All bacterial strains and plasmids used in this study are listed in Table 1. For long term storage, Campylobacter strain freezer stocks were maintained in Brucella broth (Becton Dickinson, Sparks, USA) plus 15% glycerol frozen at −80°C. Strain cultures were revived from freezer stocks directly onto Brucella agar (1.5%) plates. Kanamycin (15 μg mL¹ final concentration) or tetracycline (5 μg mL¹ final concentration) was added to Brucella agar plates or Brucella broth when growing strains possessing a cloned aph(3) or tetO gene, respectively. All C. jejuni cultures were incubated in a microaerobic (5% O₂, 10% CO₂, 85% N₂) growth chamber (Concept-M, Baker, Sanford, USA) at 42°C.

Blast analysis of the UTR directly upstream of cysM was performed as follows: the NCBI genome website was searched using “Campylobacter jejuni subsp. jejuni AND (latest OR ‘latest genbank’) AND all NOT anomalous” search words on 05-12-2023 (Altschul et al., 1990). This downloaded 326 genome fasta sequences. A 126-bp UTR sequence was used against the downloaded 326-genome sequences using Blastn + software using default parameters. The subject sequences were selected for 126 bp length and no-gaps in sequences. This produced 264 sequences of 126 bases long. The 264 sequences were aligned with MAAFT software (v7.487) with FFT-NS-i parameter (Katoh and Standley, 2013). The UTR sequence was analyzed for regulatory elements and transcription factor binding sites using bprom software (Solovyev and Salamov, 2011).¹

The C. jejuni strain RM3194ΔflaAB:tet + flaAcomp (G₂₀) had been previously constructed (Uhlich et al., 2022). In order to investigate the differences in transcription when an adenine was at the 20th position in the UTR compare to guanine we need to construct a strain the same as RM3194ΔflaAB:tet + flaAcomp (G₂₀), but with the desired single nucleotide change. Plasmid pBlueKan + cysM^Pro (G₂₀) carrying a guanine at the 20th site in the UTR was PCR amplified using two sets of primers in order to achieve a Gibson style assembly. Primers cysM3194rev + blueex and cysMfor + Kanex were used to amplify the cysM UTR and replace the guanine nucleotide with an adenine. Primers Bluefor + cysMex and Kanrev + cysMex were used to amplify the vector backbone. Using the NEBuilder HiFI DNA Assembly kit (NEB, Ipswitch, MA, USA) and the overlap extensions built into the PCR primers the plasmid pBlueKan + cysM^Pro (G₂₀) was reconstructed as plasmid pBlueKan + cysM^Pro (A₂₀) with the adenine in the desired location. Next flaA from strain RM3194 was cloned into the newly constructed plasmid pBlueKan + cysM^Pro (A₂₀), again using a Gibson style assembly. Primers 3194pBlue + FlaAREx2 and 3194FlaAR + pBlueEx2 were used to clone flaA from RM3194 and primers FlaAF + cysMPRex and 3194cysMPR + FlaAFex were used to amplify the pBlueKan + cysM^Pro (A₂₀) plasmid. Again using the NEBuilder kit and the overlap extensions built into the primer sets, flaA was cloned into plasmid pBlueKan + cysM^Pro (A₂₀) producing the new plasmid, pBlueKan + cysM^Pro (A₂₀)-flaA. The pBlueKan + cysM^Pro (A₂₀)-flaA plasmid was digested with NotI (NEB) and the fragment containing the kanamycin resistance gene, the A₂₀ cysM UTR and flaA was gel purified away from the plasmid backbone. The purified fragment was ligated with the pCJR01 vector that was also digested with NotI. The resulting plasmid was designated, pCJR01comp (A₂₀)-flaA and subsequently electroporated into strain C. jejuni RM3194ΔflaAB:tet (Davis et al., 2008). Before electroporation the plasmid pCJR01comp (A₂₀)-flaA was sequenced to ensure that no additional changes to the cysM UTR or flaA had occurred during the cloning process. Transformants were selected for on Brucella agar plus kanamycin and cells from the resulting colonies were transferred individually onto Brucella agar plus chloramphenicol and Brucella agar plus kanamycin plates. Successful double crossover integration events of the cysM UTR flaA into the bacteria chromosome were identified by the presence of kanamycin resistance and the absence of chloramphenicol resistance. All successful transformants were directly checked for motility on low agar Brucella plates, no transformants were motile. Several colonies were then chosen at random, stored as freezer stocks and designated, as RM3194ΔflaAB:tet + flaAcomp (A₂₀) strains.

