Received: 11 April 2022 | Revised: 28 June 2022 | Accepted: 17 August 2022 DOI: 10.1111/zph.12994 ORIGINAL ARTICLE Host-­associated variability of the cdtABC operon, coding for the cytolethal distending toxin, in Campylobacter jejuni Pedro Guirado1 | Yaidelis Iglesias-­Torrens2,3 | Elisenda Miró2 | Ferran Navarro2,3 | Camile Stephan-­Otto Attolini4 | Carlos Balsalobre1 | Cristina Madrid1 1 Departament de Genètica, Microbiologia i Estadística, Facultat de Biologia, Abstract Universitat de Barcelona, Barcelona, Spain Campylobacter, a major cause of food-­borne gastroenteritis worldwide, colonize the 2 Hospital de la Santa Creu i Sant Pau and Institut d'Investigació Biomèdica Sant Pau gastrointestinal tract of a wide range of animals, being birds the main reservoir. The (IIB Sant Pau), Barcelona, Spain mechanisms involved in the interaction of Campylobacter with the different hosts are 3 Departament de Genètica i poorly understood. The cytolethal distending toxin, encoded in the cdtABC operon, Microbiologia, Universitat Autònoma de Barcelona, Barcelona, Spain is considered a pivotal virulence factor during human infection. Differences in the 4 Institute for Research in Biomedicine prevalence of cdtABC genes in Campylobacter isolates from three distinct origins (wild (IRB Barcelona), The Barcelona Institute of birds, broiler chickens and humans) prompted us to further characterize their allelic Science and Technology, Barcelona, Spain variability. The sequence of cdtABC is highly conserved among broiler and human Correspondence isolates. A high diversity of cdtABC alleles was found among wild bird isolates, in- Carlos Balsalobre and Cristina Madrid, Departament de Genètica, Microbiologia cluding several alleles that do not produce any functional CDT. These results suggest i Estadística, Universitat de Barcelona, that specific variants of the cdtABC operon might define the host range of specific Avda. Diagonal 643, Barcelona 08028, Spain. Campylobacter jejuni isolates. Moreover, our data indicate that PCR methodology is Email: cbalsalobre@ub.edu and inaccurate to characterize the prevalence of the cdt genes, since negative PCR detec- cmadrid@ub.edu tion can be the result of divergences in the sequence used for primer design rather Funding information than indicating the absence of a specific gene. Catalonian government; European Regional Development Fund; Spanish Ministry of Economy and KEYWORDS Competitiveness; Spanish Ministry of allelic variability, Campylobacter, cytolethal distending toxin, PCR detection, WGS Science, Innovation and Universities; State Bureau of Investigation 1 | I NTRO D U C TI O N Campylobacter can trigger the autoimmune polyneuropathic disor- der Guillain–­Barré syndrome (Koga et al., 2006). Campylobacter je- Campylobacter is a major enteropathogen worldwide, causing the juni is the species reported to cause most of the campylobacteriosis most common zoonoses reported in humans in the European Union, cases in humans (European Food Safety Authority and European the United States and Australia (European Food Safety Authority Centre for Disease Prevention and Control, 2021). and European Centre for Disease Prevention and Control, 2021; Campylobacter jejuni pathogenicity remains poorly understood. Kaakoush et al., 2015). In most cases, campylobacteriosis is self-­ Among the array of bacterial products that contribute to its viru- limiting gastroenteritis. Severe cases, although rare, occur mostly lence, the cytolethal distending toxin (CDT) seems to play a piv- among children younger than 5 years, elderly and immunocom- otal role in the interaction of Campylobacter with the host. CDT promised patients (Kaakoush et al., 2015). Moreover, infection by blocks cell division, causing cell cycle arrest at the G2 stage prior This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. © 2022 The Authors. Zoonoses and Public Health published by Wiley-VCH GmbH. 966 |  wileyonlinelibrary.com/journal/zph Zoonoses Public Health. 2022;69:966–977. GUIRADO et al. | 967 to mitosis (Whitehouse et al., 1998). CDT is a heteromeric AB-­t ype genotoxin produced by several Gram-­n egative bacterial patho- Impacts gens, including Campylobacter, enteropathogenic Escherichia coli, • The sequence of cdtABC operon, encoding the cytole- Salmonella enterica serovar Typhi, Haemophilus duceryi and Shigella thal distending toxin in Campylobacter jejuni, is highly dysenteriae (Jinadasa et al., 2011). The toxin consists of three sub- conserved among broiler and human isolates. Instead, a units CdtA, CdtB and CdtC. CdtB is the active subunit whereas high diversity is found among wild bird isolates, includ- CdtA and CdtC are required for binding and delivery of CdtB ing several alleles that do not produce any functional within the target cell by a process that is still poorly understood CDT. (Lara-­Tejero & Galán, 2001). CdtA/CdtC binds to the lipid rafts, • Sequence analysis reveals that the cdtABC operon cholesterol-­r ich microdomains within the host cell membranes (Lai seems to accumulate DNA modifications, particularly et al., 2016). CdtB has DNaseI-­like activity producing DNA double-­ deletions, especially detected among wild bird isolates. strand breaks in the host cell that causes arrest of the cell cycle, • The data obtained