Evaluation of Various Lactic Acid Bacteria and Generic E. coli as Potential Nonpathogenic Surrogates for In-Plant Validation of Biltong Dried Beef Processing

Validation studies conducted within a food processing facility using surrogate organisms could better represent the manufacturing process than controlled laboratory studies with pathogenic bacteria on precision equipment in a BSL-2 lab. The objectives of this project were to examine potential surrogate bacteria during biltong processing, conduct biltong surrogate validation lethality studies, and measure critical factors and intrinsic parameters during processing. Beef pieces (1.9 cm × 5.1 cm × 7.6 cm) were inoculated with four-strain mixtures of Carnobacterium divergens/C. gallinarum, Pediococcus acidilactici/P. pentosaceous, and Biotype 1 E. coli ATCC BAA (-1427, -1428, -1429, and -1430), as well as a two-strain mixture of Latilactobacillus sakei and other commercially available individual bacterial cultures (P. acidilactici Saga200/Kerry Foods; Enterococcus faecium 201224-016/Vivolac Cultures). Inoculated beef was vacuum-tumbled in marinade and dried in a humidity-controlled oven for 8–10 days (24.9 °C; 55% relative humidity). Microbial enumeration of surviving surrogate bacteria and evaluation of intrinsic factors (water activity, pH, and salt concentration) were performed post inoculation, post marination, and after 2, 4, 6, 8, and 10 days of drying. Trials were performed in duplicate replication with triplicate samples per sampling time and analyzed by one-way RM-ANOVA. Trials conducted with E. faecium, Pediococcus spp., and L. sakei never demonstrated more than 2 log reduction during the biltong process. However, Carnobacterium achieved a >5 log (5.85 log) reduction over a drying period of 8 days and aligned with the reductions observed in previous trials with pathogenic bacteria (Salmonella, E. coli O157:H7, L. monocytogenes, and S. aureus) in biltong validation studies. Studies comparing resuspended freeze-dried or frozen cells vs. freshly grown cells for beef inoculation showed no significant differences during biltong processing. Carnobacterium spp. would be an effective nonpathogenic in-plant surrogate to monitor microbial safety that mimics the response of pathogenic bacteria to validate biltong processing within a manufacturer’s own facility.


Introduction
Biltong is a South African style dried beef product that is growing in popularity in the United States. This dried meat product is traditionally made using lean strips of beef that are marinated in a mixture of traditional spices (coriander and pepper), salt, and vinegar and then dried at low or ambient temperature and humidity. Dried beef processing guidelines, as issued by the United States Department of Agriculture Food Safety and Inspection Service (USDA-FSIS), require dried beef products to be heated to an internal temperature of 160 • F (71.1 • C) in a sealed oven or steam injector with a relative humidity greater than 90% during the cooking/heating process [1]. Since biltong does not have a heat lethality step during processing and deviates from these guidelines, biltong manufacturers must conduct a validation or challenge study to evaluate the ability of their Bacterial isolates obtained from previous biltong beef trials after marination and drying for 8 days at 24.9 • C (75 • F) and 55% relative humidity (RH) were identified by 16S rRNA PCR/sequencing [24] as Carnobacterium gallinarum, Carnobacterium divergens, and Latilactobacillus sakei for examination as biltong process surrogates (Table 1).
Other lactic acid bacteria used in this study included Pedicoccus acidilatcici ATCC 8042, P. acidilactici P02K5, P. pentosaceus FBB61-2, and P. pentosaceus ATCC 43200, which are maintained in our laboratory culture collection. Some of these strains have been evaluated in other surrogate studies [19,25]. Nonpathogenic E. coli ATCC BAA-1427, BAA-1428, BAA-1429, and BAA-1430 have been used as Biotype 1 surrogate strains in various process validation studies and recommended for such use by USDA-FSIS [8,20,26]. P. acidilactici Saga200, used as a protective starter culture, was obtained as a frozen slurry from Kerry Foods (Beloit, WI, USA). Enterococcus faecium 201224-016 was obtained as a freeze-dried powder from Vivolac Cultures (Indianapolis, IN, USA) and is sold as a probiotic.
Carnobacterium spp., E. faecium, and E. coli cultures were inoculated into tryptic soy broth (TSB, BD Bacto, Franklin Laes, NJ, USA) and grown at 30 • C for 24 h. L. sakei and Pedicoccus spp. were inoculated into De Man, Rogosa and Sharpe broth (MRS, BD Bacto) and grown at 30 • C for 24 h. Cultures were prepared for storage by centrifugation (7200× g, 5 • C) of 9 mL of fresh, overnight culture, and the resulting pellet was resuspended with 2-3 mL of fresh, sterile TSB or MRS broth containing 10% glycerol. The cells in freezing media were then placed in 8 mL sterile glass vials and stored in an ultralow-temperature freezer (−80 • C) until use. Prior to use, frozen stocks were revived by transferring 100 µL of partially thawed culture into 9 mL of either TSB or MRS broth and incubated overnight at 30 • C.
Several cultures were used directly after suspension from the freeze-dried or frozen state for comparison of biltong process performance with metabolically active forms grown in liquid media. Prior to use, P. acidilactici Saga200 (frozen) was resuspended by adding 0.5 g of the frozen culture to 9 mL of 0.1% buffered peptone water (BPW, BD Difco) and vortexing until completely incorporated. E. faecium 201224-016 was resuspended by adding 0.1 g of the freeze-dried culture to 9 mL of 0.1% BPW and vortexing until completely mixed.

