The food-borne pathogen Campylobacter jejuni depends on the AddAB DNA repair system to defend against bile in the intestinal environment.

Accurate repair of DNA damage is crucial to ensure genome stability and cell survival of all organisms. Bile functions as a defensive barrier against intestinal colonization by pathogenic microbes. Campylobacter jejuni, a leading bacterial cause of foodborne illness, possess strategies to mitigate the toxic components of bile. We recently found that growth of C. jejuni in medium with deoxycholate, a component of bile, caused DNA damage consistent with the exposure to reactive oxygen species. We hypothesized that C. jejuni must repair DNA damage caused by reactive oxygen species to restore chromosomal integrity. Our efforts focused on determining the importance of the putative AddAB DNA repair proteins. A C. jejuni addAB mutant demonstrated enhanced sensitivity to deoxycholate and was impaired in DNA double strand break repair. Complementation of the addAB mutant restored resistance to deoxycholate, as well as function of the DNA double strand break repair system. The importance of these findings translated to the natural host, where the AddAB system was found to be required for efficient C. jejuni colonization of the chicken intestine. This research provides new insight into the molecular mechanism utilized by C. jejuni, and possibly other intestinal pathogens, to survive in the presence of bile.

Introduction

Bile plays a critical role in the intestines of both humans and animals as a defensive barrier against colonization by pathogenic microbes. Campylobacter jejuni, which targets the intestines of susceptible hosts, is one of the leading bacterial causes of foodborne illness in the United States[1]. Numerous cases of C. jejuni illness result from the handling and consumption of foods cross-contaminated with raw poultry products. Indeed, the National Antimicrobial Resistance Monitoring System (NARMS) reported the presence of C. jejuni in 33% of retail chicken tested in 2014[2]. Alarmingly, 15% of the C. jejuni recovered from the poultry, as well as 27% of C. jejuni isolates recovered from humans, were resistant to the antibiotic ciprofloxacin. Given that poultry represent a natural reservoir for C. jejuni, research to expand our knowledge of Campylobacter colonization of chickens might prove helpful to the broiler industry in developing control methods that result in a decrease in the incidence of human campylobacteriosis. Moreover, studies are needed to further dissect the molecular mechanisms used by pathogens to overcome the toxic insult of intestinal bile.

Bile is produced in the liver and then drained into the intestine where it aids in digestion, absorption of fats, and excretion of waste products[3]. The primary bile acids (cholic acid and chenodeoxycholic acid) are synthesized in the liver from cholesterol in a multi-enzyme process. The secondary bile acids (lithocholic acid and deoxycholic acid) are produced as the result of metabolism of the primary bile acids by normal host intestinal bacteria. The average concentration of bile acids in the human intestine ranges from 0.2 to 2%[4]. The gastrointestinal tract of chickens is a natural reservoir of C. jejuni where the ceca serve as primary colonization sites. While the concentration of bile varies along the intestine of the chicken, bile concentration is greatest in the duodenum and upper jejunum[5]. Noteworthy is that bile secretion increases considerably in the duodenum of chickens during the first 21 days of life[6].

Becoming increasingly evident is that C. jejuni, as a member of the Epsilonproteobacteria, does not possess certain mechanistic and regulatory pathways used by Gammaproteobacteria, including Escherichia coli and Salmonella enterica subspecies Typhimurium. In fact, predictions for the presence and function of C. jejuni DNA repair system components has largely been based on genomic comparisons looking for characterized homologues within closely related bacteria, including Helicobacter pylori [7,8]. While C. jejuni clearly possess the gene encoding RecA, which has been demonstrated to be involved in repair of DNA double strand breaks using homologous recombination in many bacterial species, it does not harbor the recBCD genes involved in DNA repair in bacteria such as E. coli. Rather, researchers have postulated that C. jejuni encodes for two recombination repair proteins termed AddA and AddB[8,9]. Previous work with Bacillus subtilis has demonstrated that the AddA and AddB proteins function in place of the RecBCD complex to conduct the nuclease, helicase, and Chi recognition activity necessary for homologous recombination[10]. While it has been speculated that putative AddA and AddB homologues may function similarly in homologous recombination repair in C. jejuni, no experimental evidence of function or biological relevance has been described for these proteins in C. jejuni. Moreover, Campylobacter spp. lack the LexA repressor necessary to elicit a global regulatory response to DNA damage through the SOS response[11,12]. Therefore, beyond other Epsilonproteobacteria, cross species comparisons and predictions for response to DNA damage are difficult to make for C. jejuni. This lack of knowledge only reinforces the importance of more completely describing the mechanisms, regulation, and biological relevance of DNA repair in Campylobacter spp.

In a previous report, we found that the continuous growth of C. jejuni in medium with a physiological concentration of the bile salt deoxycholate induces the production of reactive oxygen species (ROS), which in turn, causes DNA damage (double strand breaks)[13]. Based on these findings, we hypothesized that C. jejuni must repair the ROS-induced DNA damage to promote survival. Given the paucity of detailed information regarding C. jejuni DNA repair systems, we decided to investigate the necessity of DNA repair in C. jejuni. Specifically, the purpose of this study was to determine if two putative DNA repair genes, designated CJH_07760 and CJH_07765 in C. jejuni strain F38011, function in DNA double strand break repair and play a role in the natural host (chickens). While RecA has been identified in C. jejuni, its partners in DNA homologous recombination have not been identified, and we suspected that the proteins encoded by the CJH_07760 and CJH_07765 genes are homologs of the AddA and AddB proteins demonstrated to work with RecA in other bacteria. In sum, this work revealed that the Campylobacter AddAB proteins contribute to DNA double strand break repair and are necessary for efficient colonization of chickens (the natural host).