indicate that PCR methodology is cellular distension and cell death (Smith & Bayles, 2006). Although inaccurate to characterize the prevalence of the cdt CDT is considered a crucial virulence factor during human infec- genes, since negative PCR detection can be the result tion, it remains unclear its role during the colonization of natural of divergences in the sequence used for primer design hosts, such as birds. rather than indicating the absence of the specific nu- Campylobacter CDT toxin is encoded in the polycistronic op- cleotide sequences of a specific gene. eron cdtABC. Most Campylobacter epidemiological studies reporting prevalence of genes coding for potential virulence factors, including CDT, are based in the use of PCR for gene detection. Very dispa- rate data on cdtABC prevalence among C. jejuni isolates from human 2.2 | PCR amplification patients suffering of gastroenteritis has been reported using this methodology. Some studies indicate a high prevalence of the cdtA, Genomic DNA was extracted from cultures grown on CBA plates cdtB and cdtC genes, whereas other reports much lower prevalence, using the InstaGene Matrix Kit (Bio-­Rad Laboratories). PCR reac- ranging from 100% (50/50) to 15% (3/20; Bang et al., 2003; de tions (PCR Master Mix x2, ThermoScientific) were performed using Melo et al., 2021; Iglesias-­Torrens et al., 2018; Koolman et al., 2015; 35 ng of DNA as a template and the specific primers for amplification Pickett et al., 1996; Weis et al., 2014). of cdtA, cdtB and cdtC (Table 1) previously used in epidemiological The allelic variability of the cdtABC genes in Campylobacter iso- studies (Martínez et al., 2006) and based in the genomic sequence lates from three distinct origins (wild birds, broiler chickens and hu- of the reference C. jejuni strain 81–­176 Amplified PCR products mans) has been characterized. The present study aimed to evaluate from cdtB gene of representative isolates showing different electro- the diversity of cdt genes among the isolates from different origins, phoretic migration were purified using the E.Z.N.A Cycle Pure Kit and the possible role that this variability could play in Campylobacter (OMEGA Bio-­tek) and sequenced using the same primers as those host range specificity. used to generate the fragment. 2 | M ATE R I A L S A N D M E TH O DS 2.3 | Genome sequencing 2.1 | Bacterial strains and culture conditions Whole genome sequencing was performed for a selection of 47 C. jejuni isolates, including 12 isolates from broiler, 16 from human The C. jejuni strains used in this study are part of a strain collec- and 19 from wild birds. The selection was performed in order to tion, previously described (Iglesias-­Torrens et al., 2018), composed sequence isolates from different origins, sequence types virulence of isolates obtained from faeces of three different sources: human and antibiotic resistance profiles (Iglesias-­Torrens et al., 2018). The suffering symptomatic gastroenteritis (50), broiler chicken (50) and wild birds (50). Wild bird isolates were obtained from different spe- TA B L E 1 Primers used in this work cies: yellow-­legged gulls (Larus michahellis) from Barcelona (YLGB) and from Medes Islands (YLGM); Audouin's gulls (Larus audouinii) Primer Sequence 5′–­3′ PCR product (bp) from Ebro Delta (AGD) and Alboran Islands (AGA); feral pigeons cdtA Fw CTATTACTCCTATTACCCCACC 422 (Columba livia) from Barcelona (FP); common ravens (Corvus corax) Rv AATTTGAACCGCTGTATTGCTC from Barcelona (CR); white storks (Ciconia ciconia) from Lleida (WS); cdtB Fw AGGAACTTTACCAAGAACAGCC 531 and norther shoveler (Spatula clypeata) from Ebro Delta (NS). Isolates Rv GGTGGAGTATAGGTTTGTTGTC were cultured onto Columbia blood agar (CBA) plates (Sharlau) cdtC Fw ACTCCTACTGGAGATTTGAAAG 339 and incubated at 42°C for 48 hr in microaerophilic conditions Rv CACAGCTGAAGTTGTTGTTGGC (CampyGen, Oxoid). 968 | GUIRADO et al. DNA extraction was carried out with the Wizard DNA-­Purification Thirteen out of the fourteen cdtA-­negative isolates belonged kit (Promega). The DNA extracted was send to LifeSequencing (Parc to the clonal complex ST-­1,275 which has been described to be Científic, Universitat de València, Spain) to proceed with the DNA highly predominant among wild birds (Hughes et al., 2009; Sheppard library preparation using the Nextera XT DNA Preparation kit and et al., 2009). These isolates were collected from Audouin's gulls the NexSeq 500 platform. The BioProject has been deposited at and yellow-­legged gull from three distant location in the Spanish NCBI GenBank, accession number PRJNA850915. To find the cd- Mediterranean coast (Medes Islands, Ebro Delta, and Alboran tABC nucleotide sequences in each genome we looked for the prim- Island). The remaining cdtA-­negative isolate belonged to the single- ers used in previous amplifications by doing the following. We used ton ST-­8514 and it was collected from a common raven. Three out the “DNAStringSet” function from the Biostrings package (Pagès of four cdtAC-­negative isolates from Audouin's gulls from Ebro Delta et al., 2022) in the R (R Core Team, 2020) programming environment and Alboran Island also belonged to the CC ST-­1275. The last cdtAC-­ (v.3.0.2) to read the primers and genomes in R. The “vmatchPat- negative isolate belonged to