Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Population Genetics and Antimicrobial Susceptibility of Canine Campylobacter Isolates Collected before and after a Raw Feeding Experiment

  • Satu Olkkola ,

    Contributed equally to this work with: Satu Olkkola, Sara Kovanen

    Affiliation Department of Food Hygiene and Environmental Health, Faculty of Veterinary Medicine, University of Helsinki, Helsinki, Finland

  • Sara Kovanen ,

    Contributed equally to this work with: Satu Olkkola, Sara Kovanen

    Affiliation Department of Food Hygiene and Environmental Health, Faculty of Veterinary Medicine, University of Helsinki, Helsinki, Finland

  • Johanna Roine,

    Affiliation Department of Equine and Small Animal Medicine, Faculty of Veterinary Medicine, University of Helsinki, Helsinki, Finland

  • Marja-Liisa Hänninen,

    Affiliation Department of Food Hygiene and Environmental Health, Faculty of Veterinary Medicine, University of Helsinki, Helsinki, Finland

  • Anna Hielm-Björkman,

    Affiliation Department of Equine and Small Animal Medicine, Faculty of Veterinary Medicine, University of Helsinki, Helsinki, Finland

  • Rauni Kivistö

    Affiliation Department of Food Hygiene and Environmental Health, Faculty of Veterinary Medicine, University of Helsinki, Helsinki, Finland


In recent years, increasing numbers of consumers have become interested in feeding raw food for their pet dogs as opposed to commercial dry food, in the belief of health advantages. However, raw meat and internal organs, possibly contaminated by pathogens such as Campylobacter spp., may pose a risk of transmission of zoonoses to the pet owners. Campylobacter jejuni is the leading cause of bacterial gastroenteritis in humans but C. upsaliensis has also been associated with human disease. In this study we investigated the effect of different feeding strategies on the prevalence of Campylobacter spp. in Finnish dogs. We further characterized the isolates using multilocus sequence typing (MLST), whole-genome (wg) MLST and antimicrobial susceptibility testing. Dogs were sampled before and after a feeding period consisting of commercial raw feed or dry pellet feed. Altogether 56% (20/36) of the dogs yielded at least one Campylobacter-positive fecal sample. C. upsaliensis was the major species detected from 39% of the dogs before and 30% after the feeding period. Two C. jejuni isolates were recovered, both from raw-fed dogs after the dietary regimen. The isolates represented the same genotype (ST-1326), suggesting a common infection source. However, no statistically significant correlation was found between the feeding strategies and Campylobacter spp. carriage. The global genealogy of MLST types of dog and human C. upsaliensis isolates revealed weakly clonal population structure as most STs were widely dispersed. Major antimicrobial resistance among C. upsaliensis isolates was against streptomycin (STR MIC > 4mg/l). Apart from that, all isolates were highly susceptible against the antimicrobials tested. Mutations were found in the genes rpsL or rpsL and rsmG in streptomycin resistant isolates. In conclusion, increasing trend to feed dogs with raw meat warrants more studies to evaluate the risk associated with raw feeding of pets in transmission of zoonoses to humans.


Campylobacteriosis is the most common bacterial gastrointestinal disease in humans worldwide. The major species causing human disease are Campylobacter jejuni and Campylobacter coli. However, Campylobacter upsaliensis, often isolated from dogs [13] has been described as a cause of human disease including gastroenteritis and bacteremia [46] and dog ownership or contact with dogs has been identified as a risk factor for human campylobacter infections [7,8]. In dogs, asymptomatic carriage is common and especially C. upsaliensis is frequently isolated from both symptomatic and asymptomatic dogs [9,10]. However, it has been suggested that C. jejuni, and especially certain genotypes (i.e. ST-45), are significantly more prevalent among diarrheic than non-diarrheic dogs [11]. Furthermore, Chaban et al. (2010) showed that diarrheic dogs were more likely to shed Campylobacter spp., C. jejuni and C. coli among others, at significantly higher concentrations compared to healthy dogs [12].

In a previous study, genotyping by randomly amplified polymorphic DNA typing could not distinguish between two human clinical C. jejuni isolates and four canine strains suggesting that dogs are significant reservoirs of Campylobacter and contribute to human enteric infections [13]. Multilocus sequence typing (MLST) has been a valuable method in studies of molecular epidemiology, population structure and source attribution of Campylobacter [1416]. However, little data on the genetic diversity of C. jejuni [8,11,17] and especially C. upsaliensis isolates in dogs worldwide exist [18].

In recent years, a growing number of consumers have become interested in offering raw meat-based feed for their dogs, considered as more natural and healthy as opposed to commercial dry food ( Raw dog food typically includes uncooked meat, edible bones and internal organs, such as liver, from various animals, for example chickens, bovines and pigs. In addition, Biologically Appropriate Raw Food or Bones and Raw Food (BARF) typically contains at least fruit and vegetables, and also possibly eggs and dairy products [19].

Campylobacters are commonly isolated from raw chicken meat [20] and also from bovine, pig and chicken livers [2123]. In a recent Canadian study, DNA-based methods further revealed the presence of various Campylobacter spp. in ground beef, used also as dog raw food, including C. jejuni (3.9%) and C. upsaliensis (2.9%) [24]. However, in two previous studies, no Campylobacter spp. were found when evaluating the bacteriological quality of commercial raw canine diets [25] or raw meat diets [26]. Further, the only study concerning canine raw feeding found one of the 42 (2.6%) raw meat-fed dogs and none of the 49 control dogs to be positive for C. jejuni [27].

