Enhancement of Campylobacter hepaticus culturing to facilitate downstream applications

Campylobacter hepaticus causes Spotty Liver Disease (SLD) in chickens. C. hepaticus is fastidious and slow-growing, presenting difficulties when growing this bacterium for the preparation of bacterin vaccines and experimental disease challenge trials. This study applied genomic analysis and in vitro experiments to develop an enhanced C. hepaticus liquid culture method. In silico analysis of the anabolic pathways encoded by C. hepaticus revealed that the bacterium is unable to biosynthesise l-cysteine, l-lysine and l-arginine. It was found that l-cysteine added to Brucella broth, significantly enhanced the growth of C. hepaticus, but l-lysine or l-arginine addition did not enhance growth. Brucella broth supplemented with l-cysteine (0.4 mM), l-glutamine (4 mM), and sodium pyruvate (10 mM) gave high-density growth of C. hepaticus and resulted in an almost tenfold increase in culture density compared to the growth in Brucella broth alone (log10 = 9.3 vs 8.4 CFU/mL). The type of culture flask used also significantly affected C. hepaticus culture density. An SLD challenge trial demonstrated that C. hepaticus grown in the enhanced culture conditions retained full virulence. The enhanced liquid culture method developed in this study enables the efficient production of bacterial biomass and therefore facilitates further studies of SLD biology and vaccine development.

The effect of culture vessel type, pH, and inoculum level on the yield of C. hepaticus. Before testing the effect of amino acids and sodium pyruvate on the growth of C. hepaticus, three factors, culture vessel types, pH, and inoculum level were examined. Figure 2A shows that C. hepaticus grew to a higher final density when growth in 24-well cell culture plates (CCP24) and 75 cm 2 cell culture flasks (TCF75) (log10 CFU/ mL = 8.54-8.62 (P ≤ 0.05) compared to that seen in 50 mL tubes (CT50) and 250 mL Erlenmeyer flasks (EF) (log CFU/mL = 7.75-8.0). The CCP24 culture plates (1 mL of culture added/well) were therefore used to test the effect of amino acid and sodium pyruvate supplementation on the growth of C. hepaticus, and TCF75 flasks (25 mL of media/flask) were used to scale-up the production of biomass of C. hepaticus.
C. hepaticus was monitored for growth in Brucella broth adjusted to a pH ranging from 6.0 to 8.0 to evaluate the pH tolerance. There was no statistically significant difference in the growth of C. hepaticus, indicating little or no effect of pH within this range (Fig. 2B). Therefore, Brucella broth without pH adjustment was used for the further culturing of C. hepaticus, as it has a neutral pH value (7.0 ± 0.2).
The growth kinetics of C. hepaticus in Brucella broth was investigated using initial inocula ranging from 10 1 to 10 7 CFU/mL (Fig. 2C). Viable cells counts were carried out after 0, 24, 48, 60, 72 and 96 h of incubation. The two highest inoculum levels (10 6 and 10 7 CFU/ml) produced maximum growth to 10 8 CFU/ml after 24 h of incubation. With the lower inoculum levels (10 5 and 10 4 CFU/mL), C. hepaticus achieved maximum growth after 48 h of incubation and for low inoculum levels of 10 1 to 10 3 CFU/ml, the maximum growth of C. hepaticus was reached after 60 h.
The effect of amino acids and sodium pyruvate on the growth of C. hepaticus. C. hepaticus cultured in Brucella broth supplemented with 0.4 mM l-cysteine, 4 mM l-glutamine, 0.8 mM l-valine, 0.4 mM l-serine and 10 mM sodium pyruvate exhibited a significantly greater growth compared to a culture grown in   (Fig. 3). l-cysteine (0.4 mM) showed highest growth enhancement to log10 = 9.26 CFU/mL, significantly higher than the density obtained with the other supplements (p ≤ 0.005). No significant difference was observed (p > 0.05, Fig. 3) when C. hepaticus was grown in Brucella broth supplemented with other amino acids including l-lysine, l-methionine, l-histidine, l-glycine, l-arginine, l-leucine, and l-threonine compared to a culture grown in Brucella broth only.
