Food Research International 113 (2018) 456–464

Contents lists available at ScienceDirect

Food Research International

journal homepage: www.elsevier.com/locate/foodres

Distribution and tracking of Clostridium difficile and Clostridium perfringens in T

a free-range pig abattoir and processing plant
⁎
Sergio Álvarez-Péreza, José L. Blancoa, , Rafael J. Astorgab, Jaime Gómez-Lagunac,
Belén Barrero-Domínguezb, Angela Galán-Relañob, Celine Harmanusd, Ed Kuijperd,
Marta E. Garcíaa
a
Department of Animal Health, Faculty of Veterinary Medicine, Complutense University of Madrid, Madrid, Spain
b
Department of Animal Health, Faculty of Veterinary Medicine, University of Cordoba, Cordoba, Spain
c
Department of Anatomy and Comparative Pathology, Faculty of Veterinary Medicine, University of Cordoba, Cordoba, Spain
d
Department of Medical Microbiology, Center of Infectious Diseases, Leiden University Medical Center, Leiden, The Netherlands

A R T I C LE I N FO A B S T R A C T

Keywords: The presence and genetic diversity of Clostridium difficile and C. perfringens along the slaughtering process of pigs
Abattoir reared in a free-range system was assessed. A total of 270 samples from trucks, lairage, slaughter line and
Clostridium difficile quartering were analyzed, and recovered isolates were toxinotyped and genotyped. C. difficile and C. perfringens
Clostridium perfringens were retrieved from 14.4% and 12.6% of samples, respectively. The highest percentage of positive samples for C.
Free-range pig
difficile was detected in trucks (80%) whereas C. perfringens was more prevalent in cecal and colonic samples
Lairage
Slaughter line
obtained in the slaughter line (85% and 45%, respectively). C. difficile isolates (n = 105) were classified into 17
PCR ribotypes (including 010, 078, and 126) and 95 AFLP genotypes. C. perfringens isolates (n = 85) belonged to
toxinotypes A (94.1%) and C (5.9%) and were classified into 80 AFLP genotypes. The same genotypes of C.
difficile and C. perfringens were isolated from different pigs and occasionally from environmental samples, sug-
gesting a risk of contaminated meat products.

1. Introduction among the most important agents of food-borne disease (Butler,

Thomas, & Pintar, 2015; Fosse et al., 2008), the classification of C.
Detection and tracking of microorganisms along the the food chain difficile as a pathogen causing food-borne outbreaks is still controversial
is of key importance to establish both pathogens' survival throughout a (Warriner, Xu, Habash, Sultan, & Weese, 2017).
particular production chain and how these microorganisms may even- Although organic and eco-friendly pig rearing systems are gaining
tually reach the consumer (Duffy, Lynch, & Cagney, 2008). Although increased importance and popularity, most surveillance studies of C.
virtually any food product can act as a reservoir of pathogenic micro- difficile and C. perfringens prevalence in swine herds, abattoirs and pig
organisms, meat and derivatives are frequently highlighted as im- carcasses published so far have focused on intensively-raised animals.
portant sources of human food-borne infection (Fosse, Seegers, & However, Keessen et al. (2011) and Susick, Putnam, Bermudez, and
Magras, 2008; Nørrung & Buncic, 2008). Accordingly, farm-to-fork Thakur (2012) found similar prevalence and strain types of C. difficile
surveillance systems have been implemented for the most prevalent among conventional and outdoor, antimicrobial-free production sys-
pathogens found in meat (Nørrung & Buncic, 2008). tems. In a previous study, a high prevalence of the epidemic C. difficile
The Gram-positive, spore-forming anaerobes Clostridium difficile and PCR ribotype 078 was detected in Iberian pigs reared in free-range
C. perfringens are frequent colonizers of the intestinal tract of diverse systems in ‘La Dehesa’, a type of human-managed Mediterranean eco-
food animals, and particularly of pigs (Songer & Uzal, 2005). Both system where they feed on acorns and fresh grass, and only peri-
bacterial species have been found in the environment of pig abattoirs parturient sows and pre-weaned (≤45-day-old) piglets are kept in
and in pork meat (Curry, Marsh, Schlackman, & Harrison, 2012; Hall & closed facilities (Álvarez-Pérez et al., 2013).
Angelotti, 1965; Metcalf, Reid-Smith, Avery, & Weese, 2010; Mooyottu In this study we determined the presence of C. difficile and C. per-
et al., 2015; Wu et al., 2017). Nevertheless, while C. perfringens ranks fringens in an abattoir and processing plant of free-range pigs, which

⁎
Corresponding author at: Departamento de Sanidad Animal, Facultad de Veterinaria, Universidad Complutense de Madrid, Avda. Puerta de Hierro s/n, 28040
Madrid, Spain.
E-mail address: [email protected] (J.L. Blanco).

https://doi.org/10.1016/j.foodres.2018.07.040
Received 31 January 2018; Received in revised form 10 July 2018; Accepted 28 July 2018
Available online 30 July 2018
0963-9969/ © 2018 Elsevier Ltd. All rights reserved.
S. Álvarez-Pérez et al. Food Research International 113 (2018) 456–464