the singleton ST-­996 and was collected tern” function was used to find all occurrences of the primers with at from a northern shoveler in Ebro Delta. The three cdtAB-­negative most 1 mismatch. The start and end positions and the corresponding belonged to the CC ST-­952 and were collected from common ra- strand were recorded. vens. The cdtC-­negative isolate belonged to the CC ST-­1034 and was collected from an Audouin's gull in Ebro Delta. Although all cdt gene negative isolates were collected from wild birds, it should be point 2.4 | Phylogenetic analysis out that C. jejuni isolates positive for the three cdt genes were also collected from wild birds at the same geographical locations. Our Evolutionary analyses were conducted with MEGA X (MEGA v10.1.8) data show that a high diversity in the prevalence of the cdt genes (Kumar et al., 2018). The evolutionary history was inferred using the was observed among C. jejuni isolates from wild birds when com- Neighbour-­Joining method (Saitou & Nei, 1987). The evolutionary pared with isolates from broilers and humans. divergence between cdtABC sequences was inferred by estimating When considering the ST-­complex determined for each isolate, the distance matrix via the Maximum Composite Likelihood method in most cases, the isolates belonging to a specific ST-­complex shared (Tamura et al., 2004). the same pattern of cdtABC prevalence (Figure 1). The only excep- tion was within the ST-­1275 complex which was the most repre- sented ST among wild birds (16 out of 50 isolates). All the ST-­1,275 2.5 | Ethics approval complex isolates were cdtA negative whereas only three of them were cdtC negative by PCR. No ethical approval was obtained because this study did not involve human subjects, animals or tissue samples and only involved non-­ invasive procedures. 3.2 | Different cdtB alleles detected among the wild bird isolates 3 | R E S U LT S A puzzling observation was done during the PCR genotyping analy- sis of the different cdt genes. Most isolates PCR-­positive for cdtB 3.1 | The prevalence of the cdtA, cdtB and cdtC depicted a band that migrated as expected for the canonical cdtB genes is highly variable among Campylobacter jejuni sequence of the reference strain 81–­176 (531 bp). However, some isolates from wild birds isolates showed a band that migrated more than expected, indicating that the PCR amplified band carried a shorter sequence (Figure 2a). The prevalence of genes coding for cdtA, cdtB and cdtC in a collec- These results suggested the presence of different cdtB alleles among tion of 150 isolates of C. jejuni from human suffering symptomatic distinct isolates. Hence, from the 46 cdtB-­positive wild bird isolates, grastroenteritis, broiler chiken and wild birds (see Materials and 17 apparently carried the shorter cdtB allele, representing 36.9% of methods for details) was determined by PCR. The results indicated the cdtB-­positive isolates. Sixteen isolates belonged to the CC ST-­ that the distribution of the cdt genes was very variable. The three cdt 1275 and the remaining isolate to the singleton ST-­8514 (Figure 2b). genes were found in all broiler chicken and human isolates. In con- Remarkably, all the isolates that carried a shorter cdtB allele were trast to what it was observed among human and broiler isolates, 22 PCR-­negative for cdtA. It should be noticed that 100% of the broiler out of 50 isolates from wild birds were negative for the detection of and human isolates carry a cdtB allele that by PCR amplification one or two of the cdt genes (Figure 1), and prevalence varies for each resembled the canonical allele of 81–­176. The amplicons from the gene, being 58%, 94% and 90% for cdtA, cdtB and cdtC, respectively reference strain 81–­176, the isolates W02, W54, carrying an appar- (Figure 1). The highest variability was detected for the cdtA gene. ently canonical allele; and W20, W25, W53, showing a shorter cdtB Hence, 14 isolates were cdtA negative, three isolates were cdtAB allele, were sequenced (Figure 2c).The amplicons from W20, W25 negative, four isolates were cdtAC negative and one was negative and W53 showed the same feature when compared to 81–­176 am- for cdtC. plicon, a deletion of 51 nucleotides, resulting in an amplicon of 480 GUIRADO et al. | 969 F I G U R E 1 Collection of Campylobacter jejuni isolates from human patients (H), broiler chicken (B) and wild birds (W). The strains are organized attending to the host and ST clonal complexes (ST-­CC). Sequence types are also indicated (ST). The species of the wild bird strains are specified (YLGB: yellow-­legged gulls from Barcelona; YLGM: yellow-­legged gulls from Medes Islands; AGD: Audouin's gulls from Ebro Delta: AGA: Audouin's gulls from Alboran Islands; FP: feral pigeons; CR: common ravens; WS: white storks; NS: northern shoveler). cdtA, cdtB and cdtC positive strains by PCR amplification are indicated in colour background (green for human isolates, red for broiler chicken isolates and blue for wild bird isolates). The different nucleotide deletions and insertions determined by WGS or PCR analysis are indicated in yellow. The shadowed strain names (clear green, pink and clear blue) indicate the strains that the complete cdtABC operon sequence have been determined. nucleotides (as compared with 