Acid Adpation of Cultures
Acid adaptation of active four-strain mixtures of Carnobacterium spp., Pediococcus spp., and E. coli BAA-strains was conducted as first described by Wilde et al. [27] and as used in previous biltong studies [4,28]. In brief, individual cultures were inoculated into TSB or MRS containing 1% glucose, incubated overnight at 30 • C, and harvested by centrifugation; cell pellets were then resuspended with 0.1% BPW. For mixed culture biltong inocula, individual strains were cultured, centrifuged, resuspended, and then combined in equal proportions to create a mixed inoculum cocktail. The commercial starter cultures (P. acidilactici Saga200 and E. faecium 201224-016) were not acid-adapted and used as a single-strain inoculum.

Beef Sample Preparation and Inoculation
USDA select-grade boneless beef rounds were obtained from a local meat processor (Ralph's Perkins, OK, USA) who obtains beef from a wholesale beef broker. Beef rounds were trimmed of fat and cut into approximately 5.1 cm wide × 1.9 cm thick × 7.6 cm long beef squares and held overnight at 5 • C on foil-lined trays wrapped in plastic bags. Beef pieces were inoculated the following morning with the respective inoculum depending on the trials being performed that day. Beef pieces were inoculated with the Carnobacterium spp. mixture (C. divergens GO-R2E-B, GO-R1B; C. gallinarum NB-R2A, NB-R2B), the L. sakei mixture (L. sakei GO-R2C, GO-R2D), the Pediococcus spp. mixture (P. acidilactici ATCC 8042, PO2K5; P. pentosaceous ATCC 43200, FBB61-2), P. acidilactici Saga200, or E. faecium 201224-016. The inoculum suspension (150 µL) was applied to each side of the beef pieces and immediately spread with a gloved finger. Inoculated beef pieces were then allowed to incubate for 30 min at 5 • C to allow for bacterial attachment prior to use.

Biltong Processing, Marination, and Drying
Biltong processing was conducted as described whereby trials were performed in duplicate and triplicate samples were harvested at each timepoint (n = 6) [4,29]. Following inoculation and attachment, the beef pieces were then dipped in sterile water to mimic rinse treatments that processors often apply using antimicrobials or water during processing. The inoculated pieces were placed in a plastic basket, dipped in sterile water in a stainless-steel tub for 30 s, and drained for 60 s to release excess liquid. The beef pieces were then placed into a chilled metal tumbling vessel containing a biltong marinade. The biltong marinade consisted of 2.2% salt, 0.8% black pepper, 1.1% coarse ground coriander, and 4% red wine vinegar (100-grain; 10% acetic acid) in relation to the total meat weight. Beef pieces were vacuum-tumbled (15 inches Hg) in a Biro VTS-43 vacuum-tumbler (Marblehead, OH, USA) for 30 min and then hung to dry in a humidity-controlled oven (Hotpack, Model 435315, Warminster, PA, USA) at 55% relative humidity and 24.9 • C (75 • F) for 8-10 days.