Results

Description of C. jejuni AddA and AddB encoding genes

Given the absence of genes encoding RecBCD in C. jejuni, we focused our efforts on the addA and addB genes that are recBCD analogs[8]. A detailed comparison of the putative C. jejuni addA and addB genes and the products they encode are presented in Supplementary Table 1. The putative addA gene in the C. jejuni F38011, 81–176, and NCTC 11168 clinical strains is 2766 nucleotides in length and is predicted to encode a protein of 921 aa, while the putative addB gene is 2367 nucleotides in length and is predicted to encode a protein of 788 aa. The deduced aa sequence of the putative AddA protein from the C. jejuni F38011 strain exhibits approximately 48% similarity and 30% identity to the AddA protein from H. pylori strain 26695 and the deduced amino acid sequence of the putative AddB protein from the C. jejuni F38011 strain exhibits approximately 46% similarity and 29% identity to the AddB protein from H. pylori strain 26695. Based on these analyses, the putative addA and addB genes are highly conserved amongst C. jejuni isolates.

The putative addA and addB genes are constitutively expressed in response to deoxycholate

Based on the analysis of RNA-Seq data and inspection of the distribution and number of reads corresponding to the putative addA and addB genes from C. jejuni strain F38011[13], the addAB genes are likely to be contained in an operon where addB is upstream of addA (Supplementary Figure 2). Given that continuous growth of C. jejuni in deoxycholate results in DNA double strand breaks and the hypothesis that C. jejuni must repair the ROS-induced DNA damage to promote survival, quantitative reverse transcription PCR (RT-qPCR) was initially performed to determine if these two putative DNA repair genes are upregulated in response to growth in deoxycholate-supplemented medium over time. No change was observed in the expression levels of the putative addA and addB genes in C. jejuni strain F38011, whether in MH with 0.05% deoxycholate or MH alone (not shown). The levels of expression of the two genes were approximately 15 to 20-fold less than that for the oxidative stress response control (rpoA)[14]. Based on the result of the RT-qPCR analysis, it was concluded that the putative addA and addB genes are constitutively expressed in response to C. jejuni growth in MH broth and in MH broth supplemented with a moderate level (0.05%) of deoxycholate.

The putative addA and addB genes from C. jejuni functionally complement an E. coli ΔrecBCD strain

The RecBCD enzyme of E. coli is essential for DNA double strand break repair and genetic recombination of linear DNA. The AddAB proteins are related to RecBCD, whereby the AddA protein essentially corresponds to RecB and the AddB protein to RecC[15]. To determine if the recombination DNA repair system from C. jejuni could functionally complement an E. coli strain with a recBCD deletion, a ciprofloxacin sensitivity assay was performed with a number of E. coli transformants. The E. coli ΔrecBCD mutant (strain V3060) was transformed with either an empty vector, a vector harboring only the C. jejuni putative recA gene, or a vector containing the C. jejuni putative addA, addB, and recA genes. All genes were driven by the T7 promoter. A ciprofloxacin sensitivity assay was then performed. Ciprofloxacin is a quinolone antibiotic that inhibits the function of DNA gyrase, eventually leading to an increase in DNA double strand breaks. Based on the literature and the fact that RecA and RecBCD are known to be involved in DNA double strand break repair, functional complementation of an E. coli RecBCD mutant was expected to demonstrate greater resistance to ciprofloxacin than the mutant transformed with an empty vector. Consistent with functional complementation, introduction of the putative recA gene into the E. coli ΔrecBCD mutant provided the isolate with increased ciprofloxacin resistance when compared to the E. coli mutant harboring the empty vector, but the E. coli ΔrecBCD isolate transformed with the C. jejuni putative addA, addB, and recA genes demonstrated even greater resistance to ciprofloxacin (Fig. 1). These findings demonstrate that the C. jejuni putative addA and addB genes functionally complemented the RecBCD enzyme of E. coli. Based on these findings, we designated CJH_07760 and its homologues in other C. jejuni strains as the gene encoding C. jejuni AddA and CJH_07765 and its homologues in other C. jejuni strains as the gene encoding AddB.

Figure 1

Addition of C. jejuni recA and putative addAB genes functionally complement an E. coli recBCD mutant. Given the high level of identity in the addA and addB gene sequences amongst the C. jejuni strains, we arbitrarily selected the putative addA and addB genes from C. jejuni strain 81–176 for these experiments. E. coli strain V3060 was transformed with an empty vector, with recA from C. jejuni strain 81–176, or with addAB and recA from C. jejuni strain 81–176. The cultures were grown in triplicate in LB broth for 18.5 hours with or without 0.0025 µg/mL of ciprofloxacin. Shown is the mean ratio (with ciprofloxacin/without ciprofloxacin) and standard deviation of the OD600 between the two culture treatments. A significant difference was observed in ciprofloxacin resistance in the E. coli strain harboring the C. jejuni recA gene alone and the E. coli transformed with the C. jejuni recA and addAB genes (*p < 0.05) compared to E. coli with the empty vector. An ANOVA followed by Tukey’s multiple comparisons test was done to evaluate significance.

