Motility defects in Campylobacter jejuni defined gene deletion mutants caused by second-site mutations

Genetic variation due to mutation and phase variation has a considerable impact on the commensal and pathogenic behaviours of Campylobacter jejuni. In this study, we provide an example of how second-site mutations can interfere with gene function analysis in C. jejuni. Deletion of the flagellin B gene (flaB) in C. jejuni M1 resulted in mutant clones with inconsistent motility phenotypes. From the flaB mutant clones picked for further analysis, two were motile, one showed intermediate motility and two displayed severely attenuated motility. To determine the molecular basis of this differential motility, a genome resequencing approach was used. Second-site mutations were identified in the severely attenuated and intermediate motility flaB mutant clones: a TA-dinucleotide deletion in fliW and an A deletion in flgD, respectively. Restoration of WT fliW, using a newly developed genetic complementation system, confirmed that the second-site fliW mutation caused the motility defect as opposed to the primary deletion of flaB. This study highlights the importance of (i) screening multiple defined gene deletion mutant clones, (ii) genetic complementation of the gene deletion and ideally (iii) screening for second-site mutations that might interfere with the pathways/mechanisms under study.


INTRODUCTION
Campylobacter jejuni is the leading bacterial cause of foodborne gastroenteritis worldwide. The impact of C. jejuni infections is significant due to their high incidence, duration of infection and possible post-infection sequelae (Ruiz-Palacios, 2007). C. jejuni has a broad range of environmental reservoirs including water, birds and other domestic animals, with chickens representing the largest source of human infection (Young et al., 2007). Although C. jejuni is considered as commensal in chickens, it has a significant impact on animal welfare in certain breeds of bird (Humphrey et al., 2014). The genome of C. jejuni is subject to considerable genetic variation, which is thought to play an important role in the survival of the species within host organisms and the environment, especially by altering surface exposed structures such as flagella, lipooligosaccharide and capsular polysaccharide (Balaban et al., 2009;Jerome et al., 2011;Mohawk et al., 2014;Thomas et al., 2014). Genetic heterogeneity arises in the C. jejuni population as a consequence of mutations due to replication errors, phase variation of homopolymeric or repetitive heteropolymeric tracts either in coding sequences or in promoter regions and to a lesser extent, recombination (Thomas et al., 2014;Wassenaar, 2011;Wilson et al., 2009). It has been suggested that the lack of a functional mismatch repair system enhances the frequency of phase variation and random mutations, and thus the overall heterogeneity of the C. jejuni population (Gaasbeek et al., 2009). Flagellar-mediated motility plays a central role in commensal and pathogenic behaviours of C. jejuni (Guerry, 2007;Lertsethtakarn et al., 2011). The flagellar system is subject to genetic variation at multiple levels, which often leads to alteration of motility (Gao et al., 2014;Guerry, 2007;Hendrixson & DiRita, 2004). C. jejuni possesses polar flagella at one or both ends of the bacterial cell which consist of a membrane-embedded basal body, a hook structure and a protruding polymeric filament that is composed of the major flagellin subunit FlaA and the minor subunit FlaB (Gao et al., 2014;Guerry et al., 1991). In C. jejuni strain 81116, inactivation of flaA led to a non-flagellated and non-motile phenotype, whereas flagella and motility were retained after the inactivation of flaB (Nuijten et al., 1990;Wassenaar et al., 1991). The motor subunits at the flagellar base, encoded by motA and motB, are responsible for flagellar rotation. Mohawk et al. (2014) reported that non-motile variants of C. jejuni 81-176 carried distinct mutations [i.e. single nucleotide polymorphisms (SNPs) and insertions and deletions (INDELs)] in the motA gene. Flagellar gene expression is tightly controlled and requires the alternative d 54 and d 28 factors, the FlgSR twocomponent system, the FlhF GTPase and the flagellar secretion system (Balaban et al., 2009;Hendrixson & DiRita, 2003;. Phase variation mechanisms acting on the FlgSR system that affect motility in C. jejuni 81-176 include reversible phase variation in (i) homopolymeric poly-A and poly-T tracts within the flgR response regulator (Hendrixson, 2006) and (ii) poly-A tracts and heteropolymeric repeats located in the flgS sensor histidine kinase (Hendrixson, 2008). It is noteworthy that the C. jejuni flagellar filament is extensively glycosylated, predominantly with pseudaminic acid and legionaminic acid and derivatives thereof. Recombination within the flagellar glycosylation gene locus and homopolymeric poly-G tract phase variation are mechanisms of variation (Howard et al., 2009;Karlyshev et al., 2002;van Alphen et al., 2008).
