Phenotypes of Campylobacter jejuni luxS Mutants Are Depending on Strain Background, Kind of Mutation and Experimental Conditions

Since the discovery that Campylobacter (C.) jejuni produces Autoinducer 2 (AI-2), various studies have been conducted to explore the function and role of AI-2 in C. jejuni. However, the interpretation of these analyses has been complicated by differences in strain backgrounds, kind of mutation and culture conditions used. Furthermore, all research on AI-2 dependent phenotypes has been conducted with AI-2 synthase (luxS) mutants. This mutation also leads to a disruption of the activated-methyl-cycle. Most studies lack sufficient complementation resulting in not knowing whether phenotypes of luxS mutants depend on disrupted metabolism or lack of AI-2. Additionally, no AI-2 receptor has been found yet. All this contributes to an intensive discussion about the exact role of AI-2 in C. jejuni. Therefore, we examined the impact of different experiment settings on three different C. jejuni luxS mutants on growth and motility (37°C and 42°C). Our study showed that differing phenotypes of C. jejuni luxS mutants depend on strain background, mutation strategy and culture conditions. Furthermore, we complemented experiments with synthetic AI-2 or homocysteine as well as the combination of both. Complementation with AI-2 and AI-2+homocysteine significantly increased the cell number of C. jejuni NCTC 11168ΔluxS in stationary phase compared to the non-complemented C. jejuni NCTC 11168ΔluxS mutant. Genetic complementation of both C. jejuni 81-176 luxS mutants resulted in wild type comparable growth curves. Also swarming ability could be partially complemented. While genetic complementation restored swarming abilities of C. jejuni 81-176ΔluxS, it did not fully restore the phenotype of C. jejuni 81-176::luxS, which indicates that compensatory mutations in other parts of the chromosome and/or potential polar effects may appear in this mutant strain. Also with neither synthetic complementation, the phenotype of the wild type-strains was achieved, suggesting yet another reason for differing phenotypes other than communication and methionine metabolism for C. jejuni luxS mutants.


Introduction
Numerous bacteria communicate via the small interspeciesspecific signalling molecule autoinducer-2 (AI-2) generated via LuxS [1]. This process is commonly known as Quorum sensing (QS). QS is a regulatory mechanism of gene expression, which enables bacteria to change their behaviour when the population reaches a particular cell-density. QS allows bacteria to communicate with each other and therefore coordinate their activities at a multicellular level. QS regulated processes are for example secretion of virulence factors, biofilm formation, motility and bioluminescence [2][3][4].
AI-2 is generated as a by-product via LuxS during the activated methyl cycle (AMC) [5,6]. The AMC is an important metabolic pathway in cells. The starting compound is S-adenosyl-methionine (SAM), which is the general methyl donor. It donates its methyl group to diverse cellular components such as DNA, RNA and proteins. SAM is thereby converted to S-adenosyl-homocysteine (SAH), which is a toxic compound and has to be recycled. For recycling of SAH, two different pathways are known so far: a onestep and a two-step pathway. Only in the two-step pathway AI-2 is produced. In the two-step pathway Pfs (59methylthioadenosine/Sadenosyl-homocysteine nucleosidase) hydrolyzes SAH to S-ribosylhomoscysteine (SRH) and adenine. LuxS catalyzes the cleavage of SRH to 4,5-dihydroxyl-2,3-pentanedion (DPD) and homocysteine [7,8]. DPD is spontaneously cyclized into AI-2, while homocysteine is converted by MetE or MetH to methionine. Methionine is then converted by MetK into SAM [9].
In V. harveyi, AI-2 binds to the periplasmic binding protein LuxP. In many other bacteria e.g. Salmonella and Escherichia coli, AI-2 binds to LsrB, the ligand binding protein of an ABC transporter. So far, no homologues of the known AI-2 receptors like LuxP or LsrB were identified in Campylobacter spp. [10,11].
Recently, Rader et al. [12] described that the chemoreceptor TlpB functiones as AI-2 receptor in Helicobacter pylori. Despite of the existence of chemoreceptors in C. jejuni, which would suggest the existence of a corresponding receptor, no TlpB receptor homolog has been found yet.
The existence of LuxS, as well as the LuxS-dependent AI-2 production in C. jejuni NCTC 11168, was first described by Elvers and Park [13]. The fact that AI-2 is a by-product of the AMC and that a receptor is yet to be found, leads to the question, if AI-2 in C. jejuni is indeed a true QS signal molecule. The disruption of luxS could lead to changing phenotypes due to the absence of AI-2 or disrupted methionine cycle. Thus, experimental analysis with luxS mutants needs to be complemented with AI-2 and/or a metabolic replacement substance like homocysteine (HC). Several studies of C. jejuni luxS mutants showed various results with diverse phenotypes in luxS mutants. For instance, motility and growth seems to be influenced through luxS disruption but in slightly different ways depending on the study design and conditions [13][14][15].
These sometimes opposing phenotypes might be due to different culture conditions. Furthermore, the authors conducted their studies with luxS mutants of different C. jejuni strains and used different mutation strategies. Additionally, most studies lack proof of complementing the luxS mutant strains with AI-2 and/or a metabolic substance to confirm whether resulting phenotypes are due to metabolic function of LuxS or a consequence of disrupting cell communication.
Therefore, we examined the impact of strain background, mutation strategy and culture condition on three different C. jejuni luxS mutants on growth and motility. Furthermore complementation experiments with synthetic AI-2 and/or homocysteine (HC) were conducted.
The mutation of luxS was genetically confirmed. Therefore all luxS mutants were verified by PCR, followed by DNA sequencing of the amplified products. Additionally, the absence of AI-2 activity of all luxS mutants, as well as the presence of AI-2 in genetically complemented mutants was routinely tested in V. harveyi bioluminescence assay [16].