The AddAB system is functional in C. jejuni

To determine if the AddAB system is functional in C. jejuni, gene deletions were generated in addA, addB, and both addA and addB in strain F38011. As mentioned previously, a mutation in addB is likely polar on addA since the two genes are contained in an operon, where addB is upstream of addA. The prRNA-cysM prom-addAB-hygromycin vector was also generated to test for AddAB functional complementation. This vector was designed to allow for wild-type copies of the addAB genes to be inserted into the chromosome whereby they are constitutively expressed from the cysM promoter. The prRNA-cysM prom-addAB-hygromycin construct was confirmed by restriction enzyme digests and sequencing. Similar to the assays performed earlier with the E. coli transformants, the C. jejuni addA, addB, and addAB deletion mutants were expected to demonstrate greater sensitivity to ciprofloxacin than the wild-type strain and complemented isolate, as treatment with this antibiotic results in DNA double strand breaks by inhibiting DNA gyrase. As predicted, the minimal inhibitory concentration (MIC) of ciprofloxacin for the C. jejuni F38011 wild-type strain was 0.125 μg/mL, whereas the MIC for the addA mutant, addB mutant, and addAB mutant was 0.03125 μg/mL. The MIC for the C. jejuni addAB mutant harboring the chromosomal copies of the addAB genes was identical to the wild-type strain (0.125 μg/mL). Collectively, these assays show that the AddAB proteins function to promote C. jejuni survival under conditions that generate DNA double strand breaks.

Deoxycholate results in DNA double strand breaks in C. jejuni

Although previous work in the laboratory had revealed that C. jejuni growth in medium supplemented with deoxycholate resulted in DNA double strand breaks, it was not clear the minimum dose required to generate the double strand breaks and whether the effect was dose-dependent. Briefly, C. jejuni were grown in MH broth alone as well as MH broth supplemented with different concentrations of deoxycholate, after which, pulsed-field gel electrophoresis (PFGE) was performed to detect DNA double strand breaks. The basis of this assay is that the fragmented DNA migrates in the agarose gel during PFGE, whereas the intact DNA does not migrate in the gel[16]. Fragmented DNA was readily visualized in samples prepared from C. jejuni grown in medium containing ≥0.1% deoxycholate (Fig. 2). In summary, C. jejuni growth in medium with 0.1% deoxycholate results in DNA double strand breaks that are readily detectable by PFGE.

Figure 2

There is a threshold concentration of deoxycholate (DOC) needed to induce DNA damage. C. jejuni strain F38011 was grown for 20 hours in MH broth with 0.025% to 0.3% (w/v) DOC. Samples were collected and subjected to pulsed-field gel electrophoresis to evaluate DNA integrity. Limited to no damage was observed at a concentration less than or equal to 0.05% DOC, while damage was apparent in 0.1% DOC. One asterisk (*) indicates intact DNA, and two asterisks (**) indicate DNA fragments. Contrast is enhanced for clarity, see Supplementary Figure 3 for original image.

The AddAB system contributes to C. jejuni resistance to deoxycholate

Relevant to the colonization of a host, C. jejuni have the ability to resist bile. Given that the MIC of deoxycholate for C. jejuni was above the critical micelle concentration, the inhibitory concentration for the C. jejuni wild-type strain and addAB mutant to deoxycholate could not be determined. Regardless, the C. jejuni wild-type strain and addAB mutant were tested for sensitivity to 0.1% deoxycholate using a plate assay. Similar to the ciprofloxacin sensitivity assays, the C. jejuni addAB mutant demonstrated much greater sensitivity to deoxycholate versus the wild-type strain (Fig. 3). To confirm that the AddAB proteins are responsible for increased resistance to deoxycholate, the C. jejuni addAB mutant harboring the rRNA-cysM prom-addAB-hygromycin insertion was tested for functional complementation as judged by its sensitivity to deoxycholate. Consistent with the ciprofloxacin sensitivity assays, the C. jejuni addAB gene deletion mutant harboring a chromosomal copy of the wild-type addAB genes restored deoxycholate resistance (Fig. 3). Noteworthy is that the colonies formed by the C. jejuni addAB gene deletion mutant appeared smaller in size than the colonies formed by the C. jejuni wild-type strain and addAB mutant harboring addAB in cis (complemented isolate). However, additional assays revealed that the growth rates of the C. jejuni wild-type strain, addAB mutant, and addAB complemented isolate were indistinguishable from one another when grown in MH broth. Based on these data, we concluded that the addAB gene deletion confers deoxycholate sensitivity. Collectively, these data suggest that exposure to deoxycholate induces stress (e.g., DNA damage), and that the C. jejuni DNA homologous recombination repair system must respond to promote bacterial survival.

Figure 3

The C. jejuni addAB mutant demonstrates enhanced sensitivity to deoxycholate. Cultures of the C. jejuni F38011 wild-type strain, addAB mutant, and addAB mutant harboring addAB in cis (complemented isolate) were adjusted to an OD540 of 0.1. Ten-fold serial dilutions were spotted on to either MH agar or MH agar with 0.1% deoxycholate (DOC), and growth was evaluated after 48 hours of incubation.

The C. jejuni AddAB proteins are involved in DNA repair

To demonstrate that the AddAB system plays a role in DNA double strand break repair, C. jejuni were treated with ciprofloxacin for 15 minutes to introduce DNA double strand breaks, washed, and then incubated in MH broth to allow DNA repair. DNA repair in the C. jejuni wild-type strain versus addAB deletion mutant was then assessed by PFGE (Fig. 4). Evident from the gels was that the C. jejuni addAB deletion mutant was impaired in DNA double strand break repair. This finding supports the notion that the AddAB proteins could protect C. jejuni from the damaging effects caused by growth in deoxycholate, presumably by repairing DNA double strand breaks.

Figure 4

The C. jejuni addAB mutant is deficient in DNA double strand break repair. DNA double strand breaks were introduced by a 15-minute treatment with ciprofloxacin, after which the bacteria were transferred to MH broth. Aliquots were collected every 45 minutes to evaluate DNA double strand break repair by pulsed-field gel electrophoresis. A bacterial sample was collected prior to ciprofloxacin treatment as a negative control (untreated sample). Intact DNA was quantitated using relative band intensity, with the untreated sample set to 100%. Three biological replicates ± standard deviation are plotted for each time point. Significance between the C. jejuni wild-type strain (•) and the addAB mutant (▪) was calculated for each time point by one-way ANOVA followed by Sidak’s multiple comparisons test, and indicated by an asterisk (*p < 0.05).