Strains RM3194ΔflaAB:tet + flaAcomp (G₂₀) and RM3194ΔflaAB:tet + flaAcomp (A₂₀) were taken directly from freezer stocks and grown on Brucella agar (1.5%) plus Kanamycin. When either strain was grown (for 16 to 48 h) on solid agar media directly from freezer stocks or passed from one Brucella agar plate to the next, RM3194ΔflaAB:tet + flaAcomp (G₂₀) was referred to as a G₂₀ solid media culture (GSM) and RM3194ΔflaAB:tet + flaAcomp (A₂₀) was referred to as an A₂₀ solid media culture (ASM). Cells from either of these strains taken from agar plates and inoculated into Brucella broth (with kanamycin) and grown for up to 48 h were still classified as GSM or ASM strains since no changes in the motility of the ASM strain was observed at this point. However, when cells were taken from the liquid cultures and placed into fresh Brucella broth plus kanamycin and incubated again for at least 48 h, changes in the motility of the formally non-motile ASM strain began to be observed. Therefore, these cultures were renamed as G₂₀ liquid media cultures (GLM) for RM3194ΔflaAB:tet + flaAcomp (G₂₀) and A₂₀ liquid media cultures (ALM) for RM3194ΔflaAB:tet + flaAcomp (A₂₀). Cells from GLM and ALM cultures after 48 h of growth were routinely transferred into fresh Brucella broth plus kanamycin to maintain the health of the bacterial cultures.

Low concentration agar (0.3%) Brucella plates were used to measure the motility of the C. jejuni strains in this study, with kanamycin or tetracycline being added to the plates where necessary (Barrero-Tobon and Hendrixson, 2012). Bacterial strains to be evaluated were initially grown overnight in 3 mL of Brucella media plus kanamycin or tetracycline if appropriate. The following day 1 μL drops of the individual C. jejuni cultures were injected into the center of the appropriate 0.3% agar Brucella plates. Plates were incubated for 16 h in a microaerobic atmosphere chamber at 42°C. Following incubation, the diameters of the resulting circular bacterial migration zones were measured to the closest millimeter. The average motility for individual strains were then calculated from at least three experimental replicates.

Fifty microliters of the appropriate C. jejuni strain, were placed on acetone cleaned 12 mm Micro-cover glass slides (Thermo Scientific Portsmouth, NH, USA) and allowed to adhere for 30 min. Samples were then covered with 2 ml of 2.5% glutaraldehyde, (Electron Microscopy Sciences, Hatfield, PA, USA) and were allowed to fix for 30 min. The samples were then rinsed twice for 30 min each with 2–3 mL of the 0.1 M imidazole, (Electron Microscopy Sciences, Hatfield, PA, USA), followed by 30 min intervals each of 50, 80, 90% ethanol (The Warner-Graham Company, Cockeysville, MD, USA), 2–3 ml each. The samples were then washed and held three times with 2 mL of 100% ethanol before being critically point dried. The samples were stacked in a wire basket, separated by cloth, and placed in a Critical Point Drying Apparatus, (Denton Vacuum, Inc., Cherry Hill, NJ, USA), using liquid carbon dioxide (Welco Co, Allentown, PA, USA) for approximately 20 min. The samples were mounted on stubs and sputter gold coated for 1 min (EMS 150R ES, EM Sciences, Hatfield, PA, USA). Samples were then viewed with a FEI Quanta 200 F Scanning Electron Microscope, (Hillsboro, OR, USA) with an accelerating voltage of 10KV in high vacuum mode.