Campylobacteriosis in humans is usually self-limiting but severe cases or immunocompromised patients are treated with antimicrobials, preferentially with macrolides or fluoroquinolones [28]. Intravenous aminoglycosides can also be used in serious campylobacter bacteremia [29]. Antimicrobial resistance of C. jejuni from dogs (and cats) has been evaluated in some studies with resistance rates varying between 0–60% for quinolones, 0–40% for tetracycline and 0–12% for erythromycin with lowest prevalence of resistant isolates found in Norway [10,11,30,31]. However, only a limited number of studies on the prevalence or mechanisms of antimicrobial resistance of canine C. upsaliensis strains exist [10,31,32]. Interestingly, in a Norwegian study, most canine C. upsaliensis isolates were resistant to streptomycin and one strain was also resistant to nalidixic acid [31] and all outbreak-associated C. upsaliensis isolates from children in day care centers in Brussels were also resistant to streptomycin [5]. Another study found that C. upsaliensis strain RM3195 isolated from a human patient was resistant to nalidixic acid, oxytetracycline and novobiocin but not to streptomycin or most of the β-lactam antibiotics. However, the authors found no known mutations, which could explain the quinolone resistance [33].

In this study our aims were i) to investigate the prevalence of Campylobacter spp. in dogs with reference to the feeding strategy before and after the feeding regimen, consisting of either raw or dry commercial dog food, (ii) to analyse the MLST and wgMLST (whole-genome MLST) types of Campylobacter spp. isolates in local and global view and iii) to study the antimicrobial resistance patterns and mechanisms of the canine Campylobacter isolates.

Materials and Methods

Sample collection

Altogether, 36 Staffordshire bull terriers originating from a total of 30 households in Southern Finland, either healthy or diagnosed with atopic dermatitis, were included in this study. The dogs were divided in two groups with 15 dogs receiving commercial dry pellet feed and 18 dogs receiving raw feed consisting of meat, bones and organs from pork, chicken and lamb and/or beef, turkey and salmon. The feeding period lasted for 4 to 5 months. Owners of three dogs did not keep to the feeding regimen and those dogs were excluded from the Campylobacter prevalence analysis. Fecal samples were collected twice, before and after the feeding period, by the owners as a three-day pooled sample and kept refrigerated prior to analysis performed within 0 to 3 days from collection. All animal work has been conducted according to relevant national and international guidelines and with the dog owners' consent. This study was authorized by a written permission from the National Animal Experiment Board (Eläinkoelautakunta ELLA, decision number ESAVI/3244/04.10.07/2013) under Regional State Administrative Agency for Southern Finland.

Bacterial isolates

Fecal samples were suspended in 0.9% saline (1g feces/1ml saline) and a 10 μl loopful of this suspension was plated on modified charcoal cefoperazone deoxycholate agar (mCCDA) (CM739, Oxoid Ltd., Basingstoke, Hampshire, UK) with the selective supplement (SR155, Oxoid Ltd.) and incubated for up to 7 days at 37°C in jars (MART, anoxomat, Netherlands) in microaerobic conditions (6% O2, 10% CO2, 5% H2). At first, samples were also enriched in Bolton broth (Oxoid), with 5% horse blood and selective supplement (SR183E, Oxoid Ltd.), and incubated microaerobically at 37°C for 48 h but since all enriched samples, unlike direct culture, were consistently negative for Campylobacter spp. we ceased using this method. Colonies showing typical growth on mCCDA and morphology in gram stain were confirmed as Campylobacter spp. using genus specific PCR [34] and as C. jejuni or C. upsaliensis with species specific PCR [34,35]. Bacterial isolates were identified at baseline with the individual dog numbers (DRXX) and after the feeding period with number two at the end (DRXX_2).

Whole genome sequencing, MLST and data analysis

Draft genome sequences of all the recovered C. jejuni and C. upsaliensis isolates were determined using Illumina MiSeq or HiSeq technology (Nextera library, Nextera XT paired end kit, 250 cycles). NGS library preparation, enrichment and sequencing were performed by the Institute for Molecular Medicine Finland (FIMM Technology Centre, University of Helsinki, Finland). The paired-end reads were assembled into contigs using SPAdes 3.1.1 [36]. MLST types were assigned using the Campylobacter MLST database ( and new allele sequences were submitted to the non jejuni/coli Campylobacter MLST database (

The draft genomes were further analysed for whole-genome MLST using Genome profiler (GeP) [37]. The genomes were annotated in RAST (Rapid annotation using subsystem technology) and the resulting gbk files were used, as suitable, as reference genomes in the GeP analysis [37]. The NeighborNet networks, representing allelic distance matrix of the shared loci of the isolates, were constructed using SplitsTree4 [38] and edited using CorelDRAW X6.

The software ClonalFrame ver. 1.2 [39] was used to generate a genealogy tree of all known C. upsaliensis STs from the MLST database ( based on the sequences of the seven housekeeping genes with 50 000 iterations, 50 000 burn-in iterations and every 100th three was sampled.

A full minimum spanning tree of all MLST profiles present in the PubMLST database and isolate data (origins, i.e. country and source combinations with more than 2 isolates were included) from this study as well as those present in the PubMLST database and published by Parsons et al. (2012) was generated using the goeBURST algorithm [40,41] and visualized using PHYLOViZ 1.1 [42].

The nucleotide sequences of the C. upsaliensis genes gyrA, rsmG, rpsL and rrs were searched by BLAST in RAST, translated using EMBOSS Transeq ( and aligned for comparison using MUSCLE ( The sequences were compared to those of C. upsaliensis RM3195.