The C. jejuni AddAB system is required for the efficient colonization of chickens

Campylobacter species frequently colonize the intestines of many wild and domestic birds. Based on the observation that the AddAB system confers deoxycholate resistance in vitro, experiments were performed to determine if the AddA and AddB proteins are necessary for C. jejuni survival in the chicken gut. Chickens were divided into 6 groups and inoculated with motile isolates of C. jejuni as outlined in ‘Materials and Methods.’ All chickens were euthanized at 14 days post-inoculation and the number of C. jejuni per gram of duodenal and cecal material was determined. Significant differences were observed in the level of colonization for a C. jejuni wild-type strain and the addAB mutant used in this study (Fig. 5). The C. jejuni wild-type strain was recovered from the duodenum of 8 of 10 birds, whereas the addAB mutant was recovered from the duodenum of only 2 of 12 birds (Fig. 5a). Similarly, the C. jejuni wild-type strain colonized the ceca of birds at a significantly higher level than the addAB mutant (Fig. 5c). Noteworthy is that the C. jejuni addAB mutant harboring the rRNA-cysM prom-addAB-hygromycin insertion was recovered from the duodenum of 9 of 12 birds and colonized the ceca of birds at a significantly higher level than the addAB mutant (Fig. 5a and c). Competition experiments showed that the C. jejuni wild-type strain and the C. jejuni addAB mutant harboring the rRNA-cysM prom-addAB-hygromycin insertion were equally recovered from both the duodenums and ceca of co-infected birds (Fig. 5b and d). Strikingly, the C. jejuni wild-type strain strongly outcompeted the addAB mutant in both duodenums and ceca of co-infected birds (Fig. 5b and d). C. jejuni was not recovered from any of the mock inoculated chickens. Based on these data, we concluded that the AddAB proteins function to promote C. jejuni colonization/survival in chickens.

Figure 5

The AddAB DNA repair system of C. jejuni is essential for efficient colonization of chickens. The experiment was performed as outlined in ‘Materials and Methods.’ Panels (a) and (b): C. jejuni recovered from the duodenum, inoculated either alone (a) or in competition (b) with equal proportion of the C. jejuni wild-type strain and addAB mutant (∆addAB), or equal proportion of the C. jejuni wild-type strain and complemented addAB mutant (∆addAB + addAB). Panels (c) and (d): C. jejuni recovered from the cecum either inoculated alone (c) or in competition (d). ND indicates that no C. jejuni were detected (limit of detection > 102 CFU/g). The horizontal bar with whiskers represents the median level of colonization and the interquartile range. The asterisk (*) indicates a significant difference (p < 0.05) in the CFU recovered versus the C. jejuni addAB mutant, as judged by a nonparametric Kruskal-Wallis test followed by Dunn’s test for multiple comparisons. C. jejuni was not recovered from any of the uninoculated chickens.

C. jejuni infection of chickens does not alter the bile concentration in the intestine

In parallel with the in vivo experiments to determine the importance of the AddAB DNA repair system in C. jejuni colonization of chickens, assays were performed to determine if C. jejuni infection alters bile concentration in the intestine. Known is that the concentration of bile varies along the intestine of the chicken, and that bile salt concentration is greatest in the duodenum and upper jejunum[5]. The release of bile into the duodenum is stimulated by the sensing of acidic pH, fatty acids, and bile salts by cells of the duodenum that subsequently release stimulating hormones, such as cholecystokinin, that triggers gallbladder contraction[3]. Therefore, bile concentration in the duodenum of an individual bird fluctuates with the presence or absence of digesta.

The concentration of bile was determined in the C. jejuni-inoculated chickens and mock inoculated chickens from samples collected when the colonization experiment was terminated (samples were collected at the time of necropsy). While the amount of total bile was observed to vary amongst the birds, a greater level of bile was observed in the duodenums than in the ceca (Fig. 6a and b). Analysis of the 20 samples (collected from the mock inoculated and C. jejuni-inoculated chickens) revealed that the bile concentration ranged from 70.6 µM to 1805.5 µM in the duodenums. Noteworthy is that ~1206 µM of deoxycholate (0.05%) was used in the in vitro assays in this study. C. jejuni colonization of the chickens had no effect on total bile concentration in either the duodenums or ceca (Fig. 6a and b). In addition, no correlation was observed between C. jejuni CFU and bile concentration in either the duodenums or ceca (Fig. 6c and d). Collectively, these findings suggest that the C. jejuni colonization of chickens does not alter intestinal bile concentration and, vice versa, the concentration of bile does not alter C. jejuni colonization of the duodenum or ceca (that is, the total number of recoverable CFU).

Figure 6

Infection of chickens does not alter the bile concentration in the intestine. Total bile concentration was determined in the duodenum (Panel a) and cecum (Panel b) of mock inoculated chickens (control) and chickens inoculated with C. jejuni wild-type strain using a bile acid assay kit (Sigma-Aldrich). The total CFU recovered from C. jejuni wild-type inoculated chickens was compared to the concentration of bile in the duodenum (Panel c) and the cecum (Panel d). The concentration of bile was independent from the number of live C. jejuni recovered. The horizontal bar with whiskers represents the median bile concentration and the interquartile range.