Selective Recovery of Inoculum Bacteria from Biltong-Inoculated Beef
The bacteria assessed in this study as potential biltong processing surrogates were inoculated onto raw beef, and initial and residual inoculum enumeration had to preclude other natural contaminants also found on raw beef, those contributed during trimming of beef, or from the marinade spice mix. Prior studies indicated that such processing conditions induce stresses, and injured cells may not be recovered on harsh selective media, thereby giving a falsely lower count [28]. To eliminate the possibility of inhibiting injured-but-viable cells, we used generic growth media (TSA, MRS agar) supplemented with antibiotics to which the strains are resistant as a selective medium to enumerate our inoculated organisms from samples taken during biltong processing [4,28]. Antibiotic resistance was determined using antibiotic susceptibility discs (BD BBL Sensi-Discs, BD Labs, Franklin Lakes, NJ, USA) to determine innate antibiotic resistance (Table 1). After identification of antibiotic resistances, cultures were then enumerated on media with and without antibiotics to ensure the absence of inhibition from the use of antibiotics in the media as described previously [4,28,30]. For some strains used as inoculum cocktails that did not have consensus of the same antibiotic resistances, antibiotic resistance was acquired by plating on low level antibiotics known to generate spontaneous antibiotic resistance (i.e., gentamycin and rifamycin).

Comparison of Commerically Available Starter Cultures as Biltong Inoculants in their
Lyophilized and Metabolically Active Forms 2.6.1. Culture Preparation Lactic acid bacteria obtained as freeze-dried cultures from starter culture companies for use in validation studies may present a facile method of use as validation inocula by simply resuspending the cells in buffer and directly inoculating beef samples [15,17]. Freeze-drying or lyophilization of bacteria exposes them to stressful conditions that can affect subsequent cell viability or activity [20,21]. Therefore, the activity of lyophilized (E. faecium 201224-016) and frozen (P. acidilactici Saga200) starter cultures and their metabolically active forms (i.e., after growth in media) were compared in their response to biltong processing.
For the lyophilized culture (E. faecium 201224-016), 0.1 g of freeze-dried powder was added to 9 mL of sterile 0.1% BPW and vortexed until completed suspended. The resuspended mixture was then used to inoculate each beef piece (300 µL; 150 µL/side) prior to marination.
For the frozen starter culture (P. acidilactici Saga200), a sterile hollow hole puncher was used to core~0.8 g of frozen Saga200 from the manufacture's container which was added to 9 mL of sterile 0.1% BPW and vortexed until mixed. The culture suspension was kept chilled on ice and used shortly thereafter to inoculate beef pieces.
Metabolically active versions of these cultures were obtained by growth in 150 mL of the appropriate media (TSB, MRS) for 24 h at 30 • C, centrifugation, and resuspension of the recovered cell pellet with 5 mL of sterile 0.1% BPW. The resuspended culture was then used to inoculate beef pieces prior to use in the validation study. The lyophilized and metabolically active forms of E. faecium 201224-016 and P. acidilactici Saga200 were used in parallel and simultaneous biltong trials to reduce any variables that might influence the observed effect of the marinade and drying process.

Lyophilization of Carnobacterium gallinarum NB-R2A
To the authors' knowledge, there is no commercially available Carnobacterium strain available in the United States. Therefore, C. gallinarum NB-R2A, isolated from biltong, was lyophilized via freeze-drying to examine a lyophilized version for comparison with the actively grown culture. Carnobacterium gallinarum NB-R2A was inoculated into 9 mL of TSB from frozen stock and incubated for 18 h at 30 • C. Following incubation, the 9 mL culture was transferred to 190 mL of TSB and incubated again for 18 h at 30 • C. The culture was then centrifuged at 7200× g for 20 min. The supernatant was removed, and the cell pellet was resuspended with 5 mL of sterile BPW and repeated. The supernatant was removed following centrifugation, and the final cell pellet was resuspended with 10 mL of autoclaved milk-based freeze-drying medium consisting of 11 g of skim milk powder, 1 g of dextrose, 1 g of trehalose, and 0.2 g of yeast extract per 100 mL. The milk/cell suspension was added to Oak Ridge tubes (5 mL each) and freeze-dried using a Heto vacuum centrifuge (Model VR-maxi) connected to a Heto freezing condensor (Model CT 60E) and a Leybold Trivac vacuum pump (Model D2.5F) setup for 24 h under vacuum. The freeze-dried powder was then stored at −80 • C until use in our biltong study. Just before use, 0.25 g of powder was added to 9 mL of sterile 0.1% BPW, vortexed until mixed, and used to inoculate beef pieces for biltong processing.

