Genotyping and antimicrobial resistance of Staphylococcus aureus isolates from diseased turkeys

Abstract

Staphylococcus aureus is a highly versatile pathogen in a large number of domestic animals, including avian species. To gain deeper insight into the epidemiology and diversity of S. aureus associated with articular disease in domestic turkeys, isolates were collected from infected foot joints of turkeys in Brittany (France). A total of 34 isolates were recovered and characterized by means of antimicrobial resistance, staphylococcal protein A typing, macrorestriction pulsed-field gel electrophoresis and micro-array analysis. Thirty isolates were identified as clonal complex (CC) 398 and methicillin-susceptible S. aureus (MSSA), one was identified as a methicillin-resistant S. aureus (MRSA) CC398 isolate, and the remaining were also MSSA and belonged to CC5, CC101, and CC121. Eleven different antimicrobial resistance patterns were detected, with most isolates resistant to penicillin and tetracycline. Based on all typing methods used, the 34 isolates could be divided into 22 different strains. Results on selected isolates, genotyped using microarrays, indicated a high homogeneity among pathogenic MSSA isolates from turkeys. Moreover, all isolates, except the unique MRSA isolate, carried specific φAvβ prophage avian-niche-specific genes, demonstrating the versatility of S. aureus to adapt to the specific ecological poultry niche.

Introduction

Staphylococcus aureus has long been recognized as an important highly versatile pathogen in both human and veterinary medicine. In poultry, as well as in many other animal species and humans, S. aureus belongs to the normal skin microflora (Devriese, 1980). S. aureus can be involved in diseases such as osteomyelitis, bumble foot, and arthritis in poultry. The pathogen can be isolated from the joints, tendon sheaths and bone of affected animals (Andreasen, 2003). Former studies have demonstrated the existence of a limited number of S. aureus genotypes associated with carriage and pathology in poultry, the majority of isolates belonging to clonal complex (CC) 5 (Fitzgerald, 2012). CC5 is also one of the most successful human-associated lineages (Lindsay, 2010). Recent phylogenetic studies have revealed that the poultry CC5 lineage has only recently diverged from this human lineage (Lowder et al., 2009). Another common lineage in poultry is CC398, which has been shown to be present also in other host species including humans, pigs, dairy cows and veal calves, horses and companion animals (Vanderhaeghen et al., 2012; Fitzgerald, 2012). Recently, whole genome sequencing has proved that humans are the ancestral host of this lineage (Price et al., 2012).

Most studies involving poultry have focused on chickens while there is little information on turkeys. Recently, studies focusing on the presence of methicillin-resistant S. aureus (MRSA) carriage among turkeys (Richter et al., 2012) and turkey meat products (de Boer et al., 2009; Feßler et al., 2011a) have been published. However there is little research on S. aureus involved in disease in turkeys, with the exception of osteomyelitis cases (Alfonso & Barnes, 2006; Corrand et al., 2012). Recently, German results showed that livestock-associated MRSA may represent a larger cause of human infection than estimated previously (Köck et al., 2013). Turkeys represent a potential source of livestock-associated MRSA, stressing the need for closed surveillance. To gain insight into the S. aureus strains present among diseased turkeys, in this study we investigated the clonal diversity and the genetic characteristics of S. aureus isolates recovered from infected joints of turkeys.

Materials and Methods

Bacterial isolates

A total of 34 S. aureus isolates were analysed in this study. The isolates were recovered by Chêne Vert Conseil (Loudéac, France) and Labofarm (Loudéac, France) and they were obtained between 2008 and 2012 from diseased breeder turkeys. Turkeys originated from three different hatcheries in the region of Brittany (France) and were subsequently raised at different farms. Each S. aureus isolate was recovered from a different bird. Twenty-three isolates were obtained from subcutaneous abscesses of feet (colloquially known as bumble foot) during the laying period, and the remaining isolates were recovered from young birds with arthritis during the rearing period. Bacteria from infection sites were cultivated on Columbia CNA agar (selective medium for Gram-positive bacteria) and Columbia sheep blood agar (Thermo Scientific, Oxoid, Basingstoke, Hampshire, UK). Suspected S. aureus colonies were identified using the Staphylococcus Latex Kit (Plasmatec Laboratory, Bridport, UK).

