Genetic characterization and biofilm formation of potentially pathogenic foodborne Arcobacter isolates

International Journal of Food Microbiology 373 (2022) 109712 Available online 9 May 2022 0168-1605/© 2022 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/bync-nd/4.0/). Genetic characterization and biofilm formation of potentially pathogenic foodborne Arcobacter isolates Irati Martinez-Malaxetxebarria a , b , * , Cecilia Girbau a , 1 , Adri´ an Salazar-S´ anchez a , Itsaso Baztarrika a , Ilargi Martínez-Ballesteros a , b , Lorena Laorden a , b , Rodrigo Alonso a , b , Aurora Fern´ andez-Astorga a a MikroIker Research Group, Department of Immunology, Microbiology and Parasitology, Faculty of Pharmacy, University of the Basque Country UPV/EHU, Paseo de la Universidad 7, 01006 Vitoria-Gasteiz, ´ Alava, Spain b Bioaraba, Microbiology, Infectious Disease, Antimicrobial Agents and Gene Therapy, 01006 Vitoria-Gasteiz, ´ Alava, Spain ARTICLE INFO Keywords: Food Prevalence MLST Virulence-associated genes Adhesion to surfaces ABSTRACT Various species of the genus Arcobacter are regarded as emerging food pathogens and can be cause of human gastroenteric illness, among others. In order to gain knowledge on the risk associated with the presence of arcobacters in retail foods, this study aimed to determine their presence in a variety of products; to evaluate the genetic diversity and the occurrence of virulence and biofilm-associated genes in the isolated strains; and to assess their biofilm activity on polystyrene, borosilicate and stainless steel. Arcobacters were detected in the 22.3% of the analysed samples and the 83 recovered isolates were identified as A. butzleri (n =53), A. cryaerophilus (n =24), A. skirrowii (n =2), A. thereius (n =3) and A. vitoriensis (n =1). They were isolated from virtually all tested food types, but mostly from squids and turkey meat (contamination levels of 60% and 40%, respectively). MLST differentiated 68 STs, most of which were novel (89.7%) and represented by a single strain (86.9%). Five novel STs were detected in various isolates derived from seafood, and the statistical analysis revealed their potential association with that type of food product (p <0,001). All the isolates except one harboured virulence-associated genes and the highest incidence was noted for A. butzleri. Nineteen isolates (23.5%) were able to form biofilms on the different surfaces tested and, of note; glass enhanced the adhesion ability of the majority of them (84.2%). The results highlight the role that common food products can have in the transmission of Arcobacter spp., the pathogenic potential of the different species, and the survival and growth ability of several of them on different food contact surfaces. Therefore, the study provides interesting information regarding the risk arcobacters may pose to human health and the food industry. 1. Introduction The genus Arcobacter, within the Campylobacteraceae family, was first described by Vandamme et al. in 1991. The taxonomy of this genus has been under debate during the last years (On et al., 2020, 2021; P´ erez-Catalu˜ na et al., 2018; Waite et al., 2017) but, at the time of writing and according to LPSN, the list of prokaryotic names with standing in nomenclature (Parte et al., 2020), the genus comprises 33 validly published species (https://lpsn.dsmz.de/genus/arcobacter; accessed on February 28th, 2022) that have been isolated from various different environments and sources. Certain species of the genus are associated with human disease. They mainly induce gastrointestinal symptoms (chronic watery diarrhoea and traveller's diarrhoea), but can also be the cause of bacteraemia, septicaemia, peritonitis and endocarditis (Collado and Figueras, 2011; Fanelli et al., 2019; Simaluiza et al., 2021). A. butzleri is the species most frequently associated with disease, followed by A. cryaerophilus, but infections due to A. skirrowii, A. thereius and A. lanthieri have also been reported (Kerkhof et al., 2021; Ramees et al., 2017; Ruíz de Alegría et al., 2021; Van den Abeele et al., 2014). Nevertheless, the mechanisms implied in the pathogenesis of these bacteria remain unclear. Various authors have demonstrated the in vitro cytotoxicity of Arcobacter species, along with their ability to adhere and * Corresponding author at: MikroIker Research Group, Department of Immunology, Microbiology and Parasitology, Faculty of Pharmacy, University of the Basque Country UPV/EHU, Paseo de la Universidad 7, 01006 Vitoria-Gasteiz, ´ Alava, Spain. E-mail address: [email protected] (I. Martinez-Malaxetxebarria). 1 Present address: Public Health Laboratory of Araba, Basque Government, Avenida Santiago 11, 01004 Vitoria-Gasteiz, Spain. Contents lists available at ScienceDirect International Journal of Food Microbiology journal homepage: www.elsevier.com/locate/ijfoodmicro https://doi.org/10.1016/j.ijfoodmicro.2022.109712 Received 7 August 2021; Received in revised form 1 May 2022; Accepted 5 May 2022 International Journal of Food Microbiology 373 (2022) 109712 2 invade different human cell lines (Buzzanca et al., 2021; Collado and Figueras, 2011; Karadas et al., 2013; Levican et al., 2013). On the other hand, the available Arcobacter genomes have shown the presence of various virulence-associated genes related to, among others, adaptation, cytotoxicity, adhesion, invasiveness and antibiotic resistance (Isidro et al., 2020; Miller et al., 2007; Müller et al., 2020a, 2020b). The correlation between the reported