Genomic Analysis Points to Multiple Genetic Mechanisms for Non-Transformable Campylobacter jejuni ST-50

Campylobacter jejuni and Campylobacter coli are well known for their natural competence, i.e., their capacity for the uptake of naked DNA with subsequent transformation. This study identifies non-transformable C. jejuni and C. coli strains from domestic animals and employs genomic analysis to investigate the strain genotypes and their associated genetic mechanisms. The results reveal genetic associations leading to a non-transformable state, including functional DNase genes from bacteriophages and mutations within the cts-encoded DNA-uptake system, which impact the initial steps of the DNA uptake during natural transformation. Interestingly, all 38 tested C. jejuni ST-50 strains from the United States exhibit a high prevalence of non-transformability, and the strains harbor a variety of these genetic markers. This research emphasizes the role of these genetic markers in hindering the transfer of antimicrobial resistance (AMR) determinants, providing valuable insights into the genetic diversity of Campylobacter. As ST-50 is a major clone of C. jejuni globally, we additionally determined the prevalence of the genetic markers for non-transformability among C. jejuni ST-50 from different regions of the world, revealing distinct patterns of evolution and a strong selective pressure on the loss of competence in ST-50 strains, particularly in the agricultural environment in the United States. Our findings contribute to a comprehensive understanding of genetic exchange mechanisms within Campylobacter strains, and their implications for antimicrobial resistance dissemination and evolutionary pathways within specific lineages.


Introduction
Campylobacter jejuni and Campylobacter coli are Gram-negative, obligately microaerophilic, spiral-shaped and highly motile bacteria that frequently colonize poultry, cattle, swine and other animals worldwide [1].Known for being the most common causative agent of bacterial gastroenteritis and the leading bacterial antecedent of the severe autoimmune complication Guillain-Barré syndrome [2][3][4], C. jejuni is an almost ubiquitous gastrointestinal pathogen.While most cases of campylobacteriosis in healthy individuals are self-resolving, immunocompromised patients may require antibiotics to prevent recurrent infections or the further progression of disease [5].However, the increased incidences of antimicrobial resistance (AMR) in C. jejuni and C. coli have made it harder to treat known infections, and thereby cause an increased threat to public health [6,7].
C. jejuni and C. coli are genetically diverse, with over 12,000 sequence types (STs) in the PubMLST database [8].C. jejuni and C. coli multilocus sequence typing (MLST), based on the alleles of seven housekeeping genes, is used to genotypically classify strains and strain lineages [9].Indeed, MLST has become an important source-attribution tool for the epidemiology of Campylobacter [10,11].
The natural competence of C. jejuni and C. coli, i.e., the ability to undergo transformation via the uptake of naked DNA from its environment, plays an important role in their pronounced genetic diversity and horizonal acquisition of AMR determinants [6,[12][13][14][15][16].However, several studies reveal a variability in the transformation capacity among Campylobacter strains [16][17][18][19].Notably, in previous studies from our laboratory, C. coli strains from turkeys were significantly more prone to be transformed to erythromycin resistance than C. coli from swine, and certain C. coli strains appeared completely unable to acquire either erythromycin or nalidixic acid resistance via transformation [20].Inability to undergo transformation may reflect either the absence of determinants required for transformation [21] or the presence of impediments to transformation, such as restriction/modification systems, DNases encoded by integrated elements, and other barriers to gene flow [12,13,17,19].Several genes have been shown to affect natural competence.The Campylobacter cts genes, which encode a DNA uptake system that is similar to type II secretion systems, have been shown to be essential for competence [21].Additionally, nucleases encoded by Campylobacter bacteriophages have been identified as key players in preventing natural competence [17,19].
Despite the importance of understanding competence variation, investigations related to the transformation-mediated transfer of AMR genes have predominantly involved a limited number of extensively characterized strains [17,18,[21][22][23].In particular, studies that reported a variation in competence of C. jejuni and C. coli largely lack cross-referenceable information on the genotypes or clonal groups of strains that appear unable to be transformed [16][17][18]20].Bridging this knowledge gap is important to further elucidate the role of competence in C. jejuni and C. coli in shaping the incidence of AMR and overall genomic diversity in different clonal groups.
In this study, we investigated the transformability of a panel of C. jejuni and C. coli strains isolated from domestic food animals in the in the United States (U.S.).From this strain panel, we identified several non-transformable C. jejuni and C. coli strains.Strikingly, we found that all the C. jejuni ST-50 strains in this strain panel were nontransformable.By analyzing the genomic sequences of these strains, we identified at least two potential mechanisms leading to non-transformability: (1) a defective cts-encoded DNA uptake system, or (2) the presence of a prophage-encoded nuclease.Evaluating C. jejuni ST-50 strains globally, we found that European and North American strains possessed considerably different prevalence of cts mutations, but comparable levels of phage-encoded DNases.Finally, our phylogenetic analysis points to the evolution of a clade of ST-50 strains throughout the U.S. which harbor ctsD and ctsE double mutations.

