OUP user menu

Deletion of the aceE gene (encoding a component of pyruvate dehydrogenase) attenuates Salmonella enterica serovar Enteritidis

Ervinna Pang, Chang Tien-Lin, Madhan Selvaraj, Jason Chang, Jimmy Kwang
DOI: http://dx.doi.org/10.1111/j.1574-695X.2011.00834.x 108-118 First published online: 1 October 2011

Abstract

Salmonella enterica serovar Enteritidis (S. Enteritidis) is a major food-borne pathogen. From a transposon insertion mutant library created previously using S. Enteritidis 10/02, one of the mutants was identified to have a 50% lethal dose (LD50) at least 100 times that of the parental strain in young chicks, with an attenuation in a poorly studied gene encoding a component of pyruvate dehydrogenase, namely the aceE gene. Evaluation of the in vitro virulence characteristics of the ΔaceE∷kan mutant revealed that it was less able to invade epithelial cells, less resistant to reactive oxygen intermediate, less able to survive within a chicken macrophage cell line and had a retarded growth rate compared with the parental strain. Young chicks vaccinated with 2 × 109 CFU of the ΔaceE∷kan mutant were protected from the subsequent challenge of the parental strain, with the mutant colonized in the liver and spleen in a shorter time than the group infected with the parental strain. In addition, compared with the parental strain, the ΔaceE∷kan mutant did not cause persistent eggshell contamination of vaccinated hens.

Keywords
  • Salmonella Enteritidis
  • aceE gene
  • mutant
  • attenuation
  • chickens
  • vaccine

Introduction

Salmonella spp. is a gram-negative, rod-shaped, motile and facultative anaerobe bacterium that normally resides in the gut of wild and domestic animals. It is one of the major causes of food-borne illnesses in humans, with poultry and eggs being the most common sources of human Salmonella infections (Perales & Audicana, 1988; Baumler et al., 2000; Hald et al., 2004). In Europe, approximately 80% of food-borne salmonellosis cases in humans are caused by Salmonella enterica serovar Enteritidis (S. Enteritidis) (World Health Organization, 2000). Salmonella enterica serovar Enteritidis is able to colonize the tissues of the ovary and oviduct of egg-laying hens without causing clinical disease (Keller et al., 1995), with the presence of S. Enteritidis in or on the eggs often caused by direct contamination of the albumen, yolk, eggshell membrane or eggshell before or during oviposition from the infected reproduction organs of laying hens or by fecal contamination (Timoney et al., 1989; Gast & Beard, 1990; Barrow & Lovell, 1991; Humphrey et al., 1991). Infection of S. Enteritidis in young chicks can lead to paratyphoid disease and results in a high mortality rate, but in adult chickens, it usually results in symptomless carriage, which constitutes a risk for public health (Dhillon et al., 1999; Wigley et al., 2002).

Salmonellosis can be treated with antibiotics; however, this is not economically feasible over the long term, and the widespread use of antibiotics encourages the emergence of multi-drug-resistant pathogens (Georges-Courbot et al., 1990; Shivaprasad et al., 1990; Ling et al., 1998). The best long-term solution is to eliminate Salmonella spp. from the food chain before it reaches humans and it can be readily achieved through the vaccination of poultry. Most of the present knowledge on S. Enteritidis pathogenesis and host immunity originates from investigations in mice or adult chickens, with little known about the systemic infection of young chicks and whether other genes additional to known virulence genes are required for virulence.

Previously, through transposon mutagenesis, we generated an insertion library of 1246 S. Enteritidis mutants, out of which 384 mutants were selected randomly for the first phase of in vivo screening in 3-day-old chicks for attenuation (Chang et al., 2008). In young chicks, 12 mutants (3.1%) were found to have a 50% lethal dose (LD50) at least 100 times that of the parental strain (Chang et al., 2008). Most of the 12 mutants have been reasonably well characterized by others, except for two mutants: one with an insertion in the yhbC gene and another with an insertion in the aceE gene (Chang et al., 2008). We have characterized the in vitro virulence characteristics as well as the persistence and serum response from a live challenge of chickens with the ΔyhbC mutant (Chang et al., 2008). Therefore, in this study, we mainly focus on the ΔaceE∷kan mutant, with its LD50 being at least 100 times higher than the parental strain. Studies were performed to determine the virulence characteristics of the ΔaceE∷kan mutant in vitro, persistence of the mutant in a live challenge of young chicks and protection against a subsequent challenge of wild-type S. Enteritidis with the induction of a serum IgG immune response, as well as investigation on eggshell contamination of the vaccinated chickens. This is the first time that a comprehensive study has been performed on the aceE gene of S. Enteritidis and is useful to provide a better understanding of the pathogenesis of S. Enteritidis in young chicks.

Materials and methods

Chickens

One-day-old Salmonella-free chicks were purchased from a commercial hatchery in Singapore. The young chicks were housed in clean cardboard boxes, and for long-duration experiments, chicks were transferred to wire mesh cages when they were 2 weeks old. Boxes and cages were kept in different rooms for different treatments. All chickens had unrestricted accessed to commercial antibiotic-free starter feed and water. At the end of the trials, the remaining chickens were euthanized humanely. All animal protocols adhered to the guidelines provided by the National Advisory Committee for Laboratory Animals in Research (NACLAR) in Singapore, and protocols were approved by the IACUC.