The growth of C. hepaticus was then investigated in Brucella broth supplemented with different combinations of each compound that could enhance the growth of C. hepaticus, as demonstrated above, including l-cysteine, l-glutamine, l-valine, l-serine, and sodium pyruvate. In general, all combinations of supplements in Brucella broth significantly enhanced the growth of C. hepaticus compared to Brucella broth alone. The highest growth of C. hepaticus observed was in Brucella broth supplemented with a mixture of l-cysteine (0.4 mM), l-glutamine (4 mM) and sodium pyruvate (10 mM) (log10 = 9.34), followed by a combination of l-cysteine (0.4 mM) and sodium pyruvate (10 mM) (log10 = 9.17), and only l-cysteine (0.4 mM) (log10 = 9.11) ( Table 1). Brucella broth supplemented with l-cysteine (0.4 mM), and l-glutamine (4 mM) and sodium pyruvate (10 mM) was then used to grow C. hepaticus in TCF75 flasks in the following experiments.
Effect of temperature and incubating conditions on growth of C. hepaticus. Because the culture vessels that provided the largest surface to volume ratio produced the highest density cultures it was proposed that gas exchange between the liquid medium and the microaerophilic atmosphere might be important. Therefore, the effects of agitation of the cultures by shaking were investigated. In the modified Brucella broth developed in this study, after 48 h of incubation C. hepaticus grew significantly better (p ≤ 0.05) in static conditions, with log10 = 9.25 CFU/mL compared to log10 = 8.76 CFU/mL when shaken. It was also found that C.   Experimental infection of laying hens with C. hepaticus culture grown in modified Brucella broth. One of the principal reasons to develop improved culturing conditions for C. hepaticus was to make the preparation of challenge inocula easier and more reproducible for experimental infection trials used to investigate the biology of infections and assess the efficacy of various SLD treatment protocols. Experience with preparing challenge inocula using the previously described plate harvesting method had suggested that the state of the inoculum was critical to the success of infection 10 . Therefore, it was important to establish that the cultures grown in the modified Brucella liquid medium were capable of eliciting disease. Hens entering peak lay were challenged with, either C. hepaticus HV10 T harvested from Brucella agar plates, or bacteria grown in the newly devised liquid media conditions and the induction of SLD lesions on the liver were scored. Birds in the  www.nature.com/scientificreports/ control group, inoculated with fresh modified Brucella broth, showed no lesion on livers, whereas 12 out of 12 hens inoculated with the agar plate derived bacteria had lesions on the liver and 11 out of 12 hens inoculated with bacteria grown in the modified Brucella liquid had liver lesions. Based on disease scores, there was no significant difference between the degree of disease elicited by bacteria grown under the two conditions (Fig. 5), demonstrating that C. hepaticus grown in the modified Brucella broth developed in this study had equal levels of virulence as those grown by the previously described plate culture method. The group sizes were sufficient to detect a one-point difference in mean scores with an alpha of 0.05 and 80% power.

Discussion
In silico pathway analysis of the C. hepaticus genome predicted that the bacterium lacks some of the genes encoding enzymes required for the biosynthesis of l-cysteine, l-lysine and l-arginine. C. hepaticus lacks the genes cysE and cysK, cysM to synthesise l-cysteine from l-serine and sulphur. cysE encodes serine acetyltransferase and cysK encodes cysteine synthase. These enzymes are required for the synthesis of l-cysteine from l-serine in bacteria and plants 19 . CysE synthesises l-cysteine from l-serine by catalysing an acyl transfer from acetyl-CoA. CysK catalyses O-acetyl-l-serine combining with hydrogen sulphide to yield l-cysteine 19 . The addition of l-cysteine to Brucella broth improved the growth of C. hepaticus significantly, to more than 10 9 CFU/mL compared to Brucella broth only. This indicates that even though Brucella broth is a rich nutrient source the l-cysteine level is insufficient to support maximal growth of C. hepaticus. Growth support of l-cysteine was also reported in other Campylobacter species. l-cysteine was defined as a vital source of sulphur for C. jejuni 20 . C. jejuni grown in minimal medium with l-cysteine added was better than their growth in the medium without cysteine 