pathogenic capabilities and the presence of specific genes has not been established yet in Arcobacter spp. However, the virulence-associated gene content of the isolates can be indicative of the risk they may pose to human health. The consumption of contaminated drinking water and/or undercooked or raw foods seems to be the main human transmission source of Arcobacter spp. They are commonly present in food products including vegetables, seafood, terrestrial animal food products and composite foods (G´ onzalez and Ferrús, 2011; Kietsiri et al., 2021; Mottola et al., 2020; Nieva-Echevarria et al., 2013); but also in different waters including continental, coastal, sea, recreational, drinking and sewage (Sciortino et al., 2021). Moreover, arcobacters are often present in food processing environments such as slaughterhouses and dairy farms and/ or plants (Ferreira et al., 2017; Giacometti et al., 2015b; Khodamoradi and Abiri, 2020) where, if the conditions are favourable, they may probably form biofilms. Arcobacters have the demonstrated ability to adhere to different surfaces and to form biofilms on them (Ferreira et al., 2013; Girbau et al., 2017; ˇ Silha et al., 2021). When formed on food contact surfaces and/or materials, biofilms increase food safety risk. Reservoirs of food spoilage and/or pathogen bacteria in food industries are an important cause of product contamination that can lead not only to a reduced shelf life of foods, but also to health problems (Abebe, 2020; Adetunji et al., 2014). Therefore, the biofilm formation by food derived Arcobacter spp. can pose a risk for public health and a problem for the food industry. Tracking the infection source and the transmission routes of arcobacters is one of the necessary steps to assess the risk related to these pathogens. Among the molecular subtyping techniques available for the species, the Multilocus Sequence Typing (MLST) scheme proposed by Miller et al. (2009) is a reliable and reproducible technique that has been successfully utilized for characterization of Arcobacter isolates from different sources (Alonso et al., 2014; Caruso et al., 2020; Kietsiri et al., 2021; Niedermeyer et al., 2020). However, partly due to the limited available data in the Arcobacter MLST database (https://pub mlst.org/organisms/arcobacter-spp), partly due to the great genetic heterogeneity shown by Arcobacter isolates, no source-associated genetic marker has been reported so far for these species. In order to increase the knowledge needed to assess the risk that arcobacters pose for human health; the purposes of this study were to confirm the presence of different Arcobacter species in retail food products; to evaluate, by MLST and virulence-associated gene detection, the genetic diversity of the isolates; and to investigate the biofilm production of all the recovered isolates. We complement, this way, our previous surveys on prevalence and characterization of arcobacters in foods (Alonso et al., 2014; Girbau et al., 2014, Girbau et al., 2017; NievaEchevarria et al., 2013) by the analysis of food products not previously surveyed. 2. Materials and methods 2.1. Sample collection and processing Two hundred and twenty samples including cockle, squid, shrimp, quail meat, rabbit meat, turkey meat, fresh cheese, spinach, Swiss chard, lettuce and carrot, were purchased from different local retail shops and supermarkets in Vitoria-Gasteiz, Spain, from May to November 2015. All samples, 20 of each type of food, were kept in coolers, transported to the laboratory and processed within 2 h of purchase. Ten grams of each sample were homogenized into 90 mL (1:10 wt/ vol) of Arcobacter-CAT broth (Oxoid) as previously described (NievaEchevarria et al., 2013), and then incubated aerobically at 30 ◦C for 48 h. After enrichment, 0.2 mL of each broth were inoculated by passive filtration with 0.45μ m nitrocellulose membrane filters (Millipore) onto blood agar plates (Columbia agar supplemented with 5% sheep blood, Oxoid) and incubated under the aforementioned conditions for 48–72 h. After incubation, four to six suspect Arcobacter colonies (small, smooth and translucent to whitish) were picked from each plate and subcultured onto blood agar plates at least three times. Upon microscopic examination, those isolates presenting a curved to spiral shape and characteristic motility were subjected to PCR identification. 2.2. Arcobacter species identification 2.2.1. Genomic DNA isolation DNA was isolated using PrepMan™ Ultra reagent (Applied Biosystems) according to the manufacturer's specifications. The concentration was determined spectrophotometrically (NanoDrop, Thermo Fisher Scientific), adjusted to 20 ng/ μ L and stored at −20 ◦C. 2.2.2. PCR and m-PCR Suspicious colonies were identified by the genus-specific PCR described by Bastyns et al. (1995). To avoid the inclusion of clones in the collection, all Arcobacter isolates recovered from the same food sample were genotyped using the enterobacterial repetitive intergenic consensus PCR (ERIC-PCR) protocol as previously described for Arcobacter (Houf et al., 2002). Patterns with at least one or more different bands were considered as different genotypes. Species identification of the isolates was carried out by two previously described methods. The m-PCR proposed by Houf et al. (2000) that targets A. butzleri, A. cryaerophilus and A. skirrowii was applied first, followed by the m-PCR proposed by Douidah et al. (2010) that simultaneously identifies A. butzleri, A. cryaerophilus, A. skirrowii, A. cibarius and A. thereius. DNA from A. butzleri RM4018, A. cryaerophilus CCUG 17801T, A. skirrowii CCUG 30483, A. cibarius CECT 7203 and A. thereius CCUG 56002, together with deionized water, were used as positive and negative controls, respectively. 