beef samples aged at 4°C, TVCs increased slowly at the beginning of aging but increased rapidly on day 6 and exceeded 5.8 log CFU/g by day 14, at which the shear force value reduced to less than 50 N. During aging at 10°C, the microbial counts exceeded 7.0 log CFU/g on day 10, and when the shear value started to drop below 50 N on day 6, the corresponding microbial count was 6.0 log CFU/g. Similarly, during aging at 15°C, the shear value of beef started below 50 N on day 5, corresponding to the microbial count of 6.3 log CFU/g that did not exceed 7.0 log CFU/g (Yang et al., 2016). A TVC total of 7.0 log CFU/g is recommended as the load limit at the end of the shelf life for red meat, According to the microbial colony counts, the shelf life of beef samples as the aged temperature increased. In order to the quality and safety of beef, it could be consider transferring the beef to 4°C storage when the shear value of beef is reduced 50 N to extend the shelf life, which is also our next research program. Pseudomonas, Brochothrix thermosphacta, Enterobacteriaceae, and Lactobacillus increased with a trend similar to the trend shown by the total number of colonies. LAB was the dominant population in the vacuum-packaged beef samples in this study. Similarly, Gribble et al. (2014b) LAB, Enterobacteriaceae, and Pseudomonas spp. populations increased when the beef was subjected to higher aging temperatures.

A characteristic fingerprint corresponding to each aging temperature was established by using gallery plots, and the difference and specific distribution of volatile compounds between different aging temperatures were compared intuitively (Figure 5). The Gallery Plot plugin of LAV software was used to compare the differences in the volatile compounds in beef samples at aged at different temperatures comprehensively. The colors of the plots indicated the signal strengths of the compounds: the darker the signal, the weaker the signal intensity and vice versa (Huang et al., 2022). The signal intensities of most volatile compounds significantly increased after aging at high temperatures, indicating that the levels of most volatile compounds were positively correlated with the aging temperature.

As shown in Figure 5, all the peaks to be analyzed in the obtained two-dimensional GC–IMS spectrum were automatically generated on the basis of fingerprints. Each row in the figure represents all signal peaks selected in a beef sample, and each column in the figure represents the signal peaks of the same volatile organic compounds in different beef samples. The substances in the red box in the picture had the highest content on 0 day compared to other groups. This substance included benzaldehyde, which decreased during aging at 4, 10, or 15°C. Benzaldehyde, is found in hyacinth, lemongrass, and rockrose and is a volatile compound that potentially results from the Strecker degradation of tyrosine, which has a bitter almond, cherry, and associated with a strong almond odor (Ma et al., 2012), if the level are raised too high, the undesirable flavor may result. The content of the substances in the yellow box in Figure 5 increased significantly in the 4°C_14D sample (p < 0.05). These substances included 2-heptanone, isobutyl butyrate, 1-octen-3-ol, and 1-pentanol. 2-Heptanone can produce a buttery, creamy, and cheesy aroma; it can be used as a spice raw material and is critical to aging flavor (Resconi et al., 2012). Isobutyl butyrate exists in pineapples and other fruits naturally and has a pleasant aroma. It is potentially a characteristic compound of fresh beef. 1-Octen-3-ol is positively correlated with umami flavor and is a volatile alcohol with a mushroom-like aroma that is found in dry-cured ham and dry sausage (Tansawat et al., 2013). Positive attributes, such as umami and juicy flavors in grain-fed steaks, are correlated with 1-octen-3-ol (Song et al., 2011). 1-Pentanol has a strong, sweet, balsamic aroma (Liu et al., 2022). Meanwhile, the dominant bacterial community was Carnobacterium, including that could produce these VOCs at low temperatures. The substances in the orange box in Figure 5 have the highest content in the 15C_3D sample. They included butanal and ethanol. Butanal is found in many essential oils, such as flowers, fruits, and dairy products, and has an ethereal fragrance when extremely diluted (Lee et al., 2019). Ethanol is the main component of wine, ethanol can also be used to make acetic acid, beverages, baked goods, and it generally have a low odour threshold, and thus they partly contribute to the flavour of cooked beef, what’s more, butanal and ethanol was found to decrease or remain stable at chill temperatures, whilst an increase was observed at elevated temperatures in vacuum packaged conditions. The substances in the green box in Figure 5 have the highest content in the 15C_5D sample compare to other groups. They included 2-methylbutanal, 3-methylbutanal, 3-methylbutanol, 2-methyl-propanal, isoamyl acetate, phenylacetaldehyde and 1-penten-3-one, and 2-butanone. 2-Methylbutanal is described as having a brothy, grainy or boiled meat aroma, which improves the sensory value of products (Lee et al., 2021). 3-Methylbutanal is a colorless and transparent liquid that is used as an intermediate in the production of flavors (Mansur et al., 2019). 2-Methylbutanal and 3-methylbutanal are Strecker degradation products of isoleucine and leucine, respectively (Saraiva et al., 2014). 2-Methyl-propanal is used in the synthesis of cellulose esters and flavors and is commonly applied in baked goods and meat products (Utama et al., 2018). Aldehydes, in general, are unstable and can easily react with other compounds to produce compounds with different flavors, it have a low odor-detection threshold; hence, even a trace amount can contribute to meat flavor, and, consequently, they are the most interesting of the lipid-derived volatiles (Legako et al., 2016). Isoamyl acetate has banana and pear aromas and is widely used in the production of various fruity edible flavors (Ercolini et al., 2009). Phenylacetaldehyde is naturally found in chicken, bread, rose oil, and citrus oil and confers beef with a clear and evocative aroma that differs in accordance with quality grade; specifically, it is higher in prime and low-choice steaks than in standard steaks (Zhu et al., 2016). 1-Penten-3-one is mainly used as a spice for food and in onion, garlic, and mustard flavoring; it has been identified as the main contributor to the integral flavor of beef due to its high odor activity values (Vilar et al., 2022). 2-Butanone is an intermediate in the preparation of pharmaceuticals, fragrances, antioxidants, and certain catalysts (Pavlidis et al., 2019). Meanwhile, the dominant bacterial community was Hafnia–Obesumbacterium, which can use the nutrients in beef to produce these compounds (Argyri et al., 2015). It could conclude that different aging temperatures produce different flavor profiles, on the one hand, these compounds could be formed protein degradation or oxidation of fatty acids; on the other hand, some bacteria studied in this work have probably contributed to different levels in the accumulation/depletion of the measured metabolic compounds.