Some coagulase negative Staphylococcus spp. isolated from buffalo can be misidentified as Staphylococcus aureus by phenotypic and Sa442 PCR methods

Staphylococcus aureus is a commonly reported cause of buffalo mastitis. However, its prevalence may be overestimated. The aim of this study was to compare S. aureus identification by conventional phenotypic and genotypic assays versus Matrix-Assisted Laser Desorption Ionization-Time of Flight Mass Spectrometry (MALDI-TOF MS) and novel real-time quantitative PCR tests for the cytochrome oxidase subunit D II (cydB) and staphylocoagulase (coa) genes. From 408 samples obtained from buffalo milk/milking environment, 32 putative S. aureus strains were identified based on characteristic growth on Baird Parker agar, positive catalase reaction, ability to clot rabbit plasma, and positive Sa442 PCR assay. However, in further testing, only 10 of these strains were positive in latex agglutination tests and by MALDI-TOF MS, only eight of the 32 strains were S. aureus while the rest were S. chromogenes (19), S. agnetis (3), S. cohnii (1), or S. xylosus (1). All eight strains identified as S. aureus by MALDI-TOF analysis and confirmed by 16S RNA gene sequencing were positive in a S. aureus-specific cydB PCR test. As well, 7/8 S. aureus strains were PCR positive in a real-time coa PCR test as were 2/69 S. chromogenes and the lone S. xylosus strain tested.


Introduction
Buffalo milk and its derivatives have become increasingly important worldwide [1] and Staphylococcus aureus is one of the most significant pathogens responsible for contagious mastitis in dairy buffaloes [2]. Antibiotic treatment of S. aureus mastitis is often unsuccessful and treatment failures can lead to spread of the infection. As a result, animals with chronic S. aureus infection are often culled [3].
The initial identification of S. aureus is based on culture and phenotype on specific media; other assays commonly used to identify S. aureus are the Sa442 PCR, nuc gene PCR, and latex agglutination tests. The Sa442 PCR test, developed by Martineau et al. [4] targets a chromosomal DNA fragment thought to be specific for S. aureus, the nuc gene encodes a species-specific thermonuclease, while commercially available latex agglutination kits such as the Staphaurex latex test are based on the interaction of S. aureus surface-anchored proteins with human IgG and fibrinogen bound to latex particles [5]. However, these tests may not be accurate and can lead to erroneous identification, and in turn, to unnecessary culling [5,6]. Here we describe how common testing approaches can lead to misidentification of "non-S. aureus" strains as S. aureus and the development of new a cydB realtime PCR assay that can be used for accurate S. aureus identification. de Almeida et al. BMC Res Notes (2018) 11:346 Main text

Sample collection
Milk samples (n = 320) were collected from 80 randomly selected female buffaloes from a private dairy farm located in Sao Paulo State, Brazil from November 2013 to April 2014. After physical examination of the mammary glands [7], teats were cleaned with 70% alcohol and milk from each quarter was evaluated by strip cup and California mastitis tests [8]. Hand samples from 16 consenting milkers and 64 samples from liners were collected using sterile swabs (Pro-Lab Diagnostics) and stored in peptone water as described previously [9].

CoPS isolation and identification
Isolation and identification of S. aureus was done according to compendium of methods for the microbiological examination of foods [10]. Strains with positive egg yolk reactions [11] were tested for Gram and catalase reactions, haemolytic activity, and ability to clot rabbit plasma using a Coagu-Plasma kit (Laborclin, Pinhais, Brazil) according to manufacturer's instructions with S. aureus ATCC 25293 and S. epidermidis ATCC 12228 as the positive and negative controls respectively. In addition, isolates were tested with the Staphyclin latex test (Laborclin, Pinhais, Brazil) according to the manufacturer's instructions.

DNA extraction
For DNA extraction, well-isolated colonies were inoculated into BHI broth, incubated at 37 °C for 18 h and extracted according to the method of Kuramae-Izioka [12] with minor modifications as described previously [13].

Sa442 detection
Isolates were characterized using the PCR assay of Martineau et al. [4] with minor modifications. PCR primers (

Matrix-Assisted Laser Desorption Ionization-Time of Flight Mass Spectrometry (MALDI-TOF MS)
Identification of putative Staphylococcus aureus strains was done using a MALDI Bruker Biotyper system (Bruker Daltonics Inc., Billerica, MA, USA) at the Animal Health Laboratory, University of Guelph, Guelph, Ontario, Canada, as described previously [13].

16S rRNA gene sequencing
16S rRNA gene sequencing (~ 1000 bp) was also done at the Animal Health Laboratory and sequences were compared with the 16S rRNA gene of Staphylococcus aureus MCRF184 (CP014791.1) and other Staphylococcus spp. sequences using blastn.

Design of cydB species-specific primers
PCR primers to the S. aureus cydB gene (cytochrome D ubiquinol oxidase subunit II; NCBI accession number NC_007795.1) were designed as described previously [13] and produced a 432 bp amplicon. The amplicons were confirmed by DNA sequencing as described above (Table 1). As well, an additional 84 putative staphylococci from buffalo milk/milking environment were evaluated using the cydB test [13].

Design of coagulase gene primers
Published coa primers [14] generate products of different sizes so a new primer pair (coaF and coaR; Table 1) was designed using PrimerQuest software (Integrated DNA  Fig. S1).

Real-time PCR
Real-time PCR primers (Table 1)

Staphylococcus spp. isolation and preliminary identification
Thirty-two putative S. aureus strains were selected based on their characteristic phenotype on Baird Parker agar. These Gram positive strains were catalase positive and were positive in the S. aureus species specific Sa442 PCR assay of Martineau et al. [4]. In further testing, 24 samples were consistently positive in the coagulase test; while eight gave at least one discordant result. Also, 21 of the 32 putative S. aureus strains were β-haemolytic, two were α-haemolytic and nine were non-haemolytic.