Discussion

Chickens are a natural host for Campylobacter species, and C. jejuni, in particular, colonizes the avian gastrointestinal tract of chicks by two to four weeks of age[17]. Based on the efficiency of C. jejuni chicken colonization (i.e., as few as 35 CFU is sufficient for the colonization of chicks), it is evident that certain strains possess strategies that promote efficient colonization and survival within this host. Over the past two decades, researchers have expended a significant amount of effort to identify and characterize proteins that facilitate C. jejuni colonization and survival in the chicken gut. While a variety of C. jejuni proteins have been identified that contribute to these processes, including proteins that are involved in motility, cell adherence, modification of bacterial cell surface structures, etc., less attention has been paid to the proteins involved in enhancing cellular survival. As such, the discovery of a previously unrecognized survival response mechanism to bile salts (e.g., maintenance and repair of the cellular DNA) could represent a strategy utilized by other pathogens to maintain homeostasis in the gut. Efficient and accurate repair of DNA damage is crucial to ensure genome stability and cell survival.

This work was based on the original finding that deoxycholate triggers production of reactive oxygen species (ROS) and causes DNA damage[18,19], including the DNA double strand breaks shown in this study. Based on the absence of any well-characterized DNA repair system in C. jejuni, and the fact that the repair of DNA double strand breaks predominantly involves homologous recombination, our efforts were focused on the characterization of the putative AddAB system. We first demonstrated that the C. jejuni addA and addB genes encode proteins that are functionally analogous to the RecBCD proteins of E. coli using well-accepted laboratory assays. More specifically, an E. coli recBCD mutant harboring C. jejuni addAB and recA was found to demonstrate greater resistance to ciprofloxacin than the E. coli mutant harboring the empty vector. The utilization of C. jejuni AddAB proteins in E. coli strongly supports the assertion that these genes function as part of the homologous DNA repair system in C. jejuni. Next, the C. jejuni addAB mutant was found to be more sensitive to both ciprofloxacin and deoxycholate treatment than the wild-type strain. Importantly, the C. jejuni addAB mutant transformed with the addAB genes was successfully complemented in vitro, as judged by the ciprofloxacin and deoxycholate sensitivity assays. The importance of the AddAB proteins were then demonstrated in vivo, as the C. jejuni addAB mutant was impaired in gut colonization when compared to the wild-type strain. Moreover, the C. jejuni addAB mutant harboring the rRNA-cysM prom-addAB-hygromycin insertion colonized the duodenum and cecum at a level that was nearly indistinguishable from that of the wild-type strain. Collectively, the data demonstrate that the AddAB system is critical for C. jejuni to colonize chickens.

The magnitude of the effect of addAB gene deletion on C. jejuni colonization is a vital observation for several reasons. First, organisms have evolved overlapping mechanisms for homologous recombination repair, and this redundancy in repair mechanisms highlights the potentially devastating impact on fitness posed by unrepaired lesions. More specifically, bacteria frequently possess two lesion dependent pathways to enable RecA loading onto single stranded DNA for homologous recombination repair. The Rec(F)OR pathway is utilized to repair DNA gaps (single strand breaks) and the AddAB (or RecBCD) pathway is utilized in the presence of DNA double strand breaks[20,21]. Based on sequence similarity to H. pylori genes, C. jejuni is believed to encode RecO and RecR homologues[7,22], so it is possible that recombination repair involving these proteins may occur in C. jejuni. However, it must also be noted that DNA single strand breaks can progress to double strand breaks if not properly repaired[20]. Accordingly, our work clearly demonstrates the critical importance of the C. jejuni AddA and AddB double strand break repair proteins. Second, due to its critical role in the natural host, one might expect the addAB genes to be upregulated in response to the stressor that is instigating the DNA double strand breaks. This is the case in Coxiella burnetii, an obligate intracellular pathogen, that upregulates addAB homologues in response to macrophage induced oxidative stress[23]. However, work on Bacillus subtilis has revealed that addAB gene expression is independent of DNA-damaging agents[24]. Although the addAB genes were found not to be upregulated in C. jejuni in response to growth in deoxycholate from 10 to 16 hours, it is possible that these genes are upregulated at other times or in response to other environmental conditions. In addition to DNA breaks, deoxycholate-mediated oxidative damage likely causes other types of DNA damage. In agreement with this proposal, the uvrC gene is upregulated in C. jejuni grown in medium containing deoxycholate[13] and ovine bile[25]. UvrC is an excision nuclease protein and part of the UvrABC complex involved with nucleotide excision repair[26]. Recent work indicates that this pathway, in addition to the base excision repair pathway, defends against oxidative damage[27,28]. Also worth noting is that C. jejuni strains demonstrate different levels of sensitivity to deoxycholate, and based on this observation, we hypothesize that an isolate with greater resistance to deoxycholate would have a selective advantage for establishing colonization and potentially causing disease. Although no correlation was observed between C. jejuni CFU and bile concentration in the duodenums of birds, curious is that the ceca, where the concentration of bile is low, are the preferential gastrointestinal colonization sites for this organism (see Fig. 5). Finally, given the critical role of the AddAB system in C. jejuni in poultry, we hypothesize that the AddAB proteins play a role in early human infection, where the pathogen survives and multiplies in the presence of bile in the upper region of the small intestine[29].

