Thermal Resistance and Gene Expression of both Desiccation-Adapted and Rehydrated Salmonella enterica Serovar Typhimurium Cells in Aged Broiler Litter

ABSTRACT

The objective of this study was to investigate the thermal resistance and gene expression of both desiccation-adapted and rehydrated Salmonella enterica serovar Typhimurium cells in aged broiler litter. S. Typhimurium was desiccation adapted in aged broiler litter with a 20% moisture content (water activity [a_w], 0.81) for 1, 2, 3, 12, or 24 h at room temperature and then rehydrated for 3 h. As analyzed by quantitative real-time reverse transcriptase PCR (qRT-PCR), the rpoS, proV, dnaK, and grpE genes were upregulated (P < 0.05) under desiccation stress and could be induced after 1 h but in less than 2 h. Following rehydration, fold changes in the levels of these four genes became significantly lower (P < 0.05). The desiccation-adapted ΔrpoS mutant was less heat resistant at 75°C than was the desiccation-adapted wild type (P < 0.05), whereas there were no differences in heat resistance between desiccation-adapted mutants in two nonregulated genes (otsA and PagfD) and the desiccation-adapted wild type (P > 0.05). Survival characteristics of the desiccation-adapted ΔPagfD (rdar [red, dry, and rough] morphotype) and ΔagfD (saw [smooth and white] morphotype) mutants were similar (P > 0.05). Trehalose synthesis in the desiccation-adapted wild type was not induced compared to a nonadapted control (P > 0.05). Our results demonstrated the importance of the rpoS, proV, dnaK, and grpE genes in the desiccation survival of S. Typhimurium. By using an ΔrpoS mutant, we found that the rpoS gene was involved in the cross-protection of desiccation-adapted S. Typhimurium against high temperatures, while trehalose synthesis or rdar morphology did not play a significant role in this phenomenon. In summary, S. Typhimurium could respond rapidly to low-a_w conditions in aged broiler litter while developing cross-protection against high temperatures, but this process could be reversed upon rehydration.

IMPORTANCE Physical heat treatment is effective in eliminating human pathogens from poultry litter used as biological soil amendments. However, prior to physical heat treatment, some populations of microorganisms may be adapted to the stressful conditions in poultry litter during composting or stockpiling, which may cross-protect them against subsequent high temperatures. Our previous study demonstrated that desiccation-adapted S. enterica cells in aged broiler litter exhibited enhanced thermal resistance. However, there is limited research on the underlying mechanisms of the extended survival of pathogens under desiccation conditions in animal wastes and cross-tolerance to subsequent heat treatment. Moreover, no information is available about the thermal resistance of desiccation-adapted microorganisms in response to rehydration. Therefore, in the present study, we investigated the gene expression and thermal resistance of both desiccation-adapted and rehydrated S. Typhimurium in aged broiler litter. This work will guide future research efforts to control human pathogens in animal wastes used as biological soil amendments.

INTRODUCTION

Poultry litter has been widely used as biological soil amendments for growing fresh produce; however, there are many species of human pathogens, such as Salmonella, that potentially may be present in poultry litter or inadequately composted poultry litter (1). Erickson et al. (2) found that Salmonella enterica serovar Typhimurium was still detectable in subsurface samples after 14 days in static composting piles composed of chicken litter and peanut hulls. These pathogens may possibly contaminate fresh produce after agricultural land application of poultry litter, as a small population of pathogens in animal wastes may survive and persist for an extended period of time (3). Islam et al. (4) reported that S. Typhimurium could persist for over 200 days in soil samples amended with poultry litter-based composts.

Physical heat treatment (heat drying after composting or without composting) is considered to be an effective process to eliminate human pathogens from animal wastes (5). However, some microbial cells may become acclimatized to the stressful environments in animal wastes, which can cross-protect them against subsequent high temperatures (6). Our previous studies demonstrated that desiccation-adapted S. enterica cells in aged broiler litter displayed enhanced thermal resistance compared to nonadapted cells (7, 8). Gruzdev et al. (9) also reported that desiccated S. enterica cells exhibited strong tolerance to dry heat at 60°C, without any significant population reduction in 1 h, in comparison to a 3-log reduction in the number of nondesiccated cells under identical conditions. Nonetheless, comparatively little is known about the underlying mechanisms of the cross-protection of desiccation-adapted Salmonella cells against subsequent high temperatures.

There have been some previously reported studies on the gene expression of S. enterica under various low-water-activity (a_w) conditions (10–15). A diverse range of responses in Salmonella can be induced in response to desiccation, depending on the experimental conditions (e.g., a_w and the method used to create a low-a_w environment) and the strain tested. Genes that have been reported to be upregulated upon desiccation in those studies are involved in some biological processes, such as the general stress response, trehalose synthesis, osmoprotectant production, fatty acid metabolism, and the heat shock response.

Extracellular components, such as thin aggregative fimbriae (tafi), are required for the long-term persistence of some microorganisms under desiccation conditions (16). S. enterica serovar Typhimurium possesses the multicellular rdar (red, dry, and rough) morphotype, which is regulated by the agfD promoter (17). The presence of rdar morphology can improve the colonization of Salmonella on and in fresh produce (18) and facilitate the dispersal of the pathogen in post-rain aerosols (19). The rdar morphotype, as well as the non-rdar saw (smooth and white) morphotype, can be isolated from poultry, produce, and clinical samples (20). If poultry litter contains the rdar morphotype, the pathogen may possibly be transferred to fresh produce after agricultural land application. In spite of the scientific importance and practical implications of the rdar morphotype, its thermal resistance remains poorly understood. Scher et al. (21) reported that pellicle cells (rdar morphotype) were significantly more resistant to sodium hypochlorite; however, the stress management of pellicle cells in response to heat or low pH was not enhanced compared to that of planktonic cells.