531 bp amplicon from 81–­176), that 3.3 | Whole-­genome sequencing reveals a high explained the shorter band detected by PCR. From now, we will refer diversity in the cdtABC locus among wild bird isolates to this cdtB allele as cdtB480. Interestingly, in the genome of the strain 81–­176, the 51 bp deleted sequence from cdtB480 allele was flanked To further characterize the diversity within the cdtABC locus, by two 10 bp direct repeats (Figure 2c), suggesting that the cdtB480 whole-­genome sequencing (WGS) was performed for 47 C. jejuni allele was most probably generated by a site-­specific recombination isolates, including 12 isolates from broiler, 16 from human and 19 event. from wild birds. The sequence of the cdtABC locus was characterized 970 | GUIRADO et al. F I G U R E 2 Identification of the cdtB480 allele. (a) PCR amplification of cdtB gene from strains 81–­176 (lane 1), W50 (lane 2), W54 (lane 3) and W54 (lane 4). The amplification of the housekeeping gltA gene was used as internal control. (b) Isolates that carry the cdtB480 allele. (c) Sequences obtained from the PCR amplification of strains carrying the cdtB81-­176 or the cdtB480 allele. Shadow square indicates the 10 bp direct repeats flanking the 51 bp deletion. (Figure S1, Appendix S1). Surprisingly, sequences of the three ORFs those isolated from broiler and human (Figure 3b). As expected, sim- of the cdtABC operon were found in all 19 isolates from wild birds, ilar clustering of the cdtABC operons as observed in the optimal tree including isolates that were negative by PCR for cdtA (5 isolates), (Figure 3a) was found. cdtAB (2 isolates), cdtAC (2 isolates) and cdtC (1 isolate; Figure 1). The single nucleotide polymorphism (SNP) distribution respect Sequence alignment of the cdtABC sequences from all 47 sequenced to the cdtABC81-­176 for the 38 isolates with a presumably intact cd- isolates and the reference strain 81–­176, revealed that 38 of the iso- tABC operon, corresponding to cluster I and II (Figure 3a), is shown lates carried a cdtABC operon of similar length to the cdtABC81-­176, in Figure 4a. Overall, it was manifest that cdtA accumulated higher whereas 9 isolates, all of them from wild birds, had suffer drastic number of SNPs as compared with cdtB and cdtC. The frequency of DNA modifications, resulting in shorter cdtABC operons (Figure S1, nucleotide substitution for C. jejuni isolates from broilers, humans Appendix S1). and wild birds, was 0.48%, 0.53% and 0.53% for cdtA, 0.25%, 0.16% The optimal tree of the evolutionary analyses of the cdtABC op- and 0.65% for cdtB, and 0.06%, 0.09% and 0.77% for cdtC, respec- erons (Figure 3a) showed that most of the operons from broiler and tively (Figure 4b). Moreover, it was also patent that a group of wild human isolates formed a very compact cluster (cluster I in Figure 3a). bird isolates (W06, W07, W13, W17, W33, W43), corresponding to Only four wild bird isolates were found scattered within this cluster cluster II, carried cdtB and cdtC genes with a higher degree of vari- I, whereas most of the cdtABC operons from wild bird isolates were ation than the isolates from cluster I. The results point out that the distant from the broiler/human cluster. The four wild bird isolates cdtABC operons from wild birds showed a higher variability (0.64%) founded within the broiler/human cluster I were W12, a ST-­4 46 as compared with the cdtABC operons from broiler and human iso- isolated from AGA; W29 a ST-­48 isolated from YLGB and W28 and lates (0.28% in both groups; Figure 4b). The nucleotide substitution W56, two ST-­354 isolated from YLGM and WS, respectively. The frequency was also calculated for the arbitrarily chosen ciaB, cadF four isolates were determined as cdtABC positive by PCR detection. and glnA genes, revealing similar divergence among the three groups The rest of cdtABC operons from wild birds were grouped in of isolates. three different clusters. Cluster II was characterized by carrying no deletions or 1 bp deletion, cluster III by having deletions shorter than 50 bp in total and cluster IV by carrying several deletions causing an 3.4 | Several Campylobacter jejuni isolates from overall deletion of more than 700 bp. wild birds carry nonfunctional cdtABC operons The evolutionary divergence between cdtABC sequences was also estimated and the resulting heat map showed the differences The analysis of the nucleotide sequence of the cdtABC operons between most of the cdtABC operons from wild bird isolates respect from wild bird isolates showed that this operon is highly variable F I G U R E 3 (a) Evolutionary relationships of 47 cdtABC sequences obtained from the WGS, inferred using the Neighbour-­Joining method. The optimal tree with the sum of branch length = 0.01161943 is shown. The evolutionary distances were computed using the Maximum Composite Likelihood method and are in the units of the number of base substitutions per site. All ambiguous positions were removed for each sequence pair (pairwise deletion option). There were a total of 2,181 positions in the final dataset. (b) Evolutionary divergence between sequences. The number of base substitutions per site from between sequences are shown. Analyses were conducted using the Maximum Composite Likelihood model. Codon positions included were 