Genotyping

Genomic DNA was purified as described previously (Vanderhaeghen et al., 2010). MRSA was identified using a triplex polymerase chain reaction (PCR) developed for detection of the staphylococcal specific 16S rDNA gene, the S. aureus specific nuc1 gene, and the methicillin resistance mecA gene (Maes et al., 2002). The strains were confirmed as belonging to the CC398 using the sau1-hsdS1 CC398 PCR reaction (Stegger et al., 2011). Strains were typed by staphylococcal protein A (spa) typing according to the Ridom StaphType standard protocol (www.ridom.de/staphtype). Some isolates were also subjected to multilocus sequence typing (http://www.mlst.net/). Briefly, amplicons of the spa and multilocus sequence typing genes were purified with the Nucleospin Extract II kit (Macherey-Nagel, Düren, Germany) and then sequenced using the CEQ 8000 Genetic Analysis System (Beckman Coulter, Suarlée, UK) according to the manufacturer's instructions. The Ridom StaphType software package (Ridom GmbH, Münster, Germany) and the S. aureus multilocus sequence typing website (http://saureus.mlst.net/) were used to assign spa and sequence types (STs), respectively. SCCmec typing of MRSA isolates were performed using the multiplex PCRs described by Kondo et al. (2007).

Macrorestriction pulsed-field gel electrophoresis analysis

Whole DNA from each S. aureus isolate was analysed by SmaI and Cfr9I macrorestriction and pulsed-field gel electrophoresis (PFGE) using a CHEF Mapper system (Bio-Rad Laboratories, Hemel Hempstead, UK). Plugs were prepared according to the protocol of Argudín et al. (2010c). Slices of the plugs were subjected to restriction with SmaI (Fermentas GmbH, Aalst, Belgium) or Cfr9I (Fermentas GmbH) following the manufacturer's instructions. The running electrophoresis conditions were 6 V/cm in 0.5× TBE (45 mM Tris, 45 mM boric acid, 1 mM ethylenediamine tetraacetic acid, pH 8) at 14°C, and runs lasted 18 h with switch times from 2.63 to 63.8 sec. PFGE profiles were compared digitally using BioNumerics software (Version 6.6; Applied Maths, Laethem-Saint-Martin, Belgium). Cluster analysis of Dice similarity indices based on the unweighted pair group method with arithmetic averages was applied to generate a dendrogram depicting the relationships among PFGE profiles. S. aureus NCTC 8325 (National Collection of Type Cultures, Salisbury, UK) was included as a control strain for PFGE analysis.

Antimicrobial susceptibility testing

Minimal inhibitory concentrations of 19 antimicrobials (penicillin, cefoxitin, kanamycin, streptomycin, gentamicin, erythromycin, clindamycin, quinupristin/dalfopristin, linezolid, tiamulin, chloramphenicol, rifampicin, ciprofloxacin, fusidic acid, tetracycline, trimethoprim, sulfamethoxazole, vancomycin and mupirocin) were determined using custom veterinary international Sensititre staphylococci plates (EUST; Trek Diagnostics System, East Grinstead, UK) according to the manufacturer's instructions. The interpretation of minimal inhibitory concentration values was according to the epidemiological cut-off values of the European Committee for Antimicrobial Susceptibility Testing (EUCAST) for S. aureus. Data from the EUCAST minimal inhibitory concentration distribution website was last accessed on 6 August 2013 (http://www.eucast.org).

