Listeria monocytogenes Growth and Sensory Properties of Ready-to-Eat Meat Strips Prepared With Sodium Nitrite, Celery Powder, and a Fruit-Spice Blend

Experimental design and data analysis

Microbiological analysis was conducted using inoculated, cooked meat strips. Three independent replicates were performed, with 5 samples per treatment, replicate, and time point, to develop L. monocytogenes growth curves. This resulted in a total sample size of 15 (N = 15) per treatment and time point, including the control. For microbiological analysis, results from the uncured control and each treatment were averaged and analyzed by sampling point.

A total of 90 untrained panelists participated in the study, with 30 panelists randomly assigned to each comparison group. Each panelist evaluated 6 triangle sets, yielding 180 triangle test evaluations per comparison group and a total of 540 responses across the study. Proximate analysis was performed with 20 samples per treatment. The control group was excluded from proximate and sensory analyses, as shown in Table 1.

All means were compared using a one-way analysis of variance (ANOVA) with a 95% confidence level, and significantly different means (P < 0.05) were determined using Tukey’s multiple comparison test. Proximate assay data, including fat, protein, and dry matter content, were also analyzed using ANOVA with Tukey’s mean separation. The triangle test was analyzed using a χ² test with a 95% CI. All statistical analyses were performed using R software (version 4.2.3).

Meat fabrication

All samples were cooked and processed at the Texas Tech University Gordon W. Davis Meat Science Laboratory. A patented, proprietary system was used to produce meat strips (Rodriguez Flores, 2019). Trimmed, denuded, and sliced inside round beef was used to fabricate meat strips. A brine solution was prepared containing water (48.46%), brown sugar (30.30%), sunflower oil (12.12%), salt (3.64%), and a blend of seasonings and spices (5.48%). Meat strips were subjected to 1 of 4 treatments: 0.52% celery powder, 0.25% NaNO₂ (resulting in a nitrite concentration of 150–155 ppm in the meat strips), 1% of FSB (NATPRE T10, Prosur, Murcia, Spain), or an uncured control with no curing agent added. Control samples were injected with the same base brine (water, salt, and sugar) as the other treatments but contained no curing agents, ensuring consistency across treatments.

Brine was injected into the meat using a multineedle injector, achieving a brine pickup of 20%. Immediately after injection, the meat was cut into strips and tumbled under vacuum (10 Hg) at 5 rpm for 20 min. Muscles were then sliced with a knife into standardized dimensions: 2.5 cm in width by 12 cm in depth by their natural length. The sliced, uncooked strips were refrigerated overnight at 4°C to allow brine equilibration. No retumbling was performed.

For cooking, strips were arranged on a plastic grid placed over a rack and loaded onto an oven truck, ensuring no direct contact between samples. The truck was then placed inside a smokehouse, where the strips underwent a controlled cooking cycle with humidity, followed by a drying phase. Water activity (a_w) was monitored throughout the drying process, and strips were removed every 35 min until the a_w reached 0.90 plus or minus 0.01. Following inoculation, a_w increased to 0.94 due to the moisture introduced with the bacterial culture, temporarily creating conditions favorable for microbial growth despite the product’s initial low a_w. The overall processing yield was 15%, indicating the extent of moisture loss during the cooking and drying phases.

As a reduced-moisture product, the final meat strips had a pH of 6.1 plus or minus 0.1, maintaining stability post-processing. Once fully cooked, strips were individually cooled and vacuum packaged in a 6-mL oxygen- and moisture-barrier film. Postpackage pasteurization (PPP) was performed by submerging the vacuum-sealed samples in a water bath at 82°C for 6 min. Packaged beef strips designated for microbial analysis were then transported to the International Center for Food Industry Excellence Food Microbiology Laboratory at Texas Tech University for further testing. The product evaluated in this study was a shelf-stable beef strip designed for room-temperature storage (similar to meat sticks or jerky), rather than a product requiring refrigeration. Therefore, the selected storage temperature (32°C) was appropriate for modeling potential temperature abuse scenarios within the expected storage environment of this product type.

