Previous Life Cycle Phase

A growth curve was developed to identify the life cycle phases of DT104 (Figure 1). One microliter of the stock culture was transferred into sterile BHI (5 ml) in a centrifuge tube (15 ml) followed by incubation at 37°C and 150 rpm from 0 to 770 h. This was carried out in three replicates and 50 μl of undiluted and appropriate 1:10 diluted (10⁻¹ to 10⁻⁵) samples from cultures were spiral plated (Whitely Automated Spiral Plater, Microbiology International, Frederick, MD) onto tryptic soy agar (TSA) plates. The TSA plates were incubated at 37°C for 24 h and then bacterial colonies were enumerated. In Figure 1, after the stationary phase, the culture had a slow and extended death phase. Since there is no time period of a stable density, it is referred to as the early death phase (EDP) instead of the LTS phase throughout the paper.

Open in a separate window

Figure 1

Growth of S. Typhimurium DT104 in brain heart infusion broth (BHI) at 37°C for different times to yield A (9 h, late log phase), B (24 h, early stationary phase), C (48 h, late log phase), and D (720 h, early death phase).

Starter Culture

One microliter of the stock culture was transferred into sterile BHI (5 ml) in a centrifuge tube (15 l) followed by incubation at 37°C and 150 rpm to obtain cells in the late log phase (LLP, 9 h), early stationary phase (ESP, 24 h), late stationary phase (LSP, 48 h) and early death phase (EDP, 720 h), respectively. The culture was centrifuged for 5 min at 2,100 × g and the pellet resuspended into 5 ml of the selected media as described below and adjusted to achieve a starting concentration of 10⁴ CFU/ml.

Growth Cultures

Analysis of cultures of DT104 for growth kinetic determinations were performed in centrifuge tubes (15 ml). Cultures grown under pH stress contained 5 ml of BHI adjusted to pH of 5, 7, or 9 with 1 N HCl or 1 N NaOH. Cultures grown under osmotic stress contained 5 ml of BHI adjusted to 2, 3 and 4% NaCl (Difco Lab, MI, USA). DT104 was also grown in ground chicken extract (GCE) prepared as previously described in Fratamico et al. (2011) with modifications. Fresh raw, 85% lean ground chicken was obtained from a local supermarket and high pressure processed (HPP) at 87,500 psi (603 MPa) for 2.5 min. Sterile water (15 ml) was added to ground chicken (50 g) in a stomacher bag and pummeled for 1 min, the liquid extract was removed by centrifugation at 2,100 × g for 5 min, and then filtered using a 0.22 μm filter, and frozen at −20°C until used. Growth cultures were incubated at 37°C and 150 rpm for 0 to 30 h.

Microbial Analysis

At selected times post inoculation, depending on the age of the culture and growth culture medium of incubation, 50 μl of undiluted and appropriate 1:10 diluted (10⁻¹ to 10⁻⁵) samples from growth cultures were spiral plated (Microbiology International, Frederick, MD) onto tryptic soy agar (TSA) plates. Sampling times for each growth condition (i.e. pH, salt concentration and GCE) were based on estimated lag time (λ), specific growth rate (μ_max), and maximum population density (Y_max) and were selected to produce a growth curve that accurately defined the lag, log and stationary phase over three log cycles of growth. The TSA plates were incubated at 37°C for 24 h.

Predictive Growth Modeling

U.S. Department of Agriculture, Integrated Pathogen Modeling Program (IPMP) 2013 was used in this study. It is a free and easy-to-use data analysis platform for fitting primary models. Primary models include common growth and inactivation models, which can be used to analyze full growth curves, incomplete growth curves and inactivation/survival curves (Huang, 2014).

The data from the samples were assigned to IPMP 2013. The primary models evaluated were the full-growth Baranyi (Baranyi and Roberts, 1994), reparametrized Gompertz (Zwietering et al., 1990), Huang (Huang, 2014), and Buchanan (Buchanan et al., 1997) three-phase linear. The performance of the models is based on the production of the Akaike information criterion (AIC). The AIC is to estimate the likelihood of a model to predict/estimate the future values. A lower AIC value indicates a better fit. The Baranyi model, results not shown, was chosen because it was the best fit for growth data and had the lowest AIC value.

The Baranyi model is the following:

where Y₀, Y_max, and Y(t) are the bacterial population, in natural logarithm of bacteria counts, at initial, maximum and time t. μ_max is the specific growth rate; h_o is the physiological state of the microorganism under consideration.