DNA microarray-based typing and detection of resistance and virulence genes

Twelve isolates were selected for micro-array analysis based on their origin and phenotypic and genotypic profiles. Strains with different antimicrobial resistance phenotypes and PFGE profiles from turkeys originating from the two main hatcheries were characterized using the Identibac S. aureus Genotyping DNA Microarray (Alere Technologies GmbH, Jena, Germany) according to the manufacturer's instructions. The DNA microarray covers 333 genetic markers for S. aureus, including species markers, capsule and agr group typing markers, as well as diverse antimicrobial resistance and virulence factors. A full list including primer and probe sequences is available online (http://identibac.com/en/home.html). Specific PCR assays were conducted for resistance and virulence genes not included in the microarray. Primer sequences and PCR conditions are shown in Table 1.

Table 1. Primers used in the present study.

Gene targeted^a	Forward and Reverse primer sequences (5′-3′)	Amplicon size (base pairs)	Annealing temperature (°C)	Reference
amp(A)	CGTTTGCTTCGTGCATTAAA TTGACACGAAGGAGGGTTTC	656	52	Feßler et al., 2011b
cop(B)	TAGTGGCCATGCACATCATC CCACCAGACAAGAACGGTTT	201	60	This study
czr(C)	TAGCCACGATCATAGTCATG ATCCTTGTTTTCCTTAGTGACTT	632	54	Cavaco et al., 2010
dfr(K)	CAAGAGATAAGGGGTTCAGC ACAGATACTTCGTTCCACTC	229	50	Argudín et al., 2011
erm(T)	ATTGGTTCAGGGAAAGGTCA GCTTGATAAAATTGGTTTTTGGA	536	45	Feßler et al., 2010
lsa(C)	GGCTATGTAAAACCTGTATTTG ACTGACAATTTTTCTTCCGT	429	50	Malbruny et al., 2011
spc	ACCAAATCAAGCGATTCAAA GTCACTGTTTGCCACATTCG'	561	52	Feßler et al., 2010
tet(L)	TCGTTAGCGTGCTGTCATTC GTATCCCACCAATGTAGCCG	267	55	Ng et al., 2001
vga(C)	AGCTGGAGAAAGCAATCCAA TCTTGCTGAAAATCCCGTTC	885	50	Argudín et al., 2011
vga(E)	ATGAAAGAATAGCAATCCCAG GGGTAGGTTGAGTTTGGAG	541	54	Schwendener & Perreten, 2011
SAAV_2008	CACCAATTACAACCGCAAGA AAGCTGGCGAATTTATCGAA	174	59	This study
SAAV_2009	TCTGAGGCAACCTTAACTGGC TGGCTTGCTTTCTTCTGCT	229	57	This study

^a The targeted genes included antimicrobial resistance genes [amp(A) (apramycin resistance), dfr(K) (trimethoprim resistance), erm(T) (macrolide-lincosamide-streptogramin B resistance), spc (spectinomycin resistance), tet(L) (tetracycline resistance), vga(C), vga(E) and lsa(C) (pleuromutilin-lincosamide-streptogramin A resistance)], metal resistance genes [czr(C) (cadmium and zinc resistance) and cop(B) (copper resistance)], and ϕAvβ prophage avian-niche-specific genes [SAAV_2008 (putative ornithine cyclodeaminase) and SAAV_2009 (putative membrane protease of the CAAX family)].

Results

Characterization of S. aureus from diseased turkeys

The 34 isolates investigated in this work were distributed among nine spa types (t150, t034, t571, t588, t645, t899, t2164, t2970, t4652), with t034 being the most common type (Table 2). Most isolates (91.2%, 31/34) showed spa types related to the CC398 (t034, t571, t588, t899, t2970, t4652), and were confirmed as CC398 using the sau1-hsdS1 PCR assay. The three isolates with spa types not related to CC398 (t150, t645, t2164) were negative for the CC398 PCR and belonged to different CCs (Table 2). The control PFGE strain NCTC 8325 has spa type t211 and belongs to ST8 (CC8). Genomic DNA of the non-CC398 isolates and the control strain NCTC 8325 was digested with both SmaI and Cfr9I. In contrast, DNA of the CC398 isolates was resistant to SmaI and susceptible to Cfr9I restriction, yielding seven different profiles (Figure 1). On the basis of the Dice similarity coefficient, a dendrogram with the Cfr9I-PFGE patterns was constructed (Figure 2). At a similarity index of 71.7%, all PFGE profiles of CC398 isolates were grouped into a single cluster, while the non-CC398 isolates and the control NCTC 8325 PFGE strain were clearly separated in the dendrogram.

