A total of 48 sirloins were obtained from approximately 48-month-old Hanwoo cows (quality grade 2). The samples were transferred in a cooler (4°C) to the Korea Institute for Animal Products Quality Evaluation (Sejong, Korea). The initial pH of all samples was measured (average 5.48±0.12) prior to the dry-aging process. As a control, Hanwoo cow meat was assigned on 2 days postmortem, vacuum-packed, and immediately frozen to −70°C (n = 24). The other 24 sirloins (n = 8 for each treatment) were dry-aged for 28 days with the different aging methods: i) TD (temperature, 1±1°C; RH, approximately 85%; air, 5±3 m/s in a specialized facility); ii) SD (temperature, 2±1°C; RH, approximately 75% in the newly developed appliance); and iii) SDB (the same condition as SD except for the aging bag [Drybagsteak LLC, Minneapolis, MN, USA]: water vapor permeability of 8,000 g [15 μ/m²/24 h and O₂ permeability of 2.3 mL/m²/d at 38°C and 50% RH]). The SDB group was vacuum-packed (HFV-600L, Hankook Fujee Co., Ltd., Siheung, Korea) to ensure that the surface of the bag was in contact with the samples. After aging, the dried surfaces (crust) of all dry-aged samples were trimmed off, and the sirloins were then vacuum-packed and frozen to −70°C for further analyses.

Ground meat (200 g) was prepared in a sample cup (FOSS, Hillerød, Denmark) and its chemical composition (moisture, fat, protein, and collagen contents) was determined using the FoodScan Lab meat analyzer (FOSS, Denmark) based on official methods of analysis of AOAC [14].

A sample of the meat (5 g) was blended in sterile saline (45 mL, 0.85%) for 2 min using a laboratory blender (BagMixer 400 P, Interscience, St. Nom la Bretèche, France). Each dilution (100 μL) was spread on plate count agar (Difco Laboratories, Detroit, MI, USA), YM agar (Difco Laboratories, USA), and eosin methylene blue agar (Difco Laboratories, USA) for enumeration of total aerobic bacteria, mold/yeast, and coliforms, respectively. The agar plates for total aerobic bacteria and coliforms were incubated at 37°C for 48 h, whereas the YM plates were incubated at 25°C for 120 h. After incubation, the microbial counts were calculated and expressed as log colony-forming unit (CFU)/g.

The meat sample was vacuum-packed (HFV-600L, Hankook Fujee Co., Ltd., Korea) and boiled in a water bath until a core temperature of 72°C was reached. The cooked samples were then cut parallel to the muscle fiber into cuboidal subsections (2.54 cm height) and placed under a Warner-Bratzler shear probe, perpendicularly to the muscle fiber. Shear force (N) was measured using a texture analyzer (TMS-Touch, Food Technology Co., Sterling, VA, USA) with a cell load of 0.1 N and a cross-head speed of 400 mm/min.

Inosine 5′-monophosphate (IMP) was extracted from the samples according to the method of Lee et al [15] with a few modifications. The extract was filtered through a membrane filter (0.2 μm; Whatman PLC., Kent, UK) into a glass vial and injected into a high-performance liquid chromatography (HPLC; Ultimate 3000, Thermo Fisher Scientific Inc., Waltham, MA, USA) system. The analytical conditions were as follows: injection volume, 10 μL; mobile phase, 20 mM potassium phosphate monobasic (pH 5.5); flow rate and time, 1.0 mL/min for 25 min; column, Synergi Hydro-RP (250×4.6 mm², 4 μm particles; Phenomenex Inc., Seoul, Korea) at 30°C; and detector, UV/Vis detector at 254 nm. The peak area was calculated from a standard curve obtained using a standard IMP (Sigma-Aldrich, St. Louis, MO, USA).

For free amino acid content, the samples were prepared according to the method of Lee et al [15] and injected into a HPLC (S 1125, Sykam GmbH, Eresing, Germany) system with post-column derivatization (Pinnacle PCX derivatization instrument, Pickering Laboratories, Mountain View, CA, USA). The analytical conditions were as follows: mobile phase, buffers A, B, C, and D (Ajoo Scientific, Gunpo, Korea); column, 4.6×150 mm² (Sykam GmbH, Germany) at 25°C; and detector, UV/V is detector at 540 nm. The standard (Sigma-Aldrich, USA) was used to generate a standard curve for calculation of the peak area.

Sensory evaluation was conducted by a consumer panel (total 30 panelists) to observe changes in the low-marbled Hanwoo cow meat after dry-aging and differences among the different dry-aging methods (TD, SD, and SDB). Each sample was cut into a similar size (50×20×6 mm³, length×width×height), grilled until the core temperature reached to 72°C, and served to the panelists. A 7-point hedonic scale (1, extremely dislike; 7, extremely like) was used to score juiciness, tenderness, flavor, and overall acceptability before and after dry-aging.

A randomized incomplete block design was applied, using the trial as the block. The control (n = 12, 24 sirloins per 2 trials) and the 3 different dry-aging methods (n = 4 for each treatment, 8 sirloins per 2 trials) were assigned, and the model was analyzed with the fixed effect (aging method) and the random effect (carcass and side of the carcass). Calculations based on the general linear model were performed using SAS 9.3 (SAS Institute Inc., Cary, NC, USA) and the results were reported as mean values with standard error of the means. Significant differences among the mean values were determined on the basis of the Student-Newman-Keuls multiple comparison test at a level of p<0.05.