Bacterial Strains and Growth Conditions
The C. jejuni and C. coli strains investigated for transformability in this study are listed in Table S1.All strains with the FSIS designation were kindly provided by Dr. Glenn Tillman and Mustafa Simmons, United States Department of Agriculture (USDA), Food Safety Inspection Service (FSIS).The strains were chosen to be from diverse food animal hosts and to lack resistance to multiple antimicrobials.Several of the strains lacked known AMR determinants and were considered pan-susceptible, while some were resistant to one or maximally two antimicrobial classes (Table S1).The strains were grown microaerobically at 42 • C on Mueller-Hinton agar (MHA), consisting of Mueller-Hinton broth (MHB) (Becton, Dickinson and Co., Franklin Lakes, NJ, USA) amended with 1.2% agar.Microaerobic conditions were established using a Bactrox Hypoxia Chamber (Shel Lab, Cornelius, OR, USA).

Determination of the Transformation and Mutation Frequencies
The transformation assays were determined as previously described [24].Briefly, DNA was extracted from donor strains C. jejuni 14980A and C. coli 14983A using the Qiagen DNeasy Blood and Tissue Kit (Qiagen, Redwood City, CA, USA), and was used for transformation of C. jejuni and C. coli, respectively, to nalidixic acid resistance (Nal R ), or to streptomycin resistance (Str R ) when the recipient was already Nal R [25].The transformation recipients were grown on MHA overnight.A colony from the MHA plate was suspended in 1 mL MHB, and 25 µL of this suspension was spotted onto a new MHA plate.Donor DNA (approx.700 ng) was added to each recipient suspension spot, and the spots were allowed to dry in a NuAire biosafety cabinet (Class 2, type A/B3; NuAire, Plymouth, MN, USA) prior to being incubated microaerobically at 42 • C overnight.For enumerations of transformants, each spot was suspended in 500 µL MHB, diluted as needed, and the dilutions spotted (10 µL, in duplicate) or spread-plated on MHA were amended with either nalidixic acid (32 µg/mL) or streptomycin (64 µg/mL).Colonies were enumerated following incubation at 42 • C microaerobically for 36-48 h.The total CFUs of the recipient at the end of the transformation period were also enumerated by plating the dilutions on MHA without nalidixic acid or streptomycin.The transformation frequency was determined as the ratio of total CFUs on MHA with the appropriate antibiotic over the total CFUs on MHA without any antibiotics.To determine the spontaneous mutation frequency of resistance to the relevant antimicrobial, cell suspensions from recipient spots with no added DNA were diluted and plated on MHA with either nalidixic acid (32 µg/mL) or streptomycin (64 µg/mL).The mutation frequency was calculated as the ratio of total CFUs in the control spot without added DNA on MHA supplemented with the appropriate antibiotic over the total CFUs on MHA with no antibiotic.All transformation assays included as a control C. jejuni Cj0264c::Cm, a derivative of C. jejuni NCTC 11168, that was confirmed to be readily transformable to Nal R and Str R using DNA from the donors employed here [24].