RNA was isolated from 2 mL aliquots of each of the designated C. jejuni strains incubated in Brucella broth supplemented with kanamycin or tetracycline where appropriate. Samples were collected for RNA isolation once they had reached optical densities of between 0.1 and 0.3 when measured at 600 nm (OD₆₀₀) using a Genesys 30 spectrometer (ThermoFisher Scientific). Cells were collected by centrifugation (7,000 × g) and RNA was isolated using the Invitrogen PureLink RNA Kit (ThermoFisher Scientific). The RNA samples were next DNase I treated with Invitrogen’s Turbo Dnase kit (ThermoFisher Scientific). Finally, cDNA synthesis was accomplished using Invitrogen’s SuperScript III First-Strand Synthesis Kit (ThermoFisher Scientific) and a Bio-Rad T100 thermocycler (Bio-Rad, Hercules, CA, USA) following manufacturer’s instructions.

Primer sets used for the quantitative PCR reactions were as follows: rpoA2B_F (forward primer) and rpoA2B_R–(reverse primer) were used to amplify the rpoA housekeeping gene transcription, which served as an internal control to normalize the different starting RNA levels between the samples. cysM2_F and cysM2_R were used to measure the native cysM transcription. flaA2D_F and flaA2D_R were used to measure transcription of the cloned flaA. RT-PCR reactions were performed in 8 tube strips using a Roche LightCycler 96 system (Roche Scientific). Reactions were 20 μL in total volume including: 10 μL of FastStart Essential DNA Green Master Mix (Roche), 1 μL of each 10 μM primer forward and reverse, and 8 μL of cDNA previously diluted 1:40. The amplification protocol was as follows: initial denaturation step at 95°C for 10 min, followed by 45 cycles of 95°C for 10 s, 58°C for 10 s, and 72°C for 10 s and concluded with a melting curve analysis. The LightCycler 96 system software was used to assess the PCR results. Each set of RT-PCR experiments also included “no reverse transcriptase” samples to control for genomic DNA contamination as well as “no template” controls lacking only the cDNA template. For each sample the relative transcription level of the target gene (cysM or flaA) was calculated as a ratio to the housekeeping gene (rpoA) to allow for comparisons between strains and conditions. Equation: ratio = E_R^Cq_R/E_T^Cq_T (E_R, quantification efficiency of reference gene, E_T, quantification efficiency of target gene, Cq_R, quantification cycle of reference, Cq_T, quantification cycle of target).

RNA was isolated from ASM and ALM samples as described above. The resulting RNA concentrations from each sample were measured using a NanoDrop spectrophotometer (ThermoFisher Scientific) and the RNA was then diluted to produce equal concentrations of starting RNA. cDNA was produced from RNA in the same manner as before except a single primer flaA_cDNA, specific to flaA mRNA was used to produce cDNA only from flaA mRNA in the isolated RNA samples. The flaA specific primer (flaA_cDNA) sequence localizes to the 3′ end of the theoretical flaA mRNA. Next the flaA specific cDNA produced in the previous step were used as template in PCR amplification using the following four forward primers flaA2D_F, flaA2Dshort_F, flaA2Dmedium_F and flaA2Dlong_F each localizing at increasing distance from the 5′ of the flaA cDNA and paired separately with the reverse primer, flaA2B_R. PCR conditions were: 95°C for 10 min followed by 35 cycles of 95°C for 30 s, 57°C for 30 s, 72°C for 90 s with a final single 72°C for 10 min. PCR amplification used the Qiagen Multiplex PCR kit following manufacturer’s recommended method. Subsequent PCR products were visualized on a 1% DNA gel subsequently stained with GelRed (Thomas Scientific).

At least three separate experimental replicates were used to produce all mean and standard error values reported in this paper. The one-way analysis of variance (ANOVA) test was used to determine if mean values compared in this study were significantly different. Subsequent to the ANOVA analysis, the Tukey HSD test was used to separate the mean values. Significant differences were defined as P-values ≤ 0.05.