have been previously identified as a common source of both clostridia the supernatants were discarded and the precipitates collected using
(Álvarez-Pérez et al., 2013; and unpublished observations). Ad- sterile cotton-tipped swabs and plated onto CLO agar (bioMérieux,
ditionally, bacterial isolates were toxinotyped and further characterized Marcy l'Étoile, France) for selective culturing of C. difficile, and Brucella
genetically to track possible sources of carcass contamination. blood agar (bioMérieux) and tryptone sulfite neomycin agar (TSN; La-
boratorios Conda, Madrid, Spain) for isolation of C. perfringens. In-
2. Material and methods oculated plates were incubated under anaerobic conditions for 48 h to
7 days at 37 °C (for C. difficile) or 46 °C (for C. perfringens).
2.1. Sampling Tonsils (5 g) and meat samples (ham, shoulder and loin, 5 g in total),
obtained at quartering, were diluted in 15 mL of the aforementioned
Sampling was performed in March 2016 (middle of the local peak enrichment broths, mechanically homogenized for 2 min using a
season for the slaughtering of free-range pigs) in an abattoir and pro- Seward 80 stomacher (Seward Medical, London, England) and further
cessing plant located in southern Spain. All the facilities complied with handled as described above. Swabs were introduced in the fecal sam-
the European Union, national and regional regulations on hygiene, food ples, and then cultured into a 10 mL tube containing 5 mL of the en-
safety and animal welfare. Air temperature in the different rooms of the richment broth for C. difficile or BHI for C. perfringens, and further
pig abattoir and processing plant sampled in this study ranged between handled as decribed above. Water samples (25 mL) were filtered
20 °C and 25 °C, except the quartering room with a temperature below through a 0.45-μm-pore-size membrane filter (Millipore Corporation,
12 °C. Systematic cleaning and disinfection of the facilities is carried out Billerica, MA, USA) using Microfil filtration funnels (Millipore) con-
following each slaughtering. In addition, in periods where no slaugh- nected to a vacuum system. Filters were then washed with 20 mL of
tering is performed, more exhaustive and meticulous cleaning and ethanol 70% (v/v) and 20 mL of sterile distilled water, introduced in the
disinfection protocols which include the dismantling of equipments are sterile 50-mL polypropylene tubes containing 15 mL of BHI or C. difficile
carried out. enrichment broth, and further handled as sponge and meat samples.
Two different batches of animals, corresponding with batches at the C. difficile isolates were identified by colony morphology, the typical
beginning and at the end of the same working day (TLSQ1 and TLSQ2, odor of this microorganism and a positive PCR reaction for the species-
respectively), were sampled to determine the prevalence of both clos- specific internal fragment of the gene encoding for triose phosphate
tridia species in Trucks, Lairage, Slaughter line and Quartering (TLSQ) isomerase (tpi) (Lemee et al., 2004). Identification of isolates as C.
(Hernández et al., 2013). The traceability of each pig was strictly fol- perfringens was achieved by observing the typical double-zone hemo-
lowed along the abattoir, and samples were obtained in the following lysis of this species when cultured on blood agar, formation of black
six stages of the production chain: i) trucks at their arrival (T1) and colonies on TSN, Gram staining reaction and microscopic morphology.
after cleaning and disinfection (T2) (floor, walls, ceiling, entrance
ramps and cabin's mat); ii) lairage, prior entry of the pigs (cleaned and 2.3. Toxinotyping of isolates
disinfected, L1) and just after departure to slaughter (dirty, L2); iii) ten
pig carcasses per batch at six different stages (pre-scalding, S1; post- For C. difficile isolates, expression of the genes which encode for
scalding, S2; post-flaming, S3; post-evisceration, S4; post-washing, S5; toxin A and toxin B (tcdA and tcdB, respectively), and the two compo-
and, chilling, S6); iv) tonsils (To), cecal (Ce) and colonic (Co) contents; nents of binary toxin (CDT) (cdtA and cdtB), was detected by PCR as
v) environmental samples from the slaughter line (ES) (scalding water, previously reported (Álvarez-Pérez et al., 2009, 2015). The genes en-
knives and saws) and from the quartering environment (EQ) (ster- coding for C. pefringens major toxins, enterotoxin and the consensus and
ilization water, tables and knives); and vi) quartering samples (Q) atypical forms of β2 toxin (cpb2) were detected as described by Álvarez-
(ham, shoulder and loin) (see details in Table 1). All samples were Pérez et al. (2016) and Álvarez-Pérez, Blanco and García (2016).
collected into sterile containers (for feces, tonsils, meat or water sam-
ples) or with sterile sponges into plastic bags (for samples from car- 2.4. Ribotyping of C. difficile isolates
casses and surfaces), and transported to the laboratory, where they
were stored at −70 °C until analyzed. PCR ribotyping of C. difficile isolates was performed according to the
high-resolution capillary gel-based electrophoresis method of Fawley
2.2. Bacterial isolation and identification et al. (2015). Ribotypes were designated according to the PHLS Anae-
robic Reference Unit (Cardiff, UK) standard nomenclature and the
As direct culturing of C. difficile and C. perfringens is unreliable (e.g. Leiden-Leeds database (The Netherlands). Non-typeable isolates were
because of the low numbers of spores/vegetative cells that may be also compared with the strain database at Leeds University (Dr. W.
present within a sample and the uneven distribution of these; Blanco, Falwey and Prof. M. Wilcox) that encompasses > 600 different types.
Álvarez-Pérez, & García, 2013), isolation of these microorganisms from
all sample types was performed by enrichment culturing. Briefly, 2.5. Amplified fragment length polymorphism (AFLP) typing
sampling sponges were defrosted and cut into half, and the pieces were
then introduced into 50-mL plastic tubes containing 15 mL of the en- Genotyping of all C. difficile and C. perfringens isolates was per-
richment broth used by Blanco et al. (2013) or brain-heart infusion formed by an AFLP method previously described (Álvarez-Pérez,
broth (BHI; TecLaim, Madrid, Spain), for enrichment of C. difficile and Blanco, & García, 2017; Álvarez-Pérez, Blanco, Harmanus, Kuijper, &
C. perfringens, respectively. After 7 days of incubation at 37 °C (for C. García, 2017). The products resulting from the selective amplification
difficile) or 72 h at 46 °C (for C. perfringens) under anaerobic conditions, step were diluted 1/10 in nuclease-free water (Biotools, Madrid, Spain)
2 mL of the liquid cultures were mixed with 2 mL of absolute ethanol and analyzed by capillary electrophoresis using the GeneScan 1200 LIZ
(Panreac) and incubated for 1 h under agitation (200 rpm) at room size standard (Applied Biosystems, Madrid, Spain). All AFLP reactions
temperature. Finally, the tubes were centrifuged at 1520g for 10 min, were performed twice on different days for each strain.