Water Activity
Uninoculated beef pieces were sampled for water activity (A w ) measurements at various stages throughout processing (in triplicate) including the initial raw beef, beef after marination, and then beef after drying for 2, 4, 6, 8, and 10 days. To obtain measurements, beef pieces were cut in half and placed in a sampling cup with the interior portion of the sample facing upward (toward the sensor). Samples were then covered with sampling cup cover containing the sensor and allowed to equilibrate to the temperature of the room. Water activity was measured using a HC2-AW-USB probe with a direct PC interface and HW4-P-Quick software (Rotronic Corp., Hauppauge, NY, USA). Measurements were taken in triplicate for each sample at each timepoint.

Moisture Loss
Following marination, each beef piece was individually weighed and labeled prior to being hung in the humidity-controlled oven. Three pieces were selected and weighed prior to processing, and then sampled every 2 days while drying. The weight at the time of sampling was compared to the initial weight of the same piece recorded prior to drying. The determination of percent moisture loss was calculated as per Equation (1).

Measurement of Biltong Beef pH
Measurements of beef pH were obtained at various points in the biltong process including raw beef, beef following marination, and beef after 2, 4, 6, 8, and 10 days of drying. At each timepoint, three pieces of uninoculated beef were collected, weighed, and then added to a laboratory blender with steel blades (Waring Commercial, New Harford, CT, USA) with sterile water of equal weight to the weight of the beef pieces. The water and beef mixtures were blended until a finely ground mixture was formed. The pH of the homogenized meat mixture was measured in triplicate using an H-series pH meter and probe (Hach, Loveland, CO, USA).

Salt Concentration
The homogenized meat mixture used to measure pH was also used to obtain salt concentrations of each sample. Horiba LAQUA Twin Pocket Meter (Horiba Instruments, Irvine, CA, USA) was used to quantify sodium ion concentration. Approximately 300 µL of the homogenized sample was placed in the sample chamber and allowed to stabilize before recording. Readings (in ppm) were taken in triplicate for each sample. To determine the salt (NaCl) concentration from the sodium ion concentration, the following equations were used: NaCl Salt g 100 g = Na mg 100 g × NaCl molar mass Na molar mass × 1 1000 . (3)

Microbial Sampling and Inoculum Enumeration of Biltong Beef
At each sampling timepoint of biltong beef processing (raw beef, after marinade, and after every 2, 4, 6, 8, and 10 days of drying), three beef pieces were selected at random and placed in a sterile Whirl-pak filter stomaching bag (Nasco, Fort Atkinson, WI, USA) in combination with 100 mL of 1% neutralizing buffered peptone water (nBPW, Criterion, Hardy Diagnostics, Santa Maria, CA, USA). Samples were stomached for 60 s in a paddleblender masticator (IUL Instruments, Barcelona, Spain). Serial dilutions were made with 1% BPW and plated on TSA containing gentamicin and rifamycin (2.5 µg/mL each) for Carnobacterium, on MRSA containing gentamicin and rifamycin (2.5 µg/mL each) for L. sakei, on MRSA containing gentamicin (10 µg/mL) and rifamycin (5 µg/mL) for Pediococcus spp., on TSA containing naldixic acid and colistin (10 µg/mL each) for E. faecium 201224-016, or on MRS containing nalidixic acid and colistin (10 µg/mL each) for P. acidilactici Saga200; the filter bag dilution was considered the 10 • dilution. Plates were incubated at 30 • C for 48 h and enumerated as log CFU/mL. Samples were collected in triplicate replication and plated in duplicate at each sampling timepoint.

Statistical Analysis
Validation trials were conducted in duplicate with triplicate sampling at each timepoint (n = 6) as per validation criteria established by the National Advisory Committee on Microbial Criteria for Foods (NACMCF) [9] and supported by the USDA-FSIS [31]. Data are presented as the mean of multiple replications with standard deviation of the mean represented by error bars. Statistical analysis of data collected over time was performed using one-way repeated-measures analysis of variance (RM-ANOVA). Pairwise multiple comparisons were performed using the Holm-Sidak test to determine significant differences. Data treatments with the same letter are not significantly different (p > 0.05); treatments with different letters are significantly different (p < 0.05).