Bacterial Strain Genomic Sequences
The genomes for the FSIS strains in Table S1 were obtained from BioProject PRJNA292668, available at NCBI (https://www.ncbi.nlm.nih.gov/bioproject/PRJNA292668/,accessed on 2 January 2024); the associated GenBank assembly accession numbers are also in Table S1.For additional genomic sequences obtained for this study, DNA was extracted from C. jejuni strains RM3405, RM3412, RM5146, RM5148, RM5149 and RM5156, as described previously [26].DNA sequencing libraries for Illumina MiSeq were prepared using an Illumina DNA Prep Tagmentation kit (Illumina, San Diego, CA, USA), following the manufacturer's instructions, except for changes that increased library insert size.To do this, the library preparation was done to a median of 631 bp and a size range between ~375 and 1100 bp by decreasing the 1st and 2nd volume of Sample Purification Beads to 40 µL and 11 µL, respectively.This modification resulted in larger inserts compared to the mostly <300 bp inserts obtained using the manufacturer's protocol.The final elution volume of the libraries was in 10 µL of the Illumina resuspension buffer.Illumina-DNA/RNA UD Indexes Plate A, B, C and D dual index adapters were ordered from Integrated DNA Technologies (Coralville, IA, USA) and used at a 1 µM final concentration.Instead of pooling equal volumes, individual libraries were quantified using the KAPA Library Quantification Kit (Roche, Indianapolis, IN, USA), since we found qPCR to be a more accurate quantification than using equal volume.Libraries were quantified in 10 µL volume reactions and 90 s annealing/extension PCR, and then pooled and normalized to 4 nM.The pooled libraries were re-quantified using ddPCR on a QX200 system (Bio-Rad, Hercules, CA, USA) using the Illumina TruSeq ddPCR Library Quantification Kit and following the manufacturer's protocols.Libraries were sequenced using a MiSeq Reagent Kit v2 (500-cycles; Illumina) on a MiSeq instrument (Illumina) at 16 pM, following the manufacturer's protocols.Genomes were assembled using a SPAdes assembler 3.15.5 within Geneious Prime v. 2023.1.2(Biomatters, Ltd., Auckland, New Zealand).The GenBank accession numbers for these genome sequences are also listed in Table S1.
For the Australian C. jejuni ST-50 strains [28], SRA reads (Table S3) were downloaded from NCBI and assembled using SPAdes assembler 3.15.5 within Geneious Prime v. 2023.1.2(Biomatters, Ltd., Auckland, New Zealand).The cts genes and the CJIE1, CJIE2 and CJIE4 nuclease genes mentioned above were identified using the annotation function within Geneious Prime v. 2023.1.2.

DNase Detection
Cultures were started on anaerobe basal agar plates (Oxoid, Thermo Fisher Scientific, Waltham, MA, USA), amended with 5% laked horse blood (Innovative Research, Novi, MI, USA) and incubated for 24-48 h at 37 • C microaerobically (using a proprietary blend of gases (Bioblend, Linde USA, Danbury, CT, USA) consisting of 10% CO 2 , 10% H 2 , and 80% N 2 to adjust the final concentration of O 2 to 3-5%) in an Oxoid 3.5 L AnaeroJar (Oxoid, Thermo Fisher Scientific, Waltham, MA, USA).Methyl green-DNA plates were prepared by adding methyl green dye (Thermo Fisher Scientific; final concentration = 50 mg L −1 ) and fish sperm DNA (Fisher Scientific; final concentration 2 g/L) to blood-free anaerobe basal agar.Cultures were spotted thickly onto methyl green-DNA media and incubated in flat = bottomed freezer bags (Ziploc, S.C. Johnson, Racine, WI, USA) at the same O 2 level.The presence of DNase was indicated by the appearance of a clear halo around the culture.

Determination of Phylogeny
A total of 63 ST-50 strains were examined phylogenetically (Table S4).The genomes of six strains (RM3405, RM3412, RM5146, RM5148, RM5149 and RM5156) sequenced for this study were uploaded to PubMLST and sequence typed to the seven gene multilocus sequence type (MLST) and the core genome MLST in silico [9].Genomes for 22 FSIS strains in Table S1 were available in PubMLST, and 41 additional ST-50 genomes in the PubMLST database were used for comparison.A multisequence alignment was created from the concatenated gene sequences of all core genes (found in 100% of isolates) shared with NCTC 11168 (AL111168) using MAFFT [29].Dendrograms were created using the neighbor-joining method.
Additionally, the 63 ST-50 strains (Table S4) were compared phylogenetically to the Australian ST-50 strains (Table S3).The assembled genomes of the 23 Australian ST-50 strains (Table S3) were uploaded to PubMLST.A neighbor-joining dendrogram was created using the Interactive Tree of Life (iTOL) plugin within PubMLST using the concatenated nucleotide sequences of the C. jejuni/C.coli cgMLSTv2 datasets for each genome.