Scanning Electron Microscope and Transmission Electron Microscope

Scanning electron microscope (SEM) and transmission electron microscope (TEM) was performed on cultures (20 μl) of DT104 in different life cycle phases as described by Wen et al. (2009). A one-way analysis of variance was used to determine the effect of the previous life cycle phase on the cell length of S. Typhimurium DT104. Pairwise comparisons were made by Tukey’s least significant difference test (α = 0.05) using Statistix 9 (Analytical Software, Tallahassee, FL).

Antibiotic Susceptibility Disk Diffusion

Cultures of DT104 in different life cycle phases were tested for antibiotic susceptibility by the disc diffusion method on Mueller-Hinton agar (Sigma-Aldrich, Munich, Germany). The antibiotic sensitivity of the cultures of DT104 in different life cycle phases were evaluated according to classification guidelines suggested by the National Committee for Clinical Laboratory Standards (NCCLS). The cultures were tested for susceptibility to a panel of antibiotics including gentamicin (GM; 10 μg), sulphamethoxazole x trimethoprim (SXT; 25 μg), kanamycin (K; 30 μg), tetracycline (TE; 30 μg), nalidixic acid (NA; 30 μg), trimethoprim (TMP; 5 μg), ciprofloxacin (CIP; 5 μg), centriaxone (CRO; 30 μg), sulfisoxazole (G; 250 μg), chloramphenicol (C; 30 μg), and streptomycin (S; 10 μg) using the disk diffusion method (Briggs and Fratamico, 1999; O’Leary et al., 2015). The zones of inhibition (ZOI) were measured and the isolates were classified as being susceptible, intermediate or resistant to the antibiotic according to the NCCLS guidelines.

Brain Heart Infusion Broth vs. Ground Chicken Extract

The cultures responded differently depending upon the phase of the culture. There was a significant (p < 0.05) interaction between the medium and previous life cycle phase on the Y_max. As seen in Figure 5, the EDP in BHI had a lower (p < 0.05) Y_max (8.95 log CFU/ml) in comparison to the LLP, ESP, and LSP (9.19, 9.16, and 9.16 log CFU/ml, respectively). In the GCE there was no significant (p > 0.05) effect of the previous life cycle phases on the Y_max. There was no significant interaction between the media and previous life cycle phase on the λ and μ_max therefore only the main effects will be discussed. In BHI, the λ significantly (p < 0.05) increased as the age of the previous life cycle phase increased except that the ESP λ equaled the LSP λ (LLP < ESP = LSP < EDP). There was also a significant (p < 0.05) effect of the previous life cycle on the λ in GCE. The EDP had the significantly (p < 0.05) longest λ at 2.47 h and there was a significant (p > 0.05) difference between LLP and LSP of 0.19 h (Figure 5). The LLP had a significantly (p < 0.05) lower μ_max than ESP in both BHI (1.03 vs. 1.18 h⁻¹, respectively) and GCE (0.96 vs. 1.10 h⁻¹) (Figure 5).

Open in a separate window

Figure 5

The lag time (λ), growth rate (μ_max), and maximum growth (Y_max) of S. Typhimurium DT104 as a function of the previous life cycle phase, late log phase (LLP), early stationary phase (ESP), late stationary phase (LSP), and early death phase (EDP) in brain heart infusion broth (BHI) and ground chicken extract (GCE). Different lowercase and uppercase letters within a cluster of bars differed significantly (p < 0.05).

pH Levels

There was a significant (p < 0.05) interaction on the Y_max between the previous life cycle phase and various pH levels (Figure 6). At pH 5 and 9, the EDP had a lower (p < 0.05) Y_max than the LLP, ESP, and LSP. While at pH 7, the previous life cycle phase had no significant (p > 0.05) effect on the Y_max. While there was a significant (p < 0.05) interaction on the Y_max, there was none (p > 0.05) on the λ and μ_max. Similar to BHI and GCE, the λ significantly (p < 0.05) increased as the age of the previous life cycle phase (LLP, ESP, LSP, and EDP) increased in both pH 5 (1.40, 1.90, 2.41, and 3.67 h, respectively) and pH 9 (1.04, 1.58, 2.13, and 3.56 h, respectively) (Figure 6). At pH 7, the previous life cycle phase also had a significant (p < 0.05) effect on the λ (LLP < ESP = LSP < EDP). The previous life cycle phase had no significant (p > 0.05) effect on the μ_max regardless of the pH level (Figure 6).

Open in a separate window

Figure 6

The lag time (λ), growth rate (μ_max), and maximum growth (Y_max) of S. Typhimurium DT104 as a function of the previous life cycle phase, late log phase (LLP), early stationary phase (ESP), late stationary phase (LSP), and early death phase (EDP) in brain heart infusion broth (BHI) at pH 5, 7, and 9. Different lowercase and uppercase letters within a cluster of bars differed significantly (p < 0.05).