V. harveyi bioluminescence assay
Overnight cultures of C. jejuni and V. harveyi BB152 (positive control) were diluted in BB or AB to a cell density of 1610 8 CFU/ ml. Culture supernatants were collected and centrifuged at 80006 g for 10 min. The supernatants were sterilized by passing through a 22 mm filter (VWR, Darmstadt, Germany) and stored at 220uC until used. In parallel the absorbance was measured at the same time point to determine cell growth.
The V. harveyi autoinducer assay was performed as described previously [16]. The reporter strain (BB170) was grown over night in AB medium and diluted (1:5000) into fresh AB medium. CFS and uninoculated AB respectively BB medium were then added to the diluted V. harveyi culture at 10% (v/v) final concentration. As another positive control AI-2 (10 mM) alone was tested. The reporter strain with CFS, AI-2 or uninoculated media were incubated at 30uC with aeration (750 rpm). After 4 hours of incubation, luminescence of 100 ml aliquots in microtiter plates were measured (10 s per well) using Luminometer (CentroPro, Berthold, Bad Wildbach). For each of three experiments, triplicates of relative light units (RLU) were measured. n-fold luminescence induction values were calculated from RLU obtained with conditioned CFS vs. RLU obtained with sterile medium.

Chemical complementation
AI-2 activity was quantified with the bioluminescence assay and compared to wild-type C. jejuni grown to an OD600 nm of 1.0, at which maximal AI-2 activity was obtained. To test for complementation of growth and motility, AI-2 (OMM Scientific, Dallas, USA) at a physiological concentration of 10 mM and non-limiting concentration of 100 mM was used. Homocysteine (HC; Sigma Aldrich, St. Louis, USA) was tested at 1 mM, 10 mM and 100 mM.
To exclude non-specific effects of AI-2 and homocysteine, the same concentrations were added to wt strains.

Genetic complementation
To exclude potential polar effects in the mutant strains, genetic complementation was performed exemplarily in the insertion and deletion mutant of strain C. jejuni 81-176. The complemented C. jejuni 81-176::luxS strain was kindly provided by Quiñ ones et al. [15].

Growth assay
For growth assays, cultures of wild type (wt) and luxS mutants of C. jejuni NCTC 11168 and C. jejuni 81-176 strains were grown overnight in BB or MH at 37uC. Precultures were inoculated in BB or MH to approx. 2610 5 CFU/ml and incubated under microaerobic conditions at 37uC and 42uC. For chemical complementation assays 10 mM AI-2, HC or AI-2+HC were added to the cultures.
Numbers of viable bacteria were determined over 48 h by plating serial dilutions of the bacterial suspensions. Results reported are the average of at least three independent assays.

