To isolate and characterize strains of Mycoplasma agalactiae from bulk tank and silo ewes' milk.
To isolate and characterize strains of Mycoplasma agalactiae from bulk tank and silo ewes' milk.
Thirteen mycoplasma isolates were obtained from samples of sheep milk taken from bulk tank and large silos and identified as Myc. agalactiae by PCR-DGGE. The isolates were typed by pulsed field gel electrophoresis (PFGE), SDS-PAGE and immunoblot. The in vitro activity of 13 antimicrobials of veterinary interest was tested against these isolates. Results showed that the most effective compounds against Myc. agalactiae in vitro were clindamycin, an antibiotic not previously described as a suitable contagious agalactia (CA) treatment, with Minimum Inhibitory Concentration (MIC) values of <0·12 μg ml−1, and quinolones, with MIC values <0·12–0·5 μg ml−1, which are used as standard treatments against CA.
Based on the in vitro assay, clindamycin, quinolones, tylosin and tilmicosin would be appropriate antimicrobials for CA treatment. The isolates were mostly resistant to erythromycin, indicating that it would not be a suitable choice for therapy. The isolates showed common molecular and protein profiles by PFGE and SDS-PAGE, with minor differences observed by immunoblot analysis, suggesting a clonal relationship among them.
This study demonstrated the importance of the appropriate selection of antimicrobials for treatment of CA.
Mycoplasma agalactiae is the main causative agent of contagious agalactia (CA), a syndrome that affects sheep and goats and is characterized by mastitis, absence or reduction of milk production, rapid spread and long persistence in affected areas. In most cases, infected animals recover rapidly from acute signs but develop chronic disease with excretion of the agent, mainly in milk and other secretions sometimes persisting for years (Bergonier et al. 1997). CA is of great economic importance in Mediterranean countries as there may be losses of up to 30% in milk production as a result of irreversible changes to the mammary gland. In addition, young animals may develop arthritis and in some cases septicaemia resulting in death or culling of the affected animal (Bergonier et al. 1997; Corrales et al. 2007).
Due to these characteristics, antimicrobials used for CA treatment should have the following properties: very low Minimum Inhibitory Concentration (MIC), long persistence in infected tissue, excretion in milk and easy diffusion from blood to mammary tissue (Bergonier et al. 1997). Antimicrobials chosen traditionally for CA treatment are tetracyclines, macrolides (tylosin and tilmicosin), fluoroquinolones, erythromycin and florfenicol, but this treatment is largely based on studies that have been carried out on Mycoplasma mycoides subspecies mycoides large colony type, now known as Myc. mycoides subspecies capri (Manso-Silvan et al. 2009), which is an important aetiological agent for CA in goats but it is rarely found in sheep. Moreover, the efficacy of antimicrobial therapy under field conditions is limited (Bergonier et al. 1997; Giguère et al. 2006). Animal health and welfare, as well as the economic importance of CA, make it necessary to determine which antimicrobial treatments are effective against Myc. agalactiae in vivo.
The description of the genome sequence of Myc. agalactiae reference strain PG2 (Sirand-Pugnet et al. 2007) has enabled the development of several diagnostic and molecular typing methods, such as VNTR analysis (McAuliffe et al. 2008) and MLST (McAuliffe et al. 2011) that are able to provide useful epidemiological information. Myc. agalactiae also has variable surface proteins (Glew et al. 2000; de la Fe et al. 2006) that may contribute to its pathogenicity and hinder the development of effective vaccines (Dybvig and Voelker 1996; Flitman-Tene et al. 2000; Glew et al. 2000).
Despite the legal and health measures taken in Spain and other European countries to prevent the spread of CA, which includes border movement controls, epidemiological monitoring, vaccination or antimicrobial treatment, a total of 931 outbreaks were reported in Spain between 2010 and 2011, according to data of the World Organization for Animal Health (http://web.oie.int/wahis/public.php?page=disease; accessed December 2012); 924 (919 sheep/5 other species) of the 931 outbreaks were declared in the Castilla y Leon region. This number of infected sheep is considerably higher than the 17 reported outbreaks in the same region in 2007.
