Growth behavior comparison of Listeria monocytogenes between Type strains and beef isolates in raw beef

Abstract

This study was conducted to compare the growth parameters of Listeria monocytogenes between beef isolates and Type strains in raw beef. Beef was artificially inoculated with 3 Log CFU/g levels and growth was measured during storage at various temperatures (5–25 °C) using conventional plating methods. The R² value for lag time (λ) and specific growth rate (μ) were determined using modified-Gompertz model, which were greater than 0.98 at all storage temperature except at 5 °C. B_f, A_f, and RMSE showed acceptable ranges, showed that the models are suitable for the modeling the growth of L. monocytogenes. At all temperatures, the λ of L. monocytogenes beef isolates was shorter than that of the L. monocytogenes Type strains, and the μ of beef isolates was higher than that of Type strains. These results showed that growth pattern prediction of beef inoculated with L. monocytogenes beef isolates gives more actual results than with Type strains.

Introduction

Beef is easily contaminated by microorganisms [1] owing to its rich nutrients content, that includes high quality proteins, essential amino acids, lipids, minerals, and vitamins. In addition, the intrinsic characteristics of beef, such as pH and water activity, are also responsible for general spoilage caused by pathogenic bacteria [2]. In beef, microbial contamination can occur during slaughtering, packaging, processing, distribution, and storage; therefore, hygiene and safety measures of beef are primary concerns.

Listeria monocytogenes is a Gram-positive and facultative anaerobic bacteria. Listeriosis caused by L. monocytogenes can be critical and fatal especially to the vulnerable groups—the elderly, pregnant women, neonates, and immunocompromised adults [3]. It is considered important because of its severity and high death rates [2]. Raw beef is refrigerated at a temperature of 2–4 °C to control and prevent microbial growth [4]. However, L. monocytogenes can grow at a temperature as low as − 0.4 °C [5], and can survive mild food preservation treatment [6].

The traditional microbiological analysis methods are time-consuming and rapid enumeration method is expensive and also requires pre-enrichment steps for the detection of food-borne pathogens [7]. The prediction of microbial growth curve for food-borne pathogen in food products can be used for faster microbiological analysis, which is less expensive and can serve as an alternative method to these traditional microbiological analysis methods [8]. It can estimate the changes in microbial population in food products in a real time situation [9]. Predictive microbiology in food products is considered to be a useful tool for product development, risk assessment, Hazard Analysis and Critical Control Points (HACCP) system, and educational purpose [10].

The predictive model of L. monocytogenes in various food samples has been studied. And many researchers used the food isolates for growth modelling [11–13]. But, no studies have compared the growth pattern between Type cultures and food isolates in beef. Therefore, the purpose of this study was to compare the growth curve characteristics between L. monocytogenes isolates and Type cultures by predictive modeling in artificially inoculated beef.

Materials and methods

Isolation of L. monocytogenes presumptive from raw beef

Raw beef in dices (approximately 1 cm × 1 cm × 1 cm) were purchased from retail markets located in northern Seoul, Korea in July 2014. Totally 300 samples were gathered and microbiological analysis was conducted. Beef sample (25 g) was mixed with 225 ml of UVM-modified Listeria enrichment broth (UVM; Oxoid Ltd., Basingstoke, Hampshire, UK) and homogenized using a stomacher (AES Stomacher; AES Chemunex, France). After incubation at 30 °C for 24 h, 100 μL of cultured media was inoculated into 10 mL of Fraser Listeria broth (Oxoid Ltd., Basingstoke, Hampshire, UK), and cultured at 37 °C for 24 h. Following incubation, the cultured Fraser broth was streaked onto Oxford agar (Oxoid) and incubated at 30 °C for 24–48 h. From the Oxford agar, black colonies with halo were selected and confirmed by PCR and 16S rRNA gene sequencing.

Preparation of L. monocytogenes culture cocktail

Three of L. monocytogenes Type strains (ATCC 7644, 19113 and 19115; all human origin) and five of L. monocytogenes beef isolates were used in this study. All the strains were maintained in stock culture at − 80 °C in TSB containing 25% glycerol. For activation, 0.1 mL of the stock culture was inoculated into 10 mL of TSB and incubated at 30 °C for 24–48 h and centrifuged at 2200×g for 25 min at 20 °C. The final pellets were resuspended in 0.85% saline and combined to construct a culture cocktail for artificial inoculation in raw beef.