Swarming assay
For swarming either BB containing 0.4% agar (BBA) or MH containing 0.4% agar (MHA) were used. The swarming ability of C. jejuni luxS mutants was investigated at 37uC and 42uC on swarming plates. For chemical complementation 10 mM each of AI-2, HC or AI-2+HC were added to the molten agar. Overnight cultures of C. jejuni strains were adjusted to 10 8 CFU/ml and 1 ml dropped on BBA or MHA. After 24 h incubation at 37uC or 42uC the diameters of the swarming halos were measured. Halos of luxS mutants were normalized to the wild type halos (100%). Results reported are the median of six independent assays.

Statistical analysis
For statistical analyses, all experiments were repeated at least three times in three independent experiments.
Statistical analyses were performed using GraphPad Prism v6.0 (GraphPad Prism, San Diego, USA). To calculate significant differences a two-tailed Mann-Whitney test was used. For all statistical analyses, a confidence level of 95% was defined.

Growth assays
Strain background. Growth of both C. jejuni wild type strains did not differ at 37uC and 42uC, whereas growth profiles of the three luxS mutants were not equal ( Fig. 1A-F). Compared to the wild type the DluxS mutant of C. jejuni NCTC 11168 showed significantly reduced cell numbers within mid-exponential (8 h) and mid-stationary phase (32 h) while cell numbers converged during late stationary phase (48 h) at both temperatures ( Fig. 1A-B). C. jejuni 81-176DluxS and::luxS mutants showed comparable cell numbers to the wild type at 37uC with a slight decrease at late stationary phase ( Fig. 1C/E). At 42uC C. jejuni 81-176DluxS showed significantly decreased cell numbers in exponential and stationary phase, but C. jejuni NCTC 11168DluxS showed much lower levels (Fig. 1D). However, the cell number of C. jejuni 81-176::luxS only decreased significantly at mid-stationary phase at 42uC (Fig. 1F). These results demonstrate that strain background has an impact on growth profiles of C. jejuni luxS mutants.
Culture conditions. When comparing growth of C. jejuni NCTC 11168 wt and mutant in MH versus BB it becomes obvious that the growth profiles were quite similar between these media at both temperatures ( Fig. 1A-B and 2A-B). In contrast, growth of C. jejuni 81-176 wt was slightly reduced in late stationary phase in MH. Both mutants of strain C. jejuni 81-176 showed growth defects in MH at 37uC that were not exhibited in BB ( Fig. 2C-F). In addition, growth profiles of C. jejuni 81-176DluxS mutants differed between MH and BB, thus the cell numbers of C. jejuni 81-176DluxS in MH were substantially reduced in the stationary phase while cell numbers in BB showed only a partially slight reduction. This indicates that composition of media could influences growth of luxS mutants. Temperature influences growth profiles of the C. jejuni 81-176::luxS mutant strain in MH. However, temperature did not have an impact on growth of the other C. jejuni luxS mutants.
Impact of mutation strategy on growth of luxS mutants. To investigate the influence of mutation strategy, three different C. jejuni luxS mutants were used. Considering growth profiles of C. jejuni in MH the influences of mutation strategy became apparent. The DluxS mutants of both C. jejuni NCTC 11168 and C. jejuni 81-176 showed growth defects which were not exhibited in the insertional mutant C. jejuni 81-176::luxS at 42uC (Fig. 2D/F). These results demonstrated the impact of mutation strategy.
Genetic complementation. Based on the varying growth profiles between C. jejuni 81-176 wt and its luxS mutant strains in MH, genetic complementation of the C. jejuni 81-176 mutant strains were conducted. Genetic complementation of both C. jejuni 81-176 luxS mutants in MH resulted in wild type comparable growth curves (Fig. 2). Introduction of the isogenic plasmid pWM1015 alone did not alter growth of C. jejuni 81-176 mutants. This indicates that no polar effects caused varying growth profiles of C. jejuni 81-176 luxS mutants in MH.
Chemical complementation. Because of the varying growth profiles between C. jejuni NCTC 11168 wt and its DluxS mutant we examined if the addition of exogenous AI-2 and HC influences growth of C. jejuni NCTC 11168DluxS in BB. Complementation with AI-2 (10 mM) and AI-2+HC (both 10 mM) significantly increased the cell number of C. jejuni NCTC 11168DluxS in stationary phase at 37uC and 42uC compared to the noncomplemented DluxS mutant ( Fig. 3A/B). In contrast, HC alone did not show any significant effect on cell numbers at any of the investigated temperatures (data not shown). Full restoration of wild type cell numbers was not achieved by the chemical complementation strategy used in luxS mutants indicating that alternative mechanisms for these different phenotypes exist. The same effects have been observed with 100 mM AI-2 and 100 mM HC (data not shown). Non-specific effects through AI-2 and homocysteine could be excluded since no alteration in phenotypes of the wild type was observed (data not shown).
With the addition of AI-2 and AI-2+HC to strain C. jejuni 81-176DluxS in MH media slightly increased cell numbers during stationary phase were observed at both temperatures, but these were not statistically significant (Fig. 3C/D). The addition of HC alone did not lead to increased cell numbers (data not shown).