457
 Table 1
Distribution of Clostridium difficile and Clostridium perfringens along the pig slaughtering process in the examined free-range pig abattoir and processing plant.
TLSQ assaya Production stage Sampleb (n) C. difficile C. perfringens Both clostridial
species

No. (%) of No. isolates Ribotypes (no. AFLP types) No. (%) of No. isolates Toxinotypesc (no. AFLP types) No. (%) of positive
S. Álvarez-Pérez et al.

positive samples positive samples samples

TLSQ1 Trucks T1 (5) 4 (80%) 11 010 (1), 078 (2), 126 (1), 572 (7) 0 0
T2 (5) 2 (40%) 5 078 (2), 126 (3) 0 0
Lairage L1 (4) 0 0 0
L2 (4) 2 (50%) 6 078 (3), U01 (1), U09 (2) 0 0
Slaughter line S1 (10) 2 (20%) 5 U02 (3), U07 (2) 0 0
S2 (10) 0 0 0
S3-S6 (40) 0 0 0
To (10) 0 2 (20%) 4 A (1), A/cpb2 + [c] (3) 0
Ce (10) 2 (20%) 4 572 (2), U06 (2) 8 (80%) 22 A (9), A/cpb2 + [a] (12), C (1) 2 (20%)
Co (10) 2 (20%) 6 572 (3), U02 (3) 3 (30%) 8 A (5), A/cpb2 + [a] (2) 0
Quartering Q (10) 0 0 0
Environment ES (7) 1 (14.3%) 3 078 (2) 0 0
EQ. (10) 0 0 0
TOTAL (135) 15 (11.1%) 40 010 (1), 078 (9), 126 (4), 572 (11), U01 (1), U02 (6), U06 (2), 13 (9.6%) 34 A (15), A/cpb2 + [c] (3), 2 (1.5%)
U07 (2), U09 (2) A/cpb2 + [a] (14), C (1)
TLSQ2 Trucks T1 (5) 5 (100%) 15 010 (1), 078 (6), 110 (2), 126 (2), U01 (2), U03 (1) 2 (40%) 2 A (2) 2 (40%)
T2 (5) 5 (100%) 13 078 (8), U03 (3), U04 (1) 1 (20%) 2 A (2) 1 (20%)
Lairage L1 (4) 2 (50%) 4 078 (4) 0 0
L2 (4) 2 (50%) 5 013 (2), 572 (3) 0 0
Slaughter line S1 (10) 7 (70%) 20 078 (5), 181 (2), 202 (6), U01 (2), U08 (3), U09 (2) 2 (20%) 2 A (2) 1 (10%)
S2 (10) 1 (10%) 2 U05 (2) 0 0
S3-S6 (40) 0 0 0

458
To (10) 0 0 0
Ce (10) 1 (10%) 3 110 (2) 9 (90%) 27 A (16), A/cpb2 + [a] (5), C (4) 1 (10%)
Co (10) 1 (10%) 3 110 (3) 6 (60%) 15 A (14), A/cpb2 + [a] (1) 1 (10%)
Quartering Q (10) 0 1 (10%) 3 A (3) 0
Environment ES (7) 0 0 0
EQ. (10) 0 0 0
TOTAL (135) 24 (17.8%) 65 010 (1), 013 (2), 078 (20), 110 (7), 126 (2), 181 (2), 202 (6), 572 21 (15.6%) 51 A (39), A/cpb2 + [a] (6), C (4) 6 (4.4%)
(3), U01 (4), U03 (4), U04 (1), U05 (2), U08 (3), U09 (2)
TLSQ1 + TLSQ2 Trucks T1 (10) 9 (90%) 26 010 (2), 078 (8), 110 (2), 126 (3), 572 (7), U01 (2), U03 (1) 2 (20%) 2 A (2) 2 (20%)
T2 (10) 7 (70%) 18 078 (10), 126 (3), U03 (3), U04 (1) 1 (10%) 2 A (2) 1 (10%)
Lairage L1 (8) 2 (25%) 4 078 (4) 0 0
L2 (8) 4 (50%) 11 013 (2), 078 (3), 572 (3), U01 (1), U09 (2) 0 0
Slaughter line S1 (20) 9 (45%) 25 078 (5), 181 (2), 202 (6), U01 (2), U02 (3), U07 (2), U08 (3), U09 (2) 2 (10%) 2 A (2) 1 (5%)
S2 (20) 1 (5%) 2 U05 (2) 0 0
S3-S6 (80) 0 0 0
To (20) 0 2 (10%) 4 A (1), A/cpb2 + [c] (3) 0
Ce (20) 3 (15%) 7 110 (2), 572 (2), U06 (2) 17 (85%) 49 A (25), A/cpb2 + [a] (17), C (5) 3 (15%)
Co (20) 3 (15%) 9 110 (3), 572 (3), U02 (3) 9 (45%) 23 A (19), A/cpb2 + [a] (3) 1 (5%)
Quartering Q (20) 0 1 (5%) 3 A (3) 0
Environment ES (14) 1 (7.1%) 3 078 (2) 0 0
EQ. (20) 0 0 0
TOTAL (270) 39 (14.4%) 105 010 (2), 013 (2), 078 (29), 110 (7), 126 (6), 181 (2), 202 (6), 572 34 (12.6%) 85 A (54), A/cpb2 + [c] (3), 8 (3%)
(14), U01 (5), U02 (6), U03 (4), U04 (1), U05 (2), U06 (2), U07 A/cpb2 + [a] (20), C (5)
(2), U08 (3), U09 (4)