2.2.3. Phylogenetic and phylogenomic analysis In order to determine the taxonomic position of one isolate which could not be identified to the species level by the aforementioned m-PCR methods, both, phylogenetic and phylogenomic analyses, were held (Alonso et al., 2020). 2.3. Genetic characterization The 83 strains recovered in this study (53 A. butzleri, 24 A. cryaerophilus, three A. thereius, two A. skirrowii and one A. vitoriensis) were genetically characterized by Multilocus Sequence Typing (MLST) and virulence-associated gene detection. 2.3.1. MLST and minimum spanning trees MLST was carried out according to the method of Miller et al. (2009) with minor modifications. The glyA gene from A. cryaerophilus was amplified with different annealing temperatures (55–59 ◦C) and the glnA gene from A. thereius was amplified by using the primers glnACR1 and glnATHF (5′-AAATGGAATGCCTTTTGATGGAG-3′). Allele numbers and sequence types (STs) were assigned using the PubMLST database (Jolley et al., 2018) at https://pubmlst.org/organisms/arcobacter-spp. New alleles and STs were submitted to the database curator to be assigned new allele or ST numbers. In order to detect possible recombination events, all the available concatenated MLST sequences were downloaded from the Arcobacter PubMLST database on January 2017 (n = 648) and subsequently analysed using five methods (RDP, Geneconv, MaxChi, Chimaera and 3Seq) implemented in the RDP3 software package (Martin et al., 2010) using default parameters. In order to visualize the relationships between the STs and their distribution among different food products, minimum spanning trees (MST) were created by I. Martinez-Malaxetxebarria et al. International Journal of Food Microbiology 373 (2022) 109712 3 the goeBURST algorithm using the PHYLOViZ v2.0a software (Nascimento et al., 2017). MSTs were constructed based on the distances between the allelic profiles of all the Arcobacter isolates with an assigned ST available at the PubMLST database on January 2021 (n =997), including those identified in the present study. 2.3.2. Detection of virulence genes The presence of ten putative virulence genes (cadF, Cj1349, ciaB, mviN, pldA, tlyA, irgA, hecA, hecB and iroE) was determined using the primer pairs designed by Douidah et al. (2012) and Karadas et al. (2013), following the protocol described by Girbau et al. (2015). DNA from A. butzleri RM4018 was used as positive control and deionized water as negative one. Additionally, DNAs from A. cryaerophilus CCUG 17801 and Ac-L7 (Girbau et al., 2015), A. skirrowii CCUG 30483 and A. thereius CCUG 56902 were also included as controls whenever any of these species were subjected to the aforementioned PCRs. 2.4. Biofilm production The ability to form biofilms of the strains recovered in this survey was investigated by biofilm-associated gene detection and in vitro phenotypic assays. 2.4.1. Biofilm-associated gene detection Genes to be detected (flaA, flaB, fliS, luxS, pta, waaF and spoT) were selected based on their association with adherence to abiotic surfaces in other campylobacteria. Based on comprehensive analyses and alignments of the published genome sequences of A. butzleri RM4018 T , A. cryaerophilus ATCC 43158 T , A. skirrowii CCUG 10374 T , A. thereius LMG 24486 T and A. vitoriensis F199 T (GenBank, accession numbers GCA_000014025, NZ_CP032823, NZ_VZOH00000000, NZ_CP035926.1 and PDKB00000000, respectively), 35 PCR primers were designed using Clone Manager 9 Professional Edition software (Sci Ed Software LLC). Once designed, the primers were tested in silico by blasting them against completed Arcobacter genome sequences available in the GenBank database. The primer sequences and the expected amplicon sizes depending on the species are listed in Table S1. All PCRs were carried out in final volumes of 50 μ L containing 100 ng of DNA as template, 1.25 U of DreamTaq DNA Polymerase (Thermo Scientific), 0.2 mM of each dNTP, 1×buffer and 0.5 μ M of each primer set. An initial denaturation step at 94 ◦C for 3 min was followed by 30 cycles of denaturation at 94 ◦C for 45 s; primer annealing at different temperatures ranging from 50 ◦C to 56 ◦C for 45 s; and elongation at 72 ◦C for 1 min. A final elongation step at 72 ◦C for 3 min was performed. 2.4.2. Motility assay In order to test the motility of those strains for which flagellin gene detection failed, individual colonies were spotted onto 0.4% thioglycolate plates (Scharlau). The plates were incubated at 30 ◦C, and the growth and expansion of colonies was examined after 16–24 h. 2.4.3. Static biofilm assays The biofilm formation ability was first assessed at 30 ◦C under aerobic conditions using polystyrene microtiter plates, as described previously (Girbau et al., 2017). Borosilicate glass tubes and stainless steel coupons were also used for further testing of the ability of those strains showing adherence to polystyrene. For each assay, every isolate was examined in three replicates and the experiments were performed at least on three separate occasions. Those biofilms formed on polystyrene and borosilicate were expressed by the biofilm formation index (BFI) according to Niu and Gilbert (2004), and subsequently categorized as strong, moderate, weak or none biofilm formation according to Naves et al. (2008). The biofilms formed on stainless steel were evaluated by plate count method on Mueller-Hinton agar (Oxoid) after gently washing the coupons with sterile distilled water (Girbau et al., 2017). 