Bacterial strains and culture

The bacterial strains used in this study and their culture were according to Chang (2008). An API 20E biochemical test kit (Biomerieux) as well as PCR (with genus- and serotype-specific primers) were used to confirm that the isolate was indeed S. Enteritidis (Chang et al., 2008).

Construction of the ΔaceE∷kan mutant and determination of its LD50

From the mutant library created previously (Chang et al., 2008), one of the mutants with an insertion in the aceE gene was identified to be attenuated for virulence in young chicks, and was selected for further study. A new aceE deletion mutant was generated by deleting the entire aceE gene, except for the first and the last 10 nucleotides, using the Lambda Red recombinase system (Datsenko & Wanner, 2000). We followed the protocol exactly, with the exception that 10–30 mM of l-arabinose was used to induce the transcription of the lambda red genes from pKD46 instead of 1 mM. Briefly, a set of primers, aceE-F (GGGACAGGTTCCAGATAACTCAACGTATTAGATAGATAAGGAATACCCCCATGTCAGAACGTGTAGGCTGGAGCTGCTTC) and aceE-R (CAGAATCTCGGTGATTTCAACTTCATCTGTCCCGATGTCCGGTACTTTGATTTCGATAGCTTCCGGGATCCGTCGACCT) was designed to generate a PCR product that was required for deleting the aceE gene through homologous recombination and carried an antibiotic resistance gene that would replace the target gene to be deleted on the chromosome. Each primer was 80 nucleotides in length and contained 60 nucleotides flanking the start or the end of the aceE gene and 20 nucleotides corresponding to the P1 or the P4 sequence on the pKD13 vector, which carried the kanamycin resistance gene. The PCR products were then introduced into S. Enteritidis 10/02 carrying the plasmid pKD46 using electroporation by standard methods (Sambrook & Russel, 2001), resulting in the aceE gene, which was replaced with the kanamycin resistance gene (ΔaceE∷kan mutant). All nucleotide sequences were based on the nucleotide sequence of S. enterica Enteritidis PT4 NCTC 13349, available from the Sanger Institute, UK. Another set of primers, aceE-F2 (CGCACAACGTGGTATTGCTTCA) and aceE-R2 (CGGTGATCAGCGACTGTTCA), were used for PCR and DNA sequencing to confirm the deletion of the target gene.

In order to determine the LD50 of the newly constructed ΔaceE∷kan mutant strain, seven groups of five, 3-day-old chicks were established at the same time. In a randomized manner, each chick of Groups 1, 2 and 3 was subcutaneously vaccinated with 2 × 109, 2 × 1010 and 2 × 1011 CFU of the ΔaceE∷kan mutant (suspended in 100 µL of phosphate-buffered saline [PBS]), respectively; each chick of Groups 4, 5 and 6 was subcutaneously infected with 2 × 106, 2 × 107 and 2 × 108 CFU of S. Enteritidis 10/02 wild type (suspended in 100 µL of PBS), respectively; each chick of Group 7 (control) was injected with 100 µL of PBS. Chicken mortalities were recorded over a 10-day period. As described in Chang (2008), most deaths occurred within 3 days after being challenged with the S. Enteritidis parental strain; thus, a 10-day observation period was sufficient to determine the LD50 of the ΔaceE∷kan mutant. Median lethal doses of S. Enteritidis 10/02 and the ΔaceE∷kan mutant were determined according to Reed & Muench (1938).

Complementation of the ΔaceE∷kan mutant

The aceE gene was amplified from the S. Enteritidis 10/02 genome using primers aceE-compF (CCGGGGATATCTAAGGAGGATATTCATATGTCAGAACGTTTCCAAAATGA)and aceE-compR (CCGGGGTCGACTTACGCCAGACGCGGGTTAA)with an EcoRV and a SalI site, respectively. The PCR product was then cloned into pBR322 and transformed directly into the ΔaceE∷kan mutant. Plasmid pBaceE (with the presence of pBR322 native promoters, but not the aceE gene promoter) was isolated from the ampicillin-resistant transformants and the presence of the aceE insert was verified by PCR. A chicken challenge experiment to access whether the expression of aceE from pBaceE plasmid could restore wild-type virulence was performed according to Chang (2008). LD50 of the complemented strain was determined by including both wild-type S. Enteritidis 10/02 and PBS control groups at the same time. The median lethal dose of the ΔaceE∷kan+pBaceE complemented strain was determined according to Reed & Muench (1938).

Bacterial growth rate

The growth rates of the S. Enteritidis 10/02 wild type, the ΔaceE∷kan mutant and the ΔaceE∷kan+pBaceE complemented strain were determined according to Chang (2008), with modifications. Briefly, the number of bacteria inoculated at the start of the assay (0 h) was 2 × 104 CFU and the growth rate was determined over a 24-h period. From the growth curve, the mean generation times of S. the Enteritidis 10/02 wild type, ΔaceE∷kan mutant and ΔaceE∷kan+pBaceE were calculated accordingly.