Swarming ability
Strain background. Swarming ability of C. jejuni wild types did not differ between the different strains (data not shown). Swarming ability of luxS mutants were normalized to the wild type (100%). C. jejuni NCTC 11168DluxS showed reduced swarming ability ( Fig. 4A/D, Fig. 5A) (approx. 42% of wt swarming in BBA). In contrast, C. jejuni 81-176DluxS showed no significant   Chemical complementation in BB. Hence, we examined if the addition of exogenous AI-2 and HC influences the reduced swarming of C. jejuni luxS mutants on BBA at 37uC and 42uC (Fig. 6). In BBA complementation with AI-2 and AI-2+HC contributed to an increased swarming ability compared to the non-complemented mutant of C. jejuni NCTC 11168DluxS at 37uC (Fig. 6) but not at 42uC (Fig. 6A/D). However, the addition of HC alone did not alter the swarming ability at both temperatures. Only the addition of both AI-2+HC to C. jejuni 81-176::luxS mutant increased the swarming motility at 37uC (Fig. 6C), while the swarming ability of C. jejuni 81-176DluxS was not significantly altered by any condition investigated (Fig. 6B/F). Our data implicate that partial complementation of luxS mutants is possible in BBA depending on temperature and strain background. Complementation only occurs if AI-2 is admitted. With neither chemical complementation of luxS mutants, the phenotype of the wild type-strains was achieved.
Chemically complementation in MH. The swarming ability of C. jejuni NCTC 11168DluxS was increased through the addition of AI-2, HC and AI-2+HC at 42uC compared to the non-complemented mutant strain on MHA (Fig. 7). At 37uC only a slightly increased swarming ability could be observed through the addition of AI-2 and AI-2+HC. However, the addition of AI-2 to C. jejuni 81-176::luxS yielded increased swarming motility at both temperatures. Furthermore, with the addition of AI-2+HC swarming motility could also be increased at 37uC in this mutant. In contrast, the swarming ability of C. jejuni 81-176DluxS was not significantly changed by any complementation investigated. With neither chemical complementation, the phenotype of the wild type-strains was achieved. Complementation in MHA is likely to occur but depends on strain background and mutation strategy.

Discussion
Since the discovery that C. jejuni produces AI-2, various studies have been conducted to explore the function and role of AI-2 in C. jejuni [18,19]. However, the interpretation of these analyses has been complicated by differences in strain background, kind of mutation and culture conditions. Furthermore most studies lack sufficient complementation resulting in not knowing whether phenotypes of luxS mutants depend on disrupted metabolism or lack of AI-2. Additionally, no AI-2 receptor has been found yet. All this contributes to an intensive discussion about the exact role of AI-2 in C. jejuni.
In the literature, various motility and growth phenotypes have been described for C. jejuni luxS mutants [14,15,20]. Therefore, we investigated if the strain background, kind of mutation and different culture conditions impact the occurring phenotypes in C. jejuni luxS mutants.
To verify that our luxS-mutant strain truly deficient in AI-2 production a V. harveyi bioluminescence reporter assay was conducted. V. harveyi only responds to the borate diester derived from (2S,4S)-THMF [9]. The wild type strain of C. jejuni as well as the synthetic AI-2 exhibit positive signals in this reporter assay, indicating that the absence of a positive signal (in the reporter assay) of the mutant strain is equatable to the absence of AI-2 production. Further the V. harveyi reporter assay indicates that the synthetic AI-2 contains a similar equilibrium of AI-2 as the one produced by C. jejuni NCTC 11168.