The objectives of this work were to study the prevalence of Myc. agalactiae in samples of bulk tank and silo sheep milk and to characterize the isolates by antimicrobial and molecular typing.
From November 2009 to March 2010, 45 samples of sheep milk from the Castilla y León region were processed for this work. Thirty-five were obtained from tanks at dairy farms suspected of having or having had CA and the other 10 were collected from cooling silos at a collection centre where milk from a total of 400 flocks from different collection routes was stored. In both cases, samples were taken aseptically in sterile containers according to the standards recommended by the American Public Health Association (White et al. 1992). Samples of 50 and 500 ml were collected in farms and silos, respectively, and stored at 4°C until processed in the laboratory within the next 24 h.
Upon receipt in the laboratory, samples were homogenized by slowly inverting the container 25 times. An aliquot of 100–200 μl of milk was inoculated in vials containing 3 ml of Eaton′s Broth (Nicholas and Baker 1998) and incubated at 37°C in a 5–10% CO2 atmosphere for 24 h with periodic agitation. After the first incubation, 1–2 ml of culture was filtered using 0·45-μm Minisart filter (Sartorius AG, Goettingen, Germany) and inoculated into a fresh broth and incubated until a pH change and opalescence was seen (Nicholas et al. 2008). Then 25 μl of each broth was subcultured on Eaton's agar (Nicholas and Baker 1998), using the same incubation conditions. Plates were inspected after 3–7 days under 40× magnification and a single typical ‘fried-egg’ colony was picked using a sterile tip and incubated in Eaton's broth vials, until turbidity, to obtain a pure culture. Aliquots in stationary phase were frozen at −80°C in 20% glycerol until required.
The DNA of each strain was extracted from 1-ml culture which was centrifuged at 13 000 g for 5 min in 1·5-ml microtubes. The pellets obtained were washed three times with Phosphate-Buffered Saline (PBS; Oxoid, Basingstoke, UK) pH 7·4. Washed pellets were subjected to a standardized protocol of extraction (GenElute DNA extraction kit; Sigma-Aldrich, Poole, UK) according to the manufacturer's instructions.
16S rDNA PCR and denaturing gradient gel electrophoresis (PCR-DGGE) were performed to identify isolates and to ensure that they were pure Myc. agalactiae cultures according to a published protocol (McAuliffe et al. 2003). The profiles obtained were compared with Myc. agalactiae PG2-type strain and other Mycoplasma species that may affect small ruminants including Myc. mycoides subsp. capri, Myc. mycoides subsp. mycoides SC, Myc. arginini, Myc. capricolum subsp. capricolum, Myc. putrefaciens and Myc. ovipneumoniae.
Minimum inhibitory concentration testing was carried out using a microbroth dilution method following the guidelines of Hannan (2000). Thirteen antimicrobials of veterinary use were tested. These antimicrobials were tylosin, tilmicosin, erythromycin, tulathromycin, danofloxacin, enrofloxacin, marbofloxacin, lincomycin, clindamycin, oxytetracycline, spectinomycin, chloramphenicol and florfenicol at specified concentrations on Sensititre plates (Trek Diagnostics, East Grinstead, West Sussex, UK). Eaton's broth medium without antimicrobials and phenol red was used to culture the Myc. agalactiae strains for the MIC inoculum (Nicholas and Baker 1998). The inoculum concentration was standardized by adjusting the optical density (OD) of the broth medium to an OD at 450 nm of 0·1, which was determined to be equivalent to approximately 108 colony-forming units (CFU) ml−1. Fresh culture was serially diluted in Eaton's broth to obtain a final concentration of 105 CFU ml−1. Wells were inoculated with 10 μl of inoculum added to 190 μl of Eaton's medium. Microplates were incubated statically for 72 h at 37°C with 5% CO2. To assess killing effect after incubation, the Sensititre plates were centrifuged at 800 g to concentrate Myc. agalactiae cells at the bottom of the wells. Cell growth was examined using an inverted mirror in a light box. Absence of growth was considered negative and the first negative well for each antimicrobial was taken as the MIC value. Results were recorded on standard worksheets, and MIC values are given as MIC range (μg ml−1) MIC50 and MIC90 (μg ml−1) in Table 2.