a
Two different batches of Iberian pigs, corresponding with the first and last batches allocated to the day of sampling (TLSQ1 and TLSQ2, respectively), were analyzed.
b
T1, trucks prior cleaning and disinfection; T2, trucks after cleaning and disinfection; L1, lairage prior entry of the pigs; L2, lairage after exit of the pigs; S1, pre-scalding; S2, post-scalding; S3, post-flaming; S4, post-
evisceration; S5, post-washing; S6, airing; To, tonsils; Ce, cecal contents; Co, colonic contents; Q, quartering samples (ham, shoulder and loin); ES, environment slaughter line (scalding water, knives and saws); EQ,
environment quartering (sterilization water, tables and knives).
Food Research International 113 (2018) 456–464

c
A, toxinotype A; C, toxinotype C; cpb2+, positive PCR result for the consensus [c] or atypical [a] form of the β2 toxin-encoding gene.
S. Álvarez-Pérez et al. Food Research International 113 (2018) 456–464

2.6. Data analysis belonged to toxinotype A and only five isolates (5.9%) were of tox-
inotype C (Table 1). None of the isolates had the enterotoxin-encoding
Dendrograms of AFLP profiles obtained for C. difficile and C. per- gene (cpe) but 20 type A isolates from diverse sample sources were
fringens isolates were created using Pearson's correlation coefficients positive for presence of an atypical form of the β2-encoding gene
and the unweighted-pair group method with arithmetic averages (cpb2), and other five type A isolates (three obtained from the tonsils of
(UPGMA) clustering algorithm, as implemented in PAST v.3.11 the same pig, one from the colonic content of another pig and the re-
(Hammer, Harper, & Ryan, 2001). Isolates clustering with ≥86% si- maining from a truck's floor) were found to carry the consensus form of
milarity were considered to belong to the same AFLP genotype (Killgore cpb2 (Table 1).
et al., 2008; Álvarez-Pérez, Blanco, & García, 2017; Álvarez-Pérez, AFLP-based fingerprinting of C. perfringens isolates yielded 85 dif-
Blanco, Harmanus et al., 2017). ferent peak profiles and 80 distinct genotypes (cp1 to cp80; Fig. 2).
Only four AFLP types grouped together two or more isolates (see details
3. Results below) and all samples from which multiple isolates of C. perfringens
could be obtained (29 in total) yielded two or more different AFLP
3.1. Prevalence of C. difficile and C. perfringens types. Notably, one AFLP type (cp55) clustered together two toxinotype
A and one toxinotype C isolates.
Clostridium difficile and C. perfringens were retrieved from 39
(14.4%) and 34 (12.6%) out of the 270 analyzed samples in total, re-
spectively. Most culture-positive samples yielded only one Clostridium 3.4. Distribution of C. difficile and C. perfringens genotypes along the
species, but eight samples (3% of total) yielded colonies of both C. production chain
difficile and C. perfringens. The distribution of positive samples per TLSQ
assay and production stage is shown in Table 1. Overall, the highest Five out of the 17 PCR ribotypes of C. difficile (29.4%) were found in
percentage of positive samples for C. difficile was detected in the trucks samples obtained at different steps of the pork production chain
(80%, considering T1 and T2) followed by the lairage stage (37.5%, (Table 2): 078 (T, L, S and environmental samples), 110 (T and S), 572
L1 + L2), whereas C. perfringens was more prevalent in the slaughter and U01 (T, L and S) and U09 (L and S).
line (16.7% of positive samples, considering all sample types obtained The tracking of individual AFLP genotypes of C. difficile and C.
at this stage) and, in particular, in the cecal and colonic content of perfringens along the different sample sources investigated is shown in
sampled pigs (85% and 45%, respectively). The overall proportion of Fig. 3. Two C. difficile genotypes were found in multiple samples from
positive samples was higher in the TLSQ2 assay than in TLSQ1 (1.6 the same TLSQ assay (cd38 from TLSQ1, and cd70 and cd79 from
times, for both C. difficile and C. perfringens) and there was some var- TLSQ2), and two additional genotypes were found in samples from both
iation in the distribution of both clostridia (Table 1). TLSQ1 and TLSQ2 (cd74 and cd89) and included isolates of PCR ribo-
types 078 and U09 (Fig. 3). In addition, a same C. perfringens genotype
3.2. Diversity of C. difficile isolates (cp23) was retrieved from TLSQ2 samples obtained from the floor of a
truck and the quartering of a carcass. However, while the truck isolate
A total of 105C. difficile isolates (x ± S.D. = 2.7 ± 0.5 isolates per yielded a positive PCR result for presence of a consensus form of the
positive TLSQ sample) were selected from the original plate cultures for cpb2 gene, the isolate from the quartering sample was cpb2-negative
ribotyping and further characterization. About 72.4% of those isolates
(76/105) could be classified into one of the already known PCR ribo-
types: 078 (34 isolates), 572 (15), 110 (9), 126 (6), 202 (6), 010 (2), Table 2
013 (2) and 181 (2). The toxin profiles and other characteristics of these Toxin profiles and AFLP types of Clostridium difficile ribotypes identified in this
ribotypes are detailed in Table 2. The remaining 29 isolates (27.6% of study.
total) belonged to nine unknown ribotypes, which will be hereafter PCR Toxin profile No. isolates No. No. samplesa
referred to as U01 to U09 (‘U' stands for ‘unknown’; Table 2). ribotype AFLP
AFLP-based fingerprinting grouped C. difficile isolates into 104 peak types Total T L S Q E
profiles and 95 distinct genotypes (designated as cd1 to cd95; Fig. 1).
010 A-B-CDT- 2 2 2 (5.1%) 2
All isolates belonging to ribotypes 078 and 126 (n = 40, in total) and to 013 A + B + CDT- 2 2 1 (2.6%) 1
the toxigenic type U09 (n = 4) were included into two well-defined 078 A + B + CDT+ 34 29 14 (35.9%) 8 3 2 1
groups that clustered apart from the isolates of the other 14 PCR ri- 110 A + B + CDT- 9 7 3 (7.7%) 1 2
botypes (Fig. 1). Although eight AFLP genotypes included multiple 126 A + B + CDT+ 6 6 3 (7.7%) 3
181 A-B-CDT- 2 2 1 (2.6%) 1
isolates, only two out of these eight AFLP genotypes clustered isolates 202 A + B + CDT- 6 6 2 (5.1%) 2
belonging to different PCR ribotypes (genotypes cd74 and cd89, both of 572 A + B + CDT- 15 14 6 (15.4%) 3 1 2
which included two type 078 isolates and one U09 isolate). All samples U01 A + B + CDT- 5 5 5 (12.8%) 2 1 2
from which multiple C. difficile isolates could be recovered (38 in total) U02 A-B-CDT- 6 6 2 (5.1%) 2
U03 A + B + CDT- 4 4 2 (5.1%) 2
yielded two or more different AFLP types, and 23.7% of these also
U04 A + B + CDT- 1 1 1 (2.6%) 1
yielded different PCR ribotypes. U05 A-B-CDT- 2 2 1 (2.6%) 1
U06 A-B-CDT- 2 2 1 (2.6%) 1
3.3. Diversity of C. perfringens isolates U07 A-B-CDT- 2 2 1 (2.6%) 1
U08 A-B-CDT- 3 3 1 (2.6%) 1
U09 A + B + CDT+ 4 4 2 (5.1%) 1 1
Eighty-five C. perfringens isolates (2.5 ± 0.8 isolates per positive
culture of TLSQ samples) were selected for detailed characterization. a
T, trucks; L, lairage; S, slaughter line; Q, quartering; E, environment of the
Toxinotyping of these isolates revealed that 80 of them (94.1%) slaughter line and processing plant.