As far as we know, there is no available information about the thermal resistance of desiccation-adapted microorganisms upon rehydration. In an effort to fill this knowledge gap, we thus investigated the thermal resistance and gene expression of both desiccation-adapted and rehydrated S. Typhimurium cells in aged broiler litter. Elucidating the response of S. Typhimurium to desiccation stress in aged broiler litter can provide insight into how this pathogen can survive under low-a_w conditions and how exposure to these conditions influences its heat resistance.

RESULTS

As shown in Table 1, the populations of 3-, 12-, and 24-h desiccation-adapted S. Typhimurium wild-type as well as nonadapted control cells in aged broiler litter decreased during heat treatment at 75°C. However, desiccation-adapted wild-type cells survived much longer than did nonadapted control cells (P < 0.05). Nonadapted cells were detectable only by enrichment (<1.30 log CFU g⁻¹) after 1.5 h, while desiccation-adapted wild-type cells could be detected by direct plating throughout the treatment, and the populations of desiccation-adapted wild-type cells were much larger than those of nonadapted control cells (P < 0.05). For example, after a 1-h heat treatment, the population of nonadapted control cells decreased rapidly to 2.14 log CFU g⁻¹, but there were still nearly 4 log CFU g⁻¹ of desiccation-adapted wild-type cells surviving in aged broiler litter. A 5-log reduction in the population of desiccation-adapted wild-type cells required >2 h of heat exposure. As a comparison, a 5-log reduction in the population of nonadapted control cells could be achieved only within 1.5 h. Nevertheless, there were no significant differences in microbial populations among wild-type desiccation-adapted cells for 3, 12, and 24 h (P > 0.05), suggesting that the heat resistance of desiccation-adapted cells was not significantly enhanced when the duration of desiccation was extended from 3 to 24 h.

TABLE 1.

Survival of wild-type S. Typhimurium ATCC 14028 and its mutants in aged broiler litter with a 20% moisture content at 75°C

Strain	Desiccation time (h) + rehydration time (h)	Mean population size (log CFU g−1) ± SD after time (h) ofa:
		0	0.5	1	1.5	2
ATCC 14028 (wild type)	0 + 0	7.06 ± 0.12 A(A)a	2.89 ± 0.13 A(B)b	2.14 ± 0.32 A(B)b	+	+
	0 + 3	7.04 ± 0.17 (A)a(a)	2.92 ± 0.23 (B)b(b)	2.01 ± 0.16 (B)b(b)	+	+
	3 + 0	7.00 ± 0.12 (A)a	3.92 ± 0.43 (A)a	3.82 ± 0.26 (A)a	3.74 ± 0.38 (A)a	3.53 ± 0.42 (A)a
	3 + 3	7.02 ± 0.01 (A)a(a)	3.88 ± 0.21 (A)a(a)	2.46 ± 0.11 (B)a(a)	2.03 ± 0.14 (B)a(a)	1.65 ± 0.22 (B)a(a)
	12 + 0	7.07 ± 0.16 (A)a	4.02 ± 0.16 (A)a	3.98 ± 0.45 (A)a	3.81 ± 0.21 (A)a	3.49 ± 0.25 (A)a
	12 + 3	7.07 ± 0.11 (A)a(a)	3.71 ± 0.42 (A)a(a)	2.57 ± 0.27 (B)a(a)	1.93 ± 0.26 (B)a(a)	1.74 ± 0.12 (B)a(a)
	24 + 0	7.09 ± 0.08 A(A)a	3.88 ± 0.33 A(A)a	3.92 ± 0.18 A(A)a	3.80 ± 0.58 A(A)a	3.58 ± 0.42 A(A)a
	24 + 3	7.00 ± 0.12 (A)a(a)	3.83 ± 0.24 (A)a(a)	2.49 ± 0.31 (B)a(a)	2.13 ± 0.19 (B)a(a)	1.60 ± 0.32 (B)a(a)
IB43 (ΔrpoS)	0 + 0	7.04 ± 0.12 Aa	2.55 ± 0.20 Bc	−	−	−
	24 + 0	7.00 ± 0.00 Aa	2.59 ± 0.19 Bc	−	−	−
XF373 (ΔotsA)	0 + 0	7.03 ± 0.01 Aa	2.91 ± 0.21 Ab	1.96 ± 0.15 Ab	+	+
	24 + 0	7.03 ± 0.06 Aa	3.99 ± 0.01 Aa	3.96 ± 0.15 Aa	3.83 ± 0.43 A	3.60 ± 0.14 A
MAE 110 (ΔPagfD)	0 + 0	7.06 ± 0.05 Aa	2.90 ± 0.25 Aa	2.09 ± 0.15 Aa	+	+
	24 + 0	7.03 ± 0.11 Aa	4.05 ± 0.27 Ab	3.95 ± 0.33 Ab	3.77 ± 0.13 A	3.54 ± 0.59 A
MAE 119 (ΔagfD)	0 + 0	7.05 ± 0.15 Aa	2.85 ± 0.22 Aa	2.13 ± 0.25 Aa	+	+
	24 + 0	7.04 ± 0.06 Aa	4.08 ± 0.43 Ab	3.90 ± 0.32 Ab	3.81 ± 0.09 A	3.59 ± 0.39 A

^aData are expressed as means ± standard deviations. According to the LSD test, in the same column, means with different uppercase letters for the same desiccation adaptation time were significantly different (P < 0.05), while means with different lowercase letters for the same strain were significantly different (P < 0.05). For the wild type in the same column, means with different uppercase letters in parentheses with the same desiccation adaptation time are significantly different (P < 0.05), while means with different lowercase letters in parentheses with different desiccation adaptation times after rehydration were significantly different (P < 0.05). +, detectable only by enrichment; −, not detectable by enrichment.