Cocktail preparation for L. monocytogenes

Frozen stock cultures of L. monocytogenes strains 19118, Scott A, and F5069 were removed from the −80°C freezer and used for cocktail preparation. Each strain was streaked onto brain heart infusion (BHI) agar and incubated at 37°C for 24 h. A 10-μL loop was used to transfer a single colony from each strain into separate 9-mL BHI broth tubes, followed by incubation at 37°C for 24 h. A subsequent 1-mL transfer into fresh BHI tubes was performed to achieve a final concentration of approximately 9 Log CFU/mL.

After 2 consecutive transfers, the individual cultures were combined into 50-mL Falcon tubes to create a multistrain cocktail. The L. monocytogenes cocktail was serially mixed in phosphate-buffered saline to the target concentration. The final inoculum (5 Log CFU/mL) was adjusted using a Thermo Scientific^™ 0.5 Polymer McFarland Standard with a nephelometer for calibration (Flynn et al., 2021).

Sample inoculation

Vacuum-sealed beef-strip packages were reopened at the laboratory following PPP, and samples were preweighed (25 ± 2 g). Each strip was then inoculated with 250 μL of the multistrain L. monocytogenes cocktail and allowed to attach for 20 min in a fresh 6-mL polybag. This inoculation step was performed after PPP to simulate post-processing contamination, ensuring that viable L. monocytogenes cells were present for subsequent microbial analysis. Water activity (a_w) increased from 0.90 to 0.94 following inoculation, temporarily creating an environment favorable for L. monocytogenes growth, as this pathogen typically requires an a_w above 0.92 to proliferate. This increase reflects realistic post-lethality contamination scenarios, where the introduction of moisture could enable pathogen growth. The bacterial cell density of the inoculum after attachment was approximately 2 Log CFU/g to 3 Log CFU/g.

This inoculum level was selected based on established recommendations for growth inhibition studies, which suggest inoculation at 2 Log CFU/g to 3 Log CFU/g to ensure accurate enumeration at time 0 and a consistent distribution of the inoculum within the food matrix, while still allowing sufficient potential for growth during the study period (Glass et al., 2025). Inoculated strips were vacuum sealed to simulate post-processing L. monocytogenes contamination in meat processing facilities, a method commonly used to assess contamination risks and evaluate the effectiveness of interventions (Merialdi et al., 2015).

Sealed packages were stored under time-temperature abuse conditions (32°C) to promote microbial growth. Samples were removed at 0 h, 6 h, 12 h, 24 h, 48 h, 72 h, and 96 h for microbial enumeration. Each sample was mixed in 225 mL of buffered peptone water, manually homogenized for 30 s, and serially diluted. A 100-μL aliquot was plated in duplicate onto an overlay of tryptic soy agar and Oxford agar base modified with Moxalactam (Remel^™). Plates were incubated at 37°C for 18 h to 24 h, and colony counts were Log₁₀ transformed for analysis. The temperature abuse condition at 32°C was intentionally selected to accelerate L. monocytogenes growth, allowing for faster detection and growth curve modeling (Osek et al., 2022). The meat strip product was shelf stable, similar to jerky, and designed for room-temperature storage, eliminating the need for refrigeration. The 32°C incubation represented a worst case scenario, simulating potential contamination of RTE meat with L. monocytogenes under abusive storage conditions.

Triangle test using untrained, independent consumer panelists

Cooked beef-strip samples were prepared following previously described protocols and cut into 2-cm by 2-cm bite-sized portions. All samples were served at room temperature (20–25°C) in a sensory evaluation laboratory under red lighting to minimize visual bias. Each sample was labeled with a randomized 3-digit code to ensure blind evaluation of the 3 treatments: celery powder, FSB, and NaNO₂. Apple juice and unsalted crackers were provided as palate cleansers between evaluations.

A sensory triangle test was conducted to evaluate whether untrained panelists could detect sensory differences between treatment groups. Three treatment comparisons were evaluated independently: (1) celery powder vs. FSB, (2) celery powder vs. NaNO₂, and (3) FSB vs. NaNO₂.

Each triangle set included 3 coded samples, 2 of which were identical and 1 different. The position of the odd sample was randomized. Panelists were instructed to select the sample they perceived as different. Responses were recorded as, “correct,” when the odd sample was accurately identified and, “incorrect,” otherwise.

For each evaluation, 2 additional data points were collected: the specific treatment identified by the panelist as different (i.e., celery powder, FSB, or NaNO₂) and the reason for the perceived difference. Reasons were classified into 3 categories: “flavor,” “tenderness,” or “flavor/tenderness.” The proportion of correct responses for each treatment comparison was analyzed using a 1-sided binomial test, with significance set at α equal to 0.05 and assuming a 33.3% chance level expected by random guessing (American Meat Science Association, 2016).