459
S. Álvarez-Pérez et al. Food Research International 113 (2018) 456–464

(caption on next page)

460
S. Álvarez-Pérez et al. Food Research International 113 (2018) 456–464

Fig. 1. Dendrogram of AFLP profiles obtained for the Clostridium difficile isolates characterized in this study (n = 105). The dendrogram was created by unweighted
pair group method with arithmetic mean (UPGMA) clustering using Pearson's correlation coefficients. Individual AFLP genotypes are distinguished at ≥86%
similarity (red dotted vertical line). The origin of isolates is indicated at the tip of branches (see legend on the lower left corner), followed by PCR ribotype, isolate
and AFLP type designations (shown in red, black and blue, respectively). The two clusters comprising all ribotype 078/126 and U09 isolates are indicated by a green
background. Abbreviations in legend: T, trucks; L, lairage; S, slaughter line; Q, quartering; E, environment of the slaughter line and processing plant; 1, TLSQ1; 2,
TLSQ2. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

(Fig. 3). No other AFLP genotype grouped together C. perfringens iso- of this has encouraged an ongoing discussion about the zoonotic and
lates from different sample sources but three AFLP types included iso- food-borne potential of the 078/126 lineage (Squire & Riley, 2013;
lates obtained from the cecal or colonic content of different pigs (cp12 Warriner et al., 2017). However, direct transmission of C. difficile (from
and cp55, and cp33, respectively). animals to humans or vice versa) has not been yet demonstrated and the
possibility of acquisition from a common environmental source cannot
be excluded (Knetsch et al., 2014; Squire & Riley, 2013).
4. Discussion Regarding the toxigenic diversity of the isolates characterized in this
study, most C. difficile isolates yielded a positive PCR result for the
Previous studies have demonstrated that C. difficile and C. perfrin- genes encoding toxins A, B and/or binary toxin, all of which are re-
gens are common environmental contaminants of abattoirs slaughtering garded as the main virulence factors of the species (Smits, Lyras, Lacy,
intensively-raised pigs (Chan et al., 2012; Hawken, Weese, Friendship, Wilcox, & Kuijper, 2016). Moreover, C. perfringens isolates were clas-
& Warriner, 2013; Rho, Chung, Lee, & Park, 2001; Rodriguez et al., sified into toxinotypes A and C, both of which are common enteric
2013; Wu et al., 2017). However, much less is known about the pre- pathogens of swine (Songer & Uzal, 2005). In addition, some C. per-
valence and diversity of these two anaerobes in abattoirs dealing with fringens isolates of diverse origins had the genes encoding for consensus
pigs raised under free-range conditions (but see Susick et al., 2012). or atypical forms of the β2 toxin, a plasmid-borne pore-forming toxin
In this study, we found that C. difficile and C. perfringens are wide- which may play a role in pathogenesis (Songer & Uzal, 2005; Uzal et al.,
spread environmental contaminants in a free-range pig abattoir and 2014). However, regardless of their origin and AFLP-type, all C. per-
processing plant. Both species were isolated from trucks (including fringens isolates yielded a negative PCR result for the gene encoding
cabin's mats which never came into direct contact with animals) and enterotoxin CPE, which is the main toxin involved in food poisoning in
lairage samples obtained after cleaning and disinfection, indicating that humans (Songer & Uzal, 2005; Uzal et al., 2014). In any case, given the
these procedures were not efficient to eliminate clostridial spores. A huge arsenal of additional toxins that C. perfringens strains can produce
similar conclusion was reached by Hernández et al. (2013) in a survey (Uzal et al., 2014), the mere presence of this species in abattoirs and
for Salmonella spp. Despite these data may be biased due to the fact that food-processing plants should be regarded as a possible threat to public
a single abattoir was sampled, they highlight the potential risk of health.
contamination by C. difficile and C. perfringens when exhaustive Finally, a 27.6% of the C. difficile isolates analyzed in this study
cleaning and disinfection protocols are not applied at every step from belonged to previously unknown PCR ribotypes. Interestingly, one of
