Polyphosphate-mediated modulation of Campylobacter jejuni biofilm growth and stability

Abstract

Biofilms increase C. jejuni’s resilience to detergents, antibiotics, and environmental stressors. In these investigations, we studied the modulation of biofilm in response to phosphate related stressors. We found that the deletion of ppk1, phoX, and ppk2 (polyphosphate associated [poly P] genes) in C. jejuni modulated different stages of biofilm formation such as attached microcolonies, air-liquid biofilms, and biofilm shedding. Additionally, inorganic phosphate also modulated attached microcolonies, air-liquid biofilms, and biofilm shedding both independently of and additively in the poly P associated mutants. Furthermore, we observed that these different biofilm stages were affected by biofilm age: for example, the adherent microcolonies were maximum on day 2, while biofilm growth at the air-liquid interface and shedding was highest on day 3. Also, we observed altered calcofluor white reactive polysaccharides in poly P-associated mutants, as well as increased secretion of autoinducer-2 (AI-2) quorum sensing molecules in the ∆ppk2 mutant. Further, the polysaccharide and flagellar biosynthesis genes, that are associated with biofilm formation, were altered in these poly P-associated mutants. We conclude that the phosphate limiting condition modulates C. jejuni biofilm formation.

Introduction

As a common bacterial pathogen, Campylobacter jejuni infects approximately 1 million people in the US each year.¹ Additionally, an estimated 1/1000 clinical cases may result in long-term neurological conditions, including Guillain-Barre syndrome.² One of C. jejuni’s unique qualities is the paradox between its stringent growth conditions and its ubiquity as an effective pathogen. Closer examination of the pathogen’s sensitivity has shown that compared with other foodborne pathogens, C. jejuni is more sensitive to osmotic stress, desiccation, oxidative stress, and other sources of environmental insult.^3,4 One of the ways that C. jejuni may overcome its fragility in the face of environmental hostility is its ability to form protective biofilms. Biofilms formation increases C. jejuni survival under unfavorable conditions.⁵ Compared with planktonic cells, C. jejuni biofilms are more resistant to antibiotics and detergents, and have increased survival in non-host environments.^6,7 The increased survivability conferred by biofilm may suggest an explanation for the unlikely contrast between the relative fragility of C. jejuni and its success as a gastrointestinal pathogen.

Poly P plays an important role in stress tolerance, and virulence in many bacteria.⁸ Poly P is reversibly synthesized by polyphosphate kinase 1 (Ppk1) and poly P in cells act as a phosphate bank for metabolic use, including the synthesis of GTP.^9,10 The Ppk2 preferentially removes inorganic phosphate from the poly P chain and adds GDP to generate GTP.¹¹ The alkaline phosphatase PhoX/PhoA is required for hydrolysis of organophosphorous compounds to yield inorganic phosphate,¹² which is necessary for poly P synthesis. In C. jejuni, deletion of ppk1 and phoX results in reduced poly P levels,^12,13 whereas the deletion of ppk2 does not affect the poly P levels.¹⁴ Interestingly, recent studies have shown that the deletion of poly P associated genes (ppk1, ppk2, and phoX) results in enhanced biofilm formation in C. jejuni 81–176.^12-14 However, it is still not known how poly P as a key stress regulator in C. jejuni, modulates this response.

The total amount of biofilms produced by C. jejuni is affected by the size, shape and environmental conditions.^5,15 The shedding of cells back into a planktonic state also determines the amount of biofilm formed.^16,17 In addition, various genes are shown to modulate C. jejuni biofilms.^6,18 Particularly deletion of waaF, the gene responsible for modification of the inner core of the lipo-oligosaccharide (LOS),¹⁸ increases both biofilm as well as calcofluor white (CFW) reactive β-1,3 and β-1,4 linked polysaccharides on cell surfaces which are important components of biofilm matrix.¹⁹ Importantly, the waaF gene is under the transcriptional regulation of PhosR/S, the C. jejuni homolog of the two-component phosphate sensor and response operon PhoR/B, which suggests a connection between changes in biofilm-affecting polysaccharide genes and phosphate metabolism.²⁰ Additionally, a study in E. coli has shown that poly P degradation is important for biofilm formation and quorum sensing (QS) by eliciting autoinducer 2 (AI-2) production.²¹ Poly P deficient mutants in Pseudomonas aeruginosa exhibit defects in QS and biofilm formation where poly P has been identified as a global regulator of virulence.^9,22 While these studies have identified phosphate metabolism as a potential regulator of protective biofilm mechanisms, little is known about the translational mechanisms that allow C. jejuni to form biofilm in response to phosphate availability in the environment.