Figure 1. Cfr9I-PFGE analysis of S. aureus isolates from turkeys. Lane N, PFGE profile of strain NCTC 8325 included as quality control; lanes C1 to C12, profiles generated by S. aureus isolates collected from turkeys. Macrorestriction with SmaI or Cfr9I yielded identical PFGE patterns in the case of profiles C1, C2 and C3. Macrorestriction with SmaI in isolates with profiles C4 to C12 yielded the typical not typeable CC398 profile. Some profiles were closely related and only differed in one additional band (profiles C4 and C5, and profiles C8 and C9).

Figure 2. Dendrogram showing the relatedness between Cfr9I macrorestriction PFGE patterns generated from S. aureus turkey isolates and the control strain NCTC 8325. At a Dice similarity index of 71.7% a cluster grouping the CC398 isolates was detected. The number of isolates is included after the profile type in those cases in which more than one isolate showed that profile.

Table 2. Features of the turkey S. aureus isolates analysed in this study.

Isolate	Year of isolation	Origin^a	Hatchery^b	spa type	CC^c	Cfr9I PFGE profile	R-pattern^d	Resistance phenotype^e	Strain^f
29-G3	2008	AT	A	t034	398	C4	R5	PEN-FUS-TET	1
30-I7	2008	BF	B	t034	398	C4	R8	PEN-ERY-CLI-FUS-TET	2
30-D8	2008	AT	C	t150	101	C1	R0	Susceptible	3
31-D7	2008	BF	C	t034	398	C5	R2	PEN-TET	4
34-G12	2009	BF	B	t034	398	C4	R6	PEN-STR-TET	5
35-D9	2009	BF	A	t034	398	C6	R2	PEN-TET	6
36-I12	2009	AT	A	t034	398	C4	R2	PEN-TET	7
37-G8	2009	BF	A	t034	398	C7	R2	PEN-TET	8
37-G9	2009	BF	A	t034	398	C4	R2	PEN-TET	7
40-B7	2009	BF	A	t034	398	C4	R2	PEN-TET	7
40-C10	2009	BF	B	t034	398	C4	R3	PEN-CIP-TET	9
40-E10	2009	BF	B	t034	398	C4	R2	PEN-TET	7
41-F5	2010	BF	A	t034	398	C4	R2	PEN-TET	7
46-H11	2010	AT	B	t571	398	C9	R3	PEN-CIP-TET	10
51-C11	2011	BF	B	t571	398	C10	R3	PEN-CIP-TET	11
51-H3	2011	BF	A	t034	398	C4	R2	PEN-TET	7
55-E2	2011	AT	B	t571	398	C8	R9	PEN-CLI-TIA-CIP-FUS-TET	12
56-B7	2011	BF	A	t588	398	C4	R2	PEN-TET	13
56-C4	2011	BF	A	t4652	398	C11	R2	PEN-TET	14
56-C5	2011	BF	A	t4652	398	C11	R2	PEN-TET	14
56-C6	2011	BF	A	t4652	398	C11	R5	PEN-FUS-TET	15
56-C7	2011	BF	A	t4652	398	C11	R2	PEN-TET	14
56-C8	2011	BF	A	t2970	398	C11	R2	PEN-TET	16
56-G6	2011	AT	C	t2164	5	C2	R7	PEN-CIP-TET-SMX	17
57-B7	2011	AT	C	t034	398	C4	R2	PEN-TET	7
57-H14	2011	AT	B	t571	398	C8	R10	PEN-ERY-CLI-CIP- FUS-TET	18
59-B4	2011	BF	A	t034	398	C4	R2	PEN-TET	7
59-G11	2012	AT	B	t2970	398	C5	R4	PEN-ERY-TET	19
60-A3	2012	BF	C	t645	121	C3	R1	TET	20
60-A9	2012	BF	B	t571	398	C8	R3	PEN-CIP-TET	21
60-B1	2012	BF	A	t571	398	C8	R3	PEN-CIP-TET	21
62-E10	2012	BF	A	t571	398	C8	R3	PEN-CIP-TET	21
62-E11	2012	AT	B	t571	398	C8	R3	PEN-CIP-TET	21
63-D10	2012	AT	B	t899	398	C12	R11	PEN-CEF-STR-CLI- SYN-TIA-CIP-TET	22