2.5. Statistical analysis Data were analysed with the SPSS 26 statistical package program (SPSS Inc., Chicago, IL, USA). The Chi-square and Fisher's exact tests were performed in order to compare the distribution of the isolates and species among samples and to assess possible associations between variables. The isolates derived from spinach (n =1), lettuce (n =1), cheese (n =1) and rabbit (n =3), along with those identified as A. thereius (n =3), A. skirrowii (n =2) and A. vitoriensis (n =1), were excluded from the analyses based on their low representation. Once tested the normality of the numerical variable “biofilm formation” using the Kolmogorov-Smirnov method, the Kruskal-Wallis test was used to compare the values obtained for biofilm formation ability on each surface among the isolates. Student t-test was used to compare the values obtained for the formed biofilms on polystyrene versus borosilicate for each strain. Results were considered significant at p values of <0.05. 3. Results 3.1. Occurrence of Arcobacter spp. in food samples The occurrence of Arcobacter spp. in the 220 food samples analysed in this study is summarized in Table 1. Overall, they were detected in all the tested types of products except chard, with a gross occurrence of 22.3% (49 out of 220 samples). The arcobacters were mostly detected in seafood products, which showed a statistically significant (p <0.001) contamination level of 43.3%. Foods of terrestrial animal and vegetable origin showed lower contamination levels, 21.3% and 7.5%, respectively. Specifically, those products from which Arcobacter spp. were mainly recovered were squid (60.0%) and turkey meat (40.0%). The recovery from squid was statistically significant (p <0.001). Among the 266 isolates identified as Arcobacter spp. by the genusspecific PCR (Bastyns et al., 1995), 83 were selected for further identification to the species level based on the ERIC-PCR results (Houf et al., 2002). Out of them, 53 isolates were identified as A. butzleri, 24 as A. cryaerophilus and two as A. skirrowii by both m-PCR (Douidah et al., 2010; Houf et al., 2000); and three as A. thereius by one of them (Douidah et al., 2010). The identification of the remaining Arcobacter isolate required phylogenetic and phylogenomic analyses, which identified it as Aliarcobacter (now Arcobacter) vitoriensis (Alonso et al., 2020). The distribution of the recovered species according to the type of food differed significantly (p =0.009) and it was as follows: A. butzleri was isolated from all type of products except chard; A. cryaerophilus from seafood and meat products; and A. skirrowii and A. thereius only from seafood. A. vitoriensis was recovered from a carrot sample that was simultaneously contaminated with A. butzleri. Three cockle samples were doubly contaminated with A. butzleri and A. cryaerophilus; a shrimp sample with A. butzleri and A. thereius; and a squid sample with A. cryaerophilus and A. thereius. A. butzleri was the most commonly isolated species from all of the food products except cockle and shrimp, where A. cryaerophilus prevailed. Approximately the half (50.9%) of the A. butzleri isolates were recovered from terrestrial animal products and the majority of the A. cryaerophilus from seafood (87.5%). Indeed, Fisher exact test established an association between seafood products and A. cryaerophilus (p <0.001). 3.2. Genotyping by MLST Eighty two isolates out of the 83 analysed were successfully typed by MLST and numerous alleles and STs were identified (Table S2). The allele sequences of the A. vitoriensis isolate could not be determined. A total of 351 alleles were identified across all seven loci, ranging from 42 alleles at gltA to 62 at glyA. Overall, 172 out of the 351 (49%) alleles were previously unreported, ranging their frequency from 33.3% (gltA) to 64.5% (glyA). Sixty eight STs (41 of A. butzleri, 22 of A. cryaerophilus, I. Martinez-Malaxetxebarria et al. International Journal of Food Microbiology 373 (2022) 109712 4 Table 1 Arcobacter species recovered from the 220 food samples purchased at the retail level in Vitoria-Gasteiz between May and November 2015. No. (%) of samples positive for: No. of genotypes identified by ERIC-PCR for: Type of sample No. Arcobacter spp. A. butzleri A. cryaerophilus A. skirrowii A. thereius A. vitoriensis A. butzleri A. cryaerophilus A. skirrowii A. thereius A. vitoriensis Seafood Cockle 20 7 (35.0) a 4 (57.1) 6 (85.7) – – – 6 9 – – – Squid 20 12 (60.0) b, * 7 (58.3) 3 (25) 1 (8.3) 2 (16.7) – 9 4 1 2 – Shrimp 20 7 (35.0) c 3 (42.9) 3 (42.9) 1 (14.3) 1 (14.3) – 3 8 1 1 – Subtotal 60 26 (43.3)* 14 (53,8) 12 (46.2) 2 (7.7) 1 (3.8) – 18 21* 2 3 Terrestrial animal products Turkey meat 20 8 (40.0) 7 (87.5) 1 (12.5) – – – 15 1 – – – Rabbit meat 20 3 (15.0) 2 (66.7) 1 (33.3) – – – 2 1 – – – Quail meat 20 5 (25.0) d 5 (100) 1 (20) – – – 9 1 – – – Fresh cheese 20 1 (5.0) 1 (100) – – – – 1 – – – – Subtotal 80 17 (21.3) 15 (88,2)* 3 (17,6) – – – 27 3 – – – Vegetables Carrot 20 4 (20.0) e 4 (100) – – – 1 (20) 6 – – – 1 Spinach 20 1 (5.0) 1 (100) – – – – 1 – – – – Lettuce 20 1 (5.0) 1 (100) – – – – 1 – – – – Chard 20 – – – – – – – – – – – Subtotal 80 6 (7.5) 6 (100)* – – – 1 (16.7) 8 – – – 1 Total 220 49 (22.3) 35 (71.4) 15 (30,6) 2 (4.1) 1 (2.0) 1 (2.0) 53 24 