Confirmed Myc. agalactiae strains were reconstituted in 3 ml of Eaton's broth and incubated until growth was observed and then subcultured into 20 ml of the same medium for a further 24–48 h to obtain an appropriate cell concentration. Cultures were then centrifuged at 6000 g and 4°C for 30 min. Pellets were carefully suspended in 2 ml of PBS pH 7·4 and centrifuged again at 13 000 g. This wash was repeated three times with the final pellet being resuspended in 250 μl of PBS pH 7·4 and stored in ice to preserve the cells integrity. Equal volumes of 2% low melting point agarose (Bio-Rad Laboratories, Hercules, CA, USA) and cell suspension were mixed and allowed to solidify in the pulsed-field moulds at 4°C. Plugs were lysed in 2-ml microtubes at 45°C for 48 h using 500 μl of lysis buffer (10 mol l−1 Tris-HCl, 1 mmol l−1 EDTA, 1% N-lauryl sarcosine and 1 mg ml−1 proteinase K). After lysis, the buffer was carefully removed and plugs were washed four times with 2 ml of TE (10 mmol l−1 Tris-HCl plus 1 mmol l−1 EDTA) and stored at 4°C until being digested with the restriction enzyme. Plug slices were digested overnight with 30 U of SmaI (Takara Bio, Otsu, Shiga, Japan) following the manufacturer's instructions. Using the CHEF DR III System (Bio-Rad Laboratories), digested plugs underwent electrophoresis for 18 h at 14°C in a 1% agarose gel in TBE buffer (0·1 mol l−1 Tris, 0·1 mol l−1 boric acid and 2 mmol l−1 EDTA, pH 8) using pulse times of 4–40 s. A Salmonella serotype Braenderup restricted with XbaI (Takara Bio) was used as a molecular weight standard (http://www.pulsenetinternational.org; accessed December 2012). Gels were stained with ethidium bromide, destained in distilled water for 30 min and observed under UV transillumination. Gel pictures were analysed using GelCompar II software (Applied Maths, St-Martens-Latem, Belgium).
Mycoplasma isolates, including reference strain PG2, were cultured in 20 ml of Eaton's broth for 7 days at 37°C with 5% CO2. Cell pellets were obtained by centrifuging cultures at 12 500 g at 4°C for 30 min. Pellets were washed three times with PBS pH 7·4 and finally resuspended in PBS at 1/200 of the original volume.
Protein concentration was estimated using the BCA protein assay (Thermo Scientific, East Grinstead, West Sussex, UK) according to the manufacturer's instructions. Ten micrograms of protein from each sample was mixed with an equal volume of 2× Laemmli buffer (Laemmli 1970), boiled for 10 min and then cooled in ice. Samples were loaded and electrophoresed in a Protean II Vertical Electrophoresis Cell (Bio-Rad Laboratories) in 1·5 mm discontinuous gels with 4% stacking gel and 10% resolving gel at a constant voltage of 200 V. Protein bands were visualized by staining with Coomassie PhastGel Blue R-350 (GE Healthcare Life Sciences, Buckinghamshire, England, UK).