In vitro assays of virulence

Assays to determine the resistance of the ΔaceE∷kan mutant to ROI, RNI, sodium deoxycholate and normal chicken serum, survival of the ΔaceE∷kan mutant in chicken macrophages, in the stationary growth phase and under acidic conditions, and the ability of the ΔaceE∷kan mutant to invade cultured HeLa cells compared with the parental strain were performed according to Chang (2008), with modifications. Other than the wild-type S. Enteritidis 10/02 and ΔaceE∷kan mutant, the ΔaceE∷kan+pBaceE complemented strain was also tested.

Chicken trials

Preliminary vaccination trial

All vaccinations were performed via a subcutaneous injection of the bacteria (suspended in 100 µL of PBS) into the chickens. In this trial, four groups of 20, 3-day-old chicks were established at the same time. Three of the four groups were vaccinated with the ΔaceE∷kan mutant, with each of the group given different bacteria doses. In a randomized manner, Group 1 was vaccinated with 2 × 106 CFU per chick, Group 2 with 2 × 107 CFU per chick and Group 3 with 2 × 108 CFU of ΔaceE∷kan mutant per chick. Group 4, which was a control group, was injected with 100 µL of PBS per chick. Up to 5 weeks, three chicks from each group were sacrificed at weekly intervals to collect their liver and spleen in order to determine the persistence of the ΔaceE∷kan mutant in the chicks. Blood (500–1500 µL) was collected from the wing vein. The liver and spleen were collected from euthanized chickens, weighed and then homogenized with a roller bottle in plastic sealable bags containing PBS. The homogenate was then serially diluted in PBS, plated onto XLD agar and incubated at 37 °C for 24–48 h. Cloacal swabs were taken and plated onto XLD agar to determine the duration of shedding of the ΔaceE∷kan mutant in the feces. Bacterial clearance was determined by a series of direct plating (serial dilution in PBS) onto XLD agar without enrichment, with the minimal detectable bacterial level determined to be 100 CFU g−1.

Vaccination and challenge trial

All vaccination, infection and challenge trials were performed via a subcutaneous injection of the bacteria (suspended in 100 µL of PBS) into the chickens. Three groups of 3-day-old chicks were established at the same time. In a randomized manner, Group 1 consisted of 70 young chicks, and each chick was vaccinated with 2 × 109 CFU of ΔaceE∷kan mutant; Group 2 consisted of 60 chicks, and each chick was infected with 2 × 107 CFU of S. Enteritidis 10/02; and Group 3 consisted of 60 chicks, and each chick was injected with 100 µL of PBS. Five chicks from each group were sacrificed at weekly intervals to collect their liver and spleen in order to determine the persistence of the ΔaceE∷kan mutant and the parental strain in the chicks. Blood, liver and spleen, and cloacal swabs were collected as described above. The mortality rate and body weight (per week) of all the groups were recorded. After 6 weeks, each chicken from the Group 1 (ΔaceE∷kan vaccinated) and Group 3 (PBS injected) was then challenged with a high dose of 1 × 1012 CFU of S. Enteritidis 10/02. Blood, liver and spleen were collected as described above. Up to 14 days, three chickens from each group were sacrificed every 2–4 days to collect their liver and spleen in order to determine the persistence of the ΔaceE∷kan mutant and the parental strain in the chickens. In this period, the mortality rate and body weight (per week) of the challenged chickens were recorded as well. Refer to Supporting Information, Fig. S1 for detailed information on the numbers of chicken that were vaccinated/infected/challenged, sacrificed and that died corresponding to the time frame of the entire trial.

Antibody response

Blood collected from the live vaccination and the challenge trial above was allowed to clot at 37 °C for 1 h and centrifuged at 13 000 g for 5–10 min. The sera collected were used to determine the antibody (IgG) response in chickens to the vaccination of different doses of ΔaceE∷kan mutants. The level of serum IgG was determined via an enzyme-linked immunosorbent assay (ELISA) as described previously (Chang et al., 2008).

Eggshells contamination of ΔaceE∷kan vaccinated hens

Three groups of 12, 200-day-old laying hens were established at the same time. In a randomized manner, Group 1 was vaccinated intravenously with 200 µL of 1 × 109 CFU of the ΔaceE∷kan mutant (suspended in PBS) per hen, Group 2 was infected intravenously with 200 µL of 1 × 107 CFU S. Enteritidis 10/02 (suspended in PBS) per hen and Group 3 (control) was injected intravenously with 200 µL of PBS per hen. Shell swabs and fecal samples were collected at weekly intervals, and were cultured in buffered peptone water for 24 h at 37oC. Three drops of the culture were then plated at the center of Modified Semi-Solid Pappaport Vassiliadis (MSRV) plates (Merck) to detect S. Enteritidis contamination. The samples were also plated on XLD agar to determine the bacterial number. Precautions were taken to ensure that fecal contamination in the environment (after oviposition) was reduced to the minimum. At week 7, the spleen and liver of laying hens from Groups 1 and 2 were collected to determine the presence of mutant and parental strains in the hens.