Water Activity, Moisture Loss, and Salt Concentrations
To complement the surrogate validation trials, we measured and recorded critical operational parameters and intrinsic factors at each key stage of processing (raw beef, inoculation, marination, and every 2 days of drying) as recommended by USDA-FSIS [1]. Water activity (A w ) is a measure of free, unbound water available for bacterial growth. USDA-FSIS considers vacuum-tumbled beef as 'nonintact beef', whereby A w is a primary safety factor as there is no heat lethality step in biltong processing and biltong is processed as thick beef samples [32,33]. Therefore, A w is a critical safety factor for control of bacteria that might be internalized due to vacuum tumbling. S. aureus that can tolerate low A w and high salt levels would be a concern for possible production of staphylococcal enterotoxin. The targeted A w for shelf-stable beef jerky is <0.85 which was achieved after 7 days of drying ( Figure 1) [1,2]. Water activity after 8 and 10 days of drying ranged from 0.82 to 0.79 respectively. Similarly, beef samples showed incremental moisture loss with 59% and 62.5% loss at 8 and 10 days, respectively ( Figure 1).
Salt concentration was also determined during the biltong process. Salt concentration was calculated from sodium readings obtained with the LAQUAtwin NA-11 sodium ion meter (Horiba Inc, Irvine, CA, USA). The initial calculated salt concentration determined on raw beef was 0.12% NaCl; then, following the marination step, the beef salt concentration shot up to 2.17% (2.17 g NaCl/100 g beef). The initial salt level falls in line with expectations given that the biltong marinade was formulated at 2.2% salt (w/w). The salt concentration increased over time and was indirectly proportional to moisture loss during the drying process ( Figure 2). As expected, as moisture loss occurred, A w was also reduced to below 0.85 A w (Figure 1) and the salt concentration increased to above 4% ( Figure 2); both conditions are inhibitory to most bacteria, helping to ensure a safe product for consumers [34]. Biltong safety involves an interplay among moisture, salt concentration, and A w since moisture loss increases salt concentration, while salt binds water and helps to draw it out of the interior of the beef, thereby reducing A w . For consumer issues regarding high sodium levels, the use of alternative salts (CaCl 2 , KCl) instead of NaCl can help lower sodium levels in finished biltong while still maintaining a 5 log reduction of pathogen (Salmonella) [29]. Salt concentration was also determined during the biltong process. Salt concentration was calculated from sodium readings obtained with the LAQUAtwin NA-11 sodium ion meter (Horiba Inc, Irvine, CA, USA). The initial calculated salt concentration determined on raw beef was 0.12% NaCl; then, following the marination step, the beef salt concentration shot up to 2.17% (2.17 g NaCl/100 g beef). The initial salt level falls in line with expectations given that the biltong marinade was formulated at 2.2% salt (w/w). The salt concentration increased over time and was indirectly proportional to moisture loss during the drying process ( Figure 2). As expected, as moisture loss occurred, Aw was also reduced to below 0.85 Aw (Figure 1) and the salt concentration increased to above 4% ( Figure 2); both conditions are inhibitory to most bacteria, helping to ensure a safe product for consumers [34]. Biltong safety involves an interplay among moisture, salt concentration, and Aw since moisture loss increases salt concentration, while salt binds water and helps to draw it out of the interior of the beef, thereby reducing Aw. For consumer issues regarding high sodium levels, the use of alternative salts (CaCl2, KCl) instead of NaCl can help lower sodium levels in finished biltong while still maintaining a 5 log reduction of pathogen (Salmonella) [29]. The initial pH of the raw meat pieces was approximately pH 5.43 (Figure 3), which was determined by blending beef samples in sterile water in a laboratory blender. The pH of the samples then decreased following the marination step to 5.02, which can be attributed to the presence of residual 100-grain red wine vinegar in the marinade. After

The pH of Beef during Biltong Processing
The initial pH of the raw meat pieces was approximately pH 5.43 (Figure 3), which was determined by blending beef samples in sterile water in a laboratory blender. The pH of the samples then decreased following the marination step to 5.02, which can be attributed to the presence of residual 100-grain red wine vinegar in the marinade. After removal from the marinade, the pH of biltong beef samples then equilibrated slightly higher to~5. 18-5.20 for the remainder of the drying process in the humidity-controlled oven (Figure 3). The pH of the marination solution was much lower (pH 2.5-2.7); during 30 min vacuum tumbling, the surface bacteria were immersed in the low-pH marinade solution, which could lead to cell death and inactivation of pathogenic bacteria [35,36], as observed in the current study and prior biltong trials where levels of inoculated pathogens were reduced after marination [4]. After removal from the vacuum tumbler, the residual marinade on the surface was absorbed, and the pH of biltong beef samples equilibrated to~5. 18-5.20 for the remainder of the drying process in the humidity oven ( Figure 3).
Microorganisms 2022, 10, x FOR PEER REVIEW 10 of 16 Figure 3. The pH of meat at each sampling timepoint during biltong processing. Samples were taken in triplicate at each timepoint following blending with sterile water in a laboratory blender (n = 6).