Culture conditions
During growth of C. jejuni wild types there were no significant differences observed in BB and MH. Also Ng et al. [21] described that there was no significant difference between cell numbers of C. jejuni among these basal media. Growth and swarming abilities of C. jejuni NCTC 11168DluxS were quite similar in BB and MH at both temperatures (37uC and 42uC), which leads to the assumption that these culture conditions do not have an impact on the luxS mutant phenotype of this strain.
Further, He et al. [14] described that the cell numbers of C. jejuni 81-176 differed between wild type and DluxS mutant at 37uC and 42uC in the mid-exponential phase, while cell numbers  converged in late stationary-phase in MH medium. In our study, we also observed reduced cell numbers of C. jejuni 81-176DluxS in exponential as well as in late stationary phase in MH at both temperatures. In contrast, growth of C. jejuni 81-176DluxS in BB only showed significant differences to growth of wild type in late stationary phase at 37uC. These results demonstrate that the choice of the culture medium has a large impact on the resulting phenotypes of C. jejuni 81-176 luxS mutants. Growth of the C. jejuni 81-176::luxS is also influenced by temperature in MH. However, growth curves at 37uC and 42uC are quite equal in C. jejuni 11168DluxS and C. jejuni 81-176DluxS, which indicate that growth differences are independent of temperature but dependent on culture media for deletion mutant strains. Motility observation confirmed this assumption as well. Like He et al. [14] we observed reduced swarming abilities of C. jejuni 81-176DluxS at 37uC on MHA media in contrast to the wild type. However, this phenotype was only observed on MHA media but not on BBA. The components of these two different basal mediums differ. For example only BB contains dextrose. Wang et al. [22] showed that glucose affected the gene expression of luxS in E. coli. Furthermore, they observed that the expression of pfs was reduced by the presence of glucose. Our findings clearly illustrate that resulting phenotypes of C. jejuni luxS mutants can be influenced by the choice of culture medium and components in media could also influence resulting phenotypes of C. jejuni luxS mutants.

Strain background
C. jejuni is a highly diverse species [23]. To investigate the influence of strain background on phenotypes of luxS mutants we used C. jejuni NCTC 11168 and C. jejuni 81-176DluxS mutants. Comparing these phenotypes we observed differences between luxS mutants of C. jejuni NCTC 11168 and C. jejuni 81-176 especially in BB (Fig. 1A-D). Also, the comparison of luxS mutant strains of C. jejuni 81-176 used by He et al. [14] and C. jejuni NCTC 11168 used by Elvers and Park [13] revealed different outcomes in terms of growth. Comparing wild types of C. jejuni in our study, cell numbers of strain C. jejuni NCTC 11168 decline under cell numbers of strain C. jejuni 81-176 in late stationary phase. Cell numbers in all other time points did not differ between these two wild type strains, whereas cell numbers of C. jejuni NCTC 11168DluxS in early stationary phase in BB are lower than cell numbers of C. jejuni 81-176DluxS. Furthermore, the ability to swarm differed between the luxS mutants of both strains. Here, we observed that C. jejuni NCTC 11168DluxS exhibits the greatest reduction of swarming ability, whereas C. jejuni 81-176DluxS did not exhibit reduced swarming abilities at all in BB. The different observed phenotypes in C. jejuni NCTC 11168 and C. jejuni 81-176 may be a consequence of genetic diversity between these strains [24]. For instance previous analysis of the complete flagellin glycosylation locus of C. jejuni strain 81-176 revealed a less complex genomic organization than the corresponding region in the genome of strain C. jejuni NCTC 11168 [25]. In addition, Dugar et al. [17] identified strain-specific transcriptome organization and sRNAs that could contribute to differential gene regulation among these strains.