1st + 2nd + 3rd + Noncoding. All ambiguous positions were removed for each sequence pair (pairwise deletion option). There were a total of 2,206 positions in the final dataset. GUIRADO et al. | 971 972 | GUIRADO et al. F I G U R E 4 Single nucleotide polymorphism. (a) Distribution of SNPs in cdtA, cdtB and cdtC as compared to the cdtABC81-­176 for the 38 isolates with a presumably intact cdtABC operon, corresponding to cluster I and II of the optimal tree (Figure 3a). Squares in light grey indicate silent mutations, dark grey squares indicate missense mutations and asterisks indicates nonsense mutations. (b) Frequency of nucleotide substitution for cdtA, cdtB, cdtC and cdtABC. The frequency determined for the indicated arbitrary chosen genes is also shown. In all cases sequences were compared to the sequence of the reference strain 81–­176. and prompted to accumulate DNA modifications. When considering as they accumulate at least five different deletions within cdtA and deletion and insertion larger than 3 bp, several modifications were cdtB, including the large D668. The variant 3 was detected in two identified within the cdtABC operon. Remarkably, the modifications isolates. The variant 4 which is identical to variant 3 but with an in- occurred within the cdtA and cdtB genes. No deletions and inser- sertion of 6 bp, is the most prevalent (detected in four isolates). The tions larger than 3 bp were found within cdtC (Figure 5a, Figure S1, variant 5, similar to variant 3 but carrying the deletion (D154) was Appendix S1). Among cdtA genes, four deletions were detected, D4, detected in one isolate. D9, D25 and D154 of 4, 9, 25 and 154 bp, respectively. An insertion The production of the CDT subunits was predicted from the nu- of 6 bp (I6) was also identified in four cdtA alleles. Within the cdtB cleotide sequence of the different cdt alleles (Table 2, Figures S2, gene three specific deletions were detected D17, D28 and the above S3 and S4, Appendix S1). All the isolates from broiler and human described D51. Interestingly, six isolates carried a large deletion of origin, with the only exception of H70, expressed apparently full 668 bp (D668) that affected both cdtA and cdtB genes. CDT subunits with a theoretical size of 268, 265 and 189 amino Considering the mentioned modifications, five cdtABC operon acids for cdtA, cdtB and cdtC, respectively. The cdtC sequence of variants were identified (Figure 5b). The cdtABC v1 has only a dele- H70 isolate carried a deletion causing a nonsense mutation, and a tion (D28) within cdtB. The cdtABC v2 seems to derive from variant predicted protein 18 AA shorter (Figure 4a and Table 2). High diver- 1 since in addition to D28, an additional deletion (D25) was found sity was again found among the wild bird isolates (Table 2). The wild in cdtA. The other three variants underwent drastic modifications, bird isolates grouped in clusters I and II seems to carry functional GUIRADO et al. | 973 F I G U R E 5 Rearrangements detected within the cdtABC operons. (a) Schematic representation of the cdtABC gene cluster. The arrows indicate the sequences of the three cdtA, cdtB and cdtC genes. The relative position of primers used for PCR amplification of each gene is indicated. The different deletions and insertions found are represented, indicating the number of nucleotides affected and the wild bird isolates that carries each of them. (b) Representation of the five variants of the cdtABC operon identified as compared with the cdtABC operon of strain 81–­176. The deletions and insertions determined in each of the variants are showed, as well as the isolates where have been characterized. cdtABC since the predicted proteins are of same length as the one sequences caused early stop codons within cdtA and cdtC, and the produced by the 81–­176 strain. The only exception was the isolates accumulated deletions did not allow generation of a CdtB protein. W06 and W07 that carried a cdtB gene encoding a slightly shorter protein of 259 amino acids. The cdtABC from the two isolates in clus- ter III encoded a shorter CdtB toxin, since had a theoretical length of 4 | DISCUSSION 145 and 179 residues for W02 and W05, respectively. Similarly, the CdtC subunit from both isolates was also shorter with 82 and 40 as In this research, the existing variability within the cdtABC operon compared with the 189 residues of CdtC18-­176. CdtA was apparently has been studied in a collection of C. jejuni isolates from three dif- expressed in W05 whereas was non-­expressed in W02 by the pres- ferent origins: broiler, human and wild birds. Our data shows that ence of a codon stop in the third codon. the cdtABC operon seems to be highly conserved among broiler The sequence of the cdtABC operon from the the seven C. jejuni and human isolates whereas a great diversity was found among isolates from cluster IV indicated that these isolates did not produce wild bird isolates as shown by the prevalence of the three cdt any functional CDT toxin since apparently any of the three subunits genes by PCR, the fragment length polymorphism of the cdtB am- could be produced. SNPs and the deletions and insertions in the plicon and the cdtABC sequence analysis from a selected group 974 | GUIRADO et al. TA B L E 2 Summary of the results of PCR