Here we have examined more closely the role of poly P limitation (Δppk1/ΔphoX) and repletion (Δppk2), low-phosphate conditions (ΔphoX), and the impact of exogenous phosphate on growth of surface attached microcolonies (adherent biofilms), growth of biofilms at the air-liquid interface, and the amount of viable cells shed from biofilms. We have also compared these mutants in their initiation of QS signals, changes in CFW reactive polysaccharide composition and assessed changes in expression of genes implicated in biofilm formation.

Results

Number of C. jejuni adherent microcolonies is modulated by poly P and the presence of inorganic phosphate

Bacterial attachment and microcolony formation is an important stage in biofilm development.^16,23,24 Since adherent microcolonies mature into biofilms at the air–liquid interface,¹⁶ assessment of microcolony formation will provide insight into how poly P and phosphate regulate biofilm synthesis in C. jejuni. Here, we assessed the microcolony formation in C. jejuni 81–176 poly P associated mutants using microscopic slides for each time point independently and the progression of the microcolony in to mature biofilm was not monitored. We found that the number of adherent microcolonies for the ppk1, ppk2, and phoX mutants increased significantly on day 1 compared with wild type (WT) when grown in the presence or absence of inorganic phosphate. The number of adherent microcolonies peaked on day 2 and decreased on day 3, without the addition of inorganic phosphate (Fig. 1). This may suggest that WT C. jejuni spend a longer time as planktonic cells compared with mutant strains, perhaps delaying biofilm formation. Moreover, the efficiency of attachment is not affected by the presence of phosphate in a poly P depletion scenario; for example ∆ppk1 mutant showed similar number of microcolonies in the presence or absence of phosphate (Fig. 1). The ∆ppk1 mutant on day 2 and 3 as well as the ∆ppk2 mutant on day 3 had significantly (P < 0.05) higher number of attached microcolonies compared with WT. However, in the ∆phoX mutant, the numbers of microcolonies were similar to the WT strain on day 2 and day 3, in the presence of phosphate (Fig. 1). Further, the observed results are not due to any growth differences in the mutants, as all mutants grow similar to WT in MH broth.^12-14 This experiment shows that phosphate plays a role in the transition between planktonic and biofilm phase of the bacterium, which may be direct or indirect. Additionally, we confirmed our results by determining the number of attached microcolonies formed over the total area of submerged portion of the slide. Results indicated that the number of microcolonies formed in WT and poly P associated mutants in the presence or absence of phosphate, overall followed a similar pattern as above (Fig. S2).

Figure 1. Number of attached C. jejuni microcolonies on glass slides. Each bar represents the mean ± SE of 3 independent experiments. Mutants were compared with the wild-type results with concurrent day and phosphate treatment (for example, phoX with phosphate was compared with wild-type with phosphate; phoX with no phosphate compared with wild type with no phosphate). In addition, each strain was also compared in the presence and absence of phosphate within itself (for example, phoX with phosphate was compared with phoX without phosphate). A P value of ≤ 0.05 was considered statistically significant. a, statistically significant when compared with wild type in the absence of phosphate; b, statistically significant when compared with wild type in the presence of phosphate; c, statistically significant when the strain grown in the presence of phosphate was compared with the same strain grown in the absence of phosphate. Phosphate supplemented condition is indicated by the strain name followed by letter p.

Complemented strains displayed phenotypes similar to those observed in the WT (Figs. S1 and S2). The data representing the results for the complementation strains in all assays are shown in Figures S1–S5.