Transformation Frequencies of C. jejuni and C. coli
An initial analysis of a panel of 25 sequenced FSIS C. jejuni and C. coli strains with few or no AMR determinants and derived from poultry processing plants in the United States (US) identified 13 that appeared to be non-transformable to Nal R , including both available C. jejuni strains of ST-50.To determine the prevalence of non-transformability in ST-50, additional ST-50 strains from diverse food animals were requested from the FSIS.Interestingly, the testing of 36 additional FSIS C. jejuni ST-50 strains from poultry, cattle and swine in the US revealed that all were similarly non-transformable (Table 1).Altogether, we examined the natural transformability of 73 Campylobacter isolated from chickens, cattle, and sheep across multiple locations in the U.S. by the USDA and FSIS from 2016 to 2019 (Table S1).This panel of strains consisted of 54 C. jejuni strains, representing six MLST sequence types and 19 C. coli strains, representing 5 sequence types (Table 1).Each of the C. jejuni and C. coli strains were used as recipient for donor DNA from the Nal R , Str R strains C. jejuni 14980A and C. coli 14983A, respectively.Certain recipient strains were already Nal R , and these were only transformed to Str R .Since Nal R and Str R strains can also arise by spontaneous point mutations within gyrA and rpsL, respectively, we performed control transformations without any donor DNA added to determine the relevant mutation frequencies (Table 1).The strains were determined to be transformable when the transformation frequency was three logs or higher above the mutation frequency.There were 9 of 18 C. coli strains that were not transformable and 46 of 55 C. jejuni strains that were not transformable, including all 38 ST-50 strains.Since 100% of the C. jejuni ST-50 strains that we examined from the U.S. were not transformable, we investigated the natural competence of six other ST-50 strains available in our collection and isolated outside of the U.S, including four from Italy and two from Canada (Table S1).Among these six ST-50 strains, one (RM5156, from Italy) was identified as transformable.

Multiple Mechanisms for Non-Transformability among ST-50 Strains
To determine whether the lack of transformability reflected the presence of specific restriction/modification systems harbored by the recipient strains, genomic DNA was extracted from spontaneous Nal R mutants identified in the control transformation assays.This DNA was then utilized as donor DNA for transformation of the corresponding parental, nalidixic acid-susceptible recipient strain to Nal R .Without exception, the recipient stains remained non-transformable.
To understand the mechanism(s) for why some strains were not naturally transformable, we queried the whole genome sequences of these strains against loci that are known to affect natural competence, specifically mutations within cts genes and the presence of bacteriophage with nuclease genes (Table 2).All of the transformable C. jejuni and C. coli strains in the panel possessed wild-type cts genes.Among the 46 C. jejuni strains that were not transformable, 19 strains had nonsense or frameshift mutations within cts genes that would prematurely disrupt the coding sequence and should disrupt the Cts DNA uptake system (Table 2).From these 19 strains, 18 were ST-50 and 15 of these ST-50 strains possessed the same double mutations within ctsD and ctsE.Strain RM3405 from Canada had nonsense mutations in ctsR and ctsF.There were also 27 non-transformable C. jejuni strains that possessed wild-type cts genes, suggesting multiple mechanisms for the non-transformable phenotype among C. jejuni strains.Among these 27 C. jejuni strains, all but one (RM3412) possessed at least one bacteriophage with a nuclease gene.Indeed, 26 ST-50 strains possessed a Mu-like bacteriophage that carried dns, which encodes an extracellular DNase.In five ST-50 strains (FSIS12030287, FSIS12030565, FSIS12031145, FSIS1608758, and FSIS1609357) and one ST-353 strain (FSIS11810577), dns possessed frameshift mutations.However, the six strains mentioned above also possessed double mutations within ctsD and ctsE and were not transformable, while the ST-353 strain with a dns nonsense mutation had wild-type cts genes and was transformable (Tables 1 and 2).The four ST-50 strains from Italy had wildtype cts genes and Mu-like bacteriophages.Among these, the one transformable strain (RM5156) harbored a nonsense mutation in dns.Bacteriophages harboring nuclease genes similar to CJE0566 and CJE1444, defined here as dns2 and dns3, respectively (Table 2), were also harbored by C. jejuni strains that were not naturally transformable.Finally, RM3412, a ST-50 strain from Canada, lacked the discernable genetic content that leads to non-transformability (Tables 1 and 2).
In the case of C. coli, all non-transformable strains possessed functional cts genes.Eleven C. coli strains were ST-3262 and eight of these were not naturally transformable.All of the ST-3262 strains harbored a Mu-like bacteriophage and CJIE2-like bacteriophage.However, the Mu-like bacteriophage did not harbor dns, while the CJIE2-like bacteriophage did possess dns2.The three transformable ST-3262 strains had nonsense mutations within dns2 of the CJIE2-like bacteriophage (Table 2).