2 3 1 a A. butzleri and A. cryaerophilus were simultaneously detected in three cockle samples. b A. cryaerophilus and A. thereius were simultaneously detected in a squid sample. c A. butzleri and A. thereius were simultaneously detected in a shrimp sample. d A. butzleri and A. cryaerophilus were simultaneously detected in a quail meat sample. e A. butzleri and A. vitoriensis were simultaneously detected in a carrot sample. * Statistically significant differences (p <0.05) between results, based on Fisher's exact test. I. Martinez-Malaxetxebarria et al. International Journal of Food Microbiology 373 (2022) 109712 5 two of A. skirrowii and three of A. thereius) were identified among the 82 genotyped isolates and their occurrence in seafood, foods of terrestrial animal origin and vegetables was 33, 28 and 7, respectively (Table 2). Most of the STs (89.7%) were previously unreported and resulted from new allele's sequences (n =50) or new combinations of known alleles (n =11). Fifty three out of the 61 novel STs (86.9%) were represented by a single strain; five STs (three of A. butzleri and two of A. cryaerophilus) by two; ST-512 by three (A. butzleri); ST-513 by four (A. butzleri) and ST517 by five (A. butzleri). The analysis of the relatedness and distribution of the identified STs and the other Arcobacter spp. STs available in de PubMLST database (Fig. 1) revealed that, overall, the STs clustered by species. The A. butzleri genotypes identified in this study distributed among all the species-specific clades except one mainly populated by isolates of human origin, and they mostly grouped in a cluster principally composed by isolates derived from food products. The A. cryaerophilus genotypes identified were more closely related to each other and they all except two grouped in the main species-specific clade. The three A. thereius and two A. skirrowii identified genotypes also clustered close to each other. No apparent host-associated STs were identified. Nevertheless, three novel STs from A. butzleri (ST-512, ST513 and ST-517, represented by two, three and five isolates, respectively) and two novel STs from A. cryaerophilus (ST-521 and ST-596, represented by two isolates each) only included isolates from seafood products (Table 2 and Fig. 1). Moreover, the Fisher exact test assessed an association (p <0,001) between seafood derived isolates of these study and the above mentioned STs. Among the previously identified STs, all except ST-16 were shared with isolates derived from different food products (Fig. 1). The recombinant analysis (RDP3) held with the 648 available concatenated MLST sequences (3341 positions) detected a potential single recombination event on ST-599, at the region between nucleotide positions 1968 and 2081, within the sequence of glyA allele. ST-609 (glyA-613, A. cryaerophilus) and ST-250 (glyA-263, A. skirrowii) were identified as the potential parents (major and minor, respectively). This event was statistically supported by all the implemented methods: RDP (1.686 ×10 −06 ), Geneconv (1.233 ×10 −10 ), MaxChi (1.602 × 10 −04 ), Chimaera (2.141 ×10 −04 ), and 3Seq (5.106 ×10 −11 ). 3.2.1. Putative virulence genes The presence and distribution of the ten putative virulence genes investigated is shown in Table 3. The profiles of virulence genes identified in the different species are shown in Table S2. Overall, all the genes were detected among the isolates analysed and, based on the virulence gene content of each isolate, 28 different profiles were identified. None of the isolates harboured all ten virulence genes, and only one isolate with apparently no virulence gene content was identified. Among the arcobacters isolated ciaB (97.6%) and mviN (94%) were the Table 2 Distribution of the STs identified among Arcobacter species. Source of isolation A. butzleri A. cryaerophilus A. skirrowii A. thereius Arcobacter spp. No. of strains No. of ST Identified ST No. of strains No. of ST Identified ST No. of strains No. of ST Identified ST No. of strains No. of ST Identified ST No. of strains No. of ST Seafood 18 9 21 19 2 2 3 3 44 33 Cockle 6 5 475, 513, 517 a (n =2), 519, 530, 9 9 415, 534, 535, 596 a , 597, 598, 599, 606, 607 Squid 9 6 18, 172, 512 a (n =2), 513 a (n = 2), 517 a (n =2), 518 4 4 521 a , 605, 608, 609 1 1 532 2 2 623, 624 Shrimp 3 3 512 a , 513 a , 517 a 8 8 521 a , 522, 596 a , 601, 602, 603, 610, 611 1 1 622 1 1 625 Terrestrial animal products 27 25 3 3 30 28 Turkey meat 15 14 16, 452, 514, 515, 516, 520 (n =2), 523, 524, 525, 527, 528, 536, 594, 595 1 1 604 Rabbit meat 2 2 406, 586 1 1 626 Quail meat 9 8 3, 506, 507, 508, 509, 510 (n = 2), 511, 587 1 1 600 Fresh cheese 1 1 533 Vegetables 8 7 8 7 Carrot 6 5 526 (n =2), 531, 588, 589, 590 Spinach 1 1 593 Lettuce 1 1 529 Total food products 53 41 24 22 2 2 3 3 82 68 Boldface entries represent STs detected in two or more isolates. a Fisher's exact test based statistically significant association (p <0.001) between STs and seafood derived A. butzleri and A. cryaerophilus. I. Martinez-Malaxetxebarria et al. International Journal of Food Microbiology 373 (2022) 109712 6 Fig. 1. Minimum spanning trees based on the MLST profiles of the isolates genotyped in this study and all other Arcobacter isolates from diverse sources available in the PubMLST, showing the relatedness and distribution of the STs among species (A) and diverse sources (B). Each circle represents an ST type and the size of the circle correlates to the number of isolates. The number next to the nodes indicates