Biofilm formation at the air-liquid interface in C. jejuni is influenced by inorganic phosphate

We observed that the thickness of air-liquid biofilms (which appeared as a discrete band at the air-liquid interface) increased in C. jejuni WT and poly P associated mutants over 3 d both in the absence or presence of phosphate (Fig. 2A). On day 1, the biofilm was both narrow and inconsistently developed for all strains both in the presence or absence of phosphate (Fig. 2A). A large increase in biofilm formation between day 2 and 3 was observed in WT and in all mutants. No difference in biofilm thickness was observed within the WT and poly P associated mutants in day 3 biofilms, in the presence or absence of phosphate. The ∆ppk1 mutant displayed increased biofilm formation compared with WT on day 2 and 3 with or without phosphate. Additionally, we observed that the ∆ppk1 mutant had a novel “hairy” phenotype on day 2 and 3, but WT cells had a smooth band of biofilm across the air-liquid interface for the same time point (Fig. 2B). The ∆ppk2 mutant had significantly lower biofilm thickness on day 1 and 2 compared with WT in the presence of phosphate; however, there was no difference in the day 3 biofilms. The biofilm thickness in the ∆phoX mutant was similar to WT on day 1 and 2 but increased on day 3, both in the presence and absence of phosphate (Fig. 2A). Differences observed in the mutants were partially restored to WT levels in the complemented strains in the absence of phosphate but were fully rescued in the presence of phosphate (Fig. S3).

Figure 2. (A) Biofilm growth at the air-liquid interface in the presence or absence of inorganic phosphate. Each bar represents the mean ± SE of 3 independent experiments. Mutants were compared with the wild-type results with concurrent day and phosphate treatment (for example, phoX with phosphate was compared with wild-type with phosphate; phoX with no phosphate compared with wild type with no phosphate). In addition, each strain was also compared in the presence and absence of phosphate within itself (for example, phoX with phosphate was compared with phoX without phosphate). A P value of ≤ 0.05 was considered statistically significant. a, statistically significant when compared with wild type in the absence of phosphate; b, statistically significant when compared with wild type in the presence of phosphate; c, statistically significant when the strain grown in the presence of phosphate was compared with same strain grown in the absence of phosphate. Phosphate supplemented condition is indicated by the strain name followed by letter p. (B) Microscopic morphology of the air liquid biofilm in the Δppk1 mutant showing a wide and hairy phenotype and the wild type strain with a smooth phenotype on day 3.

Phosphate decreases the initial biofilm shedding

Little is known about cell dispersal in C. jejuni biofilms and how this affects the bacteria’s ability to disseminate viable cells. Here, we hypothesized that phosphate metabolism may affect shedding of C. jejuni from air liquid biofilm. Biofilm shedding decreased from day 1 to day 2 by 1–2 log and increased an average of 3 logs from day 2 to day 3 in the WT and in mutants except ∆ppk2 (Fig. 3). The WT shed approximately 1 log more viable cells on day 1 than all mutants and was equally pronounced in the presence or absence of phosphate. The ∆ppk1 mutant biofilm shedding pattern was similar to that of the WT. On the other hand, the ∆phoX mutant on day 2 had decreased biofilm shedding by 1 log compared with WT. Interestingly; the ∆ppk2 mutant showed a significantly decreased shedding of cells on day 1 and 3 compared with the WT both in the presence and absence of phosphate. However, the ∆ppk2 mutant shedding between day 2 and 3 was similar, unlike the shedding pattern seen in other strains. Overall, phosphate decreased initial biofilm shedding in all mutants in the earlier stages but by day 3, the shedding pattern was similar to WT. Differences observed in the mutants were partially restored to WT levels in the complemented strains (Fig. S3).

Figure 3. Shedding of viable C. jejuni from biofilm. Biofilms were grown for 0, 24, 48, and 72 h in glass vials. The 0 time point refers to the sample taken at the time of inoculation. Each bar represents the mean ± SE of 3 independent experiments. Mutants were compared with the wild-type results with concurrent day and phosphate treatment (for example, phoX with phosphate was compared with wild-type with phosphate; phoX with no phosphate compared with wild type with no phosphate). In addition, each strain was also compared in the presence or absence of phosphate within itself (for example, phoX with phosphate was compared with phoX without phosphate). A P value of ≤ 0.05 was considered statistically significant. a, statistically significant when compared with wild type in the absence of phosphate; b, statistically significant when compared with wild type in the presence of phosphate; c, statistically significant when the strain grown in the presence of phosphate was compared with the strain grown in the absence of phosphate. Phosphate supplemented condition is indicated by the same strain name followed by letter p.