The 126 bp UTR between the cysteine synthase (cysM) gene and DNA-binding protein HU gene is well conserved between sequenced C. jejuni strains. In this study 264 sequences of this region were aligned and observed for sequence variations. Within the 264 sequences only eighteen nucleotide sites were observed to vary over the 126 bp region. Additionally, most of these nucleotide substitutions were rare with any of the nucleotide variations only occurring in a few strains in total. However, at one site a nucleotide substitution was observed in a sizable percentage of the 264 sequences, with 219 sequences having a guanine (G) nucleotide at this location and the remaining 45 sequences alternately having an adenine (A). Nucleotides within the 126 bp region were numbered starting from the nucleotide directly upstream of the first nucleotide of the cysM coding sequence (#1) and ended at the nucleotide immediately downstream of the last nucleotide of the HU DNA-binding gene (#126). The predominant SNP (G/A) was located at position number 20 with the SNPs designated as G₂₀ or A₂₀. Further analysis of the sequences identified a putative ribosomal binding site ATCCTT running from nucleotides 10 to 15 of the UTR (Figure 1). Additionally, a possible fur binding site AATAATGAT (nucleotides 24 to 32) and an integration host factor (IHF) binding site, which contained the G₂₀/A₂₀ SNP [TTTTA(G)TTT, nucleotides 16 to 23] were identified. While it is believed that C. jejuni lack most transcription factors, including IHF, it does possess the gene for HU which appears to perform many of the same functions of IHF (Stojkova et al., 2019).

Sequenced strains with the G₂₀ or A₂₀ SNP in the UTR upstream of cysM were selected from the laboratory strain collection for transcription analysis. The strains were grown in Brucella broth and collected during the exponential phase of growth, at OD₆₀₀ values of between 0.1 and 0.3. Total RNA was isolated and cDNA produced from the mRNA transcripts. Quantitative RT-PCR was used to measure cysM transcription levels for each strain (Figure 2). The values for the cysM transcription levels were presented as ratios to the transcription levels of the housekeeping gene, rpoA, allowing for normalization of concentration differences between various samples. On average the G₂₀ SNP strains (FSIS136, CDC9511, and RM1285) had higher relative transcription levels of cysM when compared to strains with the A₂₀ SNP (RM3194, RM1188). The average transcription levels of the A₂₀ SNP strains RM3194 and RM1188 did not significantly differ from one another with values of 0.22 and 0.25, respectively. Within the G₂₀ SNP strains, there was variation in the cysM transcription levels. Strain FSIS136 (0.38) differed significantly from CDC9511 (0.51) with regards to cysM transcription, while strain RM1285 (0.47) did not differ significantly from CDC9511 or FSIS136. These results suggested that during exponential growth in a rich medium, the different SNPs contributed to differences in the regulation of cysM transcription.