Researchers have spent significant time in defining C. jejuni colonization factors, because poultry products (or the chicken reservoir as a whole) are estimated to be responsible for up to 80% of human campylobacteriosis cases[17]. However, researchers have also observed that C. jejuni isolates have different colonization potentials. In brief, some isolates appear unable to colonize 14-day old chickens, whereas others colonize chickens but are eliminated from the gut, and yet others show efficient and sustained colonization of chickens. While the ability of C. jejuni to colonize birds is without a doubt multifactorial, many of the factors that contribute to colonization have likely not been identified. We initiated this study with the C. jejuni F38011 and 81–176 clinical strains. However, in our preliminary work, we observed that C. jejuni strain 81–176 poorly colonized chickens (n = 3 experiments, with a mean colonization rate = 3.06 × 103 CFU per gram of cecal content at 10 days post-inoculation). At nearly the same time, we also noted that the C. jejuni 81–176 strain is more sensitive to deoxycholate than strain F38011. Noteworthy is that C. jejuni strain 81–176 was responsible for a point-source outbreak of campylobacteriosis associated with consumption of raw milk[30], whereas the reservoir of C. jejuni strain F38011 is not known. The enhanced sensitivity of C. jejuni strain 81–176 to deoxycholate raises the possibility that the ability of an isolate to withstand and adapt to bile components may contribute to the variability in colonization capacity (or to its ability to establish a particular niche).

Pathogenic microbes that colonize the gut must cope with a wide-range of host defense mechanisms. Bile provides the host gut a formidable barrier to colonizing pathogens. Indeed, bile interacts with the cellular membranes of both mammalian and bacterial cells, causing membrane perturbation and oxidative stress. Previously, we found that the growth of C. jejuni in medium containing the bile salt deoxycholate resulted in elevated levels of intracellular ROS, which in turn, caused significant DNA damage[13]. To overcome this insult, C. jejuni adapt to the intestinal milieu (exposure to bile) through use of the RecA/AddAB DNA repair system. The importance of the DNA damage repair pathways in vivo is likely the basis for why mutations (gene deletions) in these systems have been utilized in a number of enteric bacteria for the generation of live, attenuated vaccine strains[31,32]. Regardless, based on our data, we conclude that functional DNA repair systems are necessary for organisms to maintain their chromosomal integrity. In other words, without repair of the chromosomal lesions introduced by the chemical insults found in the intestinal environment, C. jejuni would fail to replicate and transmission to a new host would cease.

Materials and Methods

Bacterial strains, vectors, and oligonucleotides

The bacterial strains, vectors, and oligonucleotides used is this study are listed in Table 1. C. jejuni were routinely cultured on Mueller-Hinton (MH) agar (Difco Brand, BD Biosciences, Sparks, MD) containing 5% citrated bovine blood (MHB agar) or in MH broth with orbital shaking under microaerobic conditions (5% O2, 10% CO2, 85% N2) at 37 °C. E. coli strains [StellarTM (Takara Bio, USA) and V3060] were routinely cultured on LB-Miller agar (Thermo Fisher Scientific, Waltham, MA) or in LB-Miller broth under aerobic conditions at 37 °C.

Table 1

Bacterial strains, vectors, and oligonucleotides used in this study.

Strain
C. jejuni 81–176	Outbreak associated with raw milk[30]
C. jejuni F38011	Human case of diarrhea[37]
E. coli Stellar™	Takara (Clontech)
E. coli V3060	8
Vector
pBSK-Kan2	38
pBSK-addA KO	This work
pBSK-addB KO	This work
pBSK-addAB KO	This work
prRNA-Hygro-cysM Pro-addAB	This work39
pacycDUETTM−1	Novagen
pacycDUET-recA	This work
pacycDUET-addAB & recA	This work
Oligo name	Sequence
Mutant Generation Primers
addA DWN T	5′-TATAGGGCGAATTGGGTACCCCTTCTTTAGCAATTTGCGCAG-3′
addA DWN B	5′-GATCGGATCCGGAAAATGAAATTCAAATTTTGGAAATATAAAC-3′
addA CAT T	5′-TTCATTTTCCGGATCCGATCTGCGCCCTTTAGT-3′
addA CAT B	5′-ATTAAAATGACTGCAGGTGTTCCTTTCCAAGTTAATTGCG-3′
addA UP T	5′-ACACCTGCAGTCATTTTAATTCCTTTTTATAAATGAGTTTATAAGG-3′
addA UP B	5′-GGGAACAAAAGCTGGAGCTCAAATTTCACAACTTTCCCTTAGTG-3′
addB DWN T	5′-TATAGGGCGAATTGGGTACCCTCTCTTACATCAAGCCTC-3′
addB DWN B	5′-GATCGGATCCGATGAAAATGCAAGCGTGG-3′
addB CAT T	5′-CATTTTCATCGGATCCGATCTGCGCCCTTTAGT-3′
addB CAT B	5′-TGCTCAATGACTGCAGGTGTTCCTTTCCAAGTTAATTG-3′
addB UP T	5′-ACACCTGCAGTCATTGAGCATTTAATTCGTAGC-3′
addB UP B	5′-GGGAACAAAAGCTGGAGCTCTGGCCTTATGGTATTTTGC-3′
Complementation Primers
cysMprom XbaI FW	5′-ATATATTCTAGACATCAGTTTTATTGGTTTTGGTACTTTTTCAACTC-3′
cysMprom SphI BamHI RV	5′-ATAGGATCCGCATGCAATTTTAATATCCTTTTTTGTTTAATAATGATAGTTTTATAAAAG-3′
addBA F38011 sphI FW	5′-ATATATGCATGCATGAAATTAAGAATTTTTAGTTCTTCAAGACAAATTAGAG-3′
addBA_F38011 BamHI RV	5′-ATATATGGATCCTTATATTTCCAAAATTTGAATTTCATTTTCCAAACAATAAAC-3′
HygR EcoRI FW	5′-ATATATGCATGCATGAAATTAAGAATTTTTAGTTCTTCAAGACAAATTAGAG-3′
HygR PstI RV	5′-ATATATCTGCAGTTATCATGCCTTTCTTTGTCTGTATTC-3′
Cipro Sensitivity primers
BamHI addB T	5′-TATATAGGATCCATGAAATTAAGAATTTTTAGCTCTTCAAGG-3′
SacI addA B	5′-TATATAGAGCTCTTATATTTCCAAAATTTGAATTTCATTTTCC-3′
NcoI addB T	5′-TATATACCATGGATGAAATTAAGAATTTTTAGCTCTTCAAGG-3′
PstI addA B	5′-TATATACTGCAGTTATATTTCCAAAATTTGAATTTCATTTTCC-3′
NdeI recA T	5′-TATATACATATGGATGATAATAAAAGAAAATCTCTAGAC-3′
KpnI recA B	5′-TATATAGGTACCTTATTCTTCTCCTTCGTCATCTTC-3′
RT-qPCR Primers
RpoA RT-F	5′-CGAGCTTGCTTTGATGAGTG-3′
RpoA RT-R	5′-AGTTCCCACAGGAAAACCTA-3′
AddA_RT1F	5′-GGGTTTGGAATTTGATCATGTG-3′
AddA_RT1R	5′-ATGGAGTTGCCAACCTTGATT-3′
AddB_RT1F	5′-CCTTTCGAGTTTCCGCTTTC-3′
AddB_RT1R	5′-CGCTTAAAACCCTAGGTTCTGC-3′