amplification, sequence homology with the primers used, deletions and insertions determined, total bp and predicted AA from cdtA, cdtB and cdtC genes of wild bird isolates, grouped in clusters as defined in Figure 3a PCR Primers Deletions Insertions bp AA cdtA WT 81–­174 POS 22/22 22/22 807 268 I W12 POS 22/22 22/22 807 268 W28 POS 22/22 22/22 807 268 W29 POS 22/22 22/22 807 268 W56 POS 22/22 22/22 807 268 II W06 POS 22/22 22/22 807 268 W07 POS 22/22 22/22 807 268 W13 POS 22/22 22/22 807 268 W17 POS 22/22 22/22 807 268 W33 POS 22/22 22/22 807 268 W43 POS 22/22 22/22 807 268 III W02 NEG 20/22 20/22 D25, 1(−1) 6 (+1) 787 1 W05 POS 22/22 22/22 807 268 IV W09 NEG 20/22 ND D4, D9, D668 1 (+1), 1 (+6) 567 11 W20 NEG 20/22 ND D4, D9, D668 1 (+1), 1 (+6) 567 11 W25 NEG 20/22 ND D4, D9, D668 1 (+1), 1 (+6) 567 11 W26 NEG 20/22 ND D4, D9, D668 1 (+1), 1 (+6) 256 11 W50 NEG 20/22 ND D4, D9, D668 1 (+1) 250 11 W52 NEG 20/22 ND D4, D9, D668 1 (+1) 250 11 W53 NEG ND/ND D4, D154, D9, D668, 94 11 1 (−1) cdtB WT 81–­174 POS 22/22 22/22 798 265 I W12 POS 22/22 22/22 798 265 W28 POS 22/22 22/22 798 265 W29 POS 22/22 22/22 798 265 W56 POS 22/22 22/22 798 265 II W06 POS 22/22 22/22 1 (−1) 797 259 W07 POS 22/22 22/00 1 (−1) 797 259 W13 POS 22/22 22/22 798 265 W17 POS 22/22 22/22 798 265 W33 POS 22/22 22/22 798 265 W43 POS 22/22 22/22 798 265 III W02 POS 22/22 22/22 D28 2 (+1) 772 145 W05 POS 22/22 21/22 D28 1 (+1) 771 179 IV W09 POS 20/22 21/22 D51, D17 1 (+1) 607 NO W20 POS 20/22 21/22 D51, D17 1 (+1) 606 NO W25 POS 20/22 21/22 D51, D17 1 (+1) 606 NO W26 POS 20/22 21/22 D51, D17, 1 (−2), 2 1 (+1) 601 NO (−1) W50 NEG 20/22 20/22 D51, D17, 1 (−1) 2 (+1) 605 NO W52 NEG 20/22 20/22 D51, D17, 1 (−1) 2 (+1) 605 NO W53 POS 21/22 20/22 2 (−1), D51, D17 602 NO GUIRADO et al. | 975 TA B L E 2 (Continued) PCR Primers Deletions Insertions bp AA cdtC WT 81–­174 POS 22/22 22/22 570 189 I W12 POS 22/22 22/22 570 189 W28 POS 22/22 22/22 570 189 W29 POS 22/22 22/22 570 189 W56 POS 22/22 22/22 570 189 II W06 POS 22/22 22/22 570 189 W07 POS 22/22 22/22 570 189 W13 POS 22/22 22/22 570 189 W17 POS 22/22 22/22 570 189 W33 POS 22/22 22/22 570 189 W43 POS 21/22 22/22 570 189 III W02 NEG 19/22 21/22 1 (−3) 2 (−1) 3 (+1) 568 82 W05 NEG 19/22 21/22 1 (−3) 2 (−1) 2 (+1) 567 40 IV W09 NEG 22/22 19/22 2 (−2) 567 44 W20 POS 22/22 19/22 2 (−2) 567 44 W25 POS 22/22 19/22 2 (−2) 567 44 W26 POS 22/22 20/22 1 (−1), 2 (−2) 1 (+1) 566 11 W50 POS 21/22 21/22 1 (−1), 2 (−2) 565 11 W52 POS 21/22 21/22 1 (−1), 2 (−2) 565 11 W53 POS 21/22 22/22 2 (−1) 1 (+2) 568 122 of isolates. The high conservation of cdtABC among broiler and suggesting that CdtA is not essential for the CDT action (Smith & human C. jejuni isolates is consistent with being the chicken meat Bayles, 2006). The frequency of nucleotide substitution for cdtA is the most common transmission route of C. jejuni to humans via similar among the isolates from clusters I and II independently of its food cross-­contamination. The cdtABC operon from most of the origin, whereas cdtB and cdtC showed low variability among human wild bird isolates showed modifications when compared with the and broiler and much higher among wild birds. These results suggest cdtABC 81-­176 operon highly prevalent among broiler, suggesting a correlation between certain cdtABC alleles and efficient coloniza- that the cdtABC 81-­176 does not seem to promote efficient coloniza- tion of specific hosts. tion among bird other than chicken. Evolutionary analyses of the Analysis of the complete sequence of the cdtABC operons pro- cdtABC sequences grouped some wild bird isolates (W12, W28, vide an explanation to the negative detection of cdt genes by PCR. W29 and W56) with the human and broiler isolates, which is not All isolates that were negative for one or more cdt genes carries cd- surprising since cross-­transmission among the three ecological tABC alleles were deletions or SNPs affect the hybridization with niches studied has been demonstrated (Hald et al., 2016). W12, the specific primers used. Our data indicate that PCR amplification W28 and W56 isolates belong to clonal complexes ST-­354 and fails when at least two SNPs are found within the primer-­hybridizing ST-­4 46 that have been previously associated with broiler chicken sequence. These results question the convenience of using PCR-­ and humans (Cobo-­D íaz et al., 2021; Jolley et al., 2018). We do genotyping to determine gene prevalence of the cdt genes. not find isolates of this clonal complexes among the wild bird iso- Several deletions were found among cdtABC operon from wild lates grouped in clusters II, III or IV. W29, also grouped with human bird isolates defining five different allelic variants. The deletions and broiler, was isolated from a sea gull (L. audouinii) in the city of D51 and D668 were previously reported in isolates from human Barcelona. Most wild birds that live in urban environment have patients (AbuOun et al., 2005; Kabir et al., 2011). In our isolate feeding habits that include feeding in urban refuse dumps. collection, we did not detect the indicated deletions among the Among the isolates carrying cdtABC operons that did not suffer isolates from broiler or human patients. In AbuOun et al. (2005), drastic DNA modifications a higher number of SNPs were detected the only