14,21 . l-cysteine has been identified as a chemotactic attractants of C. jejuni 22,23 . In contrast, although C. hepaticus was also predicted to be unable to biosynthesise l-lysine and l-arginine, the addition of these amino acids to Brucella broth did not significantly increase the culture densities that could be achieved for C. hepaticus compared to densities supported by unmodified Brucella broth. Similarly, the addition of l-arginine and l-lysine to minimal media did not enhance the growth of C. jejuni because C. jejuni may have genes that support the biosynthesis of l-arginine and l-lysine 14 . Better growth of C. hepaticus in Brucella broth supplemented with l-glutamine, l-valine or l-serine was observed although no missing genes that are responsible for the biosynthesis of these amino acids by C. hepaticus were detected. It was demonstrated that l-glutamine and l-serine are chemoattractants for C. jejuni 24 . In chickens, l-cysteine, l-glutamine, and l-serine are abundant in chicken liver. The chicken gut also has sufficent quantities of necessary amino acids such as l-cysteine, l-glutamine, l-valine and l-serine for the growth of many bacteria 25 and this likely explains why, despite its inability to synthesise a number of amino acids, C. hepaticus can colonise the intestinal tract of laying hens 10 . It was found that the reduction of amino acids such as l-cysteine, l-glutamine and l-serine in the chicken diet contributed to the reduction of Campylobacters in chickens 26 . The authors explained that these amino acid are involved in the formation of the mucin production of the intestinal mucus layer and essential for the survival and growth of Campylobacters 27 . Thus, an increase or decrease in the concentration of these amino acids results in changes in the number of Campylobacters in the chicken gut 26 . For other amino acids including l-methionine, l-histidine, l-glycine, www.nature.com/scientificreports/ l-leucine, and l-threonine, there was no significant difference in the growth of C. hepaticus. These results agree with the bioinformatic pathway analysis. This study found that the C. hepaticus showed maximum growth in Brucella broth supplemented with a mixture of l-cysteine (0.4 mM), l-glutamine (4 mM) and sodium pyruvate (10 mM) and reached 10 9.34 CFU/ mL. Sodium pyruvate provides an additional carbon source to support the growth of C. hepaticus in Brucella broth. C. hepaticus may heavily depend on the Kreb's cycle to generate energy. Many studies have demonstrated the important role of pyruvate in the growth of campylobacters. Sodium pyruvate is one of the compounds in the Campylobacter growth supplement product (FBP supplement) 28 and Campylobacter selective supplement 29 . Pyruvate was found to promote the growth of C. jejuni NCTC 11168 in MEM medium 14 . This compound plays a central role in C. jejuni metabolism and can be fermented to various products such as acetate, formate, lactate and succinate. It is linked to carbohydrate and amino acid catabolism to produce energy. The synthesis of l-leucine, l-valine, l-alanine and l-isoleucine by campylobacters use pathways in which pyruvate can play a key role 30 . Pyruvate also acts as an electron acceptor and can decrease the concentration of hydrogen peroxide to reduce the damage caused by oxygen to bacteria 31 .
C. hepaticus grew to higher densities in CCP24 and TCF75 than CT50 and EF, suggesting that the surface-area-to-volume ratio (S:V) may be important, possibly for gas exchange, as mentioned in a study of improvement of culturing of C. jejuni 32 . These authors demonstrated that TCF75 has more S:V ratio than EF and CT50 and therefore TCF75 provided better atmospheric exchange than EF and CT50. Also, tissue culture flasks were recommended to grow and study standard growth curves of C. jejuni 33 . In a study of the growth of Helicobacter pylori, Gas-Permeable Lifecell tissue culture flasks gave improved growth in Brucella broth supplemented with fetal bovine serum 34 . These authors also mentioned the effect of surface area on the growth of H. pylori and suggested that a small surface area resulted in poorer growth of H. pylori. Tissue culture flasks were used early in the history of the culture of C. jejuni and C. pylori 35,36 and have been employed in many studies of Campylobacter species. It may show that C. hepaticus requires a larger S:V ratio to grow due to better gas exchange, although it is then unexpected that agitation had a negative effect. A low level of oxygen may be an obligatory requirement