Calcofluor white reactivity indicates poly P-dependent polysaccharide variations

CFW reacts with polysaccharides and LOS possessing β-1, 3 and β-1, 4 linkages which are important components of C. jejuni biofilm matrix.^18,25 Here, we investigated whether CFW reactive polysaccharide in the biofilm is affected in poly P associated mutants and further, if these changes are modulated by phosphate. We observed that all mutants demonstrated a reduction in CFW reactivity compared with WT (Fig. 4) when grown microaerobically. The CFW hypo-reactivity was restored to WT levels in the ∆phoX mutant by the addition of 1 mM phosphate to the CFW plate and also in the complemented strain, whereas addition of phosphate partially rescued CFW reactivity in the ∆ppk2 and ∆ppk1 mutants. In addition, aerobic condition even though increased the CFW reactivity overall in all mutant strains, both in the presence and absence of phosphate, was lesser than the WT (Fig. 4).

Figure 4. Calcofluor White polysaccharide staining of wild type and poly P associated mutants in the absence (a) or presence (b) of 1 mM phosphate. CFW reactivity was visualized under long-wave (365 nm) UV light. Uniform growth was confirmed by visualization under visible light. Each condition was performed a total of three times; representative plates are shown. Complemented strains are indicated by the strain name followed by the symbol “+”.

Quorum sensing molecule (AI-2) secretion is affected in phoX, ppk1, and ppk2 deletion mutants’ biofilms

Here we measured the QS activity of the WT and poly P associated mutants in both shaking (planktonic; day 0) overnight cultures as well as standing (biofilm; day 1 to 3) cultures. We observed a similar pattern of AI-2 expression in the WT and in the ppk1 and ppk2 deletion mutants on day 2 and 3 of the biofilm cultures as well as in the day 0 shaking cultures. Further, the ΔphoX, Δppk1, and Δppk2 mutants all had increased AI-2 production in day 1 biofilms compared with the WT. AI-2 production decreased in day 2 biofilm compared with day 1 biofilm in all the mutants (Fig. 5). In general, there was no relationship between AI-2 production and biofilm phenotypes observed in the C. jejuni poly P associated mutants; therefore, the AI-2 assay was not performed in the presence of phosphate.

Figure 5. Measurement of AI-2 production using V. harveyi biolumenscence assay. The relative bioluminescence of AI-2 indicator strain BB170, compared with V. harveyi BB120 positive control, was measured from the supernatant of C. jejuni biofilm cultures in shaking (planktonic) (day 0) as well as on days 1, 2, and 3 standing (biofilm) cultures. Each column is the result of three separate experiments and represents the mean ± SE. The V. harveyi BB120 culture supernatant was used as positive control and sterile AB media was used as negative control. *, P < 0.05.

Expression of biofilm-associated genes is affected by age of biofilm and deletion of poly P associated genes

We measured changes in the transcription of genes previously shown to either enhance or inhibit biofilm formation in C. jejuni. There was a 5-fold increase in the expression of pglH in the ∆phoX mutant on day 2 biofilm compared with WT (Fig. 6). In ∆phoX 3 d biofilms, 3.3-fold higher expression of pglH was observed compared with WT. In the ∆ppk1 deletion mutant, we observed 3.2-fold higher induction of pglH in 2 d biofilms compared with WT. In the ∆ppk2 mutant, however, pglH expression was increased in both, day 2 (3.0-fold) and day 3 (3.5-fold) biofilms compared with WT. Only small transcriptional changes in kpsM were observed in all mutant strains’ biofilm (Fig. 6). However, in the ∆ppk1 mutant, the kspM expression was 3.0-fold decreased in day 3 biofilm.

Figure 6. Transcriptional changes in biofilm genes measured by qRT-PCR. The relative levels of expression of target genes were normalized using 16S rRNA gene as internal control for each strain and fold change for each target gene was presented relative to wild type control for the respective time point. The wild type expression was considered as 1 for statistical analysis. The difference in gene expression was determined by the threshold cycle (CT) method and a ± 2-fold change was considered significant. The assay was repeated three times with two replicates each time for each sample. The data represents the mean relative fold change in expression ± SE *, P < 0.05.