The UTR upstream of cysM is presumed to contain the promoter region for the gene. The G₂₀ SNP version of the UTR was previously used in research studies as a constitutive promoter to drive expression of exogenous genes cloned into C. jejuni (Uhlich et al., 2022; Gunther et al., 2023). In order to determine how the A₂₀ SNP might affect the transcription of recombinant genes cloned within the C. jejuni chromosome, the cysM UTR contained within of the cloning vector pBlueKan + cysM^Pro (G₂₀) was changed to adenine (A₂₀). This allowed for the construction of C. jejuni strain RM3194ΔflaAB:tet + flaAcomp (A₂₀). C. jejuni RM3194ΔflaAB:tet is a clinical strain with both flagellar structural genes flaA and flaB removed. The exogenous flaA cloned behind an A₂₀ UTR was inserted within the C. jejuni RM3194ΔflaAB:tet chromosome between the 16S and 23S ribosomal RNA genes. The resulting strain, RM3194ΔflaAB:tet + flaAcomp (A₂₀), was identical to the previously constructed strain, C. jejuni RM3194ΔflaAB:tet + flaAcomp (G₂₀) (Uhlich et al., 2022), with the exception of the A₂₀ SNP modification now upstream of the cloned flaA gene. The C. jejuni RM3194ΔflaAB:tet + flaAcomp (A₂₀) strain was originally believed to be non-motile, however, it was observed to become motile when incubated in Brucella broth for two or more consecutive cultures each lasting 48 h. To illustrate the motility differences produced by the two UTR SNP forms, the strains C. jejuni RM3194ΔflaAB:tet + flaAcomp (A₂₀) and C. jejuni RM3194ΔflaAB:tet + flaAcomp (G₂₀) were grown either on solid media (1.5% agar) to form cultures designated GSM (G₂₀ UTR) and ASM (A₂₀ UTR) or within liquid medium with at least two passages into fresh medium resulting in cultures designated GLM (G₂₀ UTR) and ALM (A₂₀ UTR). The motilities of each of these four cultures, were compared on low percentage agar, along with a positive control strain, C. jejuni RM3194 + express, and a negative control strain, C. jejuni RM3194ΔflaAB:tet (Figure 3). On average the positive control strain with the wild-type copies of both flagella structural genes, driven by their native promoters, produced motility zones of 28.3 mm. The GSM and GLM cultures produced average motility zones of 15.7 and 10.6 mm, respectively, with the different growth conditions having only a small influence on cell motility. This differed significantly from the influence that the growth conditions had on motility when flaA was combined with the A₂₀ UTR. The ALM culture produced an average motility zone of 18.5 mm while the ASM culture showed no motility appearing identical to the negative control strain that lacked any structural flagella genes. Additionally, cells from the ALM culture when placed onto Brucella agar plates and transferred after 24 h of growth onto fresh agar plates twice more, were observed to completely lose their motility. PCR amplification of the UTR locus from both the ASM and ALM cultures confirmed that the adenine SNP was not modified when switching between motile and non-motile states (data not shown). Additionally, since strain RM3194ΔflaAB:tet + flaAcomp (A₂₀) was shown to switch between a motile and non-motile phenotype depending on how it was sub-cultured, suggests that a mutational event in a location other than the UTR is not responsible for the observed changes.

The strains and cultures used in the previous motility experiments (Figure 3) were assessed by scanning electron microscopy to identify differences in cell structure that might affect motility. Representative images for each culture condition are presented in Figure 4. Flagella are readily visible on the cells comprising the positive control strain (Figure 4A). Conversely, no flagella were visible on cells of the negative control strain (Figure 4B). The non-motile ASM culture (Figure 4C) appears similar to the negative control without any flagella readily visible on the cells surface. The motile ALM (Figure 4D), GSM (Figure 4E), and GLM (Figure 4F) each have flagella visible on the cells within the respective cultures.

Quantitative RT-PCR was used to determine the flaA transcription levels in the positive control, RM3194 + express, and negative control strain, RM3194ΔflaAB:tet; as well as strain, RM3194ΔflaAB:tet + flaAcomp (A₂₀), grown under conditions producing motility and observable flagella (ALM) and the conditions without motility or observable flagella (ASM). The transcription of flaA was also measured for the strain, RM3194ΔflaAB:tet + flaAcomp (G₂₀), grown under GLM and GSM conditions, both of which previously produce motility and observable flagella. Results for these experiments are presented in Figure 5. Not surprisingly strain RM3194 + express had the highest level of flaA transcription since it also previously had demonstrated the greatest motility. The non-motile negative control strain with no flaA or flaB predictably did not show any measurable flaA transcription. The strain and culture conditions that previously demonstrated motility and flagellar expression, ALM, GSM, and GLM all had similar average levels of flaA transcription. Unsurprisingly, these levels of flaA transcription were significantly less than that in the positive control strain given the superior motility demonstrated by the control strain. The non-motile, no observable flagella culture, ASM, had significantly lower flaA transcription levels when compared to ALM, GSM and GLM cultures. In fact the flaA transcription of RM3194ΔflaAB:tet + flaAcomp (A₂₀) grown in the ASM conditions was not significantly different from the negative control strain which corresponded nicely with the observation that both the negative control and ASM samples were non-motile.