Statistical analysis

Statistical significance was assessed using Student's unpaired t test, and a P value of <0.05 was considered significant.

Results

LD50 of the S. Enteritidis parental strain, the ΔaceE∷kan mutant and the ΔaceE∷kan+pBaceE complemented strain

From the infection of three different groups of 3-day-old chicks with either 2 × 106, 2 × 107 or 2 × 108 CFU of the S. Enteritidis parental strain, the LD50 of parental strain was determined to be 2 × 107 CFU (over a 10-day observation period). This was in accordance to what was reported previously (Chang et al., 2008). At the same time, from the vaccination of another three different groups of 3-day-old chicks with either 2 × 109, 2 × 1010 or 2 × 1011 CFU of the ΔaceE∷kan mutant strain, vaccination of 3-day-old chicks with 2 × 109 CFU of the ΔaceE∷kan mutant did not result in any mortality; however, three (out of five) and five (out of five) chicks died when vaccinated with 2 × 1010 and 2 × 1011 CFU of the ΔaceE∷kan mutant, respectively. Therefore, the LD50 of the ΔaceE∷kan mutant was determined to be approximately 1010 CFU. The median lethal doses of the S. Enteritidis parental strain and the ΔaceE∷kan mutant calculated were 8.7 × 107 and 1.35 × 1010 CFU, respectively.

For complementation studies, in a separate animal trial from above, nine groups of 10, 3-day-old chicks were established (at the same time). The results showed that the ΔaceE∷kan mutant, when complemented with pBaceE, regained its ability to kill 3-day-old chicks at a level comparable to the parental strain, confirming that only the aceE gene was targeted for mutation. In this trial, when given a dose of 5 × 109 CFU of either the parental or the complemented strain, the mortality rate of the chicks was 100%; at a dose of 5 × 108 CFU, 70% and 90% mortality rates were observed when the young chicks were infected with 5 × 108 CFU of the parental and the complemented strain, respectively, while, at the dose of 5 × 107 CFU of the parental and complemented strains, 40% and 50% mortality rates were observed, respectively; 20% mortality rate was obtained when young chicks were infected with 5 × 106 CFU of either the parental or the complemented strain. The median lethal dose of the ΔaceE∷kan+pBaceE complemented strain calculated was 4 × 107 CFU.

Bacterial growth rate

Growth retardation was observed on the ΔaceE∷kan mutant when compared with the parental and complemented strain over a 24-h growth period (Fig. 1). The mean generation times of the wild-type S. Enteritidis, ΔaceE∷kan mutant and the ΔaceE∷kan+pBaceE complemented strain were 21.5, 25.8 and 22.4 min, respectively. The number of bacteria inoculated at the start of the assay was 2 × 104 CFU; after a 24-h incubation time under a condition of freely accessible nutrients and oxygen, the number of ΔaceE∷kan mutants was at least one log lower than the parental and complemented strains. The ΔaceE∷kan mutant clearly has a slower growth rate.

Figure 1

Growth curve of the Salmonella enterica serovar Enteritidis (S. Enteritidis) 10/02 parental strain (indicated as wild type [WT] in the graph), the ΔaceE∷kan mutant and the ΔaceE∷kan+pBaceE-complemented strain over a 24-h period. The initial inoculum size was 2 × 104 CFU. Points, mean of three independent experiments. Bars, SD. *P<0.05, **P<0.01, significantly different from the parental strain (Student's t-test).

In vitro assays of virulence

A range of in vitro assays were performed to compare the phenotypes of the ΔaceE∷kan mutant with the parental strain, which may contribute to the difference of virulence in vivo. No significant differences were determined between the ΔaceE∷kan mutant and the parental strain in their ability to grow after a prolonged stationary phase, to survive under acidic conditions, to resist killing by RNI, sodium deoxycholate and chicken serum (data not shown). However, our data show that the ΔaceE∷kan mutant was significantly less able to invade HeLa cells, with the intracellular bacterial number reaching only ∼17% that of the parental strain (Table 1). The ΔaceE∷kan mutant was also significantly less resistant to ROI compared with the parental strain, as indicated by larger zones of inhibition (Table 1). In terms of survival in chicken macrophages (HD11 cells), chicken macrophages were able to eliminate the intracellular mutant completely within 24 h postinfection as no bacteria could be recovered by plate count (Table 1). In contrast, the parental strain was still recoverable (up to 48 h postinfection), indicating that it could resist complete killing by HD11 cells (Table 1). Escherichia coli cells (negative control) were almost completely killed off after 7 h postinfection. In addition, the ΔaceE∷kan mutant, when complemented with pBaceE, was able to restore the in vitro phenotypes of the parental strain as mentioned above, confirming that the deletion of the aceE gene contributed to the difference between the ΔaceE∷kan mutant and the parental strain.