^a AT, S. aureus isolates from sample of a young turkey with arthritis during the rearing period; BF, S. aureus isolates from sample of turkey with bumble foot during laying period.

^b The three different hatcheries are indicated with letters A to C.

^c Clonal complexes on the basis of the spa typing and the sau1-hsdS1 CC398 detection PCR (Stegger et al., 2011) in positive isolates, and on the basis of MLST typing (http://saureus.mlst.net/) for isolates 30-D8 (ST101, CC101), 56-G6 (ST5, CC5) and 60-A3 (ST95, CC121).

^d Resistance (R-) patterns are numbered on the basis of the resistance phenotype, without taking into account minimal inhibitory concentration differences.

^e PEN, penicillin; FOX, cefoxitin; STR, streptomycin; ERY, erythromycin; CLI, clindamycin; SYN, quinupristin/dalfopristin; TIA, tiamulin; CIP, ciprofloxacin; FUS, fusidic acid; TET, tetracycline; SMX, sulfamethoxazole.

^f A strain shares a common spa type, PFGE profile and resistance pattern.

The isolates showed 11 resistance profiles. All strains were susceptible to kanamycin, gentamicin, linezolid, chloramphenicol, trimethoprim, vancomycin and mupirocin (Table 2). Most resistance profiles were represented by a single isolate, and only three profiles were shared by two or more CC398 isolates: R2 (16 isolates), R3 (seven isolates) and R5 (two isolates). The highest incidence of individual resistance among CC398 isolates corresponded to penicillin and tetracycline (100% each), followed by ciprofloxacin (32.4%), fusidic acid (16.1%), clindamycin (12.9%), erythromycin (9.7%), streptomycin and tiamulin (6.5% each), and quinupristin/dalfopristin (3.2%). Only one cefoxitin-resistant isolate (isolate 63-D10) with a high resistance to this antimicrobial (16 µg/µl) was found. The isolate was typed as MRSA by the triplex PCR and carried SCCmec IVa. The combination of the different typing techniques allowed the classification of the 34 isolates investigated in the present work into 21 different strains (Table 2).

Microarray typing and detection of resistance and virulence genes

Selected isolates (corresponding to strains 1, 2, 6, 8, 10, 11, 12, 15, 18, 19, 21 and 22) with different antimicrobial resistance phenotypes and PFGE profiles were characterized by microarray typing (Table 3). Most genes were homogeneously distributed in all strains, including typical S. aureus species marker and regulatory genes (23S-rRNA, gapA, katA, coA, spa, sbi, nuc, fnbA, vraS, eno, saeS), the accessory gene regulator agrI, capsule 5-related genes (capH5, capJ5, capK5), the biofilm production genes of the icaACD operon, genes encoding leukocidins (lukS-F, hlgA, lukX, lukY-variant 1), haemolysins (hla, hld), proteases (aur, sspA, sspB, sspP), adhesion factors (cflA, cflB, cna, ebh, eno, fib, ebpS, fnbA, fnbB, map, sdrC, vwb), immune-evasion factors (isaB, isdA, hysA1, hysA2), a putative transport protein (lmrP), a site-specific deoxyribonuclease subunit X (hsdSx), and staphylococcal superantigen-like proteins from the vSaα genomic islands (setB1, setB2, setB3, setC, ssl1 [set6], ssl2 [set7], ssl4 [set9/ssl4], ssl5 [set3/ssl5], ssl7 [set1/ssl7], ssl9 [set5/ssl9] and ssl10 [set4/ssl10]).