Temperature and Relative Humidity during Biltong Processing
Temperature and RH measurements were recorded by computer software connected to the handheld temperature and humidity recorders to which the probes in the oven chamber were connected (Figure 4). Two temperature probes were inserted separately into two beef pieces to measure the internal beef temperature during processing, while the humidity probe was placed midway within the chamber. Air temperature and humidity were set to 23.9 °C (75 °F) and 55% throughout the duration of each trial but cycled above and below the set points. The internal temperature of the beef was more consistent and steadily increased from their initial temperature to match the temperature of the chamber. Long-term storage at low RH helps to evaporate moisture from the beef.

Temperature and Relative Humidity during Biltong Processing
Temperature and RH measurements were recorded by computer software connected to the handheld temperature and humidity recorders to which the probes in the oven chamber were connected (Figure 4). Two temperature probes were inserted separately into two beef pieces to measure the internal beef temperature during processing, while the humidity probe was placed midway within the chamber. Air temperature and humidity were set to 23.9 • C (75 • F) and 55% throughout the duration of each trial but cycled above and below the set points. The internal temperature of the beef was more consistent and steadily increased from their initial temperature to match the temperature of the chamber. Long-term storage at low RH helps to evaporate moisture from the beef. Oven temperature and relative humidity measurements. The temperature was set to 24.9 °C (75 °F), and the relative humidity setpoint was 55% RH during the drying process over a period of 10 days. Graphical data show the typical cycling of oven control above/below setpoint. Two temperature probes were placed in various places in the chamber and two additional probes were inserted into separate pieces of beef to track the internal temperature of the biltong product over the same drying period.

Surrogate Log Reductions during Biltong Processing
Various bacteria were considered for examination as possible nonpathogenic surrogates, including strains recovered from biltong after processing. These included a twostrain mixture of L. sakei GO-R2C and GO-R2D and a four-strain mixture of C. divergens GO-R2E-B and GO-R1B and C. gallinarum NB-R2A and NB-R2B ( Figure 5). We also examined a four-strain mixture of P. acidilactici and P. pentosaceous strains (P. acidilactici ATCC 8042 and PO2K5; P. pentosaceous ATCC 43200 and FBB61-2) vs. starter cultures that were available through culture companies (E. faecium 201224-016 and P. acidilactici Saga200) as surrogate organisms ( Figure 5).
Only a slight reduction from inoculated levels was observed following vinegar/spice/salt marination (0.65, 0.58, 0.75, and 0.61 log reduction) with all cultures used, except for the four-strain mixtures of Carnobacterium spp. and E. coli ATCC BAA series ( Figure 5). A larger log reduction was observed after marination of the four-strain mixtures of Carnobacterium spp. (1.23 log) and E. coli ATCC BAA-strains (0.86 log) ( Figure 5). Trials using E. coli ATCC BAA (four-strain mix), L. sakei (two-strain mix), Pediococcus spp. (four-strain mix), E. faecium 201224-016, and P. acidilactici Saga200 failed to achieve a 5 log reduction during biltong processing with overall reductions of 4.86 log, 2.03 log, 1.87 log, 1.68 log, and 1.83 log respectively. Of all the nonpathogenic strains examined, only the four-strain mixture of Carnobacterium spp. achieved an overall reduction of greater than 5 log (5.85 log) during the 8 day drying period (Figure 5). On the basis of these results, Carnobacterium spp. were the only organisms that achieved a 5 log reduction (within 6-8 days) comparable to that observed for the pathogenic strains, and they presented the best case for use as a Salmonella, L. monocytogenes, E. coli O157:H7, or S. aureus surrogate for biltong processing ( Figure 5). Oven temperature and relative humidity measurements. The temperature was set to 24.9 • C (75 • F), and the relative humidity setpoint was 55% RH during the drying process over a period of 10 days. Graphical data show the typical cycling of oven control above/below setpoint. Two temperature probes were placed in various places in the chamber and two additional probes were inserted into separate pieces of beef to track the internal temperature of the biltong product over the same drying period.