When the effect of rehydration on the heat resistance of desiccation-adapted wild-type cells in aged broiler litter was investigated, it was found that upon rehydration, desiccation-adapted cells in aged broiler litter became less heat resistant. After a 1-h heat treatment, the numbers of surviving desiccation-adapted cells were 1 to 2 log CFU g⁻¹ higher than those of 3-h-rehydrated cells (P < 0.05). Moreover, during heat treatment, the population sizes of wild-type desiccation-adapted cells for 3, 12, and 24 h following rehydration were similar (P > 0.05) but were all larger than those of nonadapted wild-type cells (P < 0.05). For nonadapted control cells, 3-h rehydration did not result in any change in their heat resistance (P > 0.05).

The expressions of some desiccation-associated genes in wild-type S. Typhimurium were studied. Four genes in S. Typhimurium, rpoS, proV, dnaK, and grpE, were identified to be significantly upregulated (>2-fold) upon 3-, 12-, and 24-h desiccation adaptations (P < 0.05) (Fig. 1A to C). The highest upregulations were observed for the rpoS and dnaK genes, in a range of 32- to 40-fold, while the proV and grpE genes were upregulated from 2- to 9-fold. Eight other genes, otsA, otsB, agfD, kdpA, fadA, cspA, sigDE, and dps, were not differentially expressed (P > 0.05). There were no significant differences in fold changes in levels of selected genes among S. Typhimurium desiccation-adapted cells for 3, 12, and 24 h (P > 0.05). After a 3-h rehydration, fold changes in the levels of all significantly upregulated genes became much lower (P < 0.05). Some upregulated genes even became nondifferentially expressed (<2-fold) and returned to normal transcriptional levels postrehydration. For example, in response to rehydration, fold changes in the levels of the rpoS and proV genes in 24-h desiccation-adapted cells decreased from +37.6-fold to −1.0-fold and from +8.6-fold to +1.3-fold, respectively (Fig. 1C). Genes that were nondifferentially expressed after 3-, 12-, and 24-h desiccation adaptations were still nonregulated following rehydration (P > 0.05).

FIG 1.

Fold changes in levels of wild-type S. Typhimurium genes after 3-h (A), 12-h (B), and 24-h (C) desiccation adaptations and after a 3-h rehydration. Differentially expressed genes were defined as those that were significantly up- or downregulated by ≥2-fold (horizontal dotted line) between desiccation-adapted or rehydrated cells and control cells (P < 0.05). Genes underlined but not in boldface type were significantly upregulated (P < 0.05) after desiccation adaptation but nonregulated after rehydration, while genes underlined and in boldface type were significantly upregulated (P < 0.05) after both desiccation adaptation and rehydration.

To further determine if S. Typhimurium could respond to desiccation stress in aged broiler litter within an even shorter period of time (≤2 h), fold changes in the levels of four differentially expressed genes (rpoS, proV, dnaK, and grpE) were thus studied after 1- and 2-h desiccation adaptations. As shown in Fig. 2, after a 1-h desiccation adaptation, the rpoS, proV, dnaK, and grpE genes were nondifferentially expressed (<2-fold). However, when the duration of desiccation was prolonged to 2 h, these genes were significantly upregulated, by 25.5-, 5.8-, 20.6-, and 2.4-fold, respectively (P < 0.05). Fold changes in the levels of the proV and dnaK genes after a 2-h desiccation adaptation were similar to those after 3-, 12-, and 24-h desiccation adaptations (P > 0.05), while fold changes in the levels of the rpoS and grpE genes after a 2-h desiccation adaptation were still lower than those after 3-, 12-, and 24-h desiccation adaptations (P < 0.05).

FIG 2.

Fold changes in levels of differentially expressed wild-type S. Typhimurium genes after 1- and 2-h desiccation adaptations. Differentially expressed genes were defined as those that were significantly up- or downregulated by ≥2-fold (horizontal dotted line) between desiccation-adapted cells and control cells (P < 0.05).

To explore if the upregulated genes were involved in cross-protection against subsequent high temperatures, one significantly upregulated gene, rpoS, was selected for our mutation study. As controls, two nonregulated genes, otsA and agfD, were also selected. As shown in Table 1, the populations of desiccation-adapted and nonadapted IB43 (ΔrpoS mutant) cells were similar throughout the heat treatment (P > 0.05), and they were not detectable by enrichment only after 1 h, indicating that the desiccation adaptation of IB43 did not lead to an increase in its heat resistance. Additionally, desiccation-adapted IB43 cells survived for a much shorter time than did desiccation-adapted wild-type cells (P < 0.05). During 2 h of heat treatment, the populations of nonadapted ΔrpoS mutant cells were significantly smaller than those of nonadapted wild-type cells (P < 0.05), as nonadapted wild-type cells were still detected by enrichment after 2 h. In comparison, nonadapted ΔrpoS mutant cells were not detectable by enrichment only after 1 h, suggesting that the mutation in rpoS could also cause a reduction in the heat resistance of S. Typhimurium.

During heat treatment, the populations of desiccation-adapted and nonadapted XF373 (ΔotsA mutant) cells were similar to those of desiccation-adapted and nonadapted wild-type (P > 0.05) cells, respectively (Table 1). The rdar morphotype did not possess increased heat resistance, as the populations of desiccation-adapted and nonadapted MAE 110 (ΔPagfD mutant) or MAE 119 (ΔagfD mutant) cells were not significantly different from those of desiccation-adapted and nonadapted wild-type cells (P > 0.05), respectively, and the survival data for the ΔPagfD (rdar morphotype) and ΔagfD (saw morphotype) mutants were also similar throughout the entire treatment period (P > 0.05). Hence, the otsA and agfD genes do not contribute to the cross-tolerance of desiccation-adapted S. Typhimurium to heat treatment.