Immunoblot was performed by transferring electrophoresed proteins onto 0·45-μm nitrocellulose membranes (Bio-Rad Laboratories) in a semi-dry device (Wolf Laboratories, York, UK) at 15 V for 30 min. Membranes were stained with 2% Ponceau S Red in 1% acetic acid for 1 min to confirm the transfer of proteins. After washing with PBS to eliminate red staining, membranes were blocked using blocking buffer (1 mol l−1 Glycine, 1% egg albumin, 5% powdered skimmed milk, 0·1 mol l−1 PBS) overnight at 4°C. Afterwards, membranes were washed twice with PBS plus 0·1% Tween 20 and a third time only with PBS. Then, primary antibody reaction with a reference rabbit antiserum against Myc. agalactiae PG2 (University of Aarhus, Denmark; Freundt 1983) was performed at 37°C for 1 h with continuous agitation at a concentration 1/200. After three washes with PBS plus 0·1% Tween 20 and a last one with PBS, membranes were incubated as before in a 1/1000 dilution of peroxidase-conjugated Protein G (Thermo Scientific) and washed again as above. Antigenic bands were revealed by soaking the gel in the substrate 4-chloro-1-naphthol at room temperature. The reaction was stopped with deionized water when bands had developed sufficient intensity.
After several culture and isolation steps, 13 strains were obtained, thus isolation percentage from bulk-tank sheep milk constituted a 28·8% of total sampling. All the isolates were confirmed as pure cultures by specific PCR-DGGE. Strains, their source and location where samples were taken, are listed in Table 1. Additional information about isolates (complete characteristics and positive and negative sampling locations) is available in Table S1 and Figure S1.
Once the plates were incubated and centrifuged, clear deposits of cells were observed in wells where the antimicrobial was not effective. The results are summarized in Table 2. Among the 13 antimicrobials used, clindamycin was the most effective agent inhibiting the growth of 100% of strains with MIC <0·12 μg ml−1. The quinolones, danofloxacin and enrofloxacin, inhibited most isolates at a MIC ≤0·25 μg ml−1 followed by marbofloxacin which presented a MIC between 0·5 and 2 μg ml−1. Remarkably, all strains were resistant to erythromycin showing MIC >32 μg ml−1 in all cases. Intermediate MIC values between 1 and 2 μg ml−1 were obtained for the other macrolides, tylosin and tilmicosin, and for lincomycin, for most strains. Nevertheless, two strains gave MIC values of 8 μg ml−1 when tested against tilmicosin. Tulathromycin was found to have a less uniform response than the other macrolides (MIC between 1 and 8 μg ml−1). Chloramphenicol, florfenicol and spectinomycin showed quite high values (MIC90 ≤8 μg ml−1).
|ERY||8 to >32||>32||>32|
SmaI restriction profiles of 13 Myc. agalactiae isolates showed no differences in their band pattern. A single profile of six fragments of estimated sizes consistent to those reported by McAuliffe et al. (2008) for Spanish, Greek, Italian, Macedonian and Portuguese strains was observed at approximately 466, 174, 132, 91 and 11 kbp. The variable fragment described elsewhere (Tola et al. 1996, 1999; McAuliffe et al. 2008) was estimated to be 58 kbp for the isolates characterized in this work.
Protein electrophoretic profiles obtained by SDS-PAGE analysis were homogeneous within tank isolates and compared with type strain PG2 even though some slight differences of intensity were observed in low molecular weight proteins (<20 kDa) and at approximately 50 kDa for strain SP-S4 from a tank in Leon. Some minor differences were also observed in SP-2501 from a Silo in Zamora at approximately 66 and 36 kDa. Immunoblot test (Fig. 1) revealed a similar homogeneity for high molecular weight antigens. Some differences in staining intensity were observed at approximately 36–37 kDa, but in contrast to SDS profiles, an evident heterogeneity of presence or absence of bands and/or intensity was observed in antigenic molecular masses below 30 kDa. These lower molecular weight bands separated the studied strains into four different patterns: Profile 1 (strains SP-S1, SP-S2, SP-I, SP-1101, SP-0102, SP-13), profile 2 (strains SP-S4, SP-27, SP-E), profile 3 (strains SP-H, SP-2501) and profile 4 (strains SP-1801, SP-23).
Mycoplasma agalactiae isolation was obtained in a 28·8% of milk samples of which 70% (7/10) of samples from silos and 17% (6/35) from tanks were positive.