Here we show that the propensity for stochastic mutation events in C. jejuni has clear implications for gene function analysis in an experimental setting. Analysis of C. jejuni M1 flaB mutant clones revealed inconsistent motility phenotypes. Using genome resequencing and in-depth genotypic and phenotypic characterization, we identified second-site mutations and confirmed that these were responsible for the observed inconsistent motility amongst flaB mutant clones. We also describe a new versatile and user-friendly system for genetic complementation in C. jejuni.

METHODS
Bacterial strains and growth conditions. All WT strains, defined mutants, genetically complemented mutants and plasmids are summarized in Table 1. C. jejuni strain M1 was originally isolated from a researcher after visiting a poultry processing plant (Friis et al., 2010) and is fully motile and spiral-shaped. C. jejuni M1 was routinely cultured on brain heart infusion (BHI, Oxoid) agar plates supplemented with 5 % defibrinated horse blood (Oxoid) and 5 mg trimethoprim ml 21 (TrM). Gene deletion mutants were cultured in the presence of 10 mg chloramphenicol ml 21 and genetically complemented mutant strains were also supplemented with 50 mg kanamycin ml 21 . C. jejuni was cultured under microaerophilic conditions (5 % O 2 , 10 % CO 2 and 85 % N 2 ) in a MACS VA500 Variable Atmosphere Work Station (Don Whitley). Escherichia coli NEB 5a or 10-b (New England Biolabs) was used for cloning and was cultured in LB medium at 37 uC, supplemented as appropriate with antibiotics.
Electroporation and natural transformation of C. jejuni. Electroporation of C. jejuni was performed as described previously (Holt et al., 2012) with minor adjustments. C. jejuni M1 was cultured on BHI-TrM blood agar plates for *48 h then replated onto fresh BHI-TrM blood agar plates and grown for *16 h. Bacteria were harvested from plates in BHI broth, washed four times in ice-cold wash buffer [272 mM sucrose and 15 % (v/v) glycerol] and then resuspended in wash buffer. Of this suspension, 100 ml was added per 0.2 cm prechilled electroporation MicroPulser cuvette (Bio-Rad) and mixed with DNA (*3 mg). Electroporation was performed using a GenePulser Xcell system (Bio-Rad) using the following settings: 2.5 kV, 200 V and 25 mF. After a single pulse, 200 ml of pre-warmed SOC medium was added and the suspension was transferred onto 2 ml BHI-TrM blood agar in a Universal tube (Greiner). After 5 h incubation under microaerophilic conditions, 1 ml of pre-warmed BHI medium was added to resuspend the bacteria, which were plated onto selective agar plates and grown under microaerophilic conditions for 2 to 3 days.
Natural transformation of C. jejuni was performed using an adapted biphasic method (Holt et al., 2012;van Vliet et al., 1998). C. jejuni was grown on BHI-TrM blood agar plates for *48 h and replated onto fresh BHI-TrM blood agar plates and grown for *16 h. Bacteria were harvested from plates in BHI broth and the OD 600 was adjusted to *0.5. Of this suspension, 0.5 ml was pipetted onto 2 ml BHI-TrM blood agar in a Universal tube (Greiner) and incubated for 3 h under microaerophilic conditions. Thereafter, DNA was added and incubated for 3-5 h before bacteria were harvested and then plated on selective agar plates and grown for 2 to 3 days under microaerophilic conditions. Construction of defined gene deletion mutants. C. jejuni defined gene deletion mutants were generated by allelic replacement of the gene with a chloramphenicol resistance cassette (cat). The cat cassette was amplified by PCR from the plasmid pCC027 (Coward et al., 2008) and the 59 (L1 and L2 primers) and 39 (R1 and R2 primers) flanking regions of the target gene were amplified by PCR from C. jejuni M1 genomic DNA. PCRs were performed using Q5 DNA polymerase (New England Biolabs) and purified using the QIAquick PCR purification kit (Qiagen). The gene flanking region PCR primers L2 (reverse primer 59 flank) and R1 (forward primer 39 flank) contain extensions that are complementary to the primers used for amplification of the cat cassette. The cat cassette (640 ng) and gene flanking regions (200 ng each) were joined in an overlap PCR without primers, resulting in the incorporation of the cat cassette in between the 59 and 39 flanking regions. Two microlitres of the product of this overlap PCR were then used as a template in a second round of amplification, this time using the primers L1 (forward primer 59 flank) and R2 (reverse primer 39 flank), which flank the original gene sequence and thus amplify the whole fragment; a total of four 50 ml PCRs were performed for each mutant, and subsequently pooled and purified. Of this, *3 mg of the overlap PCR product was used for electroporation of C. jejuni M1 to generate first generation defined gene deletion mutants, which were selected on BHI-TrM-Cm blood agar. Next, first generation mutant DNA was isolated using the DNeasy Blood and Tissue kit (Qiagen) and checked by PCR for recombination at the desired genomic location using the L1 forward primer (59 gene flanking region) and a control reverse primer located in the gene ('A' primer, which should not generate a product) and a cat cassette primer (CC069, which should generate a product of predicted size). First generation gene deletion mutant DNA (50 ng) was used for natural transformation of C. jejuni M1 WT; the gene replacement event was selected and confirmed by PCR. For flaB deletion mutant clones, correct integration of the cat cassette was also analysed by DNA sequencing. In addition, the WT was processed in parallel through the transformation procedure without any added mutagenic DNA to obtain a 'coupled' WT strain. This was done to reduce the genetic variation between the WT and defined mutant strains. Primer sequences are given in Table S1.
Motility assay. C. jejuni was grown on BHI blood agar plates for *48 h then replated onto fresh BHI blood agar plates and grown for *16 h. BHI suspensions of OD 600 *0.5 were made and stabbed, using a 20 ml pipette tip, into the middle of a 9 cm petri dish containing 25 ml of 0.4 % select agar (Sigma Aldrich) BHI medium and grown overnight under microaerophilic conditions. The diameter of the ring of bacterial growth was measured using a ruler (n53).
Genetic complementation. For genetic complementation of defined gene deletion mutants, the target gene was inserted into a pseudogene region between CJM1_0055 and CJM1_0057 (Coward et al., 2008). For this, a new genetic complementation system was designed in silico and synthesized de novo from oligonucleotides (GeneArt Strings, Life Technologies). The genetic complementation system consists of the cat promoter region from pRY011 (Coward et al., 2008;Yao et al., 1993), a partial pUC19 multiple cloning site and the kanamycin resistance cassette including its promoter region from pRY107 (Yao et al., 1993), all of which is flanked by 400 bp of 59 and 39 sequence of the CJM1_0055-0057 pseudogene region (98-100 % identity with sequences in C. jejuni 11168, 81116 and 81-176), see Fig. 5(a) for a schematic overview. The de novo synthesized genetic complementation region was first cloned into pMiniT using the NEB PCR cloning kit (New England Biolabs) and subsequently subcloned into pUC19 using the HindIII and EcoRI restriction enzyme sites, yielding the plasmid pSV009. Of note, the 59 pseudogene region (CJM1_0055-0057) sequenced originally harboured a HindIII site, which was eliminated by introducing a synonymous mutation during the in silico design of the genetic complementation insert. The fliW coding sequence including its upstream sequence was amplified by PCR and subcloned into pSV009 using the XhoI and BamHI sites, yielding pSV012-fliW. Sequences of pSV009 and pSV012-fliW were confirmed using DNA sequencing. The fliW complementation region and the empty genetic complementation region were amplified by PCR from pSV012 or pSV009, respectively, using the primers pSV009_GCampl_FW1/RV1, which amplify the complete genetic complementation region. The fliW complementation or the empty genetic complementation region was subsequently introduced into C. jejuni by electroporation. For a schematic overview of the genetic complementation strategy, see Fig. 5(a). Primer sequences are given in Table S1.