Identification of presumptive L. monocytogenes

PCR analysis was performed to identify the presumptive L. monocytogenes with strain specific primer for detecting hlyA gene. The sequences of the primers used were 5′-GCA GTT GCA AGC GCT TGG AGT GAA-3′, and 5′-GCA ACG TAT CCT CCA GAG TGA TCG-3′, and the expected size of the amplicon was 456 bp [14].

After confirmation of L. monocytogenes presumptive by PCR analysis, 16 s rRNA identification was conducted. The sequences of the universal primers were 5′-AGA GTT TGA TCC TGG CTC AG-3′ for 27F, and 5′-GGT TAC CTT GTT ACG ACT T-3′ for 1492R. The sequencing of 16 s rDNA was performed using an ABI PRISM 3730XL DNA analyzer (Applied Biosystems). Sequence similarity was searched using GenBank BLASTn (http://www.ncbi.nlm.nih.gov/BLAST/), and assembled by Phrap (http://www.phrap.org/).

Artificial inoculation and bacterial enumeration for growth modeling

Cocktail of L. monocytogenes Type culture and isolates were inoculated onto the raw beef (10 g) surface. The initial microbial loads of L. monocytogenes were 3 log CFU/g. The inoculated samples were aerobically stored at 5, 10, 15, 20, and 25 °C. After storage, 90 mL of saline solution was added to a 10 g sample, homogenized using a stomacher, and plated onto Oxford agar. After the incubation at 30 °C for 24–48 h, black colonies on the plates were enumerated.

Application of mathematical model for L. monocytogenes growth

Calculations for the primary model at different temperature (5, 10, 15, 20, and 25 °C) for lag time (LT, λ), specific growth rate (SGR, μ), initial population (N ₀), and maximum population (N _max) values were carried out by applying the modified-Gompertz equation (Eq. 1) using a GraphPad Prism software v. 4.03 (GraphPad Software, San Diego, CA, USA). The λ, μ, and N _max at 5, 10, 15, 20, and 25 °C were obtained from a modified-Gompertz equation regressed into exponential decay (Eq. 2), and exponential growth (Eq. 3), respectively using a SigmaPlot software v. 12.0 (Systat Software Inc., Richmond, CA, USA).

$$N=N_0+A·exp\left(-exp\left(\left(\mathit{\mu }·e·\left(\mathit{\lambda }-t\right)/A+1\right)\right)\right)$$ 1

where N is the bacterial population at time (t) and A is N _max/N ₀.

$$\mathit{\lambda }={\mathit{\lambda }}_0+a_{\mathit{\lambda }}·exp\left(-b_{\mathit{\lambda }}·T\right)$$ 2

where λ ₀, a _λ, and b _λ are the regression constants.

$$\mathit{\mu }={\mathit{\mu }}_0+a_{\mathit{\mu }}·exp\left(b_{\mathit{\mu }}·T\right)$$ 3

where μ ₀, a _μ, and b _μ are the regression constants.

Validation of predictive models

The validation of the predictive models were evaluated by determination (R²), Bias factor (B_f), accuracy factor (A_f), and root mean square error (RMSE).

$$R^2=1-(\mathit{\Sigma }{e}_i^2/\mathit{\Sigma }{\left(y_i-ý\right)}^2)$$ 4

where e _i is error of predictive data, y _i is the predictive data, and $ý$ is the average of predictive data.

$$B_f={10}^{\left(\mathit{\Sigma }log\left(pred/obs\right)/n\right)}$$ 5

$$A_f={10}^{\left(\mathit{\Sigma }\left|log\left(pred/obs\right)\right|/n\right)}$$ 6

$$\text{RMSE}=\sqrt{\left(\mathit{\Sigma }{\left(obs-pred\right)}^2\right)/n}$$ 7

where obs is the observed value, pred is the predicted value, and n repetition number of the observed data.

Results and discussion

Identification of presumptive L. monocytogenes

A total of five presumptive colonies were isolated from the Oxford agar based on the distinctive black color formation. Totally 300 different samples were analyzed and all presumptive colonies were obtained from two different retail market with same food distribution company. It might be due to the poor sanitation of retail markets itself or contamination of raw material.

Figure 1 shows results of PCR amplification patterns with L. monocytogenes presumptive from raw beef. All isolated strains were amplified hlyA gene fragment (456 bp).

Fig. 1.