The ∆ppk1 mutant showed 7.6-fold increased cj0688 expression in day 2 biofilms, which diminished to WT levels in day 3 biofilms. The ∆ppk2 biofilms had cj0688 expression similar to WT on day 2, which increased to 3.6-fold in day 3 biofilm. Only the ∆ppk2 mutant had an increased expression of neuB1 on day 3 biofilms compared with WT. In all three mutants, the waaF expression was reduced in day 2 biofilm (−2.7 to −4.3-fold) and increased in day 3 biofilm (2.1- to 3.5-fold) compared with WT. The ∆ppk1 mutant 3 d biofilm had a significant decrease (−3.0-fold) in expression of fliS compared with WT. Additionally, the flagellar gene maf5 was upregulated in ∆ppk1 mutant 2 d biofilms (4.6-fold) compared with WT. Likewise, there was a smaller increase in maf5 in ∆ppk2 on 2 and 3 d biofilms (3.0-fold) compared with WT.

Discussion

Here we show that the deletion of poly P associated genes and phosphate not only affected the quantity of C. jejuni 81–176 biofilm formation; but also rate of growth, biofilm shedding, and CFW reactivity. Broadly, our observations suggest that although an increased biofilm phenotype was seen in the Δppk1, Δppk2, and ΔphoX mutants,^12-14 the resultant biofilm increase is achieved in different ways for each mutant. For instance, in the Δppk1 both the number of adherent microcolonies and air-liquid biofilm were increased. However, the ΔphoX mutant displayed increased total adherent biofilm, but the number of adherent microcolonies was comparable to WT. Similarly, the CFW reactivity was increased in the ΔphoX mutant by the addition of phosphate; however, was not affected by phosphate in the Δppk1 and Δppk2 mutants.

Our observations suggest that the optimal growth of adherent microcolonies and air-liquid biofilm is associated with poly P levels in the cell. This may be facilitated by altered expression of various genes associated with biofilm formation (Fig. 6). For example, the reduced expression of kpsM in the Δppk1 mutant in our study may contribute to increased formation of air-liquid biofilms in this mutant. Previously it was observed that in the C. jejuni kpsM mutant the biofilm formation was inhibited at later time points, but not in early stage (1 d), and this may suggest its role in biofilm maturation.²⁴ Similarly, all poly P-associated mutants exhibited a significantly decreased biofilm shedding on day 1; however, on day 3 we noticed a significant defect in biofilm shedding only in the ∆ppk2 mutant, which was not affected by the presence or absence of phosphate. Only the ∆ppk2 mutant had significantly increased expression of the LOS gene neuB1 in day 3 biofilm cultures (Fig. 6). This temporally coincides with the shedding phenotype seen in the ∆ppk2 mutant on day 3. From these results it can be interpreted that the presence of phosphate is only minimally important for biofilm shedding. As cellular shedding from mature biofilms has only been recently studied in C. jejuni; further studies are needed to understand the dynamics of shedding in the biofilm life-cycle.

Our observations provide insight into another facet of biofilm; the contribution of membrane and oligosaccharide modifications to the biofilm structure.^18,19 We found that all mutants were reduced in CFW fluorescence; however, only the ∆phoX CFW reactivity could be rescued by growth in the presence of phosphate. Both ∆spoT and ∆ppk1 mutants are diminished in poly P, but the CFW reactivity is increased in ∆spoT mutant but decreased in ∆ppk1 mutant.^10,18 This may suggests that there is not a direct connection between the diminished poly P levels and the membrane changes that induce CFW reactivity. Additionally, the waaF expression was affected in all three mutants. The waaF mutant displays increased CFW reactivity and biofilm formation in the later stages of biofilm growth.¹⁹ Both waaF and phoX are transcriptionally regulated by the PhosR/S. Further, phosR has been shown to be upregulated in ppk1 and ppk2 mutants^13,14 suggesting that the altered expression of waaF in these mutants is due to changes in phosR expression. In addition, we observed that aerobic stress rescued CFW reactivity in all mutants, suggesting that aerobic stress contributes to the formation of CFW-positive polysaccharides independent of poly P or phosphate availability.