Global Examination of the ST-50 Non-Transformability Genotype
With 42/43 non-transformable C. jejuni ST-50 strains having a likely genetic basis for the phenotype, our investigation extended to the analysis of over 2200 C. jejuni ST-50 genomes in the PubMLST database.The search focused on identifying cts mutations or the presence of bacteriophage-encoded DNase genes.The criteria for cts genes involved assessing wild-type and loss-of-function mutational states requiring the presence of the complete ctsRDPXEF operon on a single contig and the complete coding sequences of ctsG, ctsT, and ctsW.Similarly, for bacteriophage-encoded DNase assessments, the criteria included the complete coding sequences of dns (the Mu-like phage DNase gene), dns2 (CJIE2-like) and dns3 (CJIE4-like) needed to be on single contigs with frameshift and nonsense mutations indicating a nonfunctional gene.Missense mutations were disregarded, as their functional outcome was difficult to assess and beyond the scope of this study.
Table 3 illustrates that 37.8% of ST-50 strains from North America possessed cts gene mutations, while only 3.3% of European ST-50 strains exhibited such mutations.The presence of Mu-like bacteriophage dns was determined with 53.3% and 64.3% of strains possessing the gene in North America and Europe, respectively.The dns2 and dns3 genes from CJIE2 and CJIE4 were detected at lower levels than the Mu-like bacteriophage dns in both regions, with 30.3% in North America and 28.4% in Europe.Some strains featured multiple mechanisms for non-transformability, exemplified by strain FSIS1710700 with ctsD ctsE double mutations and dns, and strain FSIS1607146 with dns and dns3 (Table 2).Calculations based on presence of cts mutations, dns, dns2 and dns3 in ST-50 strains from North America and Europe revealed an approximate percentage of non-naturally transformable ST-50 strains to be 91% in North America and 75% in Europe.We also examined the genomes from 23 C. jejuni ST-50 isolated from chickens in Australia [28].We assembled these genomes from the available SRA reads in NCBI, since they were not available in PubMLST (Table 1).We found cts mutations in 8 out of 23 strains (35%), a level similar to the prevalence of cts mutations in C. jejuni ST-50 in the U.S. (Table 3).In the Australian strains, loss-of-function mutations were exclusively in a single gene, ctsE.Phage-encoded DNase genes were also identified in these strains, with dns present in five of 23 (~22%) and dns2 present in 12 out of 23 (~52%) (Supplementary Table S1).Overall, 17 out of the 23 (~74%) strains harbored genetic markers for non-transformability.
To assess whether the findings regarding non-transformability of C. jejuni ST-50 were unique to this clone, we examined the genetic markers for transformability for over 800 C. jejuni ST-8 strains to determine if non-transformability was common to C. jejuni strains in North America.C. jejuni ST-8 strains are in the same clonal complex, CC-21, as ST-50, and they are predominantly distributed in North America.For cts mutations, only 641 genome sequences harbored the operon ctsRDPXEF on a single contig.We identified that 49 of these 641 (~8%) harbored loss-of-function cts mutations.Phage-encoded DNase genes were also identified in these strains, with dns present in 62 out of 809 strains (~8%) and dns2 or dns3 present in 237 out of 809 (~29%) (Table 3).Overall, there were 641 ST-8 strains that we could examine where all genetic markers for transformability could be assessed.Collectively, our analysis indicated that 218 (~34%) of these 641 ST-8 strains harbored genetic markers for non-transformability. (Table S4).

Evidence of Extracellular DNase from dns Positive Strains
Four ST-50 strains (RM5146, RM5148, RM5149 and RM5156) possessing Mu-like bacteriophages with dns were examined for extracellular DNase activity.Three strains (RM5146, RM5148 and RM5149) were non-transformable and had wild-type dns genes.All three strains showed DNase activity.Figure 1 illustrates the DNase activity of RM5146 compared to C. jejuni strain RM1221, which also possesses dns, and strains NCTC 11168 and 81116 that do not have the dns gene.The DNase activity of strain RM5146 appeared noticeably higher than strain RM1221, while no activity was detected for strains NCTC 11168 and 81116 (Figure 1).The Campylobacter jejuni strain RM5156 with mutated dns showed no DNase activity.

Phylogenetic Comparison of North American and European ST-50 Strains
The phylogeny of 63 ST-50 strains (Table S5) was inferred from their core genomes ((genes present in all strains), using the neighbor-joining method to create a dendrogram (Figure 2).The strain collection included 30 from North America, 29 from Europe, and two each from Asia and Oceania.These strains were isolated from a variety of sources including humans, chickens, pigs, cattle and sheep.The dendrogram shows that North American ST-50 strains mostly form clusters with other North American ST-50 strains.Conspicuously, there was a large cluster of strains from the U.S. that all possessed double mutations within ctsD and ctsE (Figure 2, red-boxed cluster).The strains in this cluster were from human clinical and multiple animal sources (chicken, cattle, pig and wild bird).Moreover, the C. jejuni ST-50 strains possessing the ctsD and ctsE mutations were not confined to a particular region within the U.S., but were isolated from animals from the southeastern seaboard to the west coast of the U.S. (Table S5).Among the European ST-50 strains, 8 (PubMLST IDs 22255, 32893, 51553, 61660, 70943, 76581, 77678, 79837) possessed wild-type cts genes and no dns, dns2, or dns3.The extent to which these strains were nontransformable remains unknown, as they were not available for experimental assessments.These eight strains did not cluster together, but were dispersed throughout the dendrogram (Figure 2, see arrows, <).Finally, we performed a phylogenic analysis that included the 23 Australian ST-50 strains that were previously shown to be phylogenetically distinct from most North American and European ST-50 strains [28]