View this table:
Table 1

In vitro virulence assays in which there were differences in responses between the ΔaceE kan mutant and the parental strain, Salmonellaenterica serovar Enteritidis (S. Enteritidis) 10/02

Resistance to ROISurvival in macrophages
Log10 CFU at
In vitro assay% invasion of HeLa cells (%)Zone of inhibition (cm)0 h1 h4 h7 h24 h48 h
S. Enteritidis 10/021.44 ± 0.273.67 ± 0.036.55 ± 0.073.60 ± 0.063.57 ± 0.053.40 ± 0.052.79 ± 0.072.94 ± 0.03
ΔaceE∷kan0.24 ± 0.073.83 ± 0.036.57 ± 0.183.28 ± 0.072.90 ± 0.032.62 ± 0.1200
ΔaceE∷kan+pBaceE1.51 ± 0.303.69 ± 0.026.57 ± 0.083.72 ± 0.133.59 ± 0.043.47 ± 0.72.72 ± 0.103.02 ± 0.08
  • Percentage of invasion was calculated from the number of bacteria recovered intracellularly from the infected HeLa cells (after removing and killing of extracellular bacteria) divided by the initial inoculum.

  • P<0.05;

  • P<0.01;

  • P<0.001, significantly different from the parental strain (Student's t-test).

Vaccination and challenge trials

A preliminary trial on the vaccination of young chicks with 2 × 106, 2 × 107 and 2 × 108 CFU of the ΔaceE∷kan mutant showed that the mutant did not persist in the liver and spleen of chickens for long (Fig. 2a and b). At the same dose of 2 × 107 CFU, the amount of bacteria recovered from the liver and spleen of vaccinated chickens was significantly lower than the parental strain-infected chickens; in fact, the ΔaceE∷kan mutant was not detected in the liver and spleen of vaccinated chickens after 2 and 4 weeks of postvaccination, respectively, while the parental strain persisted in the liver and spleen for >5 weeks (Fig. 2a and b). This showed that in chickens, the ΔaceE∷kan mutant had reduced colonization abilities than the parental strain.

Figure 2

Vaccination and challenge trial — persistence and bacterial load of the ΔaceE∷kan mutant in comparison with the Salmonella enterica serovar Enteritidis (S. Enteritidis) 10/02 parental strain (indicated as wild type [WT] in the graph). Five groups of 3-day-old chicks were vaccinated with different doses of the ΔaceE∷kan mutant (2 × 106, 2 × 107, 2 × 108 or 2 × 109 CFU) or infected with 2 × 107 CFU of the parental strain. (a) Bacterial load in the liver. (b) Bacterial load in the spleen. Columns, mean of three chick samples (for the groups of chickens vaccinated with 2 × 106, 2 × 107 or 2 × 108 CFU of the ΔaceE∷kan mutant) or mean of five chick samples (for the group of chickens vaccinated with 2 × 109 CFU of the ΔaceE∷kan mutant and the group infected with 2 × 107 CFU of the parental strain). Bars, SD. Once the parental strain was not detected (below 100 CFU g−1) from the liver or the spleen, the group of chickens previously vaccinated with 2 × 109 CFU of the ΔaceE∷kan mutant and a control group previously injected with PBS were challenged with 1012 CFU of the parental strain. (c) Bacterial load in the liver after challenge. (d) Bacterial load in the spleen after challenge. Columns, mean of three chicken samples. Bars, SD. *P<0.05, **P<0.01, significantly different from the parental strain (Student's t-test).

In order to evaluate whether the ΔaceE∷kan mutant could be used as a live vaccine, another vaccination trial was conducted to measure the ability of vaccinated chickens to reduce S. Enteritidis colonization from the liver and spleen in a subsequent challenge. When two groups of 3-day-old young chicks were vaccinated either with 2 × 109 CFU of the ΔaceE∷kan mutant or infected with 2 × 107 CFU of the parental strain, the ΔaceE∷kan mutant persisted in the liver and spleen for at least 3 and 5 weeks, respectively, 1–2 weeks shorter than the group infected with the parental strain (Fig. 2a and b).

In a subsequent challenge of 1012 CFU of the parental strain into the group of chicken previously vaccinated with 2 × 109 CFU ΔaceE∷kan mutant and into a control group previously injected with PBS, the parental strain persisted in the liver of the control group for a much longer time than the ΔaceE∷kan vaccinated group (>6days) and the numbers of bacteria recovered from the control group were constantly higher than the group vaccinated with the ΔaceE∷kan mutant (Fig. 2c), while, up to 14 days, bacteria were still detected in the spleen of both ΔaceE∷kan vaccinated and control groups, and the numbers of bacteria recovered from the spleen of ΔaceE∷kan vaccinated chickens were 1–2 logs lower than the control group (Fig. 2d). In the challenge of 1012 CFU of the parental strain into adult chickens, 27% of control group chickens (previously injected with PBS) died from the challenge; this was much higher than the ΔaceE∷kan vaccinated chickens (3.3%). Immediately after being challenged, chickens from all groups began to eat less and lose body weight. After 4 days, the ΔaceE∷kan mutant-vaccinated chickens began to gain weight faster than the control group. Overall, there was a significant protective effect from wild-type S. Enteritidis in young chicks previously vaccinated with the ΔaceE∷kan mutant, in particular, with a reduction in the wild-type bacterial load in the liver, and with a shorter bacterial colonization time than the vaccinated chickens.