Latex test, identification by MALDI-TOF and 16S rRNA sequencing
Only ten of the 32 strains gave a positive latex agglutination result and MALDI-TOF MS analysis revealed that only eight of the 32 were S. aureus with the remainder being S. chromogenes (n = 19), S. agnetis (n = 3), S. xylosus (n = 1), or S. cohnii (n = 1) ( Table 2). All eight strains identified as S. aureus by MALDI-TOF analysis had 100% identity with the 16S rRNA gene of Staphylococcus aureus MCRF184 (NZ_CP014791.1). The two latex false positive strains were S. agnetis and chromogenes by 16S sequencing and MALDI_TOF.

cydB gene analysis and detection
Alignment of gene sequences (https ://www.ncbi.nlm.nih. gov/genba nk/) suggested that the cydB gene is well conserved in Staphylococcus and thus allowed for the design of species-specific primers (Additional file 2: Fig. S2). The eight S. aureus isolates were positive for the S. aureus-specific cydB primers and resultant amplicons had 99-100% identity with S. aureus NCTC 8325. Further, the S. aureus specific cydB primers did not amplify any of the other CoPS tested in this study nor the 84 strains of other putative staphylococci evaluated by Pizauro et al. [13].

Coagulase test and coa gene detection
Twenty-four strains clotted rabbit plasma ( Table 2). Eight strains gave discordant results with at least one negative and one positive. Of these strains, seven of the S. aureus (n = 8) and two of the S. chromogenes (n = 19) and one S. xylosus strain were positive for the coa gene while none of the S. agnetis (n = 3) nor the S. cohnii were positive for the coa gene. The sequence of the PCR products had 98% identity with the S. aureus coa gene from strain JCSC 7633 (accession number AB488507.1) and 99% identity with the coa gene in S. aureus strain MW2 genome (accession number BA000033.2).

Discussion
A number of typically coagulase negative Staphylococcus spp. including more than a quarter of S. chromogenes isolates (19/69) and at least some S. xylosus (1/1), S. cohnii (1/1), and S. agnetis (3/17) were coagulase positive. This finding is consistent with studies of Santos et al. [15] in which 23/42 CoNS strains clotted rabbit plasma. These authors suggested that this phenotype was related to specific PFGE-types, but not with the clumping factor test. The presence of coagulase is an indicator of pathogenicity since it enables bacteria to resist phagocytosis and cause chronic infections [16]. Host specific adaptations can be acquired though mobile genetic elements (MGEs) [16,17] from nearby S. aureus [16] or other CoPS such as S. pseudointermedis [15]. Thus, a coa gene in CoNS with the newly described coa primers may be the result of such transfer. Coagulase activity in the current study may also be related to another gene such as the one described in S. chromogenes that shares 41% identity with the predicted coagulase gene of the S. pseudintermedius [15]. In this study, non-S. aureus stains able to clot plasma (19/19 S. chromogenes, 3/3 S. agnetis, 1/1 S. cohnii and 1/1 S. xylosus) were also Sa442 positive. This is the first report of false positive reactions with these species; however, further stains/herds should be tested to know whether these findings can be generalized. When the sequences of the Sa442 primers [4] were compared with the available genome of S. chromogenes MU 970 strain (NZ_JMJF00000000.1), no significant homology was detected; however, this draft whole shotgun sequence could be missing the region containing Sa442 sequences. Apart from having been established as unique for S. aureus, the Sa442 fragment has not been further characterized [18]. As well, Klaassen et al. [18] and Heilmann et al. [19] have reported false negative results with the Sa442 test and the nuc PCR is subject to strain variation [6].
Latex agglutination tests may also be problematic. In previous studies, false positive results have been observed at relatively low frequencies (e.g., 7.9% [20] and 9.3% [21]). The greater false positive reaction in this study (20%) may have been related to the population structure and/or the relatively small sample size.
Given the impact that S. aureus can have on both human and animal health, its diagnosis is important [22,23]. Misidentifying more benign Staphylococcus species as S. aureus, though arguably less serious, is not without significant economic consequences. In the current study, the eight S. aureus strains (as identified by MALDI-TOF/16S rRNA sequencing) were positive with our novel S. aureusspecific cydB PCR test while no amplification was observed with the Sa442 and plasma clotting-positive S. chromogenes (n = 19), S. agnetis (n = 3), S. xylosus (n = 1) or S. cohnii (n = 1) strains tested. Also, it might be noted that S. caprae, S. hyicus, S. hominis, S. epidermidis, S. haemolyticus, S. warneri, S. equorum, S. sciuri, and S. pasteuri were negative in the S. aureus-specific cydB real-time PCR test in a complementary study [13] (Table 2).

Conclusions
In summary, a significant number of CoNS strains could clot rabbit plasma and were positive for the Sa442 PCR test and so, could be misclassified as S. aureus. The bases of these phenotypes remain to be determined, but they could be the result of horizontal gene transfer or to the fact that these species are less homogenous than previously thought. On the other hand, MALDI-TOF and a species-specific real-time PCR test for the cydB gene may permit accurate identification of CoNS.

Limitations
This study used samples from one buffalo herd which limits the generalisation of the results. In addition, it was beyond the scope of this study to determine the basis of the abnormal coagulase positive phenotype or to determine if there had been horizontal gene transfer to coagulase negative Staphylococcus strains.

Additional files
Additional file 1: Fig. S1. S. aureus coagulase gene (coa) alignment using CLC Sequence View 7 software. Additional file 2: Fig. S2. Alignment using CLC Sequence View 7 software of the cydB gene of 18 closely related Staphylococcus spp. in this study.