The trehalose contents in wild-type S. Typhimurium cells after 3-, 12-, and 24-h desiccation adaptations were less than 1.0 mmol mg protein⁻¹ (Fig. 3), and there were no significant differences between desiccation-adapted and nonadapted cells (P > 0.05). Therefore, trehalose was not substantially synthesized during a 3-, 12-, or 24-h desiccation adaptation.

FIG 3.

Trehalose contents in wild-type S. Typhimurium cells after 3-, 12-, and 24-h desiccation adaptations.

DISCUSSION

The stress responses in bacteria are controlled by master regulators, which include alternative sigma factors such as RpoS (22). van Hoek et al. (23) reported that a fully functional RpoS system is an advantage for the long-term survival of Escherichia coli O157 in the manure-amended soil environment. It is generally believed that these response pathways extensively overlap, and bacteria exposed to one sublethal stress may thus develop cross-protection against other stresses (24). In the present study, we observed that desiccation-adapted S. Typhimurium in aged broiler litter could develop cross-tolerance to high temperatures, which is consistent with our previously reported results (7, 8). This phenomenon is of concern in the poultry litter processing industry, where poultry litter commonly undergoes a series of interventions (e.g., composting and long-term stockpiling) in tandem to reduce the human pathogen load.

Genes that are upregulated under desiccation stress.

In Salmonella, the rpoS gene encodes an alternative sigma factor (σ^S/RpoS) that initiates the transcriptions of a series of genes and acts as a master regulator required for survival under harsh conditions (25, 26). In the present study, in agreement with the known role of the rpoS gene in the general stress response, the rpoS gene was found to be significantly upregulated following desiccation in aged broiler litter (P < 0.05). Similarly, Stasic et al. (27) observed that wild-type E. coli O157:H7 survived much longer (>28 days) in sterile bovine feces (a_w, <0.50) than did the ΔrpoS mutant (21 days). We also found that the attenuated expression of the rpoS gene in the ΔrpoS mutant could also cause significantly reduced thermal resistance in aged broiler litter (P < 0.05). As expected, desiccation adaptation of the ΔrpoS mutant could not cross-protect this mutant against subsequent heat treatment, suggesting that the rpoS gene is specifically required for the development of cross-protection. Generally, our findings are concurrent with the idea that bacteria can evoke a general stress response upon exposure to some environmental stresses (28).

To combat the loss of water during exposure to low-a_w environments, cells must balance the osmolarity of the intracellular composition with that of the external environment (29). During osmoregulation, bacteria accumulate some osmoprotective solutes known as osmoprotectants (e.g., glycine betaine) (30). ProV, encoded by the proV gene, serves the energy-coupling function in the ProU transport system for glycine betaine in Salmonella (31, 32). Li et al. (15) reported that after a 2-h desiccation in a desiccator (a_w, 0.11), the proV genes in S. enterica serovar Tennessee and S. Typhimurium LT2 were significantly induced by 12.5- and 14.9-fold, respectively. In a previous study by Finn et al. (12), the proV gene was among the most highly upregulated genes following a 4-h desiccation on a stainless steel coupon. Consistent with the above-described previous findings, we also observed a dramatic increase in the expression of the proV gene in desiccation-adapted S. Typhimurium cells in aged broiler litter. We thus further emphasized the importance of the proV gene for the survival of Salmonella in low-a_w environments.

Another gene induced during desiccation in aged broiler litter was the dnaK gene, which encodes a chaperone protein (Hsp70) that helps stabilize other proteins during heat shock (33). Gruzdev et al. (14) reported an induction of the dnaK gene in S. Typhimurium cells dehydrated in a petri dish for 22 h. When Deng et al. (11) studied the transcriptomic response of S. Enteritidis to desiccation in peanut oil (a_w, 0.30) for 72, 216, and 528 h, transcription of the dnaK gene was observed. It can thus be hypothesized that the upregulation of some heat-tolerant genes under preadaptive desiccation stress could potentially confer resistance to subsequent lethal heat stress. In contrast, Fong and Wang (13) observed that dnaK was significantly downregulated when cells were subjected to 6-day desiccation in peanut oil (a_w, 0.52). It is likely that different conditions used to desiccate cells, as well as different water activities, may contribute to different degrees of dnaK expression.

The function of the Hsp70 homologue is modified through the interaction with another heat shock protein, the GrpE protein, which is encoded by the grpE gene (34). The realization that the grpE gene was upregulated in S. Typhimurium under desiccation stress in aged broiler litter led us to hypothesize that the heat shock system could play a pivotal role when desiccation-adapted cells are exposed to high temperatures. Similarly, Deng et al. (11) also reported the transcription of the grpE gene in S. enterica serovar Enteritidis under desiccation stress in peanut oil.

Genes that are nondifferentially expressed under desiccation stress.