the transport of the animals to the lairage. these ribotypes, named U09, clustered with ribotype 078 and 126 iso-
Detailed genetic characterization of the isolates obtained in this lates in the UPGMA dendrogram built from AFLP patterns and, as these
study showed a high genetic diversity for C. difficile and C. perfringens two, included isolates with the genes encoding for toxins A, B and
and revealed the presence of some particular strain types in both en- binary toxin. Thus, U09 could be regarded as a new 078/126-like ri-
vironmental samples and pig carcasses, which agrees with the ob- botype and future studies should try to assess the prevalence and ge-
servations of other authors (Hawken et al., 2013; Wu et al., 2017). netic and phenotypic characteristics of this novel strain type.
Furthermore, high diversity of PCR ribotypes and AFLP was found even
among isolates retrieved from a same sample, thus confirming the re-
5. Conclusions
commendation of examining multiple isolates from culture-positive
clinical and environmental samples (Tanner et al., 2010; Álvarez-Pérez
In conclusion, as previously observed for abattoirs slaughtering in-
et al., 2016). Overall, these results agree with those obtained by
tensively-raised pigs, C. difficile and C. perfringens can be found in free-
Hernández et al. (2013) for Salmonella spp. but contrasts with the low
range pig abattoirs and processing plants. In addition, molecular
genetic diversity detected for Listeria monocytogenes by López et al.
tracking of individual genotypes revealed that, for both clostridia, the
(2008) in a different abattoir. Differences in the pig populations ana-
same strain types could be recovered from animal and environmental
lyzed, sampling methods, target bacterial species and/or techniques
samples, highlighting the potential for cross contamination of free-
used for molecular typing of isolates might account for these dis-
range pig carcasses.
crepancies.
Notably, we identified C. difficile PCR ribotypes which rank among
the most prevalent in outbreaks of human disease, such as ribotypes Declaration of interest
010, 078 and 126 (Davies et al., 2016). Interestingly, the PCR ribotypes
078 and 126 are close phylogenetic relatives (Schneeberg et al., 2013; None.
Stabler et al., 2012) that are frequently recovered from slaughtered
animals and meat products (Metcalf et al., 2010; Curry et al., 2012; Acknowledgments
Hawken et al., 2013; Cho et al., 2015; Mooyottu et al., 2015; Wu et al.,
2017). Moreover, high genetic relatedness between human and animal This work was supported by grant AGL2013-46116-R from the
isolates of the 078/126 ribotype complex has been repeatedly reported Spanish Ministry of Economy and Competitiveness. Jaime Gomez-
(Bakker et al., 2010, Koene et al., 2012, Schneeberg et al., 2013; Laguna is supported by a “Ramón y Cajal” contract of the Spanish
Knetsch et al., 2014; Álvarez-Pérez, Blanco, Harmanus et al., 2017). All Ministry of Economy and Competitiveness (RYC-2014-16735). The

461
S. Álvarez-Pérez et al. Food Research International 113 (2018) 456–464

Fig. 2. Dendrogram of AFLP profiles obtained for the Clostridium perfringens isolates characterized in this study (n = 85). The dendrogram was created by unweighted
pair group method with arithmetic mean (UPGMA) clustering using Pearson's correlation coefficients. Individual AFLP genotypes are distinguished at ≥86%
similarity (red dotted vertical line). The first column of colored squares at the tip of branches indicates the origin of isolates (see colour legend on the lower left
corner). Additionally, green- and violet-filled squares indicate toxinotype A and toxinotype C isolates, respectively, and the presence of the consensus or atypical form
of the β2 toxin-encoding gene (cpb2) is indicated by filled and open circles, respectively. Alphanumeric codes refer to isolate and AFLP type designations (shown in
black and blue, respectively). Abbreviations in legend: T, trucks; S, slaughter line; Q, quartering; 1, TLSQ1; 2, TLSQ2. (For interpretation of the references to colour in
this figure legend, the reader is referred to the web version of this article.)

462
S. Álvarez-Pérez et al. Food Research International 113 (2018) 456–464

(2012). The epidemiology of Clostridium perfringens type a on Ontario swine farms,