Strain RM3194ΔflaAB:tet + flaAcomp (A₂₀) grown in ALM or ASM conditions demonstrated differences in motility, observed flagella and flaA transcription levels. Since the parent strain, RM3194, is known to possess an A₂₀ UTR upstream of cysM, quantitative RT-PCR experiments were performed to determine how the ALM and ASM growth conditions affected the transcription of the native cysM. The results of these experiments are presented in Figure 6. The ALM condition, on average, resulted in significantly greater levels of cysM transcription compared to the ASM conditions, which agreed with previous results measuring the expression of recombinant flaA in strain RM3194ΔflaAB:tet + flaAcomp (A₂₀) (Figure 5). This result suggests that transcription of any gene cloned downstream of the A₂₀ UTR will be regulated differently depending on the growth conditions under which the host strain is cultured.

A flaA specific primer was used to only produce flaA cDNA from RNA collected when strain RM3194ΔflaAB:tet + flaAcomp (A₂₀) was grown in ALM or ASM conditions. The total RNA amounts present in the ALM and ASM samples were selectively diluted to equalize the starting RNA concentrations between the samples. The resulting flaA cDNAs isolated from the two different growth conditions were then used as template for PCR amplification. The forward primers used for the PCR amplifications were designed to bind to the flaA cDNA sequences at increasing distances from the 5′ end of the gene. Primer flaA2D_F binds to the sequence starting 239 bps downstream of the ATG start codon of flaA; while flaA2Dshort_F binds 438 bps downstream of the ATG start codon, flaA2Dmedium_F binds 935 bps and flaA2Dlong_F binds 1,223 bps downstream of the gene’s ATG start codon. All forward primers were paired with the same reverse primer flaA2B_R. Therefore, when paired with flaA2B_R for PCR amplification flaA2D_F produces a 1285 bp fragment, flaA2Dshort_F produces a 1,085 bp fragment, flaA2Dmedium_F produces a 588 bp fragment and flaA2Dlong_F produces a 300 bp fragment. The relative location of the primers within flaA are presented in Figure 1. When the resulting PCR amplifications were visualized (Figure 7) only extremely faint bands were visible in lanes 2 and 3, which are the results of amplifying the cDNA from the ASM sample using forward primers flaA2D_F and flaA2Dshort_F, respectively. However, in lanes 4 and 5 the forward primers flaA2Dmedium_F and flaA2Dlong_F produced easily visualized bands amplifying the same ASM sample. With the ALM sample, primers flaA2D_F, flaA2Dshort_F, flaA2Dmedium_F and flaA2Dlong_F (lanes 6–9) all produced easily visualized bands. These results suggest that there were significantly fewer full length flaA cDNAs produced from the ASM growth samples compared to the ALM samples. This would likely result from there being significantly fewer full length flaA mRNA transcripts originally in the ASM samples compared to the ALM samples. The cDNAs, produced from flaA mRNA, appear to exist in similar amounts in the ASM (lanes 4 and 5) and ALM (lanes 8 and 9) samples when the PCR reactions amplify the 3′ region of flaA cDNA present in the samples. Conversely, it appears for the most part that only the 5′ portions of the flaA mRNA sequences from the ALM sample are present and accessible to reverse transcriptase allowing for production of the full length flaA cDNA as demonstrated by the larger PCR products manufactured from the PCR reactions from ALM (lanes 6 and 7) but not ASM (lanes 2 and 3) samples. This suggests that the 5′ region of the flaA mRNA transcripts present in the ASM sample are missing or in some way not accessible to the reverse transcriptase that would normally produce full length flaA cDNA. This same type of experiment was repeated using primers specific for native cysM. Similarly, we observed that the 5′ portions of cysM mRNA sequences from the ALM samples appeared to be more accessible to reverse transcriptase; allowing for the production of more full length cysM cDNA compared to the samples from the ASM conditions. It should be noted that differences observed in full length cysM cDNA between ALM and ASM conditions were not as dramatic as that observed with flaA cDNA (data not shown).