MIC determination

All the recovered C. jejuni (n = 2) and C. upsaliensis (n = 24) isolates were screened for antimicrobial resistance for erythromycin (ERY), tetracycline (TET), streptomycin (STR), gentamicin (GEN) and for the quinolones ciprofloxacin (CIP) and nalidixic acid (NAL) with the broth microdilution method (VetMIC Camp, National Veterinary Institute, Uppsala, Sweden) according to the manufacturer’s instructions. However, due to the fastidious nature of some of the C. upsaliensis isolates, a modified method utilizing Nutrient broth (Oxoid Ltd., Basingstoke, Hampshire, UK) with 5% blood (Labema, Kerava, Finland) instead of cation-adjusted Muller-Hinton broth (Difco, Becton-Dickinson and Company, Sparks, USA) with 5% blood was used for part of the C. upsaliensis isolates [43]. The agar dilution method (Clinical and Laboratory Standards Institute M31-A3) was used to determine and confirm the STR resistance levels of C. upsaliensis isolates. The epidemiological cut-off values (ECOFFs) for C. jejuni, as determined by the European Committee on Antimicrobial Susceptibility Testing (EUCAST,, were applied to distinguish between wild type (also referred to as susceptible) and non-wild type (also referred to as resistant) populations and were as follows: MIC > 4 mg/l for ERY and STR, MIC > 0.5 mg/l for CIP, MIC > 1 mg/l for TET, MIC > 2 mg/l for GEN and MIC > 16 mg/l for NAL. Due to the lack of data concerning MIC distributions of C. upsaliensis, these ECOFFs were also applied for it.


Occurrence and MLST types of Campylobacter spp. before and after raw feeding

The main results are presented in Table 1. A total of two C. jejuni and 24 C. upsaliensis isolates were detected from the feces of the 36 dogs included in this study. Altogether 20 dogs (55.6%) were positive for Campylobacter spp. in at least one sampling and six (16.7%) were positive in both samplings. Of the 33 dogs that kept to the feeding regimen, at baseline 13 (39.4%) were positive for C. upsaliensis and after the feeding period two dogs (6.1%) carried C. jejuni and 10 dogs (30.3%) C. upsaliensis. The C. jejuni isolates were recovered from two raw-fed dogs from different households living approximately 10 kilometers apart in the Helsinki area. Both dogs were Campylobacter-negative at the beginning (Table 1). Four new C. upsaliensis-positive dogs appeared after the feeding regimen; however, seven positive dogs at baseline seemed to have cleared off their C. upsaliensis carriage. No statistically significant association was found between raw or dry pellet feed diet and prevalence of C. jejuni or C. upsaliensis.

Table 1. Feeding strategies and Campylobacter status of studied dogs at baseline sampling and after the feeding regimen.

Both C. jejuni isolates represented sequence type (ST) 1326 and the 24 C. upsaliensis isolates were assigned to 16 STs, all of which were novel i.e. not existing in the PubMLST database (Table 1). Four of the six dogs that were C. upsaliensis-positive in both samplings yielded the same ST both times. The remaining two dogs had C. upsaliensis isolates with completely different allelic profiles (DR9S, ST-158 and ST-159, Table 1) or shared sequences in only two MLST loci (DR59S, ST-181 and ST-182, Table 1) in the successive samplings.

Whole-genome multilocus sequence typing (wgMLST)

All available draft genomes representing C. jejuni ST-1326 from the PubMLST database ( and from our own collection were included in the wgMLST analysis, in addition to the canine ST-1326 isolates identified in the present study (Fig 1). Among the 1,457 shared genes between all the genomes, the least number of allelic differences (2) was seen between two Finnish chicken isolates (3719_04 and 3723_04, Fig 1), detected on the same day from different slaughter batches reared at the same farm. However, our dog isolates were more similar to each other with 53 allelic differences (DR10S_2 and DR41S_2, Fig 1), and to the two UK isolates with allelic differences ranging from 44 (DR41S_2 and OXC4736, human stool isolate) to 111 (DR10S_2 versus Dg283, isolate from an undefined animal) than to the Finnish chicken isolates.

Fig 1. wgMLST of C. jejuni ST-1326 isolates.

SplitsTree of the NeighborNet network (1,457 shared genes) of C. jejuni ST-1326 isolates using GeP (Zhang et al., 2015). Dashed lines and numbers indicate the number of allelic differences observed between the pair of isolates. Two UK isolates (OXC4736 and Dg283) obtained from PubMLST isolate database (PubMLST id 18439 and 25960) and two chicken isolates (3719_04 and 3723_04) from our own collection (Llarena et al. 2015) were included as reference strains.

The NeighborNet network of the wgMLST analysis of the Finnish canine C. upsaliensis isolates is shown in Fig 2. Altogether 664 genes were shared between the genomes. Lowest numbers of allelic differences at the wgMLST level were seen among the isolates from the same individual dogs representing the same STs, collected before and after the feeding period, ranging from 1 (DR44S, Fig 2) to 18 (DR15S, Fig 2). Furthermore, isolates collected from two different dogs living in the same household clustered closely together showing only 10 to 23 (DR15S and DR18S, Fig 2) and 21 (DR59S_2 and DR60S_2, Fig 2) allelic differences among the 664 shared genes. In addition, six isolates from the same number of dogs, originating from different households, formed two clusters, however, showing relatively high numbers of allelic differences ranging from 49 (DR36S_2 versus DR42S, Fig 2) to 120 (DR28S versus DR47S, Fig 2), while rest of the isolates showed much higher genetic diversity.

Fig 2. wgMLST of C. upsaliensis isolates.

SplitsTree of the NeighborNet network (664 shared genes) of all available C. upsaliensis whole genomes, including three reference strains and isolates from dogs that were sampled only once and thus not included in this study, using GeP (Zhang et al., 2015). A new GeP analysis was performed for all closely related isolates and the results are shown next to the pair of isolates. The number of allelic differences, observed in the primary GeP analysis among the 664 shared genes, are shown in parenthesis. The reference genomes DSM 5365, JV21 and RM3195 were obtained from GenBank (accession numbers JHZN00000000, NZ_AEPU00000000 and NZ_AAFJ00000000).