Differential motility of defined gene deletion mutant clones
For our studies into the molecular mechanisms of C. jejuni pathogenesis, e.g. invasion of gut epithelial cells (to be reported elsewhere), three defined deletion mutants were generated in C. jejuni M1 flagella genes using an overlap PCR method. Gene deletions were constructed in flaD, also referred to as flgL (CJM1_0851; hook filament junction protein) (Neal-McKinney & Konkel, 2012), pflA (CJM1_1501; paralysed flagellum protein) (Yao et al., 1994) and CJM1_1294, predicted to encode the major flagellin subunit FlaA (CJM1_1294) in the published C. jejuni M1 reference genome (Friis et al., 2010). We anticipated that deletion of these flagella genes would result in non-motile phenotypes (Neal-McKinney & Konkel, 2012;Wassenaar et al., 1991;Yao et al., 1994). As expected (Neal-McKinney & Konkel, 2012;Yao et al., 1994) three independent mutant clones for each of flaD (CJM1_0851) and pflA (CJM1_1501) were found to be non-motile (Fig. 1). Scanning EM confirmed the previously reported non-flagellated and flagellated phenotype of flaD and pflA gene deletion mutants, respectively (data not shown) (Neal-McKinney & Konkel, 2012;Yao et al., 1994). In contrast, five independent flaA (CJM1_1294) mutant clones were tested for motility with inconsistent results: mutant clones A and B displayed WT motility, clone C showed intermediate motility and clones D and E had severely attenuated motility (Fig. 1).
To unravel the basis of the inconsistent motility phenotypes of flaA (CJM1_1294) mutant clones, we first confirmed the correct insertion of the chloramphenicol (cat) resistance cassette in flaA (CJM1_1294) by Sanger sequencing. Importantly, this revealed the insertion of three A nucleotides in an 11 bp stretch at the 39 end of CJM1_1295 [annotated as 'hypothetical protein' (Friis et al., 2010)], which was also independently confirmed by Sanger sequencing of the complete CJM1_1295-1294 region. Incorporation of the A insertions into the original M1 sequence (Friis et al., 2010) led to CJM1_1295 and CJM1_1294 being merged into a larger ORF of 1731 bp (Fig. 2a). The difference between the published C. jejuni M1 reference genome (Friis et al., 2010) and the sequence of our M1 strain may reflect genotypic variation or sequencing or assembly errors in the original M1 genome sequence.
In-depth analysis of the altered flagellin locus in C. jejuni M1 revealed that: (1) the merged CJM1_1295-1294 gene actually represents the full-length flaB gene as it was identical to the published flaB sequence of C. jejuni 81116 (Nuijten et al., 1990;Pearson et al., 2007) and (2) CJM1_1296 was incorrectly annotated as flaB; it is identical to flaA of C. jejuni 81116 (Friis et al., 2010), see Fig. 2(a). Therefore, our CJM_1294 gene deletion mutant, described above, is not a flaA mutant but the cat cassette is replacing a region ranging from position 822 to 1671 bp of the flaB gene. Consequently, the motile phenotype of flaB (CJM1_1294) mutant clones (Fig. 1) confirmed the findings by Wassenaar et al., (1991), reporting a motile phenotype after the inactivation of flaB in C. jejuni 81116. Deletion of the full-length flaB gene (CJM1_1295-1294, referred to as the flaB * mutant) also resulted in WT motility levels for mutant clones B and C, although the motility of flaB * (CJM1_1295-1294) mutant clone A was slightly lower compared to the WT (Fig. 2b). In line with previous findings of Wassenaar et al. (1991) in C. jejuni 81116, deletion of flaA (CJM1_1296) and the complete flagellin locus ( flaAB; CJM1_1296 and CJM1_1295-1294) led to abolished motility (Fig. 2b). Analysis of the combined motility data for the flaA (CJM1_1296) and flaAB mutant clones showed that deletion of the complete flagellin locus resulted in more severely attenuated motility compared to deletion of flaA alone (Pv0.001), suggesting that flaB plays a minor role in motility, at least in C. jejuni strain M1.
Based on our findings, we hypothesized that the attenuated motility in flaB (CJM1_1294) mutant clones C, D and E might be caused by loss-of-function mutations in other motility-associated genes or as previously reported, by a WT WT Fig. 1. Motility of C. jejuni M1 flagella defined gene deletion mutants. Motility was analysed on BHI plates with 0.4 % (w/v) agar after overnight incubation. Three flaD (CJM1_0851) and plfA (CJM1_1501) deletion mutant clones were assayed along with their coupled WT strain and five flaB [CJM1_1294, previously annotated to encode FlaA (Friis et al., 2010)] mutant clones were tested with two coupled WT strains (A and B), with the average motility shown for the latter. All three flaD (CJM1_0851) and pflA (CJM1_1501) mutant clones were non-motile; however, motility of flaB (CJM1_1294) mutant clones was inconsistent: clones A and B showed WT motility, clone C displayed intermediate motility, whereas the motility of clones D and E was severely attenuated. Differential motility of mutant clones was tested against the coupled WT strain(s) with a Mann-Whitney test (n53). Data shown are the mean and SEM, with ***P,0.001. compensatory mechanism, for example, by increased flaB expression or by genome rearrangements, which have been shown to (partially) restore motility (Nuijten et al., 2000;Wassenaar et al., 1994Wassenaar et al., , 1995.