Salt Levels

In pertaining to the salt concentrations and the previous life cycle phase, there was a significant (p < 0.05) interaction on the μ_max (Figure 7). At 2 and 4% salt concentration, the LLP had the highest (p < 0.05) μ_max and the previous life cycle phase had no significant (p < 0.05) effect on the μ_max at 3% salt concentration. The previous life cycle phase and salt concentration had no significant (p > 0.05) interaction on the λ and Y_max. At 3 and 4% salt concentration the λ of LLP and ESP was not significantly (p > 0.05) different and then the λ significantly (p < 0.05) increased from the ESP to the EDP. However, at 2% salt the EDP had the longest (p > 0.05) λ (3.91 h). The salt concentrations had no significant (p > 0.05) effect on the Y_max regardless of the previous life cycle phase (Figure 7).

Open in a separate window

Figure 7

The lag time (λ), growth rate (μ_max) and maximum growth (Y_max) of S. Typhimurium DT104 as a function of the previous life cycle phase, late log phase (LLP), early stationary phase (ESP), late stationary phase (LSP), and early death phase (EDP) in brain heart infusion broth (BHI) at salt concentrations 2, 3, and 4%. Different lowercase and uppercase letters within a cluster of bars differed significantly (p < 0.05).

Scanning Electron Microscope and Transmission Electron Microscope

Representative photomicrographs of cells of DT104 in different phases are shown in Figure 8. The SEM photomicrographs depicts the cells decreasing in size as they transitioned from the log phase (LP) to the EDP and size increased (recovered) transitioning back to the LP and stationary phase (SP) previously from the EDP. In the LP and SP the cells were rod-shaped while a cocci (spherical) shape appeared in the EDP. ANOVA of the SEM data revealed that cell length was significantly (p < 0.05) affected by the growth phase with the LLP (2.79 μm) = LLP (previously EDP) (2.53 μm) > SP (previously EDP) (2.25 μm) = SP (2.14 μm) > EDP (1.81 μm).

Open in a separate window

Figure 8

Scanning electron microscope (SEM) and transmission electron microscope (TEM) images of S. Typhimurium DT104 at different growth phases, log phase (LP), stationary phase (SP), and early death phase (EDP) in brain heart infusion broth (BHI) at 37°C. Bars: 1.0 μm (SEM) and 500 nm (TEM).

Brain Heart Infusion and Ground Chicken Extract

BHI is a common nutrient rich broth used for the growth of S. Typhimurium and other fastidious and nonfastidious bacteria. The components (g/L) in BHI include: brain heart, infusion from solids (6), peptic digest of animal tissue (6), pancreatic digest of gelatin (14.5), dextrose (3), sodium chloride (5), and disodium phosphate (2.5). Chicken contains protein, including all of the essential amino acids, minerals (calcium, iron, magnesium, phosphorous, potassium, sodium, zinc and selenium), vitamins (niacin, pantothenic acid, vitamin B6 and B12, folate and folic acid (USDA/ARS, 2016). BHI contains 3.0 g/l of dextrose, whereas GCE does not and has a lower pH (6.4) than BHI (7.4). Overall, the GCE had a lower Y_max than BHI (8.77 vs. 9.12 log CFU/ml, respectively). This may be due to the limited nutrients in the GCE, containing no carbohydrates, while BHI contains dextrose.

The present study and previous ones (Oscar, 1999a,b, 2006) demonstrated that chicken based mediums can be used to develop models for the growth of S. Typhimurium in the place of traditional lab media. The λ of BHI after reinoculation of the ESP cells was 1.47 h which is similar to Oscar (1997) at 1.25 h with similar growth conditions. Oscar (1997) reported a slower μ_max at 0.86 log CFU/ml/h compared to 1.18 log CFU/ml/h in the present study. The λ of GCE after reinoculation of the ESP cells was 1.45 h while in Oscar (2007), a study that developed predictive growth models for DT104 in ground chicken, had a λ of 1.39 and 1.66 h at 34 and 40°C, respectively. Using IPMP 2013 we determined the significance (p < 0.05) of the previous life cycle phase and media on the growth kinetics of S. Typhimurium DT104.