Global genealogy of C. upsaliensis isolates from different sources

Similar clusters that occurred among our isolates in the wgMLST NeighborNet network were also detected in the ClonalFrame genealogy tree, based on the distinct C. upsaliensis MLST allele sequences at each locus (ST-166 and ST-170 and ST-167, ST-171 and ST-172, Fig 3). One major clonal complex (ST-42 CC, Fig 3) formed a separate cluster including only human patient isolates (mainly gastroenteritis except the isolate RM3195 from a GBS patient). In addition, some small groups, representing isolates from both dogs (from this and previous studies) and human patients (previous studies) occurred as well as a bigger cluster representing nine STs from unknown/unpublished sources (Fig 3). Otherwise most of the STs showed only little phylogenetic relatedness and at least one third of the isolates seemed quite unrelated to each other, including eight (50%) of the C. upsaliensis STs of the present study.

Fig 3. ClonalFrame genealogy tree based on all known C. upsaliensis MLST allele sequences.

Novel STs reported in this study are indicated in bold. The sources of the isolates are indicated based on the information available in the PubMLST Campylobacter non jejuni/coli database and in Parsons et al. (2012).

The evolutionary descents of C. upsaliensis isolates were further inferred using the goeBURST algorithm implemented in PHYLOViZ and visualized as an extended full Minimum Spanning Tree (MST) (Fig 4), overlaid by data representing the sources of the isolates. Although the most single locus variants originated from the same source and country combination, some were also found among canine and human isolates originating either from the same or two different countries. Furthermore, similarly as the human and dog C. upsaliensis isolates from the UK and USA, the Finnish dog isolates were distributed throughout the phylogenetic tree. The only exception was the ST-42 CC cluster, wherein most of the isolates originated from South Africa and Belgium solely from humans and the cluster representing isolates from undefined sources (Fig 4), which was also identified in ClonalFrame analysis (Fig 3).

Fig 4. goeBURST full Minimum Spanning Tree (MST) of C. upsaliensis ST allelic profiles.

Full MST of all C. upsaliensis allelic profiles present in PubMLST database overlaid by the isolation data (source and country combinations with more than 2 isolates were included), was generated using goeBURST and visualized with PHYLOViZ 1.1. The node sizes vary linearly with the number of isolates of a given ST. The links are color-coded for the number of differences i.e. darker links represent less allelic differences between the profiles than lighter links. Data used to create this figure is presented in S1 and S2 Tables.

Antimicrobial resistance of Campylobacter spp. isolates

The two C. jejuni isolates were susceptible to all the antimicrobials studied. Among C. upsaliensis, the most notable resistance trait was resistance to streptomycin with 79% (19/24) of the isolates having MICs of >4 mg/l and with seven of these with MICs of >512 mg/l. The remaining five isolates had streptomycin MICs of 0.5–2 mg/l (S3 Table). All isolates with streptomycin MIC > 4 mg/l encoded arginine (AGA) in codon 88 of rpsL, while all the five susceptible isolates encoded lysine (AAA) in the same position (S3 Table). Further, all isolates with streptomycin MIC > 512 mg/l had various deletion or insertion mutations in rsmG leading to frameshift and a premature stop codon immediately downstream of the mutation site resulting in termination at amino acid number 13, 49, 137, 144 or 155 of the encoded 7-methylguanosine methyltransferase. Also some intermediate-level resistant (MIC 16 mg/l) and susceptible isolates showed truncation of rsmG but only the last 3–5 amino acids were lost in these cases (S3 Table). No resistance associated mutations in the sequences of rrs were detected. One C. upsaliensis isolate was resistant to ciprofloxacin (MIC 1 mg/l) and nalidixic acid (MIC > 64 mg/l) and had point mutation C257T in gyrA resulting in amino acid substitution Thr-86-Met. No resistance to TET, ERY or GEN was detected.

In two cases C. upsaliensis isolates had closely related STs but differing MICs for streptomycin, and no clear pattern in the distribution of resistant or susceptible isolates was identified (Table 1). However, the C. upsaliensis isolates that originated from the same dog in the consecutive samplings and represented the same ST (DR15S and DR15S_2; DR18S and DR18S_2; DR44S and DR44S_2; DR55S and DR55S_2, Table 1) always had same streptomycin MICs.


Dogs are common pets especially in industrialized countries and approximately 50,000 new dogs are registered annually in Finland ( Dogs carrying Campylobacter spp. in their intestines may pose a risk of human infection by direct or indirect contact with fecal material of the animals [8]. In the present study, C. upsaliensis was the most common Campylobacter spp. found from dogs and C. jejuni was detected only in few cases, which is in accordance with several previous publications [3,44]. However, some studies [11,45] have found C. jejuni as the main species in canines, which could be due to the differences in the studied dog populations or isolation protocols.

Similarly to the results of Lenz et al. (2009) also we detected a low proportion of C. jejuni among the raw-fed group after the feeding period. Both C. jejuni isolates represented the same, rarely detected ST-1326 (ST-45 CC) that has previously been isolated from human patients, bovines, chickens, barnacle geese, grey seal pups and environmental water samples [46,47]. Furthermore, an association between ST-1326 and pet dog colonization was identified in a previous study from the Netherlands [8]. Since both C. jejuni isolates were detected from dogs living in different households, they likely originated from a common source, possibly from raw food. However, due to the low number of samples obtained, raw feed was not analysed in this study. Also, the option that these dogs acquired C. jejuni ST-1326 from other, possibly environmental-associated sources cannot be excluded since previous studies on wgMLST of C. jejuni have revealed that only few genetic differences (single nucleotide polymorphisms, SNPs) may occur among temporally and genetically related isolates [48,49] and in this study a total of 53 allelic differences were observed between the two C. jejuni ST-1326 isolates obtained from two unrelated dogs.