Phylogenetic Comparison of North American and European ST-50 Strains
The phylogeny of 63 ST-50 strains (Table S5) was inferred from their core genomes ((genes present in all strains), using the neighbor-joining method to create a dendrogram (Figure 2).The strain collection included 30 from North America, 29 from Europe, and two each from Asia and Oceania.These strains were isolated from a variety of sources including humans, chickens, pigs, cattle and sheep.The dendrogram shows that North American ST-50 strains mostly form clusters with other North American ST-50 strains.Conspicuously, there was a large cluster of strains from the U.S. that all possessed double mutations within ctsD and ctsE (Figure 2, red-boxed cluster).The strains in this cluster were from human clinical and multiple animal sources (chicken, cattle, pig and wild bird).Moreover, the C. jejuni ST-50 strains possessing the ctsD and ctsE mutations were not confined to a particular region within the U.S., but were isolated from animals from the southeastern seaboard to the west coast of the U.S. (Table S5).Among the European ST-50 strains, 8 (PubMLST IDs 22255, 32893, 51553, 61660, 70943, 76581, 77678, 79837) possessed wild-type cts genes and no dns, dns2, or dns3.The extent to which these strains were non-transformable remains unknown, as they were not available for experimental assessments.These eight strains did not cluster together, but were dispersed throughout the dendrogram (Figure 2, see arrows, <).Finally, we performed a phylogenic analysis that included the 23 Australian ST-50 strains that were previously shown to be phylogenetically distinct from most North American and European ST-50 strains [28] (Supplementary Table S1).This analysis verified that most Australian ST-50 strains form a distinct clade, with the inclusion of one strain from Europe (Figure S1).
(Supplementary Table S1).This analysis verified that most Australian ST-50 strains form a distinct clade, with the inclusion of one strain from Europe (Figure S1).S1) was constructed from the concatenated, aligned core genes.Bootstrap values of ≥50%, generated from 500 replicates, are shown at the nodes.Leaves are labeled with PubMLST IDs, and the strain information can be found in Table S5.Blue arrowheads indicate strains that lacked the investigated non-transformability genetic markers.Metadata related to continent (North America: NA, Europe: EU, Asia: AS, and Oceania: OC) source; cts mutations are adjacent to each leaf.The red rectangle indicates the cluster of strains that share the ctsD and ctsE mutations.

Discussion
In this study, we identified both non-transformable C. jejuni and C. coli strains isolated from domestic food animals.We employed genomic analysis to understand the mechanisms associated with Campylobacter strains that are not naturally transformable.We have established the genetic associations leading to a non-transformable state, including the presence of functional DNase genes possessed by bacteriophages in both C. jejuni and C. coli strains, which have been previously observed in some C. jejuni strains [17,19].S1) was constructed from the concatenated, aligned core genes.Bootstrap values of ≥50%, generated from 500 replicates, are shown at the nodes.Leaves are labeled with PubMLST IDs, and the strain information can be found in Table S5.Blue arrowheads indicate strains that lacked the investigated non-transformability genetic markers.Metadata related to continent (North America: NA, Europe: EU, Asia: AS, and Oceania: OC) source; cts mutations are adjacent to each leaf.The red rectangle indicates the cluster of strains that share the ctsD and ctsE mutations.