for C. hepaticus growth. Under anaerobic conditions and higher oxygen tensions (21%), C. hepaticus failed to grow 1 . Similarly, C. jejuni does not grow under anaerobic conditions 37,38 and oxygen is considered a requirement for www.nature.com/scientificreports/ DNA synthesis in C. jejuni 37 . Campylobacter spp. are sensitive to high oxygen tensions, but still need an optimal oxygen concentration (2%-10%) to grow 38 . C. hepaticus contains genes for oxidative phosphorylation 11 that need oxygen as an acceptor to generate ATP. C. hepaticus achieved maximum growth after 48-72 h, depending on the initial inoculum, showing slower growth in comparison to other Campylobacters. The maximum growth of C. jejuni was at around 30 h in both Nutrient Broth Number 2 and Mueller Hinton broth 39 . C. jejuni reached the highest densities after 24 h of growth in Brain Heart Infusion broth 40 . C. hepaticus can grow in a range of pH from 6.0 to 8.0. It has been reported that other Campylobacter species also grow well in this pH range 41 . Chickens normally have a pH of around 6.3 in the liver 42 , 6.4 in the small and large intestine 43 , 6.6-6.7 for caecum 43,44 , and 6.0 for bile 45 . These are all tissues and environments in which it has been shown that C. hepaticus can survive and colonise. C. hepaticus showed better growth at 37 °C than 42 °C while this microorganism has only been isolated from chickens, which have a body temperature of 40-42°oC 46 . The growth of C. hepaticus was reported to be somewhat slow at 42 °C, taking 7 days to form colonies in sheep blood agar 3 . Temperature differently affects Campylobacter species regarding the growth, motility, and ability to invade the host cells. A study showed that C. jejuni grew at both 37 °C and 42 °C but showed differences in motility and invasion. C. coli grew and moved better at 42 °C. C. fetus, a bacterium that is frequently detected in poultry, showed greater growth and invasion at 37 °C 47 .
Using the modified Brucella broth developed in this study, together with growth conditions including the use of a large surface area culture vessel, at 37 °C, in microaerophilic and static conditions, C. hepaticus cultures could grow to 10 9 CFU/mL and showed virulence in laying hens. This culturing method is time-saving and more cost-effective than the previously used plate harvesting method to obtain the large biomass required for SLD animal induction experiments and bacterin vaccine production. It also reduces the amount of subculturing needed, possibly minimising the effect of repeated subculture on the virulence of C. hepaticus. A study has shown that the repeated subculturing of the somewhat related organism, H. pylori, could result in a decrease of adhesion, motility, gastric inflammation and cytotoxicity, and repeated culturing is a recognised way that bacteria have been attenuated to produce live vaccines 48 . Thus, the method described in this study can facilitate further studies on C. hepaticus biology and SLD.

Materials and methods
In silico analysis of C. hepaticus metabolism. The metabolic pathways encoded within the genomes of C. hepaticus strains HV10 T , 19L and VICOCT18, were analysed using (1) RAST (Rapid Annotation using Subsystem Technology) 12 and the SEED 49 (http:// rast. nmpdr. org/ rast. cgi) for annotating the genomes and metabolic pathways prediction of C. hepaticus; and (2) BLAST 50 (Basic Local Alignment Search Tool) (https:// blast. ncbi. nlm. nih. gov/ Blast. cgi) for comparing nucleotides and protein sequences of C. hepaticus with available sequences in the gene bank. C. hepaticus HV10 T is the type strain for C. hepaticus, 19L is a representative of a clade that is distinct form HV10 T and VICOCT18 is a more recent isolate.
Bacterial strains and culture conditions. The C. hepaticus strains HV10 T , 19L 11 , VICOCT18, WEST-ERN3, NSWJUNE19, SAJULY18 and DALE3 51 were used in the study. These strains are representative isolates from independent SLD outbreaks from widely separated geographical locations in Australia. All C. hepaticus strains were stored at − 80 °C in 70% Brucella broth (BD BBL™) and 30% glycerol. C. hepaticus was routinely cultured on Brucella agar plates (Brucella broth (BD BBL™) + 1.5% agar (BD BBL™)) supplemented with 5% defibrinated horse blood (Equicel) (HBA) and cultured at 37 °C under microaerobic conditions (created using Campygen 3.5L gas generation packs (Oxoid)) in an anaerobic jar, for 96 h to recover C. hepaticus cells from − 80 °C stock or for 72 h if they were subcultured from HBA plates.