Differences in chemical composition were found between the control and the dry-aged groups (Table 1). After 28 days of dry-aging, the moisture content in the 3 dry-aged groups was significantly lower than that in the control. This typical result from the dry-aging process is attributable to moisture evaporation [10,11, 16], and is important for concentrating the flavor components in dry-aged beef [17]. As the RH can promote moisture evaporation during the dry-aging process [18], SD and SDB groups (RH, approximately 75%) were expected to have a lower moisture content than TD group (RH, approximately 85%). However, no significant difference in moisture content was detected among the 3 dry-aged groups. This was probably due to moisture evaporation in TD, promoted by air flow (5±3 m/s), whereas the bag used in SDB did not interfere with moisture evaporation as it was highly water vapor permeable (8,000 g/15 μ/m²/24 h). This phenomenon is noteworthy, given the possibility of applying SD and SDB instead of TD for aging beef at similar rate of dry-aging. The fat and protein contents showed an increasing trend during the 28 days of aging, but this was a consequence of the reduction in moisture content rather than actual changes in their content during the dry-aging process [19]. For similar reasons, the different dry-aged methods did not show differences in the fat and protein contents as their moisture contents were similar. The moisture content was negatively correlated with fat (r = −0.9) and with protein (r = −0.3). The different dry-aging methods did not cause any significant changes in the collagen content.

No mold/yeast and coliform contamination were found in the meats before and after dry-aging (Table 2). Total aerobic bacteria count in the 3 dry-aged groups was higher than that in the control. Among the dry-aged groups, SDB showed the highest number of bacteria, which was not expected as most of the previous studies have reported that the bag could inhibit microbial contamination [11,13]. This result might be attributable to the different amounts and depths of the crust and/or its formation rate as it has a role in preventing microbial penetration into the meat [20]. When we estimated the amount and depth of crust from the trimming loss among the dry-aged groups (data not shown), the trimming loss was the highest in SD group (26.31%), followed by TD (19.54%) and SDB (18.16%) groups. The results are in the same order as those of the microbial counts in Table 2. It would seem that crust formation during the early dry-aging process may be more effective in preventing microbial penetration than application of the aging bag. However, it can be a disadvantage from the economic aspect of the dry-aged beef. In this study, the saleable yield of SD group decreased approximately 13.67% relative to that of TD group, whereas SDB group maintained a similar level to TD group (p<0.05, data not shown). Thus, SDB would be a better choice for the dry-aged beef that have a relatively higher saleable yield, because the difference of microbial counts among the dry-aged groups (<1 log CFU/g), albeit statistically significant, can be ignored in practical situations.

The shear force was significantly decreased (approximately 47%) in the dry-aged groups compared with the control (Figure 1). This improvement is due to the activities of calpains, cathepsins, proteasome, and caspases during the early postmortem period [21, 22]. Meanwhile, no difference in shear force was observed among the TD, SD, and SDB groups. This seems reasonable as the differences among the dry-aging groups involved only RH, air flow, and the application of the aging bag, showing that all 3 methods have similar abilities to improve the tenderness of low-marbled beef.

IMP and free amino acids are the degradation products of ATP and protein, respectively, after slaughter [23]. The free amino acid components are influential in the flavor formation of meat and meat products, as together with reducing sugars they have a specific role in the Maillard reaction. Among these degradation products, IMP and Glu have a positive and even synergistic impact on umami taste [15,24]. In this study, the IMP content had declined by 40% to 47% during the aging period, and no significant differences were found among the different aging methods (Table 3). IMP degradation to inosine and hypoxanthine is a constant but fast reaction at the early postmortem period, and because dry-aging takes a long time to change the meat characteristics, the increases and/or concentrated levels of free amino acids would be more important than those of nucleotides. Iida et al [25] calculated the umami intensity of beef during the dry-aging process and reported a significant role of Glu at the late aging period, whereas IMP was effective only during the early aging stage. A different pattern was found between the control and dry-aged groups for various free amino acids. Whereas there were no significant changes in Glu, the contents of Ala, Arg, Cys-cys, Gly, His, Ile, Leu, Phe, Ser, Thr, Val, and total amino acids were increased during the aging period and were significantly higher than those of the control (p<0.05). The amount of free amino acids responsible for bitter taste (bitter-FAAs: Arg, His, Ile, Leu, and Phe) achieved 2.0-fold increases on average, whereas the free amino acids responsible for sweet taste (sweet-FAAs: Ala, Gly, Ser, and Thr) increased 1.5-fold on average after dry-aging. Even though the bitter-FAAs showed a higher fold increase than did sweet-FAAs, the overall changes in free amino acid content would have a positive impact on flavor formation as the absolute amounts were approximately 2-times higher in sweet-FAAs after dry-aging. On the other hand, the differences in RH, air flow, and the application of the aging bag did not cause significant changes in both IMP and most of the free amino acid content, possibly because the final moisture content was similar among the three dry-aged groups (Table 1). Kim et al [17] reported that moisture evaporation is a cause of the concentrated flavor components during the dry-aging process.

In the present study, improvement in the juiciness of low-marbled beef was expected, as Cambell et al [26] had elucidated that the fat content would concentrate with dry-aging through moisture evaporation. However, the juiciness was not significantly different among the different dry-aging groups (Figure 2), probably because their differences in moisture (62.17% to 67.01%) and fat (12.00% to 13.53%) contents were insignificant (Table 1). The palatability (juiciness, tenderness, flavor, and overall acceptability) of low-marbled beef was higher in SD, followed by SDB and TD and SD group was statistically similar to the results from SDB group (Figure 2). This is probably because TD resulted in the highest shear force and SD resulted in the highest total free amino acid content (p>0.05, Figure 1, Table 3). The present results suggest that SD and SDB can be applied instead of TD to improve the palatability of low-marbled beef.