PCR amplification of L. monocytogenes isolates from raw beef. (A) 100 bp DNA ladder (TAKARA, Bio Inc., Shiga, Japan); (B) L. monocytogenes ATCC 7644; (C) strain B-1 isolates; (D) strain B-2 isolates; (E) strain B-3 isolates; (F) strain B-4 isolates; (G), strain B-5 isolates

Secondary confirmation was performed using 16 s rRNA identification. There are several advantage to analyzing 16 s rRNA identification of bacteria in comparison with biochemical methods. The 16 s rRNA gene sequence is specific characters at the species level [15]. The closest relative species were determined by comparing with sequence database provided by the GenBank. All strains showed similarity with L. monocytogenes 07PF0776 (99% sequence similarity). The strains gathered by selective media were identified as L. monocytogenes, which differs from other studies. Barros et al. [16] reported that they isolated five different Listeria species from meat, consisting L. monocytogenes (13.2%), L. innocua (77.4%), L. seeligeri (1.3%), L. welshimeri (7.5%), and L. grayi (0.6%). Cocolin et al. [17] reported that L. monocytogenes 5UD and L. monocytogenes 9UD were isolated from sliced S. Daniele ham and De-boned S. Daniele ham, respectively. There are differences in isolated bacterial strain due to different food sample sources.

Primary models for L. monocytogenes

The predictive model can describe the L. monocytogenes growth pattern at various temperatures. Lopez [18] reported that the Gompertz model described microbial growth practically.

Figure 2 showed growth patterns of L. monocytogenes Type culture and beef isolates on raw beef at various temperatures (5–25 °C). It was reported that the growth of bacteria in food is influenced by storage temperature [19]. In this study, L. monocytogenes grew well at all conditions except at 5 °C temperature. At 5 °C, bacterial growth was observed but there was no clear lag phase and stationary phase. L. monocytogenes can growth under refrigerated condition, but as storage temperature decreased, the growth rate also decreased [12]. The N _max of L. monocytogenes Type culture was between 7.12 and 7.76, and that of L. monocytogenes beef isolates was between 7.53 and 7.80. The load of background microbiota was regarded as important factor influencing the N _max [20]. The samples used in this study were not sterilized. So various background microbiota might interrupted the growth of L. monocytogenes.

Fig. 2.

Gompertz model for L. monocytogenes type culture (filled circle) and L. monocytogenes isolates from beef in raw beef (filled triangle) at 25 °C (A), 20 °C (B), 15 °C (C), 10 °C (D), and 5 °C (E)

At higher temperature, shorter λ and higher μ were calculated (Table 1). At all temperatures, the λ of L. monocytogenes Type strains was longer than that of beef isolates. However, the μ of Type strains was lower than that of beef isolates. The λ of L. monocytogenes Type strains in raw beef were predicted to be 44.99, 23.44, 8.74, and 5.81 h and that of beef isolates were calculated as 32.07, 8.02, 4.52, and 1.77 h at 10, 15, 20, and 25 °C, respectively. The μ of L. monocytogenes Type strains stored at 10, 15, 20, and 25 °C were estimated to be 0.01, 0.02, 0.03, and 0.09 and that of beef isolates were predicted as 0.02, 0.02, 0.04, and 0.18, respectively. Compared to our result, different λ and μ values at 15 °C were reported [21]. They reported that the λ of 2-strain cocktail of L. monocytogenes containing L. monocytogenes ATCC 15313 and isolated L. monocytogenes L13-2 was 7 h and μ was 0.47 on raw pork meat at 15 °C. The different results are due to using different food and bacterial strains. Among strains, the differences in λ were significantly greater than in μ. Under lower temperature, the difference of λ and μ was more [22].

Table 1.

Estimated values of lag time (λ) and specific growth rate (μ) of L. monocytogenes Type strains and L. monocytogenes beef isolates by modified-Gompertz model in artificially inoculated raw beef

Temperature (°C)	L. monocytogenes Type strains		L. monocytogenes isolates
	Lag time, λ (h)	Specific growth rate, µ (log CFU/g/h)	Lag time, λ (h)	Specific growth rate, µ (log CFU/g/h)
5	N.Aa	N.A	N.A	N.A
10	44.99	0.01	32.07	0.02
15	23.44	0.02	8.02	0.02
20	8.74	0.03	4.52	0.04
25	5.81	0.09	1.77	0.18

^aNot available

Table 2 showed the growth parameters of modified-Gompertz model to estimate the growth of L. monocytogenes Type strains and beef isolates. The primary model with the modified-Gompertz model at all condition fit well with high R² value (R² = 0.98–1.00). In addition, RMSE was close to 0 (RMSE = 0.04–0.25) and bias was close to 1 (bias = 0.99–1.01). The evaluation of RMSE also provided goodness of fit, with value of RMSE near 0 [23]. Therefore, these results indicated that the model provide a good description of the growth behavior.