RNA extraction and quantitative reverse transcription PCR

C. jejuni strain F38011 was inoculated at an OD540 of 0.05 in 50 mL of MH broth and MH broth supplemented with 0.05% (w/v) sodium deoxycholate (Sigma-Aldrich, St. Louis, MO) in 125 mL flasks and incubated at 37 °C under microaerobic conditions with orbital shaking (220 rpm). C. jejuni was grown in MH and MH deoxycholate side-by-side in the same chamber and an aliquot of each culture was taken at 10, 12, 14, and 16 hours post inoculation. These time points were chosen because C. jejuni growth in medium with deoxycholate was inhibited after 12 hours compared with MH broth without deoxycholate (not shown). Bacteria from each aliquot were isolated by centrifugation, culture supernatants were removed, and pellets were snap frozen in liquid nitrogen. To ensure reproducible results, the time-course experiment was repeated twice. Total bacterial RNA was isolated using the Ambion Ribopure Bacteria kit and treated twice with the supplied DNAse (Thermo Fisher Scientific, Waltham, MA). All RNAs were stored in Elution Solution. First strand cDNA synthesis was conducted using the Invitrogen SuperScript™ III First-Strand Synthesis System for RT-PCR (Thermo Fisher Scientific, Waltham, MA) according to the manufacturer’s instructions. A total of 450–500 nanograms of RNA per sample was used as template, and separate reactions were conducted with and without reverse transcriptase for each sample. Applied BiosystemsTM Power SYBR™ Green PCR Master Mix (Thermo Fisher Scientific, Waltham, MA) was used according to the manufacturer’s instructions for RT-PCR reactions. Amplification was conducted using an Applied BiosystemsTM 7500 Fast Real-Time PCR System (Thermo Fisher Scientific, Waltham, MA) and data was analyzed using version 2.3 of the associated software. Reaction conditions were 95 °C for 10 min followed by 40 cycles of 95 °C for 1 minute and 60 °C for 1 minute. The RT-qPCR primers are listed in Table 1.

Sensitivity of an E. coli ΔrecBCD strain to ciprofloxacin when complemented with C. jejuni addAB

The addAB genes and recA gene were amplified from C. jejuni strain 81–176 (see Cipro Sensitivity primers, Table 1). Fragments were incorporated into the pacycDuet-1 vector to make two distinct vectors: pacycDuet-recA and pacycDuet-addAB-recA. E. coli V3060[8,9] (ΔrecBCD) was transformed by electroporation with pacycDuet-1, pacycDuet-recA, or pacycDuet-addAB-recA. Ciprofloxacin sensitivity was determined by inoculation of 5 mL of LB-Miller broth or 5 mL of LB-Miller broth with 0.0025 μg/mL ciprofloxacin-HCl (Alfa Aesar, Haverhill, MA). The cultures were grown for 18.5 hours with shaking at 37 °C, and then the OD600 was measured. Significant differences were determined by one-way ANOVA with Dunnett’s multiple comparisons test of culture OD600 ratios (with ciprofloxacin/without ciprofloxacin).

Generation of C. jejuni addA and addB gene mutants and a complemented isolate

C. jejuni addA, addB, and addAB mutants were generated in strain F38011 using standard molecular biology techniques. The suicide vectors were constructed using In-Fusion Cloning (Takara Bio, USA) to link DNA regions flanking the gene of interest, a chloramphenicol cassette, and the pBSK-Kan2 vector. The flanking regions for the addA suicide vector were amplified by PCR from C. jejuni strain F38011, and the flanking regions for the addB and addAB suicide vectors were amplified from C. jejuni strain 81–176 (see Mutant Generation Primers, Table 1).

The complementation of the C. jejuni F38011 addAB mutant was achieved using a procedure similar to that described by Karlyshev and Wren[33]. Briefly, the cysM promoter (301 bp) and addAB genes (5129 bp) were amplified from C. jejuni strain F38011 and inserted between the rRNA flanking regions of prRNA-Hygro (the order is: rRNA upstream - XbaI - cysM promoter - SphI - addAB - BamHI - FLAG tag - EcoRI - HygR Cassette - PstI - rRNA downstream) (Table 1, Complementation Primers, Supplementary Figure 1).

The vectors (for both gene deletion or complementation) were electroporated into the C. jejuni F38011 strain and addAB mutant, and transformed isolates were selected on MH blood agar with appropriate antibiotic. All electorportations were conducted using a Bio-Rad E. coli Pulser and 2 mm gap cuvettes at the 2.5 kV voltage setting. The mutant and complemented isolates were confirmed by PCR.