Two genes, otsA and otsB, are involved in trehalose biosynthesis (35). Finn et al. (12) reported that otsAB was upregulated >11-fold in S. Typhimurium under desiccation stress. A recent study by Fong and Wang (13) also showed the upregulation of otsB in S. enterica cells after a 6-day desiccation in peanut oil. Howells et al. (36) observed that after incubation at 50°C for 7 min, the number of viable cells of an S. Typhimurium ΔotsA mutant decreased to 0.6% of the initial population, whereas the wild type had an 8% survival rate. Conversely, we did not identify the upregulation of the otsA or otsB gene previously reported to be involved in the desiccation tolerance of Salmonella. Additionally, trehalose was not significantly synthesized during desiccation adaptation (P > 0.05) (Fig. 3), which further supports the idea that trehalose synthesis is not essential for the survival of S. Typhimurium cells under desiccation stress in aged broiler litter. In agreement with data from our gene expression study and trehalose content determinations, the ΔotsA mutant was not impaired in heat resistance compared to the wild type. Our results demonstrated similarity to those reported previously by Gruzdev et al. (14), who found that the habituation of Salmonella to low-a_w environments could not induce the expression of genes involved in trehalose synthesis. An early work by Hengge-Aronis et al. (37) also indicated that otsAB mutations did not impair the induction of thermal tolerance in exponentially growing E. coli cells. The different results among studies could be explained by the difference in the methods used to create low-a_w conditions. Another tentative explanation could be attributed to the fact that different strains were used. In support of this notion, Li et al. (15) found that trehalose synthesis in S. Tennessee was significantly induced by a 2-h desiccation, while no induction was detected in S. Typhimurium LT2, indicating that there are drastic differences in gene expression patterns under desiccation conditions among different strains.

In our study, the agfD gene, responsible for the biosynthesis of thin aggregative fimbriae (rdar morphology), was not upregulated during exposure to desiccation (P < 0.05), which was further confirmed by the mutation study. Our finding is consistent with data from some previous studies (12, 14, 15). These observations may indicate that while the formation of filaments may be critical for long-term persistence in low-a_w environments, this may not be the case when bacteria are desiccated under certain conditions, such as in aged broiler litter. Furthermore, we found that the desiccation-adapted rdar morphotype in aged broiler litter did not provide any benefit during heat treatment. Similarly, Scher et al. (21) also observed that the stress response of S. Typhimurium pellicle cells (rdar morphotype) to heat at 70°C was not enhanced compared to that of planktonic cells.

The β-hydroxydecanoyl ACP dehydrase, encoded by the fabA gene, is involved in the elongation of fatty acid (38). The change in the fatty acid composition of the cell membrane can affect fluidity and thermodynamics, which is believed to allow the cell to adapt to adverse conditions (39). Li et al. (15) reported that the fadA gene was upregulated 94- and 64-fold in S. Tennessee and S. Typhimurium LT2 after desiccation, respectively, which represented the greatest expression change in both strains. When Chen et al. (7) studied the expression of fatty acid biosynthesis-associated genes in Salmonella cells desiccated in granulated sugar (a_w, 0.80) for 14 days, the fadA gene was observed to be upregulated. Fong and Wang (13) also reported the significant expression of fadA in Salmonella following a 6-day desiccation in peanut oil. Conflicting results, however, were obtained in the present study; no induction of the fadA gene was detected in S. Typhimurium under desiccation conditions in aged broiler litter (P > 0.05). Similarly, some authors also observed no expression of the fadA gene in Salmonella in response to various desiccation conditions (11, 12, 14). This discrepancy could possibly be accounted for by the different water activities accompanying different methods used to develop a desiccation environment.

In the present study, we observed no induction of the kdpA gene, encoding a subunit that binds and transports K⁺ across the membrane, in desiccation-adapted S. Typhimurium cells. In contrast, in work by Gruzdev et al. (14), among all desiccation-induced genes in S. Typhimurium, the highest upregulation was observed for the kdpA gene. However, those researchers noticed that the mutation in the kdpA gene did not affect dehydration tolerance in Salmonella, but the mutant was significantly compromised during long-term persistence under desiccation conditions. These results thus implied that the Kdp high-affinity K⁺ uptake system is involved only in the early adaptation of Salmonella to a dry environment.

Deng et al. (11) reported that transcriptions of the cspA, sigDE, and dps genes were detected in S. Enteritidis in peanut oil. However, none of these genes was upregulated in our study. With respect to the difference in these findings, it can be assumed that a diverse range of stress responses could be induced in Salmonella following desiccation, depending on the precise experimental conditions involved. Accordingly, the lack of a change in the expressions of these genes may be partly due to the less stringent desiccation stress (a_w, 0.81) in aged broiler litter used in our study.

Our results revealed that in response to desiccation in aged broiler litter, S. Typhimurium showed a clear shutoff of some biological processes, such as trehalose production, rdar morphology development, and fatty acid synthesis. A conceivable explanation for this is that these processes are very energy-consuming, and a redirection of energy into other more imperative metabolic needs endows S. Typhimurium with better survival in low-a_w environments. Likewise, since invasion is not essential for the survival of S. Typhimurium under desiccation conditions in aged broiler litter, the invasion protein encoded by sigDE was not synthesized. It can thus be hypothesized that S. Typhimurium directs resources away from pathogenicity and into survival in times of desiccation stress.

Effect of duration of desiccation on thermal resistance and gene expression.

In our study, several time points (3-, 12-, and 24-h desiccation adaptations) were chosen. However, we found no significant differences in the thermal resistances or fold changes of S. Typhimurium genes among these three time points (P > 0.05). In contrast, Deng et al. (11) observed that S. Enteritidis cells desiccated in peanut oil for 216 h appeared to have increased transcriptional activity compared with cells at 72 and 528 h. More interestingly, we noticed that four upregulated genes could be induced only within a short period of time (after 1 h but in less than 2 h). This finding thus demonstrated that S. Typhimurium cells could respond rapidly to desiccation stress while producing cross-tolerance to thermal stress once they are exposed to the conditions in aged broiler litter.

Effect of rehydration on thermal resistance and gene expression.