The idea that QS is involved in different aspects of biofilm formation is not unique to C. jejuni. Quorum sensing has been shown to impact biofilm formation in E. coli, Pseudomonas, and Vibrio species.^26-28 We observed that poly P associated mutants, overall, had increased AI-2 expression in the early stages of biofilm formation (day 1) (Fig. 5); therefore, it is reasonable to speculate that this increased AI-2 production may contribute to increased overall biofilm formation. Further it is suggested that supernatant containing QS molecules from P. aeruginosa, commonly found to form heterogeneous biofilms with C. jejuni in the environment, may be sufficient to increase C. jejuni biofilms growth than mono-culture biofilms.^5,24 In conclusion, we hope that our study will advance the understanding of phosphate utilization in C. jejuni and perhaps suggest possible combinations of cellular responses to those modeling other phoX-containing bacteria.

Materials and Methods

Growth of C. jejuni and Vibrio strains

All bacterial strains used in this study are described in Table 1. C. jejuni 81–176 WT, ∆phoX, ∆ppk1, and ∆ppk2 mutant strains were cultured on Mueller-Hinton (MH) medium microaerobically at 42 °C for 24 h with or without 30 µg/mL kanamycin. Complemented strains (phoX+, ppk1+, and ppk2+) were grown on MH plates containing kanamycin (30 µg/mL) and chloramphenicol (20 µg/mL). The V. harveyi strains were grown in either LM growth media or Autoinducer Bioassay media (AB media) at 28 °C aerobically with shaking.²⁹

Table 1. Bacterial strains used in this study.

Strain	Relevant description	Source/Reference
C. jejuni 81–176 WT	Wild type strain of C. jejuni	Dr Qijing Zhang
∆phoX	C. jejuni 81–176 derivative with deletion in phoX gene; phoX::Kan	Drozd et al., 201112
phoX+	C. jejuni 81–176 phoX mutant complemented with WT copy of phoX with its RBS	Drozd et al., 201112
∆ppk1	C. jejuni 81–176 derivative with deletion in ppk1 gene; ppk1::Kan	Gangaiah et al., 200913
ppk1+	∆ppk1with pRY111 containing ppk1 coding region and the upstream promoter sequence for complementation::Cm	Gangaiah et al., 200913
∆ppk2	C. jejuni 81–176 derivative with deletion in ppk2 gene; ppk2::Kan	Gangaiah et al., 201014
ppk2+	∆ppk2 with pRY111 containing ppk2 coding region and the upstream promoter sequence for complementation::Cm	Gangaiah et al., 201014
Vibrio harveyi BB120	Sensor 1 (AI-1) and Sensor 2 (AI-2) positive strain	Bassler et al., 199729
Vibrio harveyi BB170	Sensor 1 (AI-1) negative and Sensor 2 (AI-2) positive strain	Bassler et al., 199729

Preparation of crystal violet (CV) slides

C. jejuni cultures grown overnight on MH agar were resuspended in MH broth, OD₆₀₀ was adjusted to 0.06 and 25 mL of each culture was transferred into a 50 mL conical tubes. Sterile water (control) or inorganic phosphate 100 mM was added to each tube to a 1 mM final concentration. Based on a previously described technique,¹⁶ sterile microscopic glass slides were placed in 50 mL conical tubes such that approximately 1/2 of the slide being submerged in culture (Fig. S6). Slides were incubated for 1 to 3 d microaerobically. After incubation, slides were gently removed, and stained in CV solution (1%) for 20 min. Slides were removed from CV solution; gently blotted upright on Kim wipes to remove excess CV, rinsed 3 times in dH₂O, and air-dried overnight in glass coplin staining jars. This procedure allows the quantification of adherent microcolonies as well as the attached biofilms (not pellicle) at the air liquid interface as previously described.¹⁶

Quantification of number of attached microcolonies

Attached biofilms on CV stained slides were quantified using bright field microscopy as described previously for C. jejuni and P. aeruginosa.^16,30 Crystal violet stained slides were visualized at 20× magnification and images were taken with an Axioplan2 microscope (Zeiss) equipped with AxioCamHR (Plan Neofluar 20×/0.50 Ph2 using a 23 FITC/Rhod reflector) and 48 bit RGB color, using the AxioVis40 V 4.6.3.0 software. Image size was 1388 × 1040 pixels with scaling of 0.32 μm/pixel (x- and y-axes) and 1 pixel/pixel (z-axis). Exposure length was optimized to 100% pixel intensity/slide. Adherent microcolonies were measured using the Automatic Measurement Program of the Zeiss software and microcolonies were measured in 5 different locations per slide (field of view) (Fig. S7). Measurements for each time point were averaged and standard error was calculated.