Whole-genome MLST analysis between C. upsaliensis isolates obtained from individual dogs 4–5 months apart revealed also relatively high numbers of allelic differences (range 27–68) among the isolates representing the same STs. Unfortunately, no data of the effect of long-term host colonization on the genomic variation of C. jejuni or C. upsaliensis exist. Therefore, more research should be conducted to estimate the microevolution occurring in Campylobacter genomes during long-term colonization in different hosts.

The global genealogy of the MLST types of C. upsaliensis revealed a highly diverse population, in which C. upsaliensis isolates, detected from both dogs and human patients, were dispersed throughout the phylogenetic network. MLST types showed also some degree of overlap resulting in a hypothesis that in principle, all dog isolates could be capable of causing disease also in humans. More isolates from various sources are needed to better understand the genealogy of C. upsaliensis MLST types, as only two out of five countries had deposited MLST types from both humans and dogs and other sources were lacking altogether.

Macrolides and fluoroquinolones are the first and second choice antibiotics when antimicrobial treatment of human campylobacteriosis is warranted [50]. In Finland, there are no data available on the consumption of antimicrobial agents per animal species yet but there are several fluoroquinolone containing drugs registered for small animals. However, both C. jejuni isolates were susceptible and, apart from streptomycin, most C. upsaliensis isolates were also susceptible to all the antimicrobials studied. Quinolone resistance in a small percentage among C. upsaliensis from pets has been detected previously for example in Belgium, Italy and Norway [10,31,32]. Resistance mechanisms for quinolones in C. jejuni and C. coli are well described and resistance is mediated by single point mutation in the gyrA gene and also by the increased activity of the CmeABC efflux pump [51]. The most commonly described resistance conferring mutation is C257T in gyrA, leading to amino acid substitution Thr-86-Ile in C. jejuni and C. coli and resulting in high level of quinolone resistance, while other substitutions (Thr-86-Lys, Asp-90-Asn, Asp-90-Ala, Ala-70-Thr, Thr-86-Ala) have been associated with low level of quinolone resistance or resistance to nalidixic acid alone [5153]. We describe here the same point mutation C257T in C. upsaliensis, but interestingly, this point mutation leads to Thr-86-Met substitution in this species leading to lower level of ciprofloxacin resistance (1 mg/l). To our knowledge, this is the first description of Thr-86-Met mutation in GyrA in connection to quinolone resistance.

Our finding that 79% of the C. upsaliensis isolates had increased MICs for streptomycin is in accordance with previous studies [5,31]. Streptomycin resistance in C. jejuni and C. coli can be conferred by enzymatic modification enzymes encoded in plasmids or chromosomally [54,55] and we have also shown that mutations in the rpsL gene codons 43 and 88 lead to streptomycin resistance in C. coli [56]. This latter resistance mechanism is quite well characterized also in other organisms, such as E. coli, M. tuberculosis and Helicobacter pylori [5759]. Our finding that all C. upsaliensis isolates with streptomycin MIC > 4 mg/l encode arginine in codon 88 (and lysine in 43) of the rpsL gene is consistent with some former studies: mutations in codon 43 have been associated with a higher level of STR resistance, while those in codon 88 have resulted in more variable STR MICs [56,58,60]. In addition, various (often frameshift) mutations within rsmG (previously known as gidB) encoding 7-methylguanosine methyltransferase that methylates 16S rRNA, have been associated with low level of streptomycin resistance in a number of bacterial species and high frequency of emergence of streptomycin-resistant mutants. Furthermore, rsmG rpsL double mutants have been associated with a high-level streptomycin resistant phenotype in several bacterial species, such as M. tuberculosis, Bacillus subtilis and E. coli [6163] and this was also observed in all highly resistant C. upsaliensis isolates described in this study.


C. upsaliensis, showing a weakly clonal population structure, was the most common finding among dogs before and after the feeding regimen. No statistically significant correlation was found between the feeding strategies and the prevalence of Campylobacter spp. carriage. However, C. jejuni isolates with the same ST were recovered from two raw-fed dogs, suggesting a common source of infection. The main antimicrobial resistance detected among C. upsaliensis was against streptomycin and apart from that, the isolates were highly susceptible. Further studies should be conducted to reveal the significance of C. upsaliensis in human infections and to identify its sources and reservoirs worldwide. Also the role of raw-feeding versus direct transmission of Campylobacter species between different animals should be further investigated.

Supporting Information

S1 Table. C. upsaliensis STs and corresponding allele profiles used for Full MST (Fig 4).


S2 Table. Epidemiological data associated with C. upsaliensis isolates used for Full MST image generated by PHYLOViZ (Fig 4).


S3 Table. Streptomycin MICs and rpsL and rsmG mutations in spontaneous streptomycin resistant C. upsaliensis isolates.



Mirko Rossi is acknowledged for excellent technical assistance in genome assembly.

Author Contributions

Conceived and designed the experiments: RK AHB. Performed the experiments: SO SK JR. Analyzed the data: SO SK RK. Contributed reagents/materials/analysis tools: MLH AHB. Wrote the paper: SO SK RK MLH.