A follow-up study will be performed on S. Typhimurium DT104 in GCE examining the global transcriptional profile of DT104 to provide insights on how this food matrix influences growth and survival and potentially identify specific genes that may be targeted for the development of control strategies. Previously, it has been difficult to perform transcriptomics in complex medium but due to technology advancement it has become possible instead of using a simple medium as BHI (Fratamico et al., 2011; Deng et al., 2012). GCE may be a better growth medium to mimic a poultry environment in place of BHI for transcriptomic analysis of Salmonella.

pH

Salmonella may encounter various pH levels in food production. Acid-based antimicrobial solutions such as chlorine, organic acids, peroxyacetic and peracetic acid are commonly used in the poultry industry to reduce Salmonella and Campylobacter contamination on poultry and poultry products. Studies such as Oscar (1999c) and Gibson et al. (1988) developed Salmonella growth models in broth media and demonstrated Salmonella ability to grow at low pH levels of 5.2 and 5.63, respectively. This study was carried out at a lower pH of 5 and demonstrated an ability of DT104 to grow in an unfavorable environment. Preliminary results (data not shown) demonstrated DT104 ability to grow at pH ≥ 4.5 but not at less than <4.5, this may be due to inability to maintain an adequate internal pH level. To the best our knowledge this is the first study that examined the effect of the previous life cycle and pH on the growth kinetics of a food pathogen. The starved cells in the EDP spent 2.27 h longer in the λ than LLP taking more time to adjust and adapt to pH 5 (Figure 3).

Open in a separate window

Figure 3

Baranyi model for S. Typhimurium DT104 brain heart infusion broth (BHI) at pH 5, 7, and 9 at 37°C. Observed, predicted, upper 95% confidence interval (U95CI), and lower 95% confidence interval (L95CI).

The growth kinetics have not frequently been studied at an alkaline pH. Alkaline cleaning agents are used in the food industry to remove biofilm, contamination, food, fat, and protein from processing equipment (Gibson et al., 1999; Sharma et al., 2005). Cleaners such as alkaline chlorinated cleaners work by temporarily imparting hydrophilic properties to stainless steel that decrease attachment to surfaces. While Humphrey (1987) carried out a study demonstrating the inactivation of Salmonella at high pH and temperature, this is the first study to the best of our knowledge performing predictive growth models of Salmonella at high pH (pH 9).

Salt

Salt is commonly added to food products to lower the water activity and inhibit microbial growth of pathogens and spoilage bacteria. In the presence of salt, the growth of bacteria is dependent upon the presence of osmoprotectants. In an abrupt shift of osmotic stress, cells accumulate harmful substrates such as K⁺ and glutamine. These substrates are replaced by proline, trehalose and glycine betaine, which are taken up from the medium or synthesized de novo by the cell to stimulate growth (Zhou et al., 2011). Various studies (Matches and Liston, 1972; Mellefont et al., 2003, 2004, 2005) have been carried out showing an extended λ of Salmonella as the salt concentration increases and the a_w decreases. The study by Mellefont et al. (2004) carried out at a lower a_w demonstrated that the cells from the log phase after reinoculation had a longer relative lag time than the cells from the stationary phase in contrast to our study, where cells from the LLP after reinoculation had a shorter λ than the ESP but was carried out a higher a_w until at 4% NaCl cells from the LLP had a longer λ. These studies demonstrate that at concentrations higher than 4%, the cells from the LLP are not able to adjust to abrupt osmotic shifts. The cells with shorter λ may have better defense systems which prepares them to survive various environmental stresses without prior exposure. Further research is needed to better understand these mechanisms.

Similar to other studies (Ross et al., 2003), the μ_max decreased with the addition of salt (Figure 7). This occurrence may be due to the initial response to osmotic shift with cells having three phases: dehydration, adjustment, and rehydration. After respiration and ion transport are stopped from exposure to the shock, the respiration will resume and accumulate K⁺, glutamate and solutes. Unlike, Zhou et al. (2011) there was an initial decrease in cell density in basic minimal medium at high salt concentration, DT104 did not have a decrease due to essential proteins and enzymes present in the BHI.

Scanning Electron Microscope and Transmission Electron Microscope

Like S. Typhimurium DT104 in this study, other bacteria such as L. monocytogenes and S. Virchow have also shown a decrease in cell size (McMahon et al., 2007; Wen et al., 2009). DT104 transitioning back to the rod shape from the coccoid shape suggests that the cell formation is dependent upon the available nutrients. Starvation is known to induce a change in cell shape from rods to coccoid in Arthrobacter crystallopoietes, A. globiformis, Rhizobium leguminosarum and marine vibrio (Novitsky and Morita, 1976; Hamilton et al., 1977; Watson et al., 1998; Demkina et al., 2000). The EDP cells have a condensed cytoplasmic network in comparison to the other phases. This may explain the extended lag time of the subsequent EDP when introduced to a new environment. The coccoid shape may be formed by cell shrinkage and cytoplasmic condensation illustrated in Figure 8 TEM. This may explain why the EDP cells require additional time to uptake available and essential nutrients and replenish fluidity in the cytoplasm.

We would like to thank Joan Meredith for technical assistance. Earlier version of this manuscript has been released as a pre-print at https://www.biorxiv.org/content/early/2018/05/21/326579 (Hawkins et al., 2018).