Serum response

Subcutaneous vaccination of 2 × 106, 2 × 107 and 2 × 108 CFU of the ΔaceE∷kan mutant induced a specific IgG response in a bell-shaped curve typical of primary exposure to an antigen as indicated by ELISA (Fig. S2), confirming that the young chicks were not exposed to S. Enteritidis before this trial. However, the levels of serum IgG induced from these vaccinations were very low. In contrast, a significant level of IgG was detected when the young chicks were vaccinated with 2 × 109 CFU of the ΔaceE∷kan mutant, but this is still in a level lower than the group infected with 2 × 107 CFU of the parental strain (Fig. 3a). Following a subsequent challenge with a high dose of the parental strain (1012 CFU), chickens previously vaccinated with 2 × 109 CFU of the ΔaceE∷kan mutant were able to mount a higher IgG response than the PBS control group (Fig. 3b).

Figure 3

Serum IgG response of chickens vaccinated with the ΔaceE∷kan mutant and subsequently challenged with the Salmonella enterica serovar Enteritidis (S. Enteritidis) 10/02 parental strain (indicated as wild type [WT] in the graph). Using purified recombinant flagellin of S. Enteritidis 10/02 as the target antigen, sera were diluted 1 : 50 and tested by ELISA. (a) Serum response of 3-day-old chicks vaccinated with 2 × 109 CFU of the ΔaceE∷kan mutant or infected with 2 × 107 CFU of the parental strain. Following this, up to 6 weeks, blood was collected from sacrificed chicks at weekly intervals. Points, average value from three independent ELISA assays, and mean of five different serum samples collected from different chicks. Bars, SD. (b) Serum response of approximately 6-week-old chickens (previously vaccinated with 2 × 109 CFU of the ΔaceE∷kan mutant or injected with PBS) challenged with 1012 CFU of the parental strain. Blood was collected from sacrificed chickens at 2–4-day intervals, up to 14 days. Points, average value from three independent ELISA assays, and mean of three different serum samples collected from different chickens. Bars, SD. *P<0.05, **P<0.01, significantly different from the parental strain (Student's t-test).

Eggshells contamination of ΔaceE∷kan-vaccinated hens

In the first week, from the level of eggshell contamination, the group of hens vaccinated with 109 CFU of the ΔaceE∷kan mutant started laying eggs with about a 1 log higher bacterial count than the group of hens infected with 107 CFU of the parental strain. However, in <2 weeks, the bacterial count reduced to a level that was lower than the parental strain-infected hens (Fig. 4a). The mutant strain persisted on the eggshells of hens for <3 weeks, while, for >3 weeks, a high number of bacteria was still recovered from eggshells of hens infected with the parental strain (Fig. 4a). From the fecal samples, the ΔaceE∷kan mutant was still recovered during the third week postvaccination; however, this was at least 1 week shorter than the parental strain-infected hens (Fig. 4b). The number of bacteria recovered from the fecal samples of ΔaceE∷kan-vaccinated hens was 3 logs lower than the parental strain counterpart. The mutant strain was not detected on the fourth week of postvaccination, while the parental strain was still recovered from its respective infected hens. These were confirmed, with no ΔaceE∷kan mutant detected from either the spleen or the liver of vaccinated hens at week 7 postvaccination, while the parental strain was still detected from the spleen or the liver of infected hens (data not shown). Our results suggest that in adult chickens (hens), the ΔaceE∷kan mutant had reduced bacterial colonization ability than the parental strain, and the contamination of eggshells did not persist for >3 weeks after vaccination.

Figure 4

Eggshell contamination of ΔaceE∷kan-vaccinated hens. (a) Eggshell swabs of laying hens vaccinated with 109 CFU of ΔaceE∷kan or infected with 107 CFU of the parental strain (indicated as wild type [WT] in the graph). (b) Fecal samples of laying hens vaccinated with 109 CFU of ΔaceE∷kan or infected with 107 CFU of the parental strain. Following vaccination or infection, eggshell swab and fecal sample collection were performed at weekly intervals. Columns, mean of 12 chicken samples. Bars, SD.

Discussion

The ΔaceE∷kan mutant was identified to have an LD50 at least 100 times that of the parental strain in young chicks (Chang et al., 2008). It has a deletion in the gene encoding for the E1 component of the multienzyme pyruvate dehydrogenase complex that converts pyruvate into acetyl-CoA (Chang et al., 2008; Thomson et al., 2008), acting at the gateway of the glycolysis metabolic pathway to the TCA cycle. Hence, it is reasonable to speculate that a deletion in the aceE gene will lead to defective metabolism. The effect of the aceE gene on the growth rate became evident when theΔaceE∷kan mutant was 10 times less than that of the parental strain after having grown them for 24 h with proper aeration and nutrients (Fig. 1). During the natural infection process, the ΔaceE∷kan mutant was expected to grow even slower due to the limited availability of nutrients and the harsh environment of the host.