C. jejuni strains carrying the D51 and D668 were isolated in cdtA as compared with cdtB and cdtC. Interestingly, cdtA was the from patients with underlying health problems that could induce most commonly missing gene when the prevalence of the three cdt an immunocompromised status. In fact, those patients develop genes was determined by PCR. In a previous report, it was shown Campylobacter bacteremia. The same deletions (D51 and D668) that addition of CdtB and CdtC to human culture cells has a cyto- together with D9 and D17 were detected in C. jejuni isolates toxic action as effective as when the three subunits were added, from crows (Sen et al., 2018). We detected those deletions not whereas addition of CdtA and CdtB did not induce any damage, only in crow isolates but also in isolates from different wild birds, 976 | GUIRADO et al. belonging to different clonal complex, and living in different eco- not during colonization in chickens. Infection and Immunity, 73, logical niches. 3053–­3 062. Bang, D. D., Nielsen, E. M., Scheutz, F., Pedersen, K., Handberg, K., & Our data indicates that isolates circulating among wild birds have Madsen, M. (2003). PCR detection of seven virulence and toxin specific characteristics, regarding the cdtABC operon, which differs genes of Campylobacter jejuni and Campylobacter coli isolates from from the isolates colonizing broiler and human. These findings sug- Danish pigs and cattle and cytolethal distending toxin production gest that cdtABC might be a cellular factor involved in host speci- of the isolates. Journal of Applied Microbiology, 94, 1003–­1014. Cobo-­Díaz, J. F., González del Río, P., & Álvarez-­Ordóñez, A. (2021). ficity attending to the high degree of conservation among broiler Whole resistome analysis in Campylobacter jejuni and C. coli ge- and the relevant differences detected among the wild bird isolates. nomes available in public repositories. Frontiers in Microbiology, 12, Moreover, the fact that many wild bird isolates carry apparently 1155. nonfunctional CDT toxin suggest that C. jejuni from wild bird strains de Melo, R. T., Dumont, C. F., Braz, R. F., Monteiro, G. P., Takeuchi, M. G., Lourenzatto, E. C. A., & Dos Santos, J. P. (2021). Genotypical might be less virulent to humans. relationship between human and poultry strains of Campylobacter jejuni. Current Microbiology, 788(78), 2980–­2988. AU T H O R C O N T R I B U T I O N S European Food Safety Authority and European Centre for Disease P.G, E.M., Y. I-­T., F.N., C.S-­O.A, C.M. and C.B. contributed to concep- Prevention and Control. (2021). The European Union One Health 2019 Zoonoses Report. EFSA Journal, 19, e06406. tualization and formal analysis. P.G, Y. I-­T., E.M., C.M. and C.B. per- Hald, B., Skov, M. N., Nielsen, E. M., Rahbek, C., Madsen, J. J., Wainø, M., formed the investigation. C.S-­O.A performed the WGS analysis. The Chrièl, M., Nordentoft, S., Baggesen, D. L., & Madsen, M. (2016). original draft was written by C.B. and C.M. All the authors reviewed Campylobacter jejuni and Campylobacter coli in wild birds on Danish the manuscript. C.B. and C.M. contributed to project administration livestock farms. Acta Veterinaria Scandinavica, 58, 11. Hughes, L. A., Bennett, M., Coffey, P., Elliott, J., Jones, T. R., Jones, R. and funding acquisition. All the authors have read and agreed to the C., Lahuerta-­Marin, A., Leatherbarrow, A. H., McNiffe, K., Norman, published version of the manuscript. D., Williams, N. J., & Chantrey, J. (2009). Molecular epidemiology and characterization of Campylobacter spp. isolated from wild AC K N OW L E D G E M E N T S bird populations in northern England. Applied and Environmental We deeply thank Francesc Mestres (University of Barcelona) for his Microbiology, 75, 3007–­3 015. Iglesias-­Torrens, Y., Miró, E., Guirado, P., Llovet, T., Muñoz, C., Cerdà-­ helpful assistance with the evolutionary analysis and the fruitful dis- Cuéllar, M., Madrid, C., Balsalobre, C., & Navarro, F. (2018). cussion about the results described in this manuscript. Population structure, antimicrobial resistance, and virulence-­ associated genes in campylobacter jejuni isolated from three eco- F U N D I N G I N FO R M AT I O N logical niches: Gastroenteritis patients, broilers, and wild birds. Frontiers in Microbiology, 9, 1676. This research was funded by the Spanish Ministry of Economy and Jinadasa, R. N., Bloom, S. E., Weiss, R. S., & Duhamel, G. E. (2011). Competitiveness (grant AGL2013-­ 45339R), the Spanish Ministry Cytolethal distending toxin: A conserved bacterial genotoxin that of Science, Innovation and Universities (MCIU), State Bureau of blocks cell cycle progression, leading to apoptosis of a broad range Investigation (AIE) and European Regional Development Fund of mammalian cell lineages. Microbiology, 157, 1521–­1875. Jolley, K. A., Bray, J. E., & Maiden, M. C. J. (2018). Open-­access bacterial (FEDER) (grant PGC2018-­096958-­B-­I 00) and the Catalonian gov- population genomics: BIGSdb software, the PubMLST.org website ernment (grant 2017SGR499). During this work, a trainee teaching and their applications. Wellcome Open Research, 3, 124. and research staff supported PG grant (Universitat de Barcelona). Kaakoush, N. O., Castaño-­Rodríguez, N., Mitchell, H. M., & Man, S. M. (2015). Global