Prior to physical heat treatment, water may be reintroduced into animal wastes via several routes, such as wet cleaning of the processing facility and rainfall events during outdoor storage. Water is also the vector by which bacteria can disseminate in the processing environment (40). In an effort to mimic this scenario of change from desiccation to hydration conditions, cells that had been dried in aged broiler litter were rehydrated and then subjected to heat treatment. To our knowledge, this is the first study on the impact of rehydration on the thermal resistance of desiccation-adapted microorganisms. It should be noted that after the 3-h rehydration, desiccation-adapted S. Typhimurium cells became significantly less heat resistant (P < 0.05). Therefore, the cross-tolerance to heat stress triggered under desiccation stress can be weakened or reversed to a normal stage when water is reintroduced into the low-a_w system. Four upregulated genes were expressed at lower levels (P < 0.05) or even nondifferentially expressed compared to the nonadapted control (P > 0.05), suggesting that they appeared to be redundant postrehydration, and our results confirmed their roles in desiccation tolerance. It can thus be speculated that upon desiccation, stress response requirements are greater due to an increased need for some stress-related products. As these genes are not necessary for survival at high a_w, only the expressions of a few essential genes in Salmonella remained active under these conditions. In the previous study by Finn et al. (12), otsAB in S. Typhimurium, which had been upregulated under desiccation stress, returned to basal levels following a 30-min rehydration, although speculatively, it is possible that the regulations of the dnaK and grpE genes, involved in heat shock protein synthesis, are somehow linked, in a manner that is as yet undefined. While the dnaK gene was upregulated under desiccation stress, a situation similar to that of the grpE gene, their fold changes in expression simultaneously became much lower following rehydration (P < 0.05).

In our previous study, we developed a two-step heat treatment consisting of a moist-heat treatment for 1 h at 65°C and a sequential dry-heat treatment for 1 h at 85°C to rapidly eliminate S. enterica cells in aged broiler litter (8). Therefore, our present rehydration study proved that moist heat is more efficient than dry heat, as S. Typhimurium cells became more heat sensitive upon rehydration. Our previous study also showed that S. enterica desiccation-adapted cells in aged broiler litter with a 20% moisture content (a_w, 0.87) were inactivated more slowly by a two-step heat treatment than in aged broiler litter with 40% (a_w, 0.98) and 50% (a_w, 0.99) moisture contents (8), indicating that desiccation adaptation of Salmonella cells in aged broiler litter with low moisture content could result in enhanced thermal resistance. The findings in the present study have thus proven the significant role of low-a_w conditions in aged broiler litter in developing the cross-tolerance of Salmonella to thermal stress.

Limitations of this study.

We acknowledge that there are some limitations of this study. Due to the lack of proV, dnaK, and grpE mutants, we cannot evaluate the contributions of these three upregulated genes to the cross-protection of S. Typhimurium against heat treatment. While this study provides some valid information on the response of Salmonella cells to desiccation, further characterization to elucidate the significance of these upregulated systems is warranted so as to gain a more holistic understanding of the processes involved in cross-protection. Additionally, as only 12 desiccation-associated genes in Salmonella were investigated, a more comprehensive study on global gene expression using whole-transcriptome sequencing technology (RNA-Seq) is necessary.

Conclusions.

Our data highlighted the rapid induction of several genes in S. Typhimurium, including the rpoS, proV, dnaK, and grpE genes, that could contribute to desiccation survival, and the rpoS gene was identified to be involved in the cross-protection of desiccation-adapted S. Typhimurium against high temperatures, while trehalose biosynthesis or rdar morphology was not found to play a significant role in this phenomenon. This investigation constitutes the first study on the thermal resistance of previously desiccated microbial cells upon the reintroduction of moisture. Our findings may aid organic fertilizer processors in the design of pathogen control strategies by taking cross-protection into consideration in order to optimize their thermal processing regimes. Such approaches should be aimed at the elimination of desiccation-adapted human pathogens from animal wastes used as biological soil amendments, thereby improving produce safety and protecting public health.

MATERIALS AND METHODS

Sample preparation.

Aged broiler litter was sourced from Cobb broiler chickens (Organic Farms, Livingston, CA). To prepare aged broiler litter, the litter was removed from the chicken house, followed by a partial windrow composting of 7 to 10 days, and then screened out of rice hulls. Aged broiler litter was dried under a fume hood until the moisture content was reduced to <20% and then screened to a particle size of <3 mm by using a sieve (sieve pore size, 3 by 3 mm). Sufficient samples were collected for all experiments and stored in a sealed container at 4°C.

Analysis of physical and chemical characteristics of poultry litter.

The moisture content was measured with a moisture analyzer (model IR-35; Denver Instrument, Denver, CO). The a_w was measured with a dew-point a_w meter (Aqualab series 3TE; Decagon Devices, Pullman, WA). Ammonia content and pH values were measured based on methods described previously by Weatherburn (41) and the U.S. Composting Council (42), respectively.

Bacterial strains and culture preparation.

The Salmonella enterica serovar Typhimurium ATCC 14028 wild-type, IB43 (rpoS::Tn10dCm), and XF373 (otsA::MudJ) strains were kindly provided by Ferric C. Fang (University of Washington, Seattle, WA), whereas S. Typhimurium MAE 110 (PagfD1; rdar, aggregate/multicellular phenotype) and MAE 119 (ΔagfD101; saw, smooth colony morphology) were obtained from Ariena H. C. van Bruggen (University of Florida, Gainesville, FL). MAE 110 constantly possesses rdar morphology, whereas MAE 119 (saw morphotype) has completely lost the rdar morphology. All mutants used in this study were derived from wild-type S. Typhimurium ATCC 14028. S. Typhimurium strains were grown overnight at 35°C in tryptic soy broth (TSB; Becton, Dickinson and Company, Sparks, MD). The cultures grown overnight were washed three times with sterile 0.85% saline, and the final pelleted cells were resuspended in 0.85% saline to obtain the desired cell concentrations based on the optical density at 600 nm.