Table 3. Presence of antimicrobial resistance genes, virulence, adhesion and inmunoevasion factors by DNA microarray and PCR test in selected isolates.

Negative hybridizations of genes in all strains, as well as common species or regulatory markers, are not shown. Black dots, positive results for probe hybridization; ×, ambiguous results. Gene alleles are indicated between brackets in those cases in which hybridizations with other alleles were not detected. The strain numbers are according to Table 2. BP, binding protein; MHC-II, major histocompatibility complex class II; MLS, macrolide–lincosamide–streptogramine; QAC, quaternary ammonium compounds; R, resistance.

All strains were penicillin and tetracycline resistant and carried the penicillin–ampicillin resistance operon bla (blaZ, blaI, and blaR) and the tetracycline resistance gene tetM; but three strains also carried the tetracycline resistance gene tetK. The erythromycin resistant strains 2, 18 and 19 carried the gene ermC, and the unique MRSA isolate (strain 22) carried the streptogramine resistance gene vgaA, the qacC efflux pump and some SCCmec genes (mecA, the truncated mecR, ugpQ, ccrA2 and ccrB2) related to SCCmec IV. All methicillin-susceptible S. aureus (MSSA) isolates possessed an intact beta-haemolysin gene (hlb), while the MRSA isolate (strain 22) harboured the hlb gene truncated after the probable insertion of the immune-evasion phage-borne genes sak (staphylokinase), chp (chemotaxis inhibitory protein) and scn (staphylococcal complement inhibitor). Moreover, the MRSA strain also carried the enterotoxin G gene (entG, also called seg), and other adhesion (bbp, sdrD) and immune evasion proteins (mprF), as well as the site-specific deoxyribonuclease subunit 2 (hsdS2).

The 34 isolates investigated in this work were also subjected to specific PCR assays for resistance and virulence genes not included in the microarray, with strain 29-G3 (strain 1) possessing the copper resistance gene copB, the isolate 30-D8 (strain 3) positive for the apramycin resistance gene ampA, and all isolates, except the MRSA strain 22, positive for the presence of φAvβ prophage avian-niche-specific genes SAAV_2008 and SAAV_2009. The remaining investigated genes were not detected.

Discussion

S. aureus is an opportunistic pathogen in food animals, and has been isolated from various farm animal species including pigs, poultry and cattle (Vanderhaeghen et al., 2012; Fluit, 2012). Most research performed among livestock isolates is focused on MRSA isolates, particularly from the lineage CC398, which is highly prevalent among farm animals. As such, little research on S. aureus from poultry has included MSSA isolates, despite their potential pathogenicity. In the present study, S. aureus strains, including both MRSA and MSSA, isolated from diseased turkeys from a region in France over a period of 5 years have been analysed. The results contribute to a better understanding of the epidemiology of S. aureus in this farm bird species.

The molecular characterization of the S. aureus isolates from diseased turkeys revealed that the vast majority belonged to CC398. The spa types detected in the isolates from this study were diverse, being types t034 and t899 previously reported among MRSA isolates from turkey products and healthy and diseased turkeys (Argudín et al., 2010a; Bystroń et al., 2010; Feßler et al., 2011a; Richter et al., 2012; Monecke et al., 2013). The spa type t034 is one of the most common spa types found in CC398 isolates of diverse origin, while t899 has been found in pig isolates (van Duijkeren et al., 2008), as a cause of infection in pig farmers (Pan et al., 2009) and recently in healthy carrier chickens in Belgium (Nemeghaire et al., 2013). Interestingly, the spa type t899 has also been reported in MRSA from CC9 (Wagenaar et al., 2009). In fact, comparative genomic analysis with other STs suggested that the region surrounding the spa gene in t899 isolates was acquired horizontally from an ST9 donor (Price et al., 2012). Direct phylogenetic analysis demonstrated t899 as the first lineage to diverge from the other CC398 isolates, but with the exclusion of the single nucleotide polymorphisms in the region surrounding the spa gene, the t899 lineage clustered with a more divergent clade of European isolates and a clade of human-associated MSSA isolates from France and French Guiana (Price et al., 2012). The remaining spa types found in this study have been previously reported for isolates from pigs (t571 and t2970; Argudín et al., 2010a), pig farmers and horses (t588; van Duijkeren et al., 2010; Lozano et al., 2011) or broiler flocks (t4652; Mulders et al., 2010). The three non-CC398 isolates had spa types that have been found before in pathogenic human MSSA isolates in Europe (Deurenberg et al., 2009; Pérez-Vázquez et al., 2009; Rijnders et al., 2012) and they were associated with the human lineages CC5, CC101 and CC121.