STs in the present study. I. Martinez-Malaxetxebarria et al. International Journal of Food Microbiology 373 (2022) 109712 7 most prevalent genes, and irgA (8.4%) the least one. The prevalence of the other genes varied from 75.9% for cadF to 18.1% for iroE and hecA. The statistical analyses revealed that cadF, Cj1349, pldA and tlyA genes were significantly (p <0.001) more prevalent among the strains derived from terrestrial animal products (100%, 90%, 90% and 93.3%, respectively) and vegetables (88.9% each gene) than in those derived from seafood (56.8%, 40.9%, 40.9% and 47.7%, respectively). Specifically, cadF was common to all quail and turkey derived isolates and significantly more common in carrot derived ones (85.7%); and Cj1349, pldA and tlyA were significantly more common in quail (90.0% each), turkey (93.8% Cj1349; 93.8% pldA; 100% tlyA) and carrot (90.0% Cj1349 and pldA; 93.3% tlyA) derived isolates. The genes Cj1349 and pldA were also significantly prevalent in those isolates obtained from squids (56.3% each). Significant differences (p <0.05) of the gene distribution were also observed when the species were considered. All the 53 A. butzleri isolates Table 3 Presence and distribution of putative virulence genes in Arcobacter spp. analysed in this study. Species/source No. of strains No. (%) of strains generating specific gene amplicon cadF ciaB Cj1349 hecA hecB irgA mviN pldA tlyA iroE By species A. butzleri 53 53 (100)* 53 (100) 53 (100)* 9 (17) 16 (30.2)* 6 (11.3) 52 (98.1) 53 (100)* 53 (100)* 12 (22.6) A. cryaerophilus 24 8 (33.3) 22 (91.7) 0 6 (25) 0 1 (4.2) 22 (91.7) 0 3 (12.5) 3 (12.5) A. skirrowii 2 0 2 (100) 0 0 0 0 2 (100) 0 1 (50) 0 A. thereius 3 2 (66.7) 3 (100) 0 0 0 0 1 (33.3) 0 0 0 A. vitoriensis 1 0 1 (100) 0 0 0 0 1 (100) 0 0 0 Arcobacter spp. 83 63 (75.9) 81 (97.6) 53 (63.9) 15 (18.1) 16 (19.3) 7 (8.4) 78 (94) 53 (63.9) 57 (68.7) 15 (18.1) By source of isolation Seafood Cockle 15 11 (73.3) 13 (86.7) 6 (40.0) 3 (20.0) 3 (20.0) 0 14 (93.3) 6 (40.0) 7 (46.7) 2 (13.3) Squid 16 10 (62.5) 16 (100) 9 (56.3)* 5 (31.3) 5 (31.3) 1 (6.3) 14 (87.5) 9 (56.3)* 10 (62.5) 3 (18.8) Shrimp 13 4 (56.8) 13 (100) 3 (23.1) 3 (23.1) 1 (7.7) 0 12 (92.3) 3 (23.1) 4 (30.8) 0 Subtotal 44 25 (56.8) 42 (95.5) 18 (40.9) 11 (25) 9 (20.5) 1 (2.3) 39 (88.6) 18 (40.9) 21 (47.7) 5 (11.4) Terrestrial animal products Rabbit meat 3 3 (100) 3 (100) 2 (66.7) 2 (66.7) 1 (33.3) 1 (33.3) 3 (100) 2 (66.7) 2 (66.7) 2 (66.7) Quail meat 10 10 (100)* 10 (100) 9 (90.0)* 1 (10.0) 0 3 (30.0) 10 (100) 9 (90.0)* 9 (90.0)* 1 (10.0) Turkey meat 16 16 (100)* 16 (100) 15 (93.8)* 1 (6.3) 4 (25.0) 2 (12.5) 16 (100) 15 (93.8)* 16 (100)* 3 (18.8) Fresh cheese 1 1 (100) 1 (100) 1 (100) 0 1 (100) 0 1 (100) 1 (100) 1 (100) 1 (100) Subtotal 30 30 (100)* 30 (100) 27 (90.0)* 4 (13.3) 6 (20) 6 (20.0) 30 (100) 27 (90)* 28 (93.3)* 7 (23.3) Vegetables Spinach 1 1 (100) 1 (100) 1 (100) 0 0 0 1 (100) 1 (100) 1 (100) 0 Lettuce 1 1 (100) 1 (100) 1 (100) 0 0 0 1 (100) 1 (100) 1 (100) 1 (100) Carrot 7 6 (85.7)* 7 (100) 6 (85.7)* 0 1 (14.3) 0 6 (85.7) 6 (85.7)* 6 (85.7)* 2 (28.6) Subtotal 9 8 (88.9)* 9 (100) 8 (88.9)* 0 1 (11.1) 0 8 (88.9) 8 (88.9)* 8 (88.9)* 3 (33.3) Total all food products 83 63 (75.9) 81 (97.6) 53 (63.9) 15 (18.1) 16 (19.3) 7 (8.4) 78 (94) 53 (63.9) 57 (68.7) 15 (18.1) * Statistically significant differences (p <0.05) between results, based on Fisher's exact test. Table 4 Presence and distribution of the biofilm-associated genes in Arcobacter spp. analysed in this study. Species/source No. of strains No. (%) of strains generating specific gene amplicon fliS luxS pta waaf Spot flaA flaB By species A. butzleri 53 53 (100) 53 (100) 53 (100) 53 (100) 53 (100) 36 (67.9) 36 (67.9) A. cryaerophilus 24 24 (100) 24 (100) 24 (100) 24 (100) 24 (100) 24 (100) 17 (58,3) A. skirrowii 2 2 (100) 2 (100) 2 (100) 2 (100) 2 (100) 2 (100) 2 (100) A. thereius 3 3 (100) 3 (100) 3 (100) 3 (100) 3 (100) 3 (100) 0 A. vitoriensis 1 1 (100) 1 (100) 1 (100) 1 (100) 1 (100) 1 (100) 1 (100) Arcobacter spp. 83 83 (100) 83 (100) 83 (100) 83 (100) 83 (100) 66 (79.5) 56 (67.5) By source of isolation Seafood Cockle 15 15 (100) 15 (100) 15 (100) 15 (100) 15 (100) 13 (86.7) 13 (86.7) Squid 16 16 (100) 16 (100) 16 (100) 16 (100) 16 (100) 14 (87.5) 9 (56.3) Shrimp 13 13 (100) 13 (100) 13 (100) 13 (100) 13 (100) 12 (92.3) 9 (69.2) Subtotal 44 44 (100) 44 (100) 44 (100) 44 (100) 44 (100) 39 (88.6) 31 (70.5) Terrestrial animal products Rabbit meat 3 3 (100) 3 (100) 3 (100) 3 (100) 3 (100) 3 (100) 2 (66.7) Quail meat 10 10 (100) 10 (100) 10 (100) 10 (100) 10 (100) 7 (70) 6 (60) Turkey meat 16 16 (100) 16 (100) 16 (100) 16 (100) 16 (100) 9 (56.3) 10 (62.5) Fresh cheese 1 1 (100) 1 (100) 1 (100) 1 (100) 1 (100) 0 0 Subtotal 30 30 (100) 30 (100) 30 (100) 30 (100) 30 (100) 19 (63.3) 18 (60) Vegetables Carrot 7 7 (100) 7 (100) 7 (100) 7 (100) 7 (100) 6 (85.7) 6 (85.7) Spinach 1 1 (100) 1 (100) 1 (100) 1 (100) 1 (100) 0 0 Lettuce 1 1 (100) 1 (100) 1 (100) 1 (100) 1 (100) 1 (100) 1 (100) Subtotal 9 9 (100) 9 (100) 9 (100) 9 (100) 9 (100) 7 (7.8) 7 (7.8) Total all food products 83 83 (100) 83 (100) 83 (100) 83 (100) 83 (100) 66 (79.5) 56 (67.5) I. Martinez-Malaxetxebarria et al. International Journal of Food Microbiology 373 (2022) 109712 8 were positive for cadF, ciaB, Cj1349, pldA and tlyA, while the detection rate of the other genes ranged from 11.3% for irgA to 98.1% for mviN. The most frequent combination of genes in A. butzleri was cadF, ciaB, Cj1349, mviN, pldA and tlyA, which was detected in 47.2% of the