These rates are higher than previous figures obtained for sheep and goat milk in Spain (Gonzalo et al. 2002; Contreras et al. 2008) and can be related to the high number of outbreaks recorded in Castilla y León revealing this region is an endemic area.
Regarding the antimicrobial activity, these results demonstrate the effectiveness of quinolones in vitro against Myc. agalactiae, which are included as standard treatments against CA, and show agreement with Hannan (2000); Loria et al. (2003) and Antunes et al. (2008) on the susceptibility of field isolates of Myc. agalactiae. Clindamycin was the most effective antimicrobial in all cases, but there is no previously published data to compare with this study. Erythromycin has been considered as an option for the treatment for CA (Bergonier et al. 1997; Antunes et al. 2008). Although other mycoplasmas have been reported to be susceptible, all strains of Myc. agalactiae in the present study were resistant, which is in agreement with the work of Antunes et al. (2008). In contrast to the results of previous studies, isolates in this study presented much higher oxytetracycline MIC values and slightly higher MIC values in the case of tylosin, although Loria et al. (2003) reported one isolate with an MIC of 2 μg ml−1. Both of these antimicrobials are treatments of choice against the disease. Results obtained on Sicilian strains by Loria et al. (2003) were similar for lincomycin and enrofloxacin. According to suggested published MIC breakpoint values for Myc. agalactiae (Hannan 2000), most of the strains are classified as sensitive or intermediate resistant for oxytetracycline and tylosin where the intermediate values are suggested at 8 μg ml−1 and ≤2 μg ml−1, respectively. Spectinomycin MIC values were similar to those of Antunes et al. (2008). A relatively variable response was seen to tulathromycin, which belongs to a group of new-generation macrolides, and depended on the strain (MIC range between 1 and 8 μg ml−1) giving higher MIC50 and MIC90 values compared with other antimicrobials of the same family.
It is very important to know which antimicrobials of veterinary use are effective against Myc. agalactiae isolated from sheep farms, especially in affected areas of Spain and neighbouring Mediterranean countries where CA is endemic. It is well known that in vitro sensitivity of antimicrobials does not always correspond to the effectiveness of treatment in the field, but it is highly probable that antimicrobials with high MIC values are likely to be ineffective in treating the affected animal.
Molecular typing of the isolates revealed high levels of similarity among the strains. These results are in agreement with the data published in a global VNTR and MLST typing study (McAuliffe et al. 2011) where isolates characterized in this work were also included. Strains isolated from tank milk presented a similar pattern belonging to a unique European majority VNTR profile (VNTR profile 1121). Limited variation was observed when using MLST, dividing isolates into three sequence types (ST-5, 10 strains; ST-16, 2 strains and ST-17, 1 strain) all belonging to the same clonal complex.
These results were consistent with those reported for other isolates from Spain and other European countries by Tola et al. (1996) and de la Fe et al. (2006) and were in agreement with the opinion of de la Fe et al. (2006) about the low relatedness of the reference strain PG2 compared with the isolates from Spain. In addition, a further knowledge on conserved and variable strong immunogenic proteins could be of help in the design of effective vaccines against CA.
In conclusion, the isolates of Myc. agalactiae characterized in this work came from an endemic area and shared a common molecular profile, as revealed by PFGE and SDS-PAGE, with limited antigenic variability as visualized by immunoblot analysis. Clindamycin and quinolones showed great effectiveness against the isolates and may be considered for standard treatment for CA.
This study was supported by project AGL2008-00422 of the Spanish Ministry of Science and Innovation, Government of Spain (Madrid) and by the Department for Environment, Food and Rural Affairs (DEFRA), United Kingdom. Special thanks to the Consortium of Ovine Promotion (Villalpando, Zamora, Spain) and to M. Rosa Salgado (Sersa Veterinarios) for their collaboration in the sampling process and Rafael de Garnica for his assistance in map preparation.