Mutation strategy
An explanation for the differences in growth profiles between C. jejuni NCTC 11168 and C. jejuni 81-176 luxS mutants might be the genetic differences between these two strains. However, it seems equally probable that the phenotypic differences observed were due to different mutation strategies applied. Previously Haigh et al. [26] showed that mutation design and strain background influenced phenotype of E. coli luxS mutants. The authors concluded that one explanation could be the different orientation of antibiotic resistance cassette in the mutants. The kanamycin resistance cassette of C. jejuni NCTC 11168DluxS is orientated in the same direction as the luxS gene, whereas the C. jejuni 81-176DluxS mutant has the chloramphenicol resistance cassette in the opposite direction of the deleted luxS gene. Both mutants had reduced cell numbers when cultured in MH but only the luxS mutant of C. jejuni NCTC 11168 showed lesser swarming abilities compared to the corresponding wild type. Recently, it has been demonstrated that the regulatory small RNA MicA is located closely upstream of the E. coli luxS gene [27,28] and could be an obvious target for polar effects of a luxS mutation. In C. jejuni, Dugar et al. [17] showed that a small RNA is located downstream of the luxS gene, but no function of this small RNA has been described so far. However, the expression of this small RNA could be influenced in a mutation strategy dependent manner. Since this molecule can have regulatory effects, its dysregulation can affect other pathways. Another reason for the observed different phenotypes could be the size of the deleted region. In contrast to our results, the C. jejuni NCTC 11168DluxS-mutant constructed by Elvers and Park [13] showed similar growth compared to the wt at 37uC. The DluxS mutant of Elvers and Park [13] has the same strain background and orientation of antibiotic resistance cassette as the DluxS mutant used in this study. However, the mutant used in our study has a larger deletion region, including those with functional domains from luxS (Fig. 8 illustrated the differences in mutation strategies of C. jejuni luxS mutants), whereas the mutant from Elvers and Park [13] still retains the functional domain regions. Even though the functionality relating to AI-2 production is disabled, it cannot be ruled out that other regions within this sequence exert an influence on other processes. Also the C. jejuni NCTC 11168::luxS mutant described by Plummer et al. [20] (an insertion mutant up-stream of the functional domains) showed similar growth like the wt. However, the mutant of Plummer et al. [20] and Elvers and Park [13] exhibited decreased motility. Examining the mutant in the current study, decreased motility haloes in semisolid media have been observed, which indicates that disruption of luxS causes a reduction of swarming ability in C. jejuni NCTC 11168 whether or not the functional domain regions are deleted.
To investigate the influence of mutation strategy on phenotypes of C. jejuni luxS mutants we additionally conducted our study with an insertion mutant of C. jejuni 81-176. The resistance cassette of this mutant strain is also (like in C. jejuni 81-176DluxS) orientated in the opposite direction of the luxS gene. By examining growth in MH at 37uC and 42uC cell numbers of the deletion mutant are smaller than cell numbers of insertion mutant in stationary phase. Again, one reason could be the lack of the region among the functional domains within luxS in the deletion mutant, whereas this region is still present in the insertion mutant of C. jejuni 81-176. Additionally, Quinones et al. [15] replaced the luxS gene of strain 81-176 by a mutated luxS gene of the C. jejuni strain RM1221, which is another possible explanation for differing phenotypes of luxS mutants. Furthermore, Quiñ ones et al. [15] used larger up-and downstream regions of luxS for their mutation construct. Even though there is a high DNA sequence identity (96.8%) between the luxS genes of strains RM1221 and 81-176, there are some differences which might influence phenotypes of luxS mutants.
The observation of swarming ability additionally indicates the importance of mutation strategy. Like Quiñ ones et al. [15] we observed a significantly reduced swarming ability of C. jejuni 81-176::luxS mutant compared to the wild type on BBA, whereas the swarming ability of C. jejuni 81-176DluxS is not reduced on BBA in contrast to the wt.
While genetic complementation restored swarming abilities of C. jejuni 81-176DluxS, genetic complementation restored swarming abilities of C. jejuni 81-176::luxS only partial. The phenotype of wt strain C. jejuni 81-176 was not fully achieved. The incomplete restoration was probably caused by polar effects in this insertion mutant strain. Also Haigh et al. [26] argued that insertion may always result in adverse polar effects.
Combined, these findings suggest that growth deficits of C. jejuni luxS mutants might be associated with deletion of a larger region of luxS, while motility might be influenced by polar or compensatory mutation effects of the luxS mutation.