isolates. The gene content of the A. cryaerophilus isolates was notably lower: only five out of the 24 tested strains showed four genes or more. The most prevalent genes in this species were ciaB and mviN (91.7% each), whose combination was noted in 37.5% of the isolates. Cj1349, hecB and pldA were not detected and the prevalence of the remaining genes was variable, ranging from 4.2% (irgA) to 33.3% (cadF). Both A. skirrowii isolates were positive for ciaB and mviN, being one of them also positive for tlyA. All the three A. thereius isolates possessed ciaB, two of them cadF and the third one also mviN. Arcobacter vitoriensis was positive just for ciaB and mviN. 3.3. Biofilm production All the Arcobacter isolates were positive for fliS, luxS, pta, waaF and spoT genes, but 17 out of the 83 (20.5%) isolates resulted negative for flaA and flaB and ten (12%) for flaB (Table 4). Among these, all were motile upon examination on thioglycolate soft agar plates (data not shown). The biofilm activity could only be tested with 81 of the 83 strains included in the study (we were unable to recover two A. cryaerophilus isolates, Ac-BER3 and Ac-CH1, from the strain collection). The initially measured adherence by microtitter assay is shown in Table S3. Table 5 summarizes the distribution and categorization of the tested isolates among food products and surfaces. Overall, 19 isolates (23.5%) were able to form biofilms on polystyrene surfaces under the experimental conditions. Among them, eight (42.1%) were categorized as weakly adherent, another eight (42.1%) as moderate, and three (15.8%) as strongly adherent. Nevertheless, based on Kruskal-Wallis, no significantly higher adhesion ability was identified among these adherent isolates. The proportion of adherent isolates differed significantly (p = 0.037) among species: 32.1% (17 isolates) in A. butzleri, 9.1% (2 isolates) in A. cryaerophilus and 0% in A. skirrowii, A. thereius and Table 5 Distribution and categorization of the isolates based on their adhesion ability among food sources and surfaces. Polystyrene Borosilicate No. () of tested strains No. (%) of adherent strains on: No. (%) of strains categorized as: No. (%) of adherent strains on polystyrene categorized as: WA b MA c SA d NA a WA b MA c SA d By species A. butzleri (53) 17 (32.1)* 8 (47.1) 6 (35.3) 3 (17.6) 1 (5.9) 2 (11.8) – 14 (82.3) A. cryaerophilus (22) 2 (9.1) – 2 (100) – – – 1 (50) 1 (50) A. skirrowii (2) – – – – – – – – A. thereius (3) – – – – – – – – A. vitoriensis (1) – – – – – – – – Arcobacter spp. (81) 19 (23.5) 8 (42.1) 8 (42.1) 3 (15.8) 1 (5.3) 2 (10.5) 1 (5.3) 15 (78.9) By source of isolation Arcobacter Ab e Ac f As g At h Av i Seafood (42) Cockle (6 Ab, 8 Ac) 4 (9.5) 4 (66.7)* – – – – 1 Ab (25) 2 Ab (50) 1 Ab (25) – 1 Ab (25) – 3 Ab (75) Squid (9 Ab, 3 Ac, 1 As, 2 At) 5 (11.9) 5 (55.6)* – – – – 3 Ab (60) 2 Ab (40) – 1 Ab (20) – – 4 Ab (80) Shrimp (3 Ab, 8 Ac, 1 As, 1 At) 3 (7.1) 1 (33.3) 2 (25) – – – 1 Ab (100) 2 Ac (100) – – – 1 Ac (50) 1 Ab (100), 1 Ac (50) Subtotal (18 Ab, 19 Ac, 2 As, 3 At) 12 (28,6) 10 (55.6)* 2 (10.5) – – – 5 (41.7) 6 (50) 1 (8.3) 1 (8.3) 1 (8.3) 1 (8.3) 9 (75) Terrestrial animal products (30) Rabbit meat (2 Ab, 1 Ac) 1 (3.3) 1 (50) – – – – 1 Ab (100) – – – – – 1 Ab (100) Quail meat (9 Ab, 1 Ac) 3 (10) 3 (33.3) – – – – 2 Ab (66.7) – 1 Ab (33.3) – – – 3 Ab (100) Turkey meat (15 Ab, 1 Ac) 1 (3.3) 1 (6.7) – – – – – – 1 Ab (100) – 1 Ab (100) – – Fresh cheese (1 Ab) – – – – – – – – – – – – – Subtotal 5 (16,7) 5 (18.5) – – – – 3 (60) – 2 (40) – 1 (20) – 4 (80) Vegetables (8) Spinach (1) 1 (12.5) 1 (100) – – – – – 1 Ab (100) – – – – 1 Ab (100) Lettuce (1) – – – – – – – – – – – – Carrot (6 Ab, 1 Av) 1 (12.5) 1 (16.7) – – – – – 1 Ab (100) – – – 1 Ab (100) Subtotal 2 (25) 2 (25) – – – – – 2 (100) – – – – 2 (100) Total all food products (81) 19 (23.5) 17 (32.1)* 2 (9.1) – – – 8 (42.1) 8 (42.1) 3 (15.8) 1 (5.3) 2 (10.5) 1 (5.3) 15 (78.9) a NA, no adherent. b WA, weakly adherent. c MA, moderately adherent. d SA, strongly adherent. e Ab, A. butzleri. f Ac, A. cryaerophilus. g As, A. skirrowii. h At, A. thereius. i Av, A. vitoriensis. * Statistically significant differences (p <0.05) between results, based on Fisher's exact test. I. Martinez-Malaxetxebarria et al. International Journal of Food Microbiology 373 (2022) 109712 9 A. vitoriensis. Regarding the source of isolation, adherent isolates were detected in all type of food products except lettuce and fresh cheese. The distribution of the A. butzleri adherent isolates varied significantly (p = 0.03) among sources, and was as follows: ten adherent isolates derived from seafood (58.8%), five from foods of terrestrial animal origin (29.4%) and two from vegetables (11.8%). Specifically, the number of adherent strains among cockle and squid-derived isolates was significantly higher (p =0.034). Both A. cryaerophilus adherent isolates derived from seafood. The influence of the material on the adhesion ability of the 19 adherent strains was also tested, and the results are shown in Table 6. Eighteen isolates (94.7%) were able to adhere to borosilicate surfaces, and all of them to stainless steel. The adhesion ability of the majority of the strains (84.2%) was higher on borosilicate than on polystyrene, ranging the increase in BFI values from approximately double to more than 14 times higher. In fact, six weakly and seven moderately adherent isolates were categorized as strongly adherent when tested on borosilicate. The increased biofilm formation capability on borosilicate surface was significant for Ab-CH8 (p =0.02), Ab-CH11 (p =0.037) and Ac-G2 (p =0.044) according to the Student t-test. Based on Kruskal-Wallis, no strain was significantly more adherent than other on borosilicate. Regarding stainless steel, the number of viable cells that adhered to the coupons ranged from 0.69 ±0.43 to 3.66 ±0.26 log CFU per cm 2 , and based on Kruskal-Wallis, the adhesion ability of Ab-CH11 was significantly higher (p =0.002). 