Discussion
In this study, we identified both non-transformable C. jejuni and C. coli strains isolated from domestic food animals.We employed genomic analysis to understand the mechanisms associated with Campylobacter strains that are not naturally transformable.We have established the genetic associations leading to a non-transformable state, including the presence of functional DNase genes possessed by bacteriophages in both C. jejuni and C. coli strains, which have been previously observed in some C. jejuni strains [17,19].Additionally, we identified C. jejuni strains possessing cts mutations that would disrupt the DNA uptake system involved in natural transformation [21,22].
An especially interesting and unexpected finding was the non-transformability of C. jejuni ST-50, a major, globally-distributed clone with intriguing phylogeny [28].Among the Campylobacter strain panel investigated for transformation capacity, we identified that all but one (43/44) strains of C. jejuni ST-50 were non-transformable.The sole ST-50 strain that was found to be transformable was from Italy and did not harbor the established genetic markers.We further determined that globally >75% of approximately 2400 C. jejuni ST-50 strains from the PubMLST database would be expected to be non-transformable, as they possess the genetic markers associated with the non-transformable phenotype.The remaining ST-50 strains not harboring these genetic markers may be non-transformable via alternative mechanisms not investigated in this study, or they may be transformable.Thus, we defined the genetic markers for a non-transformable genotype among field isolates of C. jejuni ST-50 and other non-transformable field isolates of C. jejuni and C. coli, specifically the presence of functional bacteriophage DNase genes or mutations within cts genes.
These genetic markers harbored by non-transformable C. jejuni and C. coli strains are expected to have a negative impact on the initial steps of DNA uptake during natural transformation.The natural competence for transformation of C. jejuni with naked chromosomal DNA was first detailed in 1990 [16].The identification of cts genes as encoding components of a DNA uptake apparatus was elucidated more than a decade later [21].We identified nonsense and frameshift mutations in ctsD, ctsE, ctsR and ctsF.These mutations cause prematurely truncated components and result in a nonfunctional DNA uptake system, as previously observed following insertion mutations within the same genes [21,22].The involvement of the bacteriophage DNases in natural competence was reported for three different bacteriophage families, CJIE-1-like bacteriophage (Mu-like phage), CJIE2-like bacteriophage and CJIE4-like bacteriophage [17,19].We identified the non-transformable strains possessing dns from the Campylobacter Mu-like phage and dns2 or dns3 from CJIE2 and CJIE4 bacteriophages, respectively.Additionally, we observed that strains with lossof-function mutations in dns and dns2 were transformable, further supporting the role of dns and dns2 in non-transformability.The impact of these DNases on natural transformation would be to degrade the naked donor DNA prior to its interaction with the DNA uptake system or once in the cytoplasm nd before recombination.We demonstrated that the strains possessing dns degraded extracellular DNA.Further research is needed to determine mechanisms via which dns2 or dns3 inhibit transformation.
The ability to acquire and recombine DNA between strains of C. jejuni and C. coli plays a significant role in the genetic diversity of C. jejuni and C. coli, including the movement of chromosomal AMR markers in these species [9,16,30].Our study along with others [20,24] establishes that the transfer of AMR determinants through natural transformation is markedly more frequent, by several orders of magnitude, than the occurrence of mutations leading to resistance.Although most natural transformation studies are performed in vitro using purified donor DNA [16,20,24], it is noteworthy that natural transformation has been demonstrated in vivo within chickens and turkeys, which are prevalent hosts for this pathogen [31,32].
All 38 experimentally-investigated C. jejuni ST-50 strains isolated from domestic food animals by the U.S. Department of Agriculture Food Safety Inspection Service exhibited the non-transformable phenotype and corresponding associations with the cts and DNase non-transformability genetic markers.Strikingly, the non-transformable genotypes varied among the C. jejuni ST-50 strains that we investigated.Among the entire panel (n = 44, including 38 from the US and six from elsewhere) strains of C. jejuni ST-50 that were experimentally assessed for transformation, 9 harbored cts mutations only, 14 possessed a complete dns only, 5 had a complete dns2 only, 2 harbored a complete dns3 only, and 12 harbored multiple markers.One strain was non-transformable but lacked the investigated genetic markers.In this case, it is possible that a cts missense mutation impaired the DNA uptake system, or that the strain possesses a novel DNase that caused the nontransformable phenotype.Intriguingly, as indicated above several of these ST-50 strains possessed more than one genetic marker, i.e., cts mutations as well as a complete dns.
Upon examining these non-transformable genotypes in a broader context within the global set of C. jejuni ST-50, a discernible pattern emerged.Specifically, cts loss-of-function mutations, particularly the ctsD and ctsE double mutations, appear to have evolved and dispersed throughout the agricultural environment within the U.S.These mutations seem to have occurred independent of the bacteriophage DNases, which are present at high levels in both North American and European C. jejuni ST-50 strains.Our phylogenetic results examining a global set C. jejuni ST-50 strains support the independent evolution of the C. jejuni ST-50 possessing the ctsD and ctsE double mutations in the U.S.This observation aligns with a study by Wallace et al. [28], who performed a similar phylogenetic analysis on 162 C. jejuni ST-50 isolates, comparing those from Australia with other parts of the world.Wallace et al. [28] highlighted the independent evolution of Australian ST-50 strains, which formed a distinct cluster from ST-50 isolates from Europe and North America.Inclusion of the Australian ST-50 strains not only verifies the findings of this previous study [28], but demonstrates that the evolution of strains in the U.S. and Australia is truly independent.
As mentioned above, we have identified that the genetic markers for non-transformability vary in occurrence.Notably, cts mutations were much more common in strains from North America and Australia (~37%) when compared to those from Europe (~3%).However, the prevalence of the different bacteriophages was comparable between North America and Europe, whereas the Mu-like phage was present in >50% of the ST-50 strains from the U.S. and Europe but in <25% of the ST-50 strains from Australia.Nevertheless, the overall prevalence of the non-transformable genotypes was ~75% in Europe and Australia and ~90% in North America.Finally, we determined that the prevalence of genetic markers associated with non-transformability within C. jejuni ST-8 strains was ~34%, demonstrating that nontransformability is not common to all C. jejuni strains in North America.