  1. 1. Parsons BN, Porter CJ, Ryvar R, Stavisky J, Williams NJ, Pinchbeck GL, et al. Prevalence of Campylobacter spp. in a cross-sectional study of dogs attending veterinary practices in the UK and risk indicators associated with shedding. Vet J. 2010; 184: 66–70. pmid:19223212
  2. 2. Hald B, Pedersen K, Waino M, Jorgensen JC, Madsen M. Longitudinal study of the excretion patterns of thermophilic Campylobacter spp. in young pet dogs in Denmark. J Clin Microbiol. 2004; 42: 2003–2012. pmid:15131162
  3. 3. Engvall EO, Brandstrom B, Andersson L, Baverud V, Trowald-Wigh G, Englund L. Isolation and identification of thermophilic Campylobacter species in faecal samples from Swedish dogs. Scand J Infect Dis. 2003; 35: 713–718. pmid:14606609
  4. 4. Couturier BA, Hale DC, Couturier MR. Association of Campylobacter upsaliensis with persistent bloody diarrhea. J Clin Microbiol. 2012; 50: 3792–3794. pmid:22915607
  5. 5. Goossens H, Giesendorf BA, Vandamme P, Vlaes L, Van den Borre C, Koeken A, et al. Investigation of an outbreak of Campylobacter upsaliensis in day care centers in Brussels: analysis of relationships among isolates by phenotypic and genotypic typing methods. J Infect Dis. 1995; 172: 1298–1305. pmid:7594667
  6. 6. Shimizu Y, Ishii A, Takahata A, Kajiyama T, Yamahatsu A, Io H, et al. Campylobacter bacteremia in hemodialysis patients by eating raw meat—the importance of sanitary education. Case Rep Nephrol Urol. 2012; 2: 145–151. pmid:23197970
  7. 7. Kapperud G, Skjerve E, Bean NH, Ostroff SM, Lassen J. Risk factors for sporadic Campylobacter infections: results of a case-control study in southeastern Norway. J Clin Microbiol. 1992; 30: 3117–3121. pmid:1452694
  8. 8. Mughini Gras L, Smid JH, Wagenaar JA, Koene MG, Havelaar AH, Friesema IH, et al. Increased risk for Campylobacter jejuni and C. coli infection of pet origin in dog owners and evidence for genetic association between strains causing infection in humans and their pets. Epidemiol Infect. 2013; 141: 2526–2535. pmid:23445833
  9. 9. Carbonero A, Torralbo A, Borge C, Garcia-Bocanegra I, Arenas A, Perea A. Campylobacter spp., C. jejuni and C. upsaliensis infection-associated factors in healthy and ill dogs from clinics in Cordoba, Spain. Screening tests for antimicrobial susceptibility. Comp Immunol Microbiol Infect Dis. 2012; 35: 505–512. pmid:22640550
  10. 10. Rossi M, Hänninen ML, Revez J, Hannula M, Zanoni RG. Occurrence and species level diagnostics of Campylobacter spp., enteric Helicobacter spp. and Anaerobiospirillum spp. in healthy and diarrheic dogs and cats. Vet Microbiol. 2008; 129: 304–314. pmid:18164874
  11. 11. Amar C, Kittl S, Spreng D, Thomann A, Korczak BM, Burnens AP, et al. Genotypes and antibiotic resistance of canine Campylobacter jejuni isolates. Vet Microbiol. 2014; 168: 124–130. pmid:24210812
  12. 12. Chaban B, Ngeleka M, Hill JE. Detection and quantification of 14 Campylobacter species in pet dogs reveals an increase in species richness in feces of diarrheic animals. BMC Microbiol. 2010; 10: 73-2180-10–73.
  13. 13. Workman SN, Mathison GE, Lavoie MC. Pet dogs and chicken meat as reservoirs of Campylobacter spp. in Barbados. J Clin Microbiol. 2005; 43: 2642–2650. pmid:15956378
  14. 14. Kittl S, Heckel G, Korczak BM, Kuhnert P. Source attribution of human Campylobacter isolates by MLST and fla-typing and association of genotypes with quinolone resistance. PLoS One. 2013; 8: e81796. pmid:24244747
  15. 15. Miller WG, On SL, Wang G, Fontanoz S, Lastovica AJ, Mandrell RE. Extended multilocus sequence typing system for Campylobacter coli, C. lari, C. upsaliensis, and C. helveticus. J Clin Microbiol. 2005; 43: 2315–2329. pmid:15872261
  16. 16. Dingle KE, Colles FM, Wareing DR, Ure R, Fox AJ, Bolton FE, et al. Multilocus sequence typing system for Campylobacter jejuni. J Clin Microbiol. 2001; 39: 14–23. pmid:11136741
  17. 17. Parsons BN, Cody AJ, Porter CJ, Stavisky JH, Smith JL, Williams NJ, et al. Typing of Campylobacter jejuni isolates from dogs by use of multilocus sequence typing and pulsed-field gel electrophoresis. J Clin Microbiol. 2009; 47: 3466–3471. pmid:19794053
  18. 18. Parsons BN, Porter CJ, Stavisky JH, Williams NJ, Birtles RJ, Miller WG, et al. Multilocus sequence typing of human and canine C. upsaliensis isolates. Vet Microbiol. 2012; 157: 391–397. pmid:22266159
  19. 19. Billinghurst I. The BARF Diet: Raw Feeding for Dogs and Cats Using Evolutionary Principles. 1st ed. N.S.W: Australia: N.S.W. Australia: Ian Billinghurst.; 2001.
  20. 20. Katzav M, Isohanni P, Lund M, Hakkinen M, Lyhs U. PCR assay for the detection of Campylobacter in marinated and non-marinated poultry products. Food Microbiol. 2008; 25: 908–914. pmid:18721681