The different antimicrobial usages between countries and animal species can influence the antimicrobial resistance patterns observed in bacterial populations. The S. aureus isolates analysed in this study showed rather homogeneous resistance profiles, and about one-half showed a multi-resistance pattern (resistance to three or more antimicrobials). Most strains were MSSA, and showed typical resistance traits (penicillin and tetracycline resistance) of both MRSA and MSSA CC398 isolates from different origins (Pantosti, 2012, Crombé et al., 2013). Interestingly, the isolates from this study showed low resistance percentages to erythromycin (8.8%) and clindamycin (11.8%), while other studies, particularly those on S. aureus from turkey flocks, turkey meat and products from Germany, have recorded high percentages of erythromycin and/or clindamycin resistance, up to 87%. However, these strains were MRSA and had the typical MRSA CC398 multi-resistance pattern (Feßler et al., 2011a; Richter et al., 2012; Monecke et al., 2013). The MRSA CC398 strain isolated in this study showed a similar multi-resistance pattern, but also amongst MSSA from turkeys, high percentages of erythromycin resistance (93%) were found in Germany (Monecke et al., 2013). Similarly, in MRSA isolates from broiler chickens in Belgium (Nemeghaire et al., 2013) and the Netherlands (Wendlandt et al., 2013) high resistance percentages (up to 50%) to macrolides and lincosamides were reported. Conversely, a study in earlier poultry S. aureus Belgian isolates (recovered between 1970 and 1972) showed a lower percentage of resistance (<20%) to these antimicrobials (Nemati et al., 2008). Regarding the other antimicrobials examined, the prevalence of resistance to ciprofloxacin (32.4%) was similar to a study of MRSA isolates from birds and staff in flocks of turkeys examined in Germany (Richter et al., 2012). Other studies in Germany showed fluoroquinolone resistance percentages of between 13 and 59% in MRSA isolates (Feßler et al., 2011a; Monecke et al., 2013) and of 86% in MSSA isolates (Monecke et al., 2013). Lower fluoroquinolone resistance percentages (about 17%) were also recorded among broiler chickens in the Netherlands (Wendlandt et al., 2013) and MRSA CC398 from different origin in Belgium (Jamrozy et al., 2012). These country and regional variations underlie the importance of the antimicrobial resistance monitoring studies to determine and substantiate treatment regimens and strategies.