Desiccation adaptation and rehydration of S. Typhimurium.

As described previously by Chen et al. (7), washed S. Typhimurium wild-type cultures were mixed (1:10, vol/wt) with 100 g of aged broiler litter with a low ammonia content (72.66 μg g⁻¹) at a final concentration of ca. 10 log CFU g⁻¹ by using a sterile blender (KitchenAid Inc., St. Joseph, MI). The moisture content of the inoculated litter was then adjusted to 20% (a_w, 0.81) with sterile tap water, and the litter was incubated in a sterile container covered loosely with aluminum foil at room temperature for 1-, 2-, 3-, 12-, and 24-h desiccation adaptations. Salmonella cells added to aged broiler litter and immediately used for heat treatment (0-h desiccation adaptation) served as controls (nonadapted cells).

To perform rehydration after 3-, 12-, and 24-h desiccation adaptations, the inoculated litter (10 g) was mixed with 30 ml of sterile water in a Whirl-Pak bag (Nasco, Fort Atkinson, WI) (12) and homogenized by using a Seward 400 Circulator Lab blender (Seward Ltd., Worthing, West Sussex, UK). The homogenate was kept at room temperature for 3 h.

Thermal inactivation of desiccation-adapted and rehydrated S. Typhimurium cells.

To prepare desiccation-adapted cells as an inoculum in aged broiler litter for thermal inactivation, 10 g of the litter sample was homogenized with 30 ml of sterile water immediately before heat treatment, allowing its initial population to be consistent with that of rehydrated cells. With respect to the inoculation of aged broiler litter, the homogenate with desiccation-adapted or rehydrated cells was centrifuged at 1,500 rpm for 1 min to remove large litter particles. The supernatant was then collected and mixed (1:100, vol/wt) with 500 g of aged broiler litter. The moisture content of the inoculated litter was then adjusted to 20% with sterile tap water for subsequent heat treatment. The initial population of Salmonella was ca. 7 log CFU g⁻¹.

After mixing with desiccation-adapted or rehydrated cells, 20 g of litter samples in triplicate was distributed evenly into an aluminum pan (10-cm internal diameter [ID]), placed at three different locations (close to the door, in the center, and far away from the door) on the shelf of a controlled convectional oven (Binder Inc., Bohemia, NY), and then exposed to dry heat at 75°C for up to 2 h. The temperature was continuously monitored with T-type thermocouples (DCC Corp., Pennsauken, NJ), with one cord inside the oven chamber and others in litter samples. The temperature of the oven was initially set at 80°C to reduce the come-up time (0.87 h). When the interior temperature of the litter sample reached the target temperature, the temperature setting was readjusted to 75°C. Chicken litter samples were withdrawn from the oven every 0.5 h to determine Salmonella populations. Samples were transferred into a Whirl-Pak bag and placed immediately in an ice water bath to cool down the samples.

Bacterial enumeration.

Salmonella cells were counted by using a modified two-step overlay method with xylose-lysine-Tergitol 4 (XLT-4) agar (Becton, Dickinson and Company, Sparks, MD) as the selective medium and tryptic soy agar (TSA; Becton, Dickinson and Company, Sparks, MD) as the nonselective medium to allow heat-injured cells to resuscitate (7). Samples negative for Salmonella by a direct plating method were preenriched in universal preenrichment broth (UPB; Neogen Corp., Lansing, MI), followed by secondary enrichment in Rappaport-Vassiliadis (RV) broth (Becton, Dickinson and Company, Sparks, MD). After a 24-h incubation at 42°C, enriched cultures were plated onto XLT-4 agar. The detection limits of direct plating and enrichment were 1.30 and 1.00 log CFU g⁻¹, respectively.

Gene expression study on desiccation-adapted and rehydrated S. Typhimurium cells.

For the gene expression study, in order to extract high-quality RNA from S. Typhimurium, autoclaved aged broiler litter (dry cycle at 121°C for 15 min) was used for desiccation adaptation to reduce interference from background microflora (43), and desiccation adaptations and rehydration were conducted as mentioned above.

After desiccation adaptation or rehydration, the inoculated litter (10 g) was mixed with 50 ml of phosphate-buffered saline (PBS) in a Whirl-Pak bag and homogenized with the Seward 400 Circulator Lab blender. The mixture was centrifuged at 1,500 rpm for 1 min to remove large particles. S. Typhimurium cells were then separated by using anti-Salmonella Dynabeads (Thermo Fisher Scientific Inc., Asheville, NC), as described previously by Singh and Jiang (43). Afterwards, the cells attached to Dynabeads were washed twice in PBS, and 2 volumes of RNAProtect Bacteria reagent (Qiagen, Valencia, CA) were then added to 1 volume of PBS with beads for the stabilization of Salmonella RNA.

Afterwards, the total RNA in S. Typhimurium cells was extracted by using an RNeasy minikit (Qiagen) according to the manufacturer's instructions. Contaminating DNA was removed with a Turbo DNA-free kit (Thermo Fisher Scientific Inc.). A NanoDrop spectrophotometer (Thermo Fisher Scientific Inc.) was used to determine the RNA concentration and purity (A₂₆₀/A₂₈₀). The overall RNA integrity was analyzed by gel electrophoresis in a 1.2% nondenaturing agarose gel in Tris-acetate buffer at 60 V for 1 h. RNA was heated at 70°C for 1 min and chilled on ice to denature the secondary structures before loading onto the gel. The gel was stained in an ethidium bromide solution and visualized with a Gel Doc 1000 system (Bio-Rad Laboratories Inc., Hercules, CA).