with special reference to cpb2-positive isolates. BMC Veterinary Research, 8, 156.
Curry, S. R., Marsh, J. W., Schlackman, J. L., & Harrison, L. H. (2012). Prevalence of
Clostridium difficile in uncooked ground meat products from Pittsburgh, Pennsylvania.
Applied Environmental Microbiology, 78, 4183–4186.
Davies, K. A., Ashwin, H., Longshaw, C. M., Burns, D. A., Davis, G. L., Wilcox, M. H., &
EUCLID study group (2016). Diversity of Clostridium difficile PCR ribotypes in Europe:
results from the European, multicentre, prospective, biannual, point-prevalence study
of Clostridium difficile infection in hospitalised patients with diarrhoea (EUCLID),
2012 and 2013. Euro Surveillance, 21 (pii=30294).
Duffy, G., Lynch, O. A., & Cagney, C. (2008). Tracking emerging zoonotic pathogens from
farm to fork. Meat Science, 78, 34–42.
Fawley, W. N., Knetsch, C. W., MacCannell, D. R., Harmanus, C., Du, T., Mulvey, M. R., ...
Wilcox, M. H. (2015). Development and validation of an internationally-standar-
dized, high-resolution capillary gel-based electrophoresis PCR-ribotyping protocol for
Clostridium difficile. PLoS One, 10, e0118150.
Fosse, J., Seegers, H., & Magras, C. (2008). Foodborne zoonoses due to meat: A quanti-
tative approach for a comparative risk assessment applied to pig slaughtering in
Europe. Veterinary Research, 39, 1.
Hall, H., & Angelotti, R. (1965). Clostridium perfringens in meat and meat products. Applied
Microbiology, 13, 352–357.
Hammer, Ø., Harper, D. A. T., & Ryan, P. D. (2001). PAST: Paleontological Statistics
Software Package for Education and Data Analysis. Palaeontologia Electronica, 4
(9pp).
Hawken, P., Weese, J. S., Friendship, R., & Warriner, K. (2013). Carriage and dis-
semination of Clostridium difficile and methicillin resistant Staphylococcus aureus in
pork processing. Food Control, 31, 433–437.
Hernández, M., Gómez-Laguna, J., Luque, I., Herrera-León, S., Maldonado, A., Reguillo,
L., & Astorga, R. J. (2013). Salmonella prevalence and characterization in a free-range
pig processing plant: Tracking in trucks, lairage, slaughter line and quartering.
Fig. 3. Tracking of individual AFLP genotypes of Clostridium difficile and International Journal of Food Microbiology, 162, 48–54.
Clostridium perfringens along the pork production chain. Detection of each Keessen, E. C., van den Berkt, A. J., Haasjes, N. H., Hermanus, C., Kuijper, E. J., & Lipman,
genotype from the different sample sources is indicated by shaded boxes, which L. J. (2011). The relation between farm specific factors and prevalence of Clostridium
difficile in slaughter pigs. Veterinary Microbiology, 154, 130–134.
also include the PCR ribotype (for C. difficile) or toxin profile (for C. perfringens).
Killgore, G., Thompson, A., Johnson, S., Brazier, J., Kuijper, E., Pepin, J., Frost, E. H.,
Only genotypes detected in two or more sample sources are included. Savelkoul, P., Nicholson, B., Berg, R. J., Kato, H., Sambol, S. P., Zukowski, W., Woods,
Abbreviations for sample sources: T1, trucks prior cleaning and disinfection; T2, C., Limbago, B., Gerding, D. N., & McDonald, L. C. (2008). Comparison of seven
trucks after cleaning and disinfection; L1, lairage prior entry of the pigs; L2, techniques for typing international epidemic strains of Clostridium difficile: restric-
lairage after exit of the pigs; S1, pre-scalding; S2, post-scalding; S3, post- tion endonuclease analysis, pulsed-field gel electrophoresis, PCR-ribotyping, multi-
locus sequence typing, multilocus variable number tandem-repeat analysis, amplified
flaming; S4, post-evisceration; S5, post-washing; S6, chilling; To, tonsils; Ce,
fragment length polymorphism, and surface layer protein A gene sequence typing.
cecal contents; Co, colonic contents; Q, quartering samples (ham, shoulder and Journal of Clinical Microbiology, 46, 431–437.
loin); ES, environment slaughter line (scalding water, knives and saws); EQ, Knetsch, C. W., Connor, T. R., Mutreja, A., van Dorp, S. M., Sanders, I. M., Browne, H. P.,
environment quartering (sterilization water, tables and knives). ... Lawley, T. D. (2014). Whole genome sequencing reveals potential spread of
Clostridium difficile between humans and farm animals in the Netherlands, 2002 to
2011. Euro Surveillance, 19, 20954.
funder had no role in study design, data collection and interpretation, Koene, M. G., Mevius, D., Wagenaar, J. A., Harmanus, C., Hensgens, M. P., Meetsma, A.
M., ... Kuijper, E. J. (2012). Clostridium difficile in Dutch animals: Their presence,
or the decision to submit the work for publication. The staff of the
characteristics and similarities with human isolates. Clinical Microbiology and
Genomics Service at Universidad Complutense de Madrid is gratefully Infection, 18, 778–784.
acknowledged for providing excellent technical assistance. Lemee, L., Dhalluin, A., Testelin, S., Mattrat, M. A., Maillard, K., Lemeland, J. F., & Pons,
J. L. (2004). Multiplex PCR targeting tpi (triose phosphate isomerase), tcdA (toxin A),
and tcdB (toxin B) genes for toxigenic culture of Clostridium difficile. Journal of Clinical
References Microbiology, 42, 5710–5714.
López, V., Villatoro, D., Ortiz, S., López, P., Navas, J., Dávila, J. C., & Martínez-Suárez, J.
Álvarez-Pérez, S., Blanco, J. L., Bouza, E., Alba, P., Gibert, X., Maldonado, J., & García, M. V. (2008). Molecular tracking of Listeria monocytogenes in an Iberian pig abattoir and