epidemiology of Campylobacter infection. Clinical Microbiology Reviews, 28, 687–­720. C O N FL I C T O F I N T E R E S T Kabir, S. M. L., Kikuchi, K., Asakura, M., Shiramaru, S., Tsuruoka, N., Goto, The authors declare that the research was conducted in the absence A., Hinenoya, A., & Yamasaki, S. (2011). Evaluation of a cytolethal of any commercial or financial relationship that could be construed distending toxin (cdt) gene-­based species-­specific multiplex PCR as a potential conflict of interest. assay for the identification of Campylobacter strains isolated from diarrheal patients in Japan. Japanese Journal of Infectious Diseases, 64, 19–­27. DATA AVA I L A B I L I T Y S TAT E M E N T Koga, M., Gilbert, M., Takahashi, M., Li, J., Koike, S., Hirata, K., & Yuki, Research data are not shared. N. (2006). Comprehensive analysis of bacterial risk factors for the development of Guillain-­Barré syndrome after Campylobacter jejuni ORCID enteritis. The Journal of Infectious Diseases, 193, 547–­555. Koolman, L., Whyte, P., Burgess, C., & Bolton, D. (2015). Distribution of Ferran Navarro https://orcid.org/0000-0002-4302-2838 virulence-­associated genes in a selection of Campylobacter isolates. Carlos Balsalobre https://orcid.org/0000-0002-4147-219X Foodborne Pathogens and Disease, 12, 424–­432. Cristina Madrid https://orcid.org/0000-0002-5861-3980 Kumar, S., Stecher, G., Li, M., Knyaz, C., & Tamura, K. (2018). MEGA X: Molecular evolutionary genetics analysis across computing plat- forms. Molecular Biology and Evolution, 35, 1547–­1549. REFERENCES Lai, C. K., Chen, Y. A., Lin, C. J., Lin, H. J., Kao, M. C., Huang, M. Z., Lin, AbuOun, M., Manning, G., Cawthraw, S. A., Ridley, A., Ahmed, I. H., Y. H., Chiang-­Ni, C., Chen, C. J., Lo, U. G., Lin, L. C., Lin, H., Hsieh, Wassenaar, T. M., & Newell, D. G. (2005). Cytolethal distending J. T., & Lai, C. H. (2016). Molecular mechanisms and potential clini- toxin (CDT)-­negative Campylobacter jejuni strains and anti-­ CDT cal applications of Campylobacter jejuni cytolethal distending toxin. neutralizing antibodies are induced during human infection but Frontiers in Cellular and Infection Microbiology, 6, 9. GUIRADO et al. | 977 Lara-­Tejero, M., & Galán, J. E. (2001). CdtA, CdtB, and CdtC form a tripar- Tamura, K., Nei, M., & Kumar, S. (2004). Prospects for inferring very large tite complex that is required for cytolethal distending toxin activity. phylogenies by using the neighbor-­joining method. Proceedings of Infection and Immunity, 69, 4358–­4365. the National Academy of Sciences of the United States of America, 101, Martínez, I., Mateo, E., Churruca, E., Girbau, C., Alonso, R., & Fernández-­ 11030–­11035. Astorga, A. (2006). Detection of cdtA, cdtB, and cdtC genes in Weis, A. M., Miller, W. A., Byrne, B. A., Chouicha, N., Boyce, W. M., & Campylobacter jejuni by multiplex PCR. International Journal of Townsend, A. K. (2014). Prevalence and pathogenic potential Medical Microbiology, 296, 45–­48. of Campylobacter isolates from free-­ living, human-­ commensal Pagès, H., Aboyoun, P., Gentleman, R., DebRoy, S. (2022). Biostrings: American crows. Applied and Environmental Microbiology, 80, Efficient manipulation of biological strings. R package version 2.64.0. 1639–­1644. https://bioco​nduct​or.org/packa​ges/Biost​rings Whitehouse, C. A., Balbo, P. B., Pesci, E. C., Cottle, D. L., Mirabito, P. M., Pickett, C. L., Pesci, E. C., Cottle, D. L., Russell, G., Erdem, A. N., & Zeytin, & Pickett, C. L. (1998). Campylobacter jejuni cytolethal distending H. (1996). Prevalence of cytolethal distending toxin production toxin causes a G2-­phase cell cycle block. Infection and Immunity, 66, in Campylobacter jejuni and relatedness of Campylobacter sp. cdtB 1934–­1940. gene. Infection and Immunity, 64, 2070–­2078. R Core Team. (2020). R: A language and environment for statistical comput- ing. R Foundation for Statistical Computing. https://www.R-­proje​ S U P P O R T I N G I N FO R M AT I O N ct.org/ Saitou, N., & Nei, M. (1987). The neighbor-­joining method: A new method Additional supporting information can be found online in the for reconstructing phylogenetic trees. Molecular Biology and Supporting Information section at the end of this article. Evolution, 4, 406–­425. Sen, K., Lu, J., Mukherjee, P., Berglund, T., Varughese, E., & Mukhopadhyay, A. K. (2018). Campylobacter jejuni colonization in the crow gut in- How to cite this article: Guirado, P., Iglesias-­Torrens, Y., volves many deletions within the cytolethal distending toxin gene Miró, E., Navarro, F., Attolini, C.-O., Balsalobre, C., & cluster. Applied and Environmental Microbiology, 84, e01893-­17. Sheppard, S. K., Dallas, J. F., Macrae, M., Mccarthy, N. D., Gormley, F. Madrid, C. (2022). Host-­associated variability of the cdtABC J., Strachan, N. J. C., Ogden, I. D., Maiden, M. C., & Forbes, K. J. operon, coding for the cytolethal distending toxin, in (2009). Campylobacter genotypes from food animals, environmen- Campylobacter jejuni. Zoonoses and Public Health, 69, tal sources and clinical disease in Scotland 2005/6, 134, 96–­103. 966–977. https://doi.org/10.1111/zph.12994 Smith, J. L., & Bayles, D. O. (2006). The contribution of cytolethal distend- ing toxin to bacterial pathogenesis. Critical Reviews in Microbiology, 32, 227–­248.