Quantification of air–liquid biofilm

Air liquid biofilms were quantified using a previously described technique with slight modification.^16,30 The bacteria were cultured on a glass slide in 50 mL conical tubes containing the OD₆₀₀ adjusted bacterial culture in MH media as described above. The attached air liquid biofilm was stained with CV as above and quantified using the caliper tool on bright field images taken with an Axioplan2 (Zeiss)/AxioCamHR on magnification 5× and 48 bit RGB color. No binning was used on these images, which was acquired using AxioVis40 V 4.6.3.0 software. Biofilm thickness was measured at least 5 times in different locations/slide. Measurements for each time point were averaged and standard error was calculated.

Shedding of viable C. jejuni cells from biofilm

C. jejuni cultures grown overnight on MH plates were resuspended in MH broth, OD₆₀₀ was adjusted to 0.06 and 2 mL of each culture were transferred into a 4 mL glass vial. Inorganic phosphate was added to a final concentration of 1 mM or sterile distilled H₂O (control). Each condition (presence or absence of phosphate) was performed in duplicate for each time point. Cultures were incubated for 0, 24, 48 and 72 h microaerobically at 42 °C. Shedding for viable C. jejuni from biofilm was assessed as described previously¹⁶ with minor modifications. Biofilms attaching to the glass surface were rinsed gently, once with 4 mL MH broth and then incubated for 4 h in fresh MH broth. Samples were 10-fold serially diluted, 100 µL of diluted sample was plated on MH agar, grown for 36 h microaerobically, and CFU/mL determined. Each sample was plated in duplicate and the experiments were performed 3 times independently for all conditions.

Calcofluor white (CFW) reactive polysaccharide detection

The CFW polysaccharide detection was performed as described earlier.¹⁸ MH plates containing CFW were prepared by adding 2% CFW to a final concentration of 0.02%. One millimolar phosphate was added as required. Overnight cultures for each strain were diluted to 0.05 OD₆₀₀ nm and 10 µL of each culture was spotted in duplicate on CFW plates. Plates were incubated microaerobically at 42 °C for 48 h. For aerobic experiments, the plates were incubated overnight microaerobically and then transferred to a 42 °C aerobic environment for 24 h. All CFW plate growth and incubations were performed in the dark. CFW reactivity was visualized under long-wave (365 nm) UV light. Uniform growth was confirmed by visualization under visible light. Experiment for each condition was performed a total of three times.

Quantification of quorum sensing molecule (AI-2) secretion using bioluminescent V. harveyi reporter

Cell-free culture (containing AI-2) supernatant was obtained from C. jejuni WT and poly P associated mutant strains grown for up to 3 d. Supernatant was centrifuged at 20 000 × g for 5 min, passed through a 0.2-μm filter and stored at −20 °C. C. jejuni culture supernatants were harvested under the following growth conditions: shaking (planktonic), 1, 2, and 3 d biofilms (static). Planktonic cultures were tested after 22 h of growth at 42 °C with shaking, which was determined to be the optimal for AI-2 expression in C. jejuni in previous studies.³¹ For static biofilm cultures: 3 mL of cultures were grown in borosilicate glass vials at 42 °C in duplicates and their supernatants pooled. Cultures were tested at 2 and 3 d of static biofilm growth.