  21. 21. Edwards DS, Milne LM, Morrow K, Sheridan P, Verlander NQ, Mulla R, et al. Campylobacteriosis outbreak associated with consumption of undercooked chicken liver pate in the East of England, September 2011: identification of a dose-response risk. Epidemiol Infect. 2014; 142: 352–357. pmid:23711104
  22. 22. Noormohamed A, Fakhr MK. A higher prevalence rate of Campylobacter in retail beef livers compared to other beef and pork meat cuts. Int J Environ Res Public Health. 2013; 10: 2058–2068. pmid:23698698
  23. 23. Sasaki Y, Haruna M, Murakami M, Hayashida M, Ito K, Noda M, et al. Prevalence of Campylobacter spp., Salmonella spp., Listeria monocytogenes, and hepatitis E virus in swine livers collected at an abattoir. Jpn J Infect Dis. 2013; 66: 161–164. pmid:23514917
  24. 24. Trokhymchuk A, Waldner C, Chaban B, Gow S, Hill JE. Prevalence and diversity of Campylobacter species in Saskatchewan retail ground beef. J Food Prot. 2014; 77: 2106–2110. pmid:25474057
  25. 25. Weese JS, Rousseau J, Arroyo L. Bacteriological evaluation of commercial canine and feline raw diets. Can Vet J. 2005;46: 513–516. pmid:16048011
  26. 26. Strohmeyer RA, Morley PS, Hyatt DR, Dargatz DA, Scorza AV, Lappin MR. Evaluation of bacterial and protozoal contamination of commercially available raw meat diets for dogs. J Am Vet Med Assoc. 2006; 228: 537–542. pmid:16478425
  27. 27. Lenz J, Joffe D, Kauffman M, Zhang Y, LeJeune J. Perceptions, practices, and consequences associated with foodborne pathogens and the feeding of raw meat to dogs. Can Vet J. 2009; 50: 637–643. pmid:19721784
  28. 28. Blaser MJ, Engberg J. Clinical aspects of Campylobacter jejuni and Campylobacter coli infections. In: Nachamkin I, Szymanski CM, Blaser MJ, editors. Campylobacter. Washington DC, USA: ASM Press; 2008. pp. 99–121.
  29. 29. Aarestrup FM, Engberg J. Antimicrobial resistance of thermophilic Campylobacter. Vet Res. 2001; 32: 311–321. pmid:11432422
  30. 30. Acke E, McGill K, Quinn T, Jones BR, Fanning S, Whyte P. Antimicrobial resistance profiles and mechanisms of resistance in Campylobacter jejuni isolates from pets. Foodborne Pathog Dis. 2009; 6: 705–710. pmid:19580444
  31. 31. Sandberg M, Bergsjo B, Hofshagen M, Skjerve E, Kruse H. Risk factors for Campylobacter infection in Norwegian cats and dogs. Prev Vet Med. 2002; 55: 241–253. pmid:12392875
  32. 32. Vandenberg O, Houf K, Douat N, Vlaes L, Retore P, Butzler JP, et al. Antimicrobial susceptibility of clinical isolates of non-jejuni/coli campylobacters and arcobacters from Belgium. J Antimicrob Chemother. 2006; 57: 908–913. pmid:16533825
  33. 33. Fouts DE, Mongodin EF, Mandrell RE, Miller WG, Rasko DA, Ravel J, et al. Major structural differences and novel potential virulence mechanisms from the genomes of multiple Campylobacter species. PLoS Biol. 2005; 3: e15. pmid:15660156
  34. 34. Denis M, Soumet C, Rivoal K, Ermel G, Blivet D, Salvat G, et al. Development of a m-PCR assay for simultaneous identification of Campylobacter jejuni and C. coli. Lett Appl Microbiol. 1999; 29: 406–410. pmid:10664985
  35. 35. Lawson AJ, Linton D, Stanley J, Owen RJ. Polymerase chain reaction detection and speciation of Campylobacter upsaliensis and C. helveticus in human faeces and comparison with culture techniques. J Appl Microbiol. 1997; 83: 375–380. pmid:9351218
  36. 36. Bankevich A, Nurk S, Antipov D, Gurevich AA, Dvorkin M, Kulikov AS, et al. SPAdes: a new genome assembly algorithm and its applications to single-cell sequencing. J Comput Biol. 2012; 19: 455–477. pmid:22506599
  37. 37. Zhang J, Halkilahti J, Hanninen M, Rossi M. Refinement of whole-genome multilocus sequence typing analysis by addressing gene paralogy. J Clin Microbiol. 2015 May; 53(5):1765–7. pmid:25788543
  38. 38. Huson DH, Bryant D. Application of phylogenetic networks in evolutionary studies. Mol Biol Evol. 2006; 23: 254–267. pmid:16221896
  39. 39. Didelot X, Falush D. Inference of bacterial microevolution using multilocus sequence data. Genetics. 2007; 175: 1251–1266. pmid:17151252
  40. 40. Feil EJ, Li BC, Aanensen DM, Hanage WP, Spratt BG. eBURST: inferring patterns of evolutionary descent among clusters of related bacterial genotypes from multilocus sequence typing data. J Bacteriol. 2004; 186: 1518–1530. pmid:14973027
  41. 41. Francisco AP, Bugalho M, Ramirez M, Carrico JA. Global optimal eBURST analysis of multilocus typing data using a graphic matroid approach. BMC Bioinformatics. 2009; 10: 152-2105-10–152.
  42. 42. Francisco AP, Vaz C, Monteiro PT, Melo-Cristino J, Ramirez M, Carrico JA. PHYLOViZ: phylogenetic inference and data visualization for sequence based typing methods. BMC Bioinformatics. 2012; 13: 87-2105-13–87.
  43. 43. Zanoni RG, Rossi M, Giacomucci D, Sanguinetti V, Manfreda G. Occurrence and antibiotic susceptibility of Helicobacter pullorum from broiler chickens and commercial laying hens in Italy. Int J Food Microbiol. 2007; 116: 168–173. pmid:17303278