C. jejuni ciprofloxacin sensitivity assays

The minimum inhibitory concentration (MIC) of ciprofloxacin was determined by inoculating 5 mL of MH broth containing a 2-fold serial dilution (from 0.5 μg/mL to 0.00390625 μg/mL) of ciprofloxacin with the C. jejuni F38011 wild-type strain, addA mutant, addB mutant, addAB mutant, and addAB mutant harboring the prRNA-cysM prom-addAB-hygromycin insertion. Cultures were grown with orbital shaking under microaerobic conditions (5% O2, 10% CO2, 85% N2) at 37 °C for 18 hours. The MIC was defined as the lowest concentration of ciprofloxacin that impaired growth (less culture density) compared to the MH only (no ciprofloxacin) control.

C. jejuni deoxycholate sensitivity assays

C. jejuni was grown overnight in MH broth, pelleted, adjusted to an OD540 of 0.1 in PBS. Ten-fold serial dilutions were then performed in PBS. Four µL drops were spotted on to MH agar plates and MH agar plates containing 0.1% (w/v) sodium deoxycholate. Growth on the plates was evaluated after 48 hours of incubation.

Pulsed-field gel electrophoresis and DNA-repair assays

DNA damage in response to deoxycholate was evaluated by growing C. jejuni strain F38011 in MH broth and MH broth with the indicated concentrations of sodium deoxycholate for 20 hours. Samples were collected and embedded in agarose plugs.

Ciprofloxacin inhibits the function of DNA gyrase and increases the likelihood of DNA double strand breaks. The ability to repair ciprofloxacin-induced damage was evaluated by inoculating an overnight culture of the C. jejuni F38011 wild-type strain and addAB mutant in MH broth at an OD540 of 0.4. A sample was collected immediately after dilution as a negative control sample. Ciprofloxacin was added at a concentration of 1 µg/mL and the cultures were incubated with orbital shaking under microaerobic conditions at 37 °C for 15 minutes. Cultures were pelleted at 10,000 x g for 10 minutes, washed in MH broth, pelleted again, and then suspended in fresh MH broth without ciprofloxacin. A sample was taken immediately after culture resuspension as the zero time point. Cultures were incubated with orbital shaking under microaerobic conditions at 37 °C and samples were taken every 45 minutes until 225 minutes post ciprofloxacin treatment. All samples were embedded in agarose plugs and prepared for Pulsed-Field Gel Electrophoresis (PFGE). PFGE was performed using a CHEF-DR III variable angle system (Bio-Rad, Hercules, CA), as outlined previously[34]. Typical run parameters consisted of a reorientation angle of 120° with a constant voltage of 120 V at 14 °C. The pulse times were ramped from 6.8 to 35.4 s over 19 hours. Gels were stained for 30 minutes in 3 µg/mL ethidium bromide and destained in water. Images were captured using a GE Healthcare ImageQuant LAS 4000 Mini system and processed using ImageJ (version 2.0.0-rc34).

Leghorn chickens

Two-day old straight run Pearl-White Leghorn chickens (catalogue number PEAS) were obtained from Murray McMurray Hatchery (Webster City, IA). Chickens were housed in PLAS labs isolator units and provided food (Purina Flock Raiser Crumbles) and water ad libitum. Water was fortified with Quik Chik (Murray McMurray Hatchery) vitamin and electrolyte supplement for one week after receipt. All efforts were made to raise the animals humanely.

C. jejuni colonization of chickens

Motility assays were performed as outlined elsewhere[35]. The chickens were divided into 6 treatment groups of 10–12 broiler chickens, with each group housed in a separate isolator unit. The treatment groups consisted of: 1) one negative control group (mock inoculated); 2) one positive control group inoculated with a wild-type strain of C. jejuni, strain F38011 (C. jejuni-inoculated group); 3) one group inoculated with the addAB mutant; 4) one group inoculated with the addAB mutant harboring the prRNA-cysM prom-addAB-hygromycin insertion; 5) an equal mix of the C. jejuni wild-type strain and addAB mutant; and 6) an equal mix of the C. jejuni wild-type strain and addAB mutant harboring the prRNA-cysM prom-addAB-hygromycin insertion. At 12 days of age, the chickens were orally inoculated with 1–2 × 107  C. jejuni in 0.2 mL of MH broth or were mock inoculated with 0.2 mL of MH broth without bacteria. The chickens were housed for an additional 14 days after inoculation and then euthanized by carbon dioxide asphyxiation followed by cervical dislocation. A 1–2 inch section of the duodenum and one cecum were dissected from each bird. Each section of the duodenum and cecum was individually weighed, stomached in an equal volume of MH broth (1 mL broth per gram of content), and then serially diluted for enumeration on Campy Cefex Agar[36]. The samples from groups five and six were also plated on Campy Cefex Agar with 8 μg/mL chloramphenicol to enable determination of the colonization of the addAB mutant and addAB mutant harboring the prRNA-cysM prom-addAB-hygromycin insertion. Chicken experiments were performed in accordance with the National Research Council’s Guide for the Care and Use of Laboratory Animals (8th edition) and approved by the Washington State University Institutional Animal Care and Use Committee under protocol ASAF 4847.

Measurement of total bile contents in the chicken intestine

Bile acid levels in the intestines of negative (uninoculated) and positive (inoculated with a wild-type C. jejuni strain) control groups were measured using the Bile Acid Assay Kit as outlined by the manufacturer (Sigma-Aldrich, St. Louis, MO). Fluorescence intensity (λex = 530 nm/λem = 595 nm) was read using a VICTOR ×5 reader (Perkin-Elmer, Waltham, MA).

Supplementary Figures