Surrogate Log Reductions during Biltong Processing
Various bacteria were considered for examination as possible nonpathogenic surrogates, including strains recovered from biltong after processing. These included a two-strain mixture of L. sakei GO-R2C and GO-R2D and a four-strain mixture of C. divergens GO-R2E-B and GO-R1B and C. gallinarum NB-R2A and NB-R2B ( Figure 5). We also examined a four-strain mixture of P. acidilactici and P. pentosaceous strains (P. acidilactici ATCC 8042 and PO2K5; P. pentosaceous ATCC 43200 and FBB61-2) vs. starter cultures that were available through culture companies (E. faecium 201224-016 and P. acidilactici Saga200) as surrogate organisms ( Figure 5).
Only a slight reduction from inoculated levels was observed following vinegar/spice/salt marination (0.65, 0.58, 0.75, and 0.61 log reduction) with all cultures used, except for the four-strain mixtures of Carnobacterium spp. and E. coli ATCC BAA series ( Figure 5). A larger log reduction was observed after marination of the four-strain mixtures of Carnobacterium spp. (1.23 log) and E. coli ATCC BAA-strains (0.86 log) ( Figure 5). Trials using E. coli ATCC BAA (four-strain mix), L. sakei (two-strain mix), Pediococcus spp. (four-strain mix), E. faecium 201224-016, and P. acidilactici Saga200 failed to achieve a 5 log reduction during biltong processing with overall reductions of 4.86 log, 2.03 log, 1.87 log, 1.68 log, and 1.83 log respectively. Of all the nonpathogenic strains examined, only the four-strain mixture of Carnobacterium spp. achieved an overall reduction of greater than 5 log (5.85 log) during the 8 day drying period (Figure 5). On the basis of these results, Carnobacterium spp. were the only organisms that achieved a 5 log reduction (within 6-8 days) comparable to that observed for the pathogenic strains, and they presented the best case for use as a Salmonella, L. monocytogenes, E. coli O157:H7, or S. aureus surrogate for biltong processing ( Figure 5). Microorganisms 2022, 10, x FOR PEER REVIEW 12 of 16 Figure 5. Composite graph of biltong processing data of nonpathogenic bacteria attempting to mimic the biltong process log reduction of pathogenic bacteria (light-gray lines) to be considered a possible 'biltong processing surrogate' organism for in-plant validation. Log reduction curves of various lactic acid bacteria (Carnobacterium spp., Pediococcus spp., L. sakei, and E. faecium) and Biotype I E. coli strains tested as potential surrogate organisms for biltong processing over a period of 8-10 days. Strains were compared to the log reduction curves observed during previous biltong validation studies using pathogenic bacteria including Salmonella serovars [4], S. aureus, E. coli O157:H7, and L. monocytogenes [5]. Data points are the mean of duplicate trials sampled in triplicate (n = 6).
Statistical analysis was performed using one-way repeated-measures analysis of variance (RM-ANOVA) of the entire time course of data; curves with the same letter are not significantly different (p > 0.05); isolates with different letters are significantly different (p < 0.05).

Comparison of Lyophilized/Frozen Starter Cultures with Metabolically Active (Grown) Versions in Biltong Processing Trials
Several reports in the literature have used freeze-dried or frozen cultures, resuspended directly in buffer, to inoculate food samples in process trials for direct comparison to pathogens grown in microbiological media (which we describe as 'active cultures') [15,17]. The ease of availability of freeze-dried/frozen cultures from culture companies would facilitate the use of such cultures for in-plant validation studies; however, we were interested to see if they could provide the same response in a biltong process as the actively grown cultures ( Figure 6). The comparisons were between two commercially available starter cultures, E. faecium 201224-016 (Vivolac Cultures; freeze-dried) and P. acidilactici Saga200 (Kerry Foods; frozen), and a lyophilized C. divergens NB R2A, which was Figure 5. Composite graph of biltong processing data of nonpathogenic bacteria attempting to mimic the biltong process log reduction of pathogenic bacteria (light-gray lines) to be considered a possible 'biltong processing surrogate' organism for in-plant validation. Log reduction curves of various lactic acid bacteria (Carnobacterium spp., Pediococcus spp., L. sakei, and E. faecium) and Biotype I E. coli strains tested as potential surrogate organisms for biltong processing over a period of 8-10 days. Strains were compared to the log reduction curves observed during previous biltong validation studies using pathogenic bacteria including Salmonella serovars [4], S. aureus, E. coli O157:H7, and L. monocytogenes [5]. Data points are the mean of duplicate trials sampled in triplicate (n = 6). Statistical analysis was performed using one-way repeated-measures analysis of variance (RM-ANOVA) of the entire time course of data; curves with the same letter are not significantly different (p > 0.05); isolates with different letters are significantly different (p < 0.05).