  44. 44. Procter TD, Pearl DL, Finley RL, Leonard EK, Janecko N, Reid-Smith RJ, et al. A cross-sectional study examining Campylobacter and other zoonotic enteric pathogens in dogs that frequent dog parks in three cities in south-western Ontario and risk factors for shedding of Campylobacter spp. Zoonoses Public Health. 2014; 61: 208–218. pmid:23802765
  45. 45. Tsai HJ, Huang HC, Lin CM, Lien YY, Chou CH. Salmonellae and Campylobacters in household and stray dogs in northern Taiwan. Vet Res Commun. 2007; 31: 931–939. pmid:17285243
  46. 46. Llarena AK, Huneau A, Hakkinen M, Hanninen ML. Predominant Campylobacter jejuni Sequence Types Persist in Finnish Chicken Production. PLoS One. 2015; 10: e0116585. pmid:25700264
  47. 47. Baily JL, Meric G, Bayliss S, Foster G, Moss SE, Watson E, et al. Evidence of land-sea transfer of the zoonotic pathogen Campylobacter to a wildlife marine sentinel species. Mol Ecol. 2015; 24: 208–221. pmid:25401947
  48. 48. Revez J, Llarena AK, Schott T, Kuusi M, Hakkinen M, Kivisto R, et al. Genome analysis of Campylobacter jejuni strains isolated from a waterborne outbreak. BMC Genomics. 2014; 15: 768-2164-15–768.
  49. 49. Kivistö RI, Kovanen S, Skarp-de Haan A, Schott T, Rahkio M, Rossi M, et al. Evolution and comparative genomics of Campylobacter jejuni ST-677 clonal complex. Genome Biol Evol. 2014; 6: 2424–2438. pmid:25193305
  50. 50. EFSA (European Food Safety Authority) and ECDC (European Centre for Disease Prevention and Control). EU summary report on antimicrobial resistance in zoonotic and indicator bacteria from humans, animals and food in 2013. EFSA J. 2015; 4036: 178.
  51. 51. Luo N, Sahin O, Lin J, Michel LO, Zhang Q. In vivo selection of Campylobacter isolates with high levels of fluoroquinolone resistance associated with gyrA mutations and the function of the CmeABC efflux pump. Antimicrob Agents Chemother. 2003; 47: 390–394. pmid:12499221
  52. 52. Wang Y, Huang WM, Taylor DE. Cloning and nucleotide sequence of the Campylobacter jejuni gyrA gene and characterization of quinolone resistance mutations. Antimicrob Agents Chemother. 1993; 37: 457–463. pmid:8384814
  53. 53. Jesse TW, Englen MD, Pittenger-Alley LG, Fedorka-Cray PJ. Two distinct mutations in gyrA lead to ciprofloxacin and nalidixic acid resistance in Campylobacter coli and Campylobacter jejuni isolated from chickens and beef cattle. J Appl Microbiol. 2006; 100: 682–688. pmid:16553723
  54. 54. Nirdnoy W, Mason CJ, Guerry P. Mosaic structure of a multiple-drug-resistant, conjugative plasmid from Campylobacter jejuni. Antimicrob Agents Chemother. 2005; 49: 2454–2459. pmid:15917546
  55. 55. Qin S, Wang Y, Zhang Q, Chen X, Shen Z, Deng F, et al. Identification of a novel genomic island conferring resistance to multiple aminoglycoside antibiotics in Campylobacter coli. Antimicrob Agents Chemother. 2012; 56: 5332–5339. pmid:22869568
  56. 56. Olkkola S, Juntunen P, Heiska H, Hyytiainen H, Hanninen ML. Mutations in the rpsL gene are involved in streptomycin resistance in Campylobacter coli. Microb Drug Resist. 2010; 16: 105–110. pmid:20370506
  57. 57. Funatsu G, Wittmann HG. Ribosomal proteins XXXIII. Location of amino-acid replacements in protein S12 isolated from Escherichia coli mutants resistant to streptomycin. J Mol Biol. 1972; 68: 547–550. pmid:4560854
  58. 58. Meier A, Sander P, Schaper KJ, Scholz M, Bottger EC. Correlation of molecular resistance mechanisms and phenotypic resistance levels in streptomycin-resistant Mycobacterium tuberculosis. Antimicrob Agents Chemother. 1996; 40: 2452–2454. pmid:8913445
  59. 59. Torii N, Nozaki T, Masutani M, Nakagama H, Sugiyama T, Saito D, et al. Spontaneous mutations in the Helicobacter pylori rpsL gene. Mutat Res. 2003; 535: 141–145. pmid:12581531
  60. 60. Fukuda M, Koga H, Ohno H, Yang B, Hirakata Y, Maesaki S, et al. Relationship between genetic alteration of the rpsL gene and streptomycin susceptibility of Mycobacterium tuberculosis in Japan. J Antimicrob Chemother. 1999; 43: 281–284. pmid:11252336
  61. 61. Nishimura K, Johansen SK, Inaoka T, Hosaka T, Tokuyama S, Tahara Y, et al. Identification of the RsmG methyltransferase target as 16S rRNA nucleotide G527 and characterization of Bacillus subtilis rsmG mutants. J Bacteriol. 2007; 189: 6068–6073. pmid:17573471
  62. 62. Okamoto S, Tamaru A, Nakajima C, Nishimura K, Tanaka Y, Tokuyama S, et al. Loss of a conserved 7-methylguanosine modification in 16S rRNA confers low-level streptomycin resistance in bacteria. Mol Microbiol. 2007; 63: 1096–1106. pmid:17238915
  63. 63. Benitez-Paez A, Cardenas-Brito S, Corredor M, Villarroya M, Armengod ME. Impairing methylations at ribosome RNA, a point mutation-dependent strategy for aminoglycoside resistance: the rsmG case. Biomedica. 2014; 34 Suppl 1: 41–49. pmid:24968035