Conclusions
Campylobacter strains with reduced natural transformation frequencies were frequently observed in a panel of C. jejuni and C. coli strains from domestic food animals in the U.S and exhibited resistance to no or few antimicrobials (maximally two antimicrobial classes).Genomic characterization of the non-transformable strains allowed us to identify specific genomic markers/genotypes that likely accounted for non-transformability, including (1) loss-of-function mutations in the cts genes encoding a DNA uptake system, and/or (2) the presence of bacteriophage-encoded DNases.Notably, these non-transformabilityassociated genomic markers were detected, individually or in combination, in all but 1 of the 43 non-transformable C. jejuni ST-50 strains on our panel.Examining the prevalence of non-transformable genotypes of ST-50 strains from different regions of the world, we observed distinct patterns of evolution, particularly in the agricultural environment of the U.S. The findings suggest strong selection pressures for the evolution and maintenance of the non-transformable state in ST-50, a major, globally distributed clone of C. jejuni.The potential impacts of non-transformability on the natural fitness, and possibly the pathogenicity, of this clone remain to be elucidated.Moreover, our phylogenetic analysis supports the independent evolution of specific non-transformability-associated genotypes within C. jejuni ST-50 strains, possibly reflecting the independent evolutionary pathways in different geographic locations described by Wallace et al. [28].This study provides a deeper understanding of the genetic mechanisms involved in the natural transformation of Campylobacter and highlights the complex interplay between natural transformation and bacterial evolution, genetic diversity, and AMR dissemination.

Figure 1 .
Figure 1.Extracellular DNase activity of C. jejuni.Methyl green-DNA media were inoculated with four different strains C. jejuni: RM1221 (dns + ), NCTC 11168 (dns − ), RM5146 (dns + ), and 81116 (dns − ).The presence of DNase was indicated by the appearance of a clear halo beneath and around the culture.

Figure 1 .
Figure 1.Extracellular DNase activity of C. jejuni.Methyl green-DNA media were inoculated with four different strains C. jejuni: RM1221 (dns + ), NCTC 11168 (dns − ), RM5146 (dns + ), and 81116 (dns − ).The presence of DNase was indicated by the appearance of a clear halo beneath and around the culture.

Figure 2 .
Figure 2. Neighbor-joining dendrogram of C. jejuni ST-50 strains.The neighbor-joining dendrogram of 63 C. jejuni ST-50 strains (Supplementary TableS1) was constructed from the concatenated, aligned core genes.Bootstrap values of ≥50%, generated from 500 replicates, are shown at the nodes.Leaves are labeled with PubMLST IDs, and the strain information can be found in TableS5.Blue arrowheads indicate strains that lacked the investigated non-transformability genetic markers.Metadata related to continent (North America: NA, Europe: EU, Asia: AS, and Oceania: OC) source; cts mutations are adjacent to each leaf.The red rectangle indicates the cluster of strains that share the ctsD and ctsE mutations.

Figure 2 .
Figure 2. Neighbor-joining dendrogram of C. jejuni ST-50 strains.The neighbor-joining dendrogram of 63 C. jejuni ST-50 strains (Supplementary TableS1) was constructed from the concatenated, aligned core genes.Bootstrap values of ≥50%, generated from 500 replicates, are shown at the nodes.Leaves are labeled with PubMLST IDs, and the strain information can be found in TableS5.Blue arrowheads indicate strains that lacked the investigated non-transformability genetic markers.Metadata related to continent (North America: NA, Europe: EU, Asia: AS, and Oceania: OC) source; cts mutations are adjacent to each leaf.The red rectangle indicates the cluster of strains that share the ctsD and ctsE mutations.

Table 1 .
Campylobacter strains used for natural transformation and mutation frequencies.

Table 2 .
Campylobacter genetic markers affecting natural competence.