Table 2.

Validation of primary model to predict the growth of L. monocytogenes Type strains and L. monocytogenes beef isolates in raw beef

Temperature (°C)	L. monocytogenes Type strains			L. monocytogenes isolates
	R2a	RMSEb	Bcf	R2	RMSE	Bf
5	N.Ad	N.A	N.A	N.A	N.A	N.A
10	0.99	0.12	1.00	0.99	0.19	1.01
15	0.99	0.20	1.00	1.00	0.04	1.00
20	0.99	0.17	1.00	0.98	0.25	1.00
25	0.99	0.18	1.01	0.99	0.20	0.99

^aDetermination coefficient

^bRoot-mean-square error

^cBias factor

Secondary models for L. monocytogenes

The λ and μ calculated using the primary predictive model of L. monocytogenes in raw beef were applied to secondary model as a function of temperature using exponential equation. These results showed that λ is inversely proportional to μ and is influenced by temperature change. Secondary model was developed to define the primary model parameters that were affected by temperature, including λ and μ [24].

Table 3 showed the evaluations of secondary model of L. monocytogenes Type strains. For the λ and μ equation, R² values were both 1.00. B_f and A_f of λ equation were 0.99 and 1.12, respectively. In addition, in the μ equation, B_f and A_f were 1.02 and 1.12, respectively. These results showed that both B_f and A_f values of λ and μ equation were close to 1.00. The bias and accuracy factor equal to 1.0 indicated a perfect agreement between predicted and observed values [25]. The R² values of secondary model of L. monocytogenes beef isolates were both 1.00. In the λ equation, B_f and A_f were 1.03 and 1.18, respectively. For μ equation, B_f and A_f were 1.00 and 1.03, respectively. As mentioned above, both B_f and A_f values of λ and μ equation were also close to 1.00. Hence, it was found that the predictions were similar to the experimental results.

Table 3.

Evaluation of exponential equation to predict the growth of L. monocytogenes Type strains and L. monocytogenes beef isolates in raw beef

Cultures	Variables	Exponential equation	Regression analysis
			R2a	Bbf	Acf
L. monocytogenes Type strains	λ (h)	$\mathit{\lambda }=-1.0385+182.9051·{exp}^{-0.1374·T}$	1.00	0.99	1.12
	μ (log CFU/g/h)	$\mathrm{\mu }=0.0111+0.00009059·{exp}^{0.2707·T}$	1.00	1.02	1.12
L. monocytogenes isolates	λ (h)	$\mathit{\lambda }=2.2416+709.2921·{exp}^{-0.3170·T}$	1.00	1.03	1.18
	μ (log CFU/g/h)	$\mathrm{\mu }=0.0186+0.0000061977·{exp}^{0.4067·T}$	1.00	1.00	1.03

^aDetermination coefficient

^bBias factor

^cAccurate factor

Under higher temperature, the difference of λ and μ between L. monocytogenes Type strains and beef isolates was more. Therefore, the growth rate of L. monocytogenes beef isolates is probably much higher than Type strains with increases in the storage temperature.

The Type strains used in this study is widely used to construct culture cocktail in many studies [26–28]. But these strains were not isolated from beef, all Type strains used in this study were isolated from human. So if when these strains were used to construct culture cocktail for measuring growth pattern by artificial inoculation, it might not represent the actual growth pattern in raw beef.

In conclusion, based on the differences of growth parameters it was suggested that more realistic information on L. monocytogenes growth behaviors in beef can be obtained by using beef isolates, and can be used in subsequent quantitative microbial risk assessments for compliance with food safety criteria.

Author’s contribution

So-Yeon Lee: Execution of experiment, and preparation of manuscript; Gi-Hyeon Gwon: Discussion of the results; Changhoon Chai: Discussion of results relating to shelf life; Se-Wook Oh: Advice for the study.

Compliance with ethical standards

Competing interest

The authors declare that have no competing interests.