Effect of type of culture vessel on the growth of C. hepaticus. Costar ® 24-well cell culture plates (CCP24), Corning ® 50 mL centrifuge tubes (CT50) with a vented cap (0.2 μm pore size), Corning ® 75cm 2 cell culture flask (TCF75) with a vented cap (0.2 μm pore size) and Erlenmeyer flasks (250 mL) (EF) were used to compare the growth of C. hepaticus in Brucella broth (BD BBL™). The volume of culture media used in CCP24 was 1 mL, 25 mL (CT50 and TCF) and 40 mL (EF). All vessels were placed in BD GasPak™ EZ container, charged with CampyGen 3.5 L (Oxoid) to produce microaerobic conditions and then incubated at 37 °C for 48 h. The growth rate of C. hepaticus HV10 was determined by the plate count method on HBA plates.
Effect of pH and initial inoculum on C. hepaticus growth. The pH of Brucella broth was adjusted from 6.0 to 8.0 in increments of 0.5 units using 1 M NaOH or 1 M HCl. C. hepaticus was cultured in CCP24 plates to examine growth at different pH levels. Each well of CCP24 was inoculated with 10 6 CFU/mL of C. hepaticus HV10 T . Plates were incubated under microaerobic conditions at 37 °C. Growth was enumerated after 48 h of incubation by plating serial dilution on HBA. The experiment was performed in triplicate.
The effect of initial inoculum size on the growth kinetics and final culture yields was analysed using seven initial inoculum levels (10 1 to 10 7 CFU/ml) of C. hepaticus HV10 T in a CCP24 plate. The plates were incubated at 37 °C under microaerobic conditions. The growth of C. hepaticus HV10 T was examined after 24, 48, 60, 72, and 96 h of incubation using the plate count method on HBA.
Growth of C. hepaticus in static and shaking conditions. C. hepaticus HV10 T was suspended into Brucella broth to achieve acell density of 10 5 CFU/ml and then 1 mL of bacterial suspension was incubated into each well of CCP24 at static condition. For shaking conditions, the anaerobic jars were shaken in a shaking incubator with a speed of 100 rpm. After  www.nature.com/scientificreports/ Effects of sodium pyruvate and amino acids on the growth of C. hepaticus. Based on the results of in silico pathway analysis of C. hepaticus and results from Alazzam et al. 14 , the following supplements were added to Brucella broth to examine the growth of this bacterium: amino acids (l-cysteine, l-lysine, l-methionine, and l-leucine, l-glutamine, l-valine, l-histidine); carbon source (sodium pyruvate) (Sigma). Each compound was completely dissolved in Milli-Q water at a concentration recommended in MCLMAN, a new minimal medium for C. jejuni 14  A total of 36 Hy-Line brown laying hens were used in the experiment. Only healthy birds laying eggs regularly were included in the study. Chickens were housed in groups of 4 birds per pen, with 3 pens per group: a total of 12 birds in each group (n = 12). Birds were randomly allocated to groups and cages by stratified rank order based on weight. Unchallenged control birds were orally inoculated with 1 ml of Brucella broth only whereas 1 ml of Brucella broth containing 1 × 10 9 CFU of C. hepaticus HV10 T was orally administered to the birds for the plate and broth methods groups. Birds were sacrificed after 5 days and SLD lesions on the surface of the liver were counted to measure the severity of the induced disease. The experienced and trained scorers were blinded to the treatment groups. Scores were based on a logarithmic scale from 0 to 4, where 0 = no visible lesions, 1 = 1-9 lesions, 2 = 10-99 lesions, 3 = 100-999 lesions and 4 = more than 1000 lesions.

Statistical analysis.
All experiments to study the growth of C. hepaticus HV10 in different conditions and supplements were repeated a minimum of three times with biological replicates. Campylobacter cell counts in all tests were converted to log 10 CFU/mL. Statistical comparison of all parameters was performed by t-test, one-way ANOVA using Graphpad Prism version 8 for Windows, GraphPad Software (San Diego California USA, www. graph pad. com). The significance level was set at 5% (p ≤ 0.05). Sample size calculation for the SLD animal trial was performed using the online calculator at https:// clinc alc. com/ stats/ sampl esize. aspx.
Declarations. All the methods were carried out in accordance with relevant guidelines and regulations and the animal trial is reported in accordance with the ARRIVE Essential 10 guidelines.