Since its discovery, the CC398 lineage has been considered a livestock-associated commensal, but recent findings based on whole genome sequencing (Price et al., 2012) suggest that this lineage originated from humans and spread to livestock. Microarray and sequence-based studies support that CC398 isolates essentially carry the same core genome and have different secreted genes located on mobile genetic elements depending on the host and country of origin (Stegger et al., 2010; McCarthy et al., 2011, 2012; Price et al., 2012). For example, the φSa3 prophage genes (sak, chp and scn) were detected only in the typical human isolates and not in those from pigs or humans in contact with pigs (McCarthy et al., 2011, 2012). In the study by Price et al. (2012), different livestock isolates from 19 countries were included and the loss of these human-niche-specific genes has also been confirmed for CC398 turkey isolates. Moreover, these turkey isolates have acquired the avian-niche-specific genes from the φAvβ prophage, which also belongs to the accessory gene pool of broiler chicken-associated S. aureus ST5 (Price et al., 2012). In accordance with this observation, this study demonstrated the presence of the avian-niche-specific genes SAAV_2008 and SAAV_2009 from the φAvβ prophage in all MSSA isolates including both CC398 and non-CC398 isolates. However, the φAvβ prophage is an hlb-converting phage (Lowder et al., 2009) and the MSSA strains investigated in this study carried an intact hlb gene. This indicates the possibility of a new non-hlb-converting phage carrying the specific SAAV_2008 and SAAV_2009 genes or another location for these determinants. This possibility has already been described for sak-containing prophages. The sak gene is usually related to φSa3 hlb-converting phages, which also carry chp and scn genes (Goerke et al., 2009); but a non-hlb-converting phage (φ6390) with sak has also been described (Goerke et al., 2006). Similar to the sak gene, the SAAV_2008 and SAAV_2009 determinants could also be located in a new non-hlb converting phage. The MRSA strain 22 was negative for the presence of the φAvβ prophage genes, but contained the three typical φSa3 prophage genes sak, chp and scn. These results differed from the study of Monecke et al. (2013), where none of the CC398 S. aureus strains isolates from diseased turkeys in Germany carried the φSa3 prophage genes chp and scn, but 65% of the MSSA and 9% of the MRSA strains examined carried the staphylokinase gene sak together with the enterotoxin gene sea, which is also associated with φSa3 prophages (Monecke et al., 2013). Interestingly, the microarray results on German turkey isolates provided by Monecke et al. (2013) included isolates with intact hlb together with scn or sak and scn, indicating the possibility of other new non hlb-converting phages. The carriage of other virulence factors (including leukocidins, proteases, staphylococcal superantigen-like proteins and the haemolysin genes hla and hld), adhesion and immune-evasion factors were similar between the strains of this study and those of other German turkey isolates (Monecke et al., 2013). Interestingly, the core genome seems stable, and these turkey isolates have a similar genetic background to CC398 isolates from poultry, pigs, bovines and horses (Nemati et al., 2009; Walther et al., 2009; Hallin et al., 2011; Jamrozy et al., 2012). Only minor differences were detected, particularly among the CC398 isolates from this study that were negative for the presence of bbp and sdrD genes, while the MRSA strain 22 carried both genes. These genes might be important for S. aureus pathogenicity in a human host. The Bbp protein is important in the invasion of bone tissue and thus might be of relevance in the pathogenicity of osteomyelitis (Tung et al., 2000), while SdrD might play a role in the interactions between the pathogen and the human immune system and serum or may specifically react to nutrients and other factors present in human blood (Sitkiewicz et al., 2011). The absence of both genes (bbp and sdrD) was also common in German MSSA turkey isolates (Monecke et al., 2013), but also in Belgian MRSA isolates from poultry, pigs, horses, chicken and bovines (Nemati et al., 2009; Jamrozy et al., 2012). Conversely, these genes appeared common among German MRSA isolates from poultry (Monecke et al., 2013) and swine (Kadlec et al., 2009). These results confirm that the mobile genetic elements distributed among CC398 isolates are influenced by host and country.

The unique MRSA isolate (strain 22) carried the enterotoxin G gene, which is carried by the egc cluster of the vSaß genomic island (Argudín et al., 2010b). The egc cluster is considered a nursery of enterotoxins and enterotoxin-like genes, and some variants developed through recombination processes have been described (Argudín et al., 2010b). Interestingly, the enterotoxin genes of the egc locus are usually more common among nasal than invasive isolates, and it has been proposed that the egc-encoded toxins could enhance the potential carriage of S. aureus (van Belkum et al., 2006). An incomplete egc cluster, with only the seg gene, has also been described in one human nasal isolate in Japan (Omoe et al., 2005) and one ST398 MRSA from pig in France (Laurent et al., 2009).

This study provides evidence that S. aureus CC398 has an extraordinary ability to adapt from one reservoir to another, and highlights the need for further surveillance and research on this potentially versatile and pathogenic pathogen.

Acknowledgements

This research was funded by the CODA-CERVA. M. A. Argudín is supported by a research grant from the Fundación Alfonso Martín Escudero.