Complementation
It remains unclear whether phenotypes of luxS mutants are due to the lack of AI-2 or result from the metabolic deficits caused by disruption of this enzyme in the activated methyl cycle. Additionally, occurring phenotypes may be appearing through polar effects depending on the kind of mutation. Genetic complementation of both C. jejuni 81-176 luxS mutants resulted in wild type comparable growth curves, which indicates that no polar effects influence growth of C. jejuni 81-176 luxS mutants.
Therefore, we investigated the phenotypes of C. jejuni luxS mutants chemically complemented with AI-2, HC and AI-2+HC (Fig. 3). Growth of luxS mutants could be partially complemented by AI-2 but not to wt level, implicating that altered phenotypes of luxS mutants not solely occur as a consequence of lacking AI-2. The addition of AI-2 and AI-2+HC to mutant strain C. jejuni 81-176DluxS leads to slightly increased cell numbers, too. Addition of HC to the mutant strains did not alter the luxS mutant phenotype indicating that disruption of the activated methyl cycle downstream of LuxS might not be responsible for the observed phenotypes (data not shown). The accumulation of components upstream of LuxS within the AMC as well as other unknown functions of LuxS could also have an impact on the observed phenotypes.
A peculiarity of AI-2 signaling is that diverse bacteria have different AI-2 receptors which recognize distinct forms of AI-2. For example, V. harveyi responds to the borate diester derived from (2S,4S)-THMF [30,31], whereas Salmonella Typhimurium [9], Sinorhizobium meliloti [32] and Yersinia pestis [33] respond to (2R,4S)-THMF. Thereby it is possible that the DPD used in this study might not have harboured adequate amounts of the relevant DPD variant for C. jejuni. However, as the synthetic DPD also induced luminescence in the V. harveyi assay, it can be concluded that AI-2 produced by C. jejuni and synthetic DPD contained a similar variation of AI-2. The response observed following  Table 2. Phenotypes (growth and swarming ability) of luxS mutants and its chemical complementation. chemical complementation would suggest the existence of the adequate structure of AI-2 for receptor recognition. Nevertheless it remains unclear if the lack of complete complementation could be caused by an inappropriate chemically equilibrium of the AI-2 structures. We observed significantly increased swarming ability at 37uC by chemical complementation with AI-2 and AI-2+HC in C. jejuni NCTC 11168DluxS on BBA and in C. jejuni 81-176::luxS on MHA. While genetic complementation restored swarming abilities of C. jejuni 81-176DluxS, it did not completely restore phenotype of C. jejuni 81-176::luxS. By the addition of AI-2, the same swarming abilitiy as shown for the genetic complemented mutant was achieved. These data indicate that reduced swarming abilities are partially due to polar mutation effects in C. jejuni 81-176::luxS, but could be partially complemented by exogenous AI-2.

Conclusions
Our study provides a clue why literature about phenotypes of C. jejuni luxS mutants is extremely contradictory. Our analyses demonstrated that occurring phenotypes of C. jejuni luxS mutants are depending on strain background, kind of mutations and experimental conditions (Table 2).
Further, some luxS mutant phenotypes could be partially complemented by AI-2, even though not to wild type levels, suggesting that C. jejuni can regulate its behaviour by AI-2 dependent Quorum sensing. Further studies should clarify which kind of AI-2 structure is recognized by C. jejuni.
Future studies could also clarify how the mutation strategy influences gene expression. Variability in the different mutants examined in this study reflects the likely presence of compensatory mutations in other parts of the chromosome. Also polar effects on down-and upstream genes are possible. Further, the influence of mutation strategy on small RNAs cannot be excluded.
One option for further research could be to investigate the expression of the small RNA located downstream of luxS in the settings used in this study.