Comparison of Lyophilized/Frozen Starter Cultures with Metabolically Active (Grown) Versions in Biltong Processing Trials
Several reports in the literature have used freeze-dried or frozen cultures, resuspended directly in buffer, to inoculate food samples in process trials for direct comparison to pathogens grown in microbiological media (which we describe as 'active cultures') [15,17]. The ease of availability of freeze-dried/frozen cultures from culture companies would facilitate the use of such cultures for in-plant validation studies; however, we were interested to see if they could provide the same response in a biltong process as the actively grown cultures ( Figure 6). The comparisons were between two commercially available starter cultures, E. faecium 201224-016 (Vivolac Cultures; freeze-dried) and P. acidilactici Saga200 (Kerry Foods; frozen), and a lyophilized C. divergens NB R2A, which was chosen from among the Carnobacterium mixed strains demonstrating >5 log reduction in Figure 5. spectively ( Figure 6). The lyophilized single strain C. divergens NB R2A also showed no significant difference from the metabolically active culture and again achieved 5 log reduction during the biltong process ( Figure 6). The data show that lyophilized or frozen versions of E. faecium, P. acidilactici, or C. gallinarum do not respond differently than actively grown cultures to biltong processing conditions and, when possible, their use might facilitate inoculated studies.

Conclusions
The lethality observed in the biltong process with Carnobacterium spp. aligned with that observed with four major pathogenic organisms indicating that Carnobacterium spp. could be an effective in-plant surrogate organism to monitor the effectiveness of biltong processing within a manufacturer's facility. Enterococcus faecium, L. sakei, and Pediococcus spp. were not reduced much (<2 log) and were resilient toward the acid, salt, and low Aw experienced during 10 days of biltong processing. The use of lyophilized/frozen cells as inoculum for validation of biltong processing was not significantly different than using actively grown cells. This work helps to fill USDA-FSIS knowledge gaps in air-dried shelf- Figure 6. Biltong processing of beef inoculated with lyophilized/frozen cells vs. metabolically active cells (freshly grown) of E. faecium 201224-016, P. acidilactici, and C. gallinarum NB-R2A. Lyophilized C. gallinarum NB-R2A was compared to a four-strain cocktail of metabolically active C. divergens/gallinarum. Graph curves of frozen or lyophilized cultures have hollow symbols. Statistical analysis was performed using one-way repeated-measures analysis of variance (RM-ANOVA) over the entire time course of the datasets; graphs with the same letter are not significantly different (p > 0.05); isolates with different letters are significantly different (p < 0.05).
Neither the lyophilized version of E. faecium 201224-016 (1.43 log reduction) nor the frozen version of P. acidilactici Saga200 (1.54 log reduction) achieved the 5 log reduction target; survival curves of the lyophilized/frozen forms were also not significantly different when compared to their metabolically active forms, i.e., 1.68 and 1.83 log reduction, respectively ( Figure 6). The lyophilized single strain C. divergens NB R2A also showed no significant difference from the metabolically active culture and again achieved 5 log reduction during the biltong process ( Figure 6). The data show that lyophilized or frozen versions of E. faecium, P. acidilactici, or C. gallinarum do not respond differently than actively grown cultures to biltong processing conditions and, when possible, their use might facilitate inoculated studies.

Conclusions
The lethality observed in the biltong process with Carnobacterium spp. aligned with that observed with four major pathogenic organisms indicating that Carnobacterium spp. could be an effective in-plant surrogate organism to monitor the effectiveness of biltong processing within a manufacturer's facility. Enterococcus faecium, L. sakei, and Pediococcus spp. were not reduced much (<2 log) and were resilient toward the acid, salt, and low A w experienced during 10 days of biltong processing. The use of lyophilized/frozen cells as inoculum for validation of biltong processing was not significantly different than using actively grown cells. This work helps to fill USDA-FSIS knowledge gaps in air-dried shelfstable dried beef (biltong) processing with regard to potential surrogate organisms and critical factors involved in the biltong process. Future studies on biltong processing may include whether pathogens such as Salmonella, known to survive long periods of low water activity, can survive the extended shelf-life of biltong products to ensure that this does not become a possible (overlooked) problem.