4. Discussion Various members of the genus Arcobacter are regarded as emerging food and waterborne pathogens (Collado and Figueras, 2011; Ramees et al., 2017); and their distribution in foods has been widely studied worldwide (Cruzado-Bravo et al., 2020; Fern´ andez et al., 2015; G´ onzalez et al., 2017; Hsu and Lee, 2015; Kietsiri et al., 2021; Laishram et al., 2016; Mottola et al., 2016a, 2016b, 2020, 2021; Uljanovas et al., 2021; Zhang et al., 2019). However, these studies are not so abundant in Spain, being even less common those combining various products in the same survey. In this study, different types of seafood, meat, vegetables and fresh cheese purchased in local markets in the city of Vitoria-Gasteiz were examined for Arcobacter spp., and the recovered isolates were genetically characterized by MLST and virulotyping. Additionally, the biofilm activity of the isolates was studied. Arcobacter spp. was isolated from virtually all the tested food products with an overall prevalence of 22.3%. The most frequently isolated species among the five identified was A. butzleri (71.4%), but the enrichment broth employed for the procedure may have favoured this result, as it is known to benefit the recovery of this species over others (Levican et al., 2016). In line with a previous study carried out in the same geographical area (Nieva-Echevarria et al., 2013), the most highly contaminated products (p <0.001) were those coming from the sea (43.3%), especially the squids (60%). This result confirms the cephalopods to be an important reservoir of Arcobacter spp. (Rathlavath et al., 2017; Zhang et al., 2019) and highlights the importance of squids as a potential source of human infection if consumed raw or poorly cooked. It is known that shellfish such as bivalves constitute natural reservoirs of various marine Arcobacter species (Collado and Figueras, 2011). The seafood derived isolates of this study were identified as A. butzleri, A. cryaerophilus, A. skirrowii and A. thereius. The isolation of three A. thereius strains from squid and shrimp samples was an interesting finding that, in accordance with previous observations (Levican et al., 2014; Zhang et al., 2019), confirms that A. thereius can also be isolated from other sources apart from animal faeces and abortions. Not in line with other reports (Levican et al., 2014; Morej´ on et al., 2017; Mottola et al., 2016b; Nieva-Echevarria et al., 2013; Rathlavath et al., 2017; Zhang et al., 2019), A. cryaerophilus prevailed above A. butzleri among seafood derived isolates. In contrast, A. butzleri was the most common species isolated from meat (90%) and vegetable (88.9%) products, especially from poultry and carrots. The predominance of these species in meat products has been frequently reported (Collado and Figueras, 2011; Khodamoradi and Abiri, 2020; Kim et al., 2019; Nieva-Echevarria et al., 2013; Ohnishi and Hara-Kudo, 2021). In addition, the observed prevalence of arcobacters in turkey (40%) and rabbit (15%) meats is consistent with that reported by Collado et al. (2009) for the same products in the same country (33.3% and 10%, respectively); regardless of the differences noted in comparisons between our results and those obtained Table 6 Biofilm formation ability of the adherent A. butzleri and A. cryaerophilus isolates on different abiotic surfaces. Isolate Polystyrene Borosilicate Stainless steel BFI a Classification b BFI a Classification b log CFU/cm 2 Ab-BER1 0.57 ±0.12 Weak 0.69 ±0.42 Weak 0.69 ±0.43 Ab-BER4 1.05 ±0.15 Moderate 1.86 ±1.06 Strong 1.92 ±0.16 Ab-BER6 0.96 ±0.05 Moderate 3.29 ±2.21 Strong 2.12 ±0.3 Ab-BER7 2.48 ±1.16 Strong 10.56 ±8.75 Strong 1.92 ±0.23 Ab-CH8 0.46 ±0.04 Weak 3.67 ±0.17* Strong 1.79 ±0.24 Ab-CH9 0.74 ±0.22 Moderate 1.94 ±0.25 Strong 1.59 ±0.03 Ab-CH10 0.41 ±0.34 Weak 1.95 ±1.1 Strong 3.28 ±0.23 Ab-CH11 0.76 ±0.13 Moderate 2.55 ±0.48* Strong 3.66 ±0.26 • Ab-CH12 0.41 ±0.22 Weak 0.04 ±0.503 None 2.29 ±0.2 Ab-CZ3 0.53 ±0.26 Weak 2.74 ±3.04 Strong 1.11 ±0.03 Ab-CZ5 0.62 ±0.62 Weak 1.96 ±1.46 Strong 1.31 ±0.34 Ab-CZ6 3.00 ±2.9 Strong 1.90 ±0.77 Strong 2.99 ±0.66 Ab-PV7 1.50 ±1.01 Strong 0.40 ±0.44 Weak 1.83 ±0.15 Ab-CN1 0.65 ±0.64 Weak 3.78 ±1.76 Strong 1.07 ±0.26 Ab-E1 0.74 ±0.39 Moderate 1.82 ±0.83 Strong 2.84 ±0.42 Ab-G1 0.65 ±0.44 Weak 2.96 ±2.45 Strong 1.99 ±0.48 Ab-Z7 0.78 ±0.13 Moderate 11.18 ±11.45 Strong 2.06 ±0.18 Ac-G2 0.73 ±0.13 Moderate 5.11 ±1.35* Strong 3.15 ±2.66 Ac-G4 0.81 ±0.23 Moderate 0.83 ±0.9 Moderate 3 ±2.68 a BFI, biofilm formation index. Values are expressed as means ±standard errors. b Biofilm formation pattern according to Naves et al. (2008). • Kruskal-Wallis based statistically significant (p =0.002) differences obtained when comparing the values obtained for biofilm formation ability on stainless steel for each isolate. * Student t–based statistically significant (p <0.05) differences obtained when comparing biofilm formation on polystyrene versus borosilicate. The higher BFI value, representing the most suitable surface for biofilm formation, is indicated. I. Martinez-Malaxetxebarria et al.