Genotypic and phenotypic analysis of motility defects
A genome resequencing approach was used to assess whether second-site mutations might be responsible for the attenuated motility of flaB (CJM1_1294) mutant clones C, D and E. Three pflA, three flaD and five flaB mutant clones and their 'coupled' M1 WT strains were genome sequenced. Eighty-one variants, i.e. SNPs and INDELs, were detected relative to the C. jejuni M1 reference genome [GenBank CP001900 (Friis et al., 2010)]. Of these, 34 SNPs and INDELs were present in all sequenced gene deletion mutant clones and their 'coupled' WTs, whereas 47 SNPs and INDELs were only detected in a subselection of them (Fig. 3). Although the flaB mutant  clones subclustered, limited mutant-specific hierarchical clustering was detected, indicating that variation may be introduced throughout all growth steps during the construction of the gene deletion mutants. The SNPs and INDELs nucleotide position numbers correspond to the position in the original C. jejuni M1 reference genome sequence (Friis et al., 2010). Interestingly, the genome resequencing approach revealed the presence of SNPs and INDELs unique to the severely attenuated motility flaB (CJM1_1294) mutant clones D and E. These were an insertion of a G at position 689 014 within CJM1_0719 (encoding a putative periplasmic protein) and deletion of a TA-dinucleotide at position 1 021 140 within CJM1_1052 ( fliW). The TA-deletion in fliW is predicted to result in a three amino acid C-terminal truncation of the FliW protein. FliW encodes a predicted flagellin filament-stabilizing chaperone, although it has not yet been demonstrated whether FliW interacts with FlaA and/or FlaB flagellin subunits (Barrero-Tobon & Hendrixson, 2014;Titz et al., 2006). Previously it was shown that deletion of fliW in C. jejuni 81-176 resulted in a phenotype consisting of severely attenuated motility with the presence of a probable flagellar hook structure but without flagellin filaments (Barrero-Tobon & Hendrixson, 2014). Scanning EM analysis revealed that the majority of the flaB (CJM1_1294) mutant clones D and E also lacked flagellar filaments and instead possessed hook-like structures, but with very few bacteria displaying a possible filament structure (Fig. 4d, e). In contrast, the majority of the bacteria imaged for the flaB (CJM1_1294) coupled WTs (Fig. 4f,g) and flaB (CJM1_1294) mutant clones A, B and C harboured a flagellar filament at one or both poles (Fig. 4a-c). Notably, the flaB (CJM1_1294) mutant clones D and E differed at two variant positions (Fig. 3), which suggests that they did not originate from a single transformation event. Further, the intermediate motile flaB (CJM1_1294) mutant clone C possessed a unique deletion of an A at position 69 410 in CJM1_0051 ( flgD), encoding a flagellar hook assembly protein. This deletion was predicted to result in a severely truncated flgD ORF, encoding a protein of 55 aa instead of 294 aa. In C. jejuni 81-176 a flgD transposon mutant was previously reported to be non-motile (Hendrixson et al., 2001).
The genome sequencing data obtained in this study for the M1 WTs were used to generate a reference genome sequence (designated CJM1cam) for the C. jejuni M1 strain used in our laboratory. The CJM1cam genome sequence and annotation are available via GenBank accession no. CP012149; for more information see Methods.