Proximate analysis

Fat, moisture, ash, and protein content were analyzed using standardized methods. Lipid extraction was performed using the Bligh and Dyer method (Bligh and Dyer, 1959). Moisture and ash contents were determined via oven drying at 105°C and 550°C, respectively, following AOAC protocols (AOAC, 2007). Protein content was measured using a LECO FP928 combustion nitrogen determinator (LECO Corporation, 2022). Results were calculated based on standardized equations for fat, moisture, and ash percentages (FAO, 2003).

Plate counts of L. monocytogenes in beef strips

The growth of L. monocytogenes in beef strips across different treatments is presented in Table 2 and visualized in Figure 1. No significant differences (P > 0.05) were observed among treatments at 0 h, 6 h, and 12 h. However, at later time points, treatments showed different effects on microbial growth.

Figure 1.

L. monocytogenes count (Log CFU/g) in beef strips subjected to 4 different treatments: celery powder, control, sodium nitrite, and fruit- and spice-based blend over 96 h of storage at 32°C.

ANOVA, analysis of variance; FSB, fruit- and spice-based blend; HSD, honestly significant difference; NaNO₂, sodium nitrite.

The y-axis shows Log CFU/g, and the x-axis shows time (h).

P values (P < 0.05) represent statistically significant differences between treatments at each time point. Statistical differences were determined using 1-way ANOVA followed by Tukey’s HSD.

At 24 h, the FSB exhibited significantly different (P = 0.01) L. monocytogenes counts compared to the celery powder treatment but not compared to NaNO₂ or the control. By 48 h and 72 h, the FSB maintained lower L. monocytogenes counts compared to NaNO₂, celery powder, and the control (P < 0.05) but was not significantly different from NaNO₂ at 96 h. At 96 h, the FSB had significantly lower L. monocytogenes counts than celery powder and the control but did not differ significantly from NaNO₂ (Table 2, Figure 1). Figure 1 illustrates the growth trends of L. monocytogenes under various treatments. Growth entered the exponential phase at hour 6 and peaked between 24 h and 72 h, after which it transitioned into the death phase, as reflected in the decline in Log CFU/g values. Celery powder-treated samples supported the highest microbial growth, maintaining higher Log CFU/g values across all time points. The FSB slowed the growth of L. monocytogenes, as reflected by the delayed increase in counts over time compared to the control and celery powder treatments. However, this effect was not bactericidal, as L. monocytogenes was still able to grow under these conditions. Therefore, the observed effect is better described as growth suppression rather than true bacteriostasis.

The findings are consistent with previous studies, demonstrating the antimicrobial properties of natural fruit extracts and spice-derived compounds. Nieto et al. (2023) reported similar reductions in L. monocytogenes growth when applying vinegar and natural ingredients in RTE ham. These antimicrobial effects are likely attributed to organic acids, flavonoids, and terpenoids, which have been reported to inhibit bacterial proliferation in meat products (Olvera-Aguirre et al., 2023). Additionally, natural curing alternatives have been explored for their potential to maintain meat quality while controlling microbial growth (Pateiro et al., 2018; Karwowska and Dolatowski, 2016).

One limitation of this study is the absence of a compositional analysis of the FSB used. Although the antimicrobial effects observed are attributed to components commonly found in the natural blend’s composition (e.g., polyphenols, flavonoids, organic acids) and listed on the product label.

Proximate composition of ready-to-eat meat strips

The proximate composition of beef strips subjected to different curing treatments is presented in Table 3. Meat strips had a varied fat content (P = 0.03), with samples treated with celery powder showing the highest values. No significant differences (P = 0.08) were observed in moisture content across treatments, whereas ash content showed statistical differences (P < 0.001), with celery powder-treated samples presenting the highest values. Protein content was significantly different (P = 0.03) in the FSB compared to celery powder, but no significant difference was found between the FSB and NaNO₂.