To further investigate the characteristics of the ΔaceE∷kan mutant, a series of in vitro assays were performed to compare the phenotypes of the ΔaceE∷kan mutant with the S. Enteritidis parental strain. It was found that there were no significant differences between the ΔaceE∷kan mutant and the parental strain in their ability to grow after a prolonged stationary phase or to resist low-pH conditions, RNIs, sodium deoxycholate and chicken serum (data not shown). However, the ΔaceE∷kan mutant was less able to invade epithelial cells (HeLa cells), was less resistant to ROI and was unable to resist killing by chicken macrophages (Table 1). From these, we propose that, although not directly associated with virulence, mutation of the gene affects the growth rate, epithelial cells' invasion ability and resistance to macrophages, resulting in decreased pathogenicity of the mutant due to the defect of function in glycolysis. These were somehow in accordance with other reports in the studies of Salmonella Typhimurium (Yimga et al., 2006; Bowden et al., 2009; Mercado-Lubo et al., 2009), Listeria monocytogenes (O'Riordan et al., 2003) and Streptococcus mutans (Korithoski et al., 2008), as glycolysis and other central metabolic pathways play an important role in successful bacterial infection and intracellular survival. Moreover, the disruption of aceAB genes encoding components of PDH has been reported to inactivate the type III secretion system of Pseudomonas aeruginosa (Dacheux et al., 2002).

While studies by Dacheux (2002) suggested that an intact PDH aceAB operon was required for the activation of a P. aeruginosa type III secretion system, in S. Enteritidis, in which aceE is located at the same operon with aceF, complementation of pBaceE into ΔaceE∷kan mutant restores the LD50 as well as in vitro phenotypes of the complemented strain in a level similar to the parental strain. These ensured that any effects of the mutation were solely due to the loss of the aceE gene, with a polar effect on downstream genes not detected.

In order to evaluate the potential of the ΔaceE∷kan mutant as a live vaccine, vaccination trials were conducted to study the ability of ΔaceE∷kan vaccinated chickens to resist and reduce a high dose of the S. Enteritidis parental strain from the liver and spleen, to induce IgG immune protection, as well as to determine eggshell contamination of ΔaceE∷kan-vaccinated hens. Similar to the initial in vivo screenings of the S. Enteritidis mutant library (Chang et al., 2008), a subcutaneous injection was administered in all of the chicken trials mentioned in this study. This method bypassed the initial phase of entry of bacteria into the host; therefore, we did not expect to find a mutant that was attenuated for invasion, but to determine the ability of the mutant to survive and replicate inside the host, which are important characteristics for a systemic infection caused by S. Enteritidis (Chang et al., 2008). In this study, the young chicks were vaccinated with 2 × 109 CFU of the ΔaceE∷kan mutant or infected with 2 × 107 CFU (LD50) of the parental strain. Previously, in Chang (2008), mutants were identified as attenuated for virulence if they had an LD50, which was at least 100 times that of the parental strain, with the ΔaceE∷kan mutant identified as one of the attenuated strains. The young chicks were not vaccinated with the LD50 of the ΔaceE∷kan mutant, as it is not reasonable to vaccinate the young chicks with this dose at which 50% of the vaccinated chicks may die afterwards. However, in the case of infection of young chicks with an LD50 of the parental strain, this could be the scenario that occurs in the poultry industry, causing the death of young chicks, and subsequently, contamination of poultry products from the remaining S. Enteritidis carriage chickens. In addition, the results on vaccination of young chicks with a ΔaceE∷kan mutant at a dose similar to (2 × 107 CFU), 10 times higher (2 × 108 CFU) and 10 times lower (2 × 106 CFU) than the LD50 of the parental strain (Fig. 2a and b) showed that when young chicks were vaccinated with these doses, the mutant persisted in the liver and spleen of chickens in a shorter time than the group infected with 2 × 107 CFU of the wild-type strain; however, a low level of serum IgG from the vaccinated chicks was induced (Fig. S2). These trials were terminated after following the chickens for 5 weeks when the mutants were not detected (below 100 CFU g−1) from both the liver and the spleen, as it was expected that these vaccinations (with a dose lower than 2 × 109 CFU) would not confer any protection if the chickens were subsequently challenged with a high dose of the parental strain.