V. harveyi BB170 reporter (sensor 1⁻, sensor 2⁺) was grown overnight at 28 °C, the culture was diluted 1:5000 in fresh AB media and 180 µL of diluted culture was added to each well containing 20 µL of C. jejuni cell-free culture supernatants in a 96-well black opaque culture plate (Corning Inc.). The negative control consisted of V. harveyi BB170+AB media; positive control was V. harveyi BB170+ supernatant from BB120 grown in AB media. All measurements were reported after 1.5 h incubation at 28 °C with shaking, when the difference between negative and positive controls reached maximal levels based on our preliminary time-course experiments. Bioluminescence was measured using the in vivo imaging system (IVIS) with the following setting; acquisition time was set to “auto” and maximum signal set to 50 000 at a constant focus range of B (10 cm). Mean luminescence/well was quantified. Each plate contained 8 replicates of each strain and each growth condition was repeated a minimum of 3 times. Quorum sensing signal was calculated as a % luminescence of the positive control BB120 strain (100 × (C. jejuni strain average luminescence / mm³) / (positive control average luminescence / mm³) for each well such that the control was compared with the experimental wells on the plate. The results were averaged and standard error was calculated.

qRT-PCR of biofilm associated genes

All the oligonucleotides used in the present study are listed in Table 2 and were synthesized at the Integrated DNA Technologies. The WT, ∆phoX, ∆ppk1, and ∆ppk2 strains were examined for changes in biofilm related genes in day 2 and 3 biofilms. Day 1 biofilms were not analyzed as biofilm growth was not sufficient for RNA extraction. Each culture was prepared in MH to an initial OD₆₀₀ of 0.1 and was incubated at 42 °C microaerobically. RNA extraction, cDNA synthesis, and quantitative RT-PCR (qRT-PCR) was performed as described.^12-14 qRT-PCR was performed targeting genes involved in biofilm formation (Table 2).^6,19 Gene specific primers were designed using Beacon Designer 7.0. The 16S rRNA gene was used as an internal control.³² The relative levels of expression of target genes was normalized with 16SrRNA amplified from same sample and fold change in expression of target genes was presented relative to the expression levels in the WT for the respective time points. The difference in expression of the genes was calculated using the comparative threshold cycle (CT)^12-14 and a ±2-fold cut off was considered significant. For statistical analysis, expression of genes in WT C. jejuni was normalized to 1. The qRT-PCR was performed in duplicates and assay was performed a total of three times.

Table 2. Oligonucleotides used for quantitative RT-PCR.

Name	Description	Sequence
16sRNA F 16sRNA R	rRNA gene	GTCTCTTGTG AAATCTAATG GTATTCTTGG TGATATCTAC
kpsM F kpsM R	Capsular polysaccharide synthesis	CCCTAAAGCA AAAGCTGAGC TTTGCCTATA AACCTGTAAA ACCTATAC
waaF F waaF R	Heptosyltransferase	ATCACAAATG ACAGTGGACC T GCCAAGGTGA AGTTTGAGTA AAT
pglH F pglH R	Protein glycosylation	CCTTGACATT TTCAATGCGT CC AAACCCTTGT CATTTTAGCG ATG
neuB1 F neuB1 R	Lipo-oligosaccharide	GTTTCAACGG GCATTGCTAC TCCAAGTGCT ACTGCCATAA C
fliS F fliS R	Flagella protein synthesis	TGCTTTATGA GGGAATTTTG CG GAATTTCTCT TGTATAAAGC CCGC
Cj0688 F Cj0688 R	Phosphate acetyltransferase	GCAGTTGATT AAGCGTAGCA C AAACAAAACG CCACAAGACG
maf5 F maf5 R	Flagella formation	GCTAGACATC TACCCTTTGC TC CTTTCAACCT CTCCTTCTCC G

Statistical analysis

Statistical significance of data was determined using Student t test (unpaired, two-tailed). Mutants were compared with the WT results with concurrent day and phosphate treatment (For example; phoX with phosphate was compared with WT with phosphate; phoX with no phosphate compared with WT with no phosphate). In addition, each strain was also compared in the presence and absence of phosphate within itself (for example, phoX with phosphate was compared with phoX without phosphate). A P value of ≤ 0.05 was considered significant.

Glossary

Abbreviations:

poly P

polyphosphate

Ppk

polyphosphate kinase

PhoX

alkaline phosphatase

AI-2

autoinducer-2

CFW

calcofluor white

QS

quorum sensing

CV

crystal violet