E. (2009). Prevalence of Clostridium difficile in diarrhoeic and non-diarrhoeic piglets. processing plant. Meat Science, 78, 130–134.
Veterinary Microbiology, 137, 302–305. Metcalf, D., Reid-Smith, R. J., Avery, B. P., & Weese, J. S. (2010). Prevalence of
Álvarez-Pérez, S., Blanco, J. L., & García, M. E. (2017). Clostridium perfringens type a Clostridium difficile in retail pork. Canadian Veterinary Journal, 51, 873–876.
isolates of animal origin with decreased susceptibility to metronidazole show ex- Mooyottu, S., Flock, G., Kollanoor-Johny, A., Upadhyaya, I., Jayarao, B., &
tensive genetic diversity. Microbial Drug Resistance, 23, 1053–1058. Venkitanarayanan, K. (2015). Characterization of a multidrug resistant C. difficile
Álvarez-Pérez, S., Blanco, J. L., Harmanus, C., Kuijper, E., & García, M. E. (2017). meat isolate. International Journal of Food Microbiology, 192, 111–116.
Subtyping and antimicrobial susceptibility of Clostridium difficile PCR ribotype 078/ Nørrung, B., & Buncic, S. (2008). Microbial safety of meat in the European Union. Meat
126 isolates of human and animal origin. Veterinary Microbiolgy, 199, 15–22. Science, 78, 14–24.
Álvarez-Pérez, S., Blanco, J. L., Peláez, T., Astorga, R. J., Harmanus, C., Kuijper, E., & Rho, M. J., Chung, M. S., Lee, J. H., & Park, J. (2001). Monitoring of microbial hazards at
García, M. E. (2013). High prevalence of the epidemic Clostridium difficile PCR ri- farms, slaughterhouses, and processing lines of swine in Korea. Journal of Food
botype 078 in Iberian free-range pigs. Research in Veterinary Science, 95, 358–361. Protection, 64, 1388–1391.
Álvarez-Pérez, S., Blanco, J. L., Peláez, T., Lanzarot, M. P., Harmanus, C., Kuijper, E., & Rodriguez, C., Avesani, V., Van Broeck, J., Taminiau, B., Delmée, M., & Daube, G. (2013).
García, M. E. (2015). Faecal shedding of antimicrobial-resistant Clostridium difficile Presence of Clostridium difficile in pigs and cattle intestinal contents and carcass
strains by dogs. Journal of Small Animal Practice, 56, 190–195. contamination at the slaughterhouse in Belgium. International Journal of Food
Álvarez-Pérez, S., Blanco, J. L., Peláez, T., Martínez-Nevado, E., & García, M. E. (2016). Microbiology, 166, 256–262.
Water sources in a zoological park harbor genetically diverse strains of Clostridium Schneeberg, A., Neubauer, H., Schmoock, G., Baier, S., Harlizius, J., Nienhoff, H., ...
perfringens type a with decreased susceptibility to metronidazole. Microbial Ecology, Seyboldt, C. (2013). Clostridium difficile genotypes in piglet populations in Germany.
72, 783–790. Journal of Clinical Microbiology, 51, 3796–3803.
Bakker, D., Corver, J., Harmanus, C., Goorhuis, A., Keessen, E. C., Fawley, W. N., ... Smits, W. K., Lyras, D., Lacy, D. B., Wilcox, M. H., & Kuijper, E. J. (2016). Clostridium
Kuijper, E. J. (2010). Relatedness of human and animal Clostridium difficile PCR ri- difficile infection. Nature Review. Disease Primers, 2, 16020.
botype 078 isolates determined on the basis of multilocus variable-number tandem- Songer, J. G., & Uzal, F. A. (2005). Clostridial enteric infections in pigs. Journal of
repeat analysis and tetracycline resistance. Journal of Clinical Microbiology, 48, Veterinary Diagnostic Investigation, 17, 528–536.
3744–3749. Squire, M. M., & Riley, T. V. (2013). Clostridium difficile infection in humans and piglets: A
Blanco, J. L., Álvarez-Pérez, S., & García, M. E. (2013). Is the prevalence of Clostridium ‘one health’ opportunity. Current Topics in Microbiology and Immunology, 365,
difficile in animals underestimated? The Veterinary Journal, 197, 694–698. 299–314.
Butler, A. J., Thomas, M. K., & Pintar, K. D. (2015). Expert elicitation as a means to Stabler, R. A., Dawson, L. F., Valiente, E., Cairns, M. D., Martin, M. J., Donahue, E. H., ...
attribute 28 enteric pathogens to foodborne, waterborne, animal contact, and person- Wren, B. W. (2012). Macro and micro diversity of Clostridium difficile isolates from
to-person transmission routes in Canada. Foodborne Pathogens and Disease, 12, diverse sources and geographical locations. PLoS One, 7, e31559.
335–344. Susick, E. K., Putnam, M., Bermudez, D. M., & Thakur, S. (2012). Longitudinal study
Chan, G., Farzan, A., Soltes, G., Nicholson, V. M., Pei, Y., Friendship, R., & Prescott, J. F. comparing the dynamics of Clostridium difficile in conventional and antimicrobial free

463
S. Álvarez-Pérez et al. Food Research International 113 (2018) 456–464

pigs at farm and slaughter. Veterinary Microbiology, 157, 172–178. community-acquired infection? Journal of Applied Microbiology, 122, 542–553.
Uzal, F. A., Freedman, J. C., Shrestha, A., Theoret, J. R., Garcia, J., Awad, M. M., ... Wu, Y. C., Chen, C. M., Kuo, C. J., Lee, J. J., Chen, P. C., Chang, Y. C., & Chen, T. H.
McClane, B. A. (2014). Towards an understanding of the role of Clostridium perfringens (2017). Prevalence and molecular characterization of Clostridium difficile isolates
toxins in human and animal disease. Future Microbiology, 9, 361–377. from a pig slaughterhouse, pork, and humans in Taiwan. International Journal of Food
Warriner, K., Xu, C., Habash, M., Sultan, S., & Weese, S. J. (2017). Dissemination of Microbiology, 242, 37–44.
Clostridium difficile in food and the environment: Significant sources of C. difficile

464