The young chicks vaccinated with the ΔaceE∷kan mutant at a dose 2 log higher than the parental strain were able to eliminate the mutant faster than the parental strain, presumably as a result of clearance by macrophages from the liver and spleen, where Salmonella is thought to reside during systemic infection (Fig. 2) (He et al., 2006). In a subsequent challenge with a high dose of the parental strain to the ΔaceE∷kan-vaccinated chickens, a significant reduction (reduced by >8 times) of chicken mortality was obtained in the ΔaceE∷kan-vaccinated chicken compared with the control group. The results of the serum response suggest that ΔaceE∷kan is able to induce IgG B-cell immunity (Fig. 3); however, only when young chicks were vaccinated with 2 × 109 CFU of the ΔaceE∷kan mutant was a relatively high level of IgG obtained. While Beal (2006) report that the clearance of Salmonella infection in chickens is B-cell independent, some studies suggest that vaccine-induced protection against infection involves both cell-mediated and humoral responses, the latter in the later stages of infection (Mastroeni et al., 2000; McSorley & Jenkins, 2000; Mittrucker et al., 2000; Mastroeni & Menager, 2003). Altogether, by taking into consideration the colonization and clearance efficacy of the ΔaceE∷kan mutant from young chicks, and subsequent protection against the wild-type strain, at this stage, it is suggested that vaccination of young chicks with at least 2 × 109 CFU of the ΔaceE∷kan mutant would be necessary in order to confer a significant level of protection. Whether this protection is cell- or humoral mediated, more investigations would have to be performed in the future. Even though the clearance efficacy of the mutant from chickens might not provide a drastic improved feature compared with the established attenuated live vaccine strains of S. Enteritidis, vaccination of newly hatched chicks with the ΔaceE∷kan mutant might provide competitive exclusion of wild-type S. Enteritidis infection in nature, while will enable the development and maturation of the immune system in young chicks against S. Enteritidis, concurrent with the clearance of the ΔaceE∷kan mutant.

As suggested by epidemiological analyses, the major risk factors for S. Enteritidis infection in human are eggs and egg products (De Buck et al., 2004). Vaccination of laying hens with established vaccine strains confers a good level of protection on the chickens; however, it takes time for the vaccine strains to reduce egg contamination by S. Enteritidis. In fact, sporadic isolation of S. Enteritidis from eggs indicates that vaccination does not fully eliminate S. Enteritidis contamination (Davison et al., 1999; Parker et al., 2001; De Buck et al., 2004). Salmonella enterica serovar Enteritidis has been commonly recovered from the eggshells of infected hens and this represents a potential threat to public health, just as contamination of the contents of the egg (Poppe et al., 1992; de Louvois, 1993; Humphrey, 1994; Schutze et al., 1996; De Buck et al., 2004); moreover, eggshell contamination has been reported as one of the most infected sites of contaminated eggs (Bichler et al., 1996; Miyamoto et al., 1997; Okamura et al., 2001; De Buck et al., 2004). Therefore, in this study, surface contamination (eggshell) of the eggs from the vaccinated hens was investigated. Checking on the contamination of eggshell itself also served as a good indicator of bacterial infection on the lower reproductive tract of hens (De Buck et al., 2004). Even though eggshell contamination by feces in cloaca could also occur, this has been reported to be less likely, as when a hen lays an egg, its bearing everts the vagina beyond the alimentary tract, which protects the emerging egg from fecal contamination and the stretching of the cloacal lining effectively makes the intestinal tract somewhat slit-like, further reducing the chance for contamination of the eggshell (De Buck et al., 2004). The results of this work suggested that the ΔaceE∷kan mutant did not persist on the eggshell for a long time, with the level and duration of contamination significantly lower and shorter than those in the hens infected by the parental strain, and were not inferior to the established vaccine strains in terms of vertical transmission to eggs (Fig. 4) (Davison et al., 1999; Parker et al., 2001; De Buck et al., 2004).

While several studies have reported on the role of central metabolic pathways in S. Typhimurium virulence and intracellular survival (Yimga et al., 2006; Bowden et al., 2009; Mercado-Lubo et al., 2009), little is known about their roles in S. Enteritidis infection in particular. Here, we report that deletion of the aceE gene (encoding for the E1 component of PDH) resulted in a reduced growth rate and reduced capacity to colonize in chickens, while retaining some of the virulence characteristics of the wild-type strain. Our works cover both in vitro studies to determine the virulence characteristics of the ΔaceE∷kan mutant as well as in vivo studies to investigate the persistence of the mutant and its immunogenicity in young chicks. Further studies on the role of aceE in relation to the bacterial pathogenicity as well as detailed investigations on the immune protection offered by the ΔaceE∷kan mutant in chickens will be performed in the future. So far, these works have generated considerable useful information about the possible functions of aceE gene in chicken pathogenesis, and have suggested that there are as yet uncharacterized features of S. Enteritidis that could determine their lethality in young chicks, which could provide valuable insights for the design of new vaccine strains against Salmonella.

Supporting Information

Additional Supporting Information may be found in the online version of this article:

Fig. S1. Diagrams of chicken trials.

Fig. S2. Serum IgG response of young chicks vaccinated with different doses of ΔaceE::kan mutant.

Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

Acknowledgements

This work was supported by a grant from the Temasek Life Sciences Laboratory, Singapore. We would like to thank the Agri-Food and Veterinary Authority of Singapore, for the generosity in providing bacterial samples. Also, we thank Dr He Haiqi, from Southern Plain Agricultural Research Center, USDA-ARS, College Station, TX, USA, for his generosity in providing the HD11 chicken macrophage cell line.

Footnotes

  • Editor: Eric Oswald

References

View Abstract