Despite statistical differences across treatments, the mean values remained relatively close, suggesting that these variations are unlikely to have a significant impact on the overall nutritional composition. Similar findings have been reported in other studies, comparing natural and synthetic curing agents in processed meats. Jin et al. (2018) observed that sausages formulated with alternative curing agents, including celery powder and natural extracts, exhibited minor differences in protein, fat, and ash content but were overall comparable to those treated with nitrite. Similarly, Osakue et al. (2016) evaluated RTE fried chicken and found that different formulations had no major impact on proximate composition, reinforcing the idea that curing ingredients do not drastically alter the fundamental nutritional profile of processed meats. Additionally, McKeith (2014) highlighted that fruit- and vegetable-derived nitrates can be used as natural curing alternatives without compromising proximate composition.

Sensory analysis

Figure 2 illustrates the proportion of correct responses obtained from the triangle test conducted on beef strips formulated with celery powder, NaNO₂, and a FSB. A total of 180 comparisons per treatment set were evaluated. Results were analyzed using a binomial test to determine whether the proportion of correct identifications significantly exceeded the chance level of 1 out of 3 (33.3%). None of the comparisons yielded statistically significant results (P > 0.05), indicating that panelists could not consistently distinguish between the treatments. Specifically, the comparisons of celery powder vs. FSB (P = 0.11), celery powder vs. NaNO₂ (P = 0.992), and FSB vs. NaNO₂ (P = 0.336) all fell short of the threshold for perceptible sensory differences.

Figure 2.

Proportion of correct responses in triangle test comparisons for beef strips treated with celery powder, sodium nitrite, and a fruit- and spice-based blend.

FSB, fruit- and spice-based blend; NaNO₂, sodium nitrite.

Bar plot represents the proportion of panelists who correctly identified the odd sample across 3 treatment comparisons: celery powder vs. FSB, celery powder vs. NaNO₂, and FSB vs. NaNO₂. Statistical significance was assessed using binomial probability (α = 0.05), with P values shown above each bar.

Figure 3 provides a visual breakdown of the treatment selected as “different” in each triangle test. While panelists were asked to choose the sample they perceived as different, their selections were evenly distributed among the treatment options, suggesting that any selection was likely based on guesswork rather than a true sensory distinction. This reinforces the binomial test results, further supporting the conclusion that the sensory profiles of the treatments were indistinguishable under the test conditions.

Figure 3.

Panelist treatment selections from triangle tests comparing beef strips treated with celery powder, sodium nitrite, and a fruit- and spice-based blend.

FSB, fruit- and spice-based blend; NaNO₂, sodium nitrite.

Stacked bar chart displays the percentage of responses identifying each treatment as different in 3-pairwise comparisons. Each bar represents 180 responses, distributed by the treatment selected as odd. Color coding reflects treatment identification by panelists.

In addition, Figure 4 presents the reported reasons why panelists believed a sample was different when they selected one. Among those who perceived a difference, the majority cited “flavor” as the distinguishing characteristic, followed by combined responses of “flavor/tenderness” and then “tenderness” alone. Although these insights provide context into what attributes panelists were focusing on, they must be interpreted cautiously given the absence of significant discrimination between treatments. This suggests that perceived differences may have stemmed from panelist bias or random variation rather than actual formulation effects.

Figure 4.

Panelist-reported reasons for identifying treatment as different in triangle test comparisons.

Responses were categorized based on panelist comments and grouped by attribute: flavor, tenderness, or both. The majority of responses attributed perceived differences to flavor. Percentages are based on total responses from all triangle test comparisons.

The lack of perceptible differences aligns with previous findings that naturally cured meat products can mimic the sensory qualities of those cured with synthetic nitrites. For instance, Jin et al. (2018) reported that sausages made with celery powder, fruit extract powder, or purple sweet potato powder showed no significant sensory differences when compared to NaNO₂-cured sausages. Similarly, Sindelar et al. (2007) observed that hams cured with low concentrations (0.2%) of vegetable juice powder had sensory properties comparable to nitrite-cured hams, though higher concentrations introduced vegetal notes.

Given the increasing demand for natural curing agents, these findings highlight that both celery powder and the FSB evaluated here are viable alternatives to synthetic nitrites in terms of flavor and texture preservation. While Alahakoon et al. (2015) indicated that plant-based nitrite alternatives can introduce slight changes in taste or odor, our results suggest that the FSB used in this study does not compromise sensory quality and remains indistinguishable from both celery powder and traditional curing agents under the conditions of a triangle test.