Desiccation-associated genes were selected based on data reported in the literature (10–15). Quantitative real-time reverse transcriptase PCR (qRT-PCR) was performed in a Mastercycler ep gradient S thermal cycler (Eppendorf, Hamburg, Germany) using the SuperScript III Platinum SYBR green one-step qRT-PCR kit (Thermo Fisher Scientific Inc.). Primer sequences and qRT-PCR procedures (temperatures and incubation times) for 12 desiccation-associated genes and a reference gene were based on data reported in the references shown in Table 2. PCRs were performed with RNA extracts from two independent trials, each with three replicates. The Pfaffl method was used to calculate the relative expression (fold change) of gene targets (44). Primer efficiency was determined by 10-fold serial dilution of the extracted RNA and calculated as described previously by Pfaffl (44). Gene expression was normalized against 16S rRNA as the reference gene. Differentially expressed genes were defined as being significantly up- or downregulated by ≥2-fold between desiccation-adapted or rehydrated cells and nonadapted control cells (P < 0.05) (15).

TABLE 2.

Selected genes and their primers used for qRT-PCR^a

Gene	Primer sequence (5′–3′)	Primer efficiency (%)	Reference
rpoS (regulator of the general stress response)	F, CAAGGGGAAATCCGTAAACCC	100	48
	R, GCCAATGGTGCCGAGTATC
otsA (trehalose-6-phosphate phosphatase)	F, GGAGTGGCGAGACAGGTAAC	98	49
	R, AGAACCGCATTGGAAAATTG
otsB (trehalose-6-phosphate synthase)	F, ACCTTGATGGCACATTGGCAGA	96	50
	R, ACGCCCTGAAATCAATGCCA
agfD (positive regulator of thin aggregative fimbriae production)	F, GTGCTCGAGGGACTTCATTAAACATGATG	103	16
	R, GCCGGATCCTGTTTTTCATGCTGTCAC
kdpA (K+-transporting ATPase subunit A)	F, GGCGCTACTGACGCTCAATC	97	51
	R, AGGCTTGCCAGTTGGTATTGG
proV (osmoprotectant transporters)	F, CCACAATGGTACGCCTTCTCA	96	12
	R, GCATGAGCGCAAATGACTGGA
fadA (fatty acid metabolism)	F, ATCTCTCCGCCCACTTAATGCGTA	101	15
	R, AGCCTTGCTCCAGCGTTTGTTGTA
dnaK (chaperone protein)	F, CGATTATGGATGGAACGCAGG	104	52
	R, GGCTGACCAACCAGAGTT
cspA (cold shock protein)	F, GTTCAACGCTGATAAAGGCTTCGG	96	53
	R, CAGGCTGGTTACGTTGCCAGC
sigDE (invasion protein)	F, TGGCATAAAGGGACAGCAC	99	54
	R, AGCGGCAAAGATCGTACAG
dps (starvation/stationary-phase protection protein)	F, CCCGTAACGATGTATCAGAG	103	55
	R, GCGCTCGGCCATAGTATCCA
grpE (heat shock protein)	F, CAGAAAACGCCTGAGGGGCA	96	55
	R, CGCAGGTTTTCCATTTCCGC
16S rRNA (reference gene)	F, CCTCAGCACATTGACGTTAC	98	56
	R, TTCCTCCAGATCTCTACGCA

^aF, forward; R, reverse.

Mutation study.

The survival data for S. Typhimurium IB43, XF373, MAE 110, and MAE 119 in aged broiler litter with a 20% moisture content at 75°C were compared with those for the wild type by direct plating and enrichment methods. Desiccation adaptation and thermal inactivation were carried out as described above.

Trehalose content determination.

As trehalose synthesis was previously observed under desiccation conditions (15), the intracellular trehalose content was determined. Desiccation-adapted cells were separated by using Dynabeads as described above. Cells attached to Dynabeads were suspended in 5 ml of PBS and disrupted by subjecting cells to vortexing for 5 min in the presence of 5 g of disruptor beads with a diameter of 0.1 mm (Electron Microscopy Sciences, Hatfield, PA). Cell debris, disruptor beads, and Dynabeads were then removed by centrifugation at 5,000 rpm for 10 min. The supernatant was used for the subsequent determination of the trehalose content.

The trehalose content was measured by using an enzymatic colorimetric assay by converting trehalose to glucose with trehalase and then determining the glucose content (45, 46). Briefly, 50 μl of the extract was incubated with 150 μl of 0.2 mol liter⁻¹ sodium acetate buffer (pH 5.5) and 50 μl of 0.3 U ml⁻¹ trehalase at 37°C for 6 h. The trehalose content was calculated from the amount of liberated glucose, which was quantified by using a glucose assay kit (Abnova Corporation, Taipei, Taiwan). Standard trehalose solutions were used to generate a trehalose standard curve, with concentrations ranging from 0 to 10 nmol of trehalose (Sigma-Aldrich Corp., St. Louis, MO) per reaction mixture. To nullify the interfering effect of glucose from other sources, the background glucose content in the extract was determined with a parallel control containing the above-described reaction mixture components in which trehalase was replaced with distilled water and then subtracted from the total glucose content to obtain the net glucose content (47). The trehalose content was expressed as trehalose weight per total water-soluble protein weight (nanomoles trehalose per milligram protein). The protein content was determined by using a Pierce bicinchoninic acid (BCA) protein assay kit (Thermo Fisher Scientific Inc.).

Statistical analysis.

All results were obtained from two independent trials. Plate count data were converted to log CFU per gram (dry weight). SigmaPlot 12.3 (Systat Software Inc., San Jose, CA, USA) was applied for data analysis. Analysis of variance (ANOVA), followed by the least-significant-difference (LSD) test, was performed to determine whether there were significant differences (P < 0.05) among different groups.