Second-site mutations in fliW responsible for motility defects in flaB mutants We hypothesized that the predicted C-terminal truncation of the FliW protein by 3 aa may be responsible for the observed attenuated motility of flaB (CJM1_1294) mutant clones D and E. To confirm that fliW is also required for motility in C. jejuni M1, a fliW gene deletion mutant was generated, which showed abolished motility (Fig. 5b). To exclude that this was a clone-specific phenotype, we assayed two other fliW gene deletion mutant clones; both clones were found to be non-motile (Fig. S1). Genome sequencing of fliW mutant clone A and its coupled WT showed that 14 SNPs and INDELs were shared between them, five variants were unique to the WT and two were only found in the fliW mutant clone A (Table S3). Unique variants to the fliW mutant clone A were an intergenic A insertion between CJM1cam_0120 (putative metalloprotease) and CJM1cam_0121 (hypothetical protein) and a synonymous SNP (A to G) in CJM1cam_0791 (hypothetical protein), indicating that the non-motile phenotype is not caused by second-site mutations. Electron micrographs of the C. jejuni M1 fliW mutant were similar to those reported previously of a C. jejuni 81-176 fliW mutant, i.e. lack of flagellar filaments and presence of a hook-like structure (Barrero-Tobon & Hendrixson, 2014) (Fig. 4h). The motility of the fliW mutant was slightly less than that of the flaB (CJM1_1294) mutant clones D and E, although this difference was not statistically significant. It is possible that the predicted 3 aa truncated FliW protein in flaB (CJM1_1294) mutant clones D and E could still possess limited functional activity.
To confirm that the TA-deletion in fliW was responsible for the severely attenuated motility of flaB (CJM1_1294) mutant clones D and E, the WT fliW gene including its native promoter was inserted into the CJM1_0055-0057 pseudogene region. For this, the new genetic complementation plasmid pSV009 was developed, which enables: (1) flexible cloning options through a whole range of restriction sites, (2) the possibility to drive the expression from the cat promoter or replace this with the native promoter of the gene of interest and (3) pseudogene region insertion in well-studied C. jejuni strains, e.g. M1, 81-176, 81116 and 11168, via homologous recombination regions with w98 % identity (Fig. 5a). Using the newly developed complementation system, insertion of fliW into the CJM1_0055-0057 pseudogene region in the fliW mutant as well as in flaB (CJM1_1294) mutant clones D and E was shown to restore motility to WT levels, suggesting that the flagella are restored, although this was not confirmed by EM. Insertion of the empty genetic complementation backbone did not restore motility (Fig. 5b). These results confirm that the TA-deletion in fliW was responsible for the attenuated motility phenotype of flaB (CJM1_1294) mutant clones D and E.

Concluding remarks
Many studies aimed at investigating the role of genes in C. jejuni make use of targeted deletion of genes by replacing them with an antibiotic resistance cassette. Here we have demonstrated that it is of critical importance to assess the phenotype of  multiple mutant clones and/or genetically complement the gene deletion. Ideally, genome sequencing may be used to screen for second-site mutations that may interfere with the pathways/mechanism under study, especially when mutant clones show inconsistent behaviour or when genetic complementation fails. As shown in this study, genome sequencing can be used as a troubleshooting tool and may also yield novel genes associated with the phenotypes under study.

ACKNOWLEDGEMENTS
This work was funded by BBSRC grant RG66581.We thank Gemma Murray from the Department of Genetics, University of Cambridge for providing the R script used for Cluster of Orthologous Genes assignment and Diane Newell for providing the C. jejuni M1 strain. Jeremy Skepper at the Cambridge Advanced Imaging Centre, Department of Physiology, Development and Neuroscience, University of Cambridge kindly performed the scanning electron microscopy. Xho I site of pSV009, yielding pSV012-fliW. This replaced the pSV009 cat promoter and enabled fliW expression from its native promoter. The genetic complementation region is flanked by regions for homologous recombination to facilitate insertion into the CJM1_0055-0057 pseudogene region, which is predicted to be compatible with other commonly used C. jejuni isolates (e.g. 11168, 81116 and 81-176). For insertion of fliW (CJM1_1052) into the pseudogene region, the complementation region was amplified by PCR using the pSV009_GCamplif primers and subsequently introduced by electroporation. (b) As expected, insertion of fliW (CJM1_1052) in the CJM1_0055-0057 pseudogene region in a fliW (CJM1_1052) mutant restored motility. Insertion of fliW (CJM1_1052) in the pseudogene region also restored motility in the severely attenuated motility flaB (CJM1_1294) mutant clones D and E, whereas insertion of the empty genetic complementation ('empty GC') region did not restore motility. The white bars indicate the coupled WT strains for flaB (CJM1_1294) and fliW (CJM1_1052) gene deletion mutants. Differential motility of mutant clones was tested against coupled WT isolates using a Mann-Whitney test (n53). Data shown are the mean and SEM, with *P,0.05 and **P,0.01.