Multiplex polymerase chain reaction to identify and determine the toxigenicity of Corynebacterium spp with zoonotic potential and an overview of human and animal infections

Corynebacterium diphtheriae , Corynebacterium ulcerans and Corynebacterium pseudotuberculosis constitute a group of potentially toxigenic microorganisms that are related to different infectious processes in animal and human hosts. Currently, there is a lack of information on the prevalence of disease caused by these pathogens, which is partially due to a reduction in the frequency of routine laboratory testing. In this study, a multiplex polymerase chain reaction (mPCR) assay that can simultaneously identify and determine the toxigenicity of these corynebacterial species with zoonotic potential was developed. This assay uses five primer pairs targeting the following genes: rpoB ( Corynebacterium spp), 16S rRNA ( C. ulcerans and C. pseudotuberculosis ), pld ( C. pseudotuberculosis ), dtxR ( C. diphtheriae ) and tox [diphtheria toxin (DT)]. In addition to describing this assay, we review the literature regard-ing the diseases caused by these pathogens. Of the 213 coryneform strains tested, the mPCR results for all toxigenic and non-toxigenic strains of C . diphtheriae , C. ulcerans and C. pseudotuberculosis were in 100% agreement with the results of standard biochemical tests and PCR-DT. As an alternative to conventional methods, due to its advantages of specificity and speed, the mPCR assay used in this study may successfully be applied for the diagnosis of human and/or animal diseases caused by potentially toxigenic corynebacterial species.

an animal shelter for abandoned dogs found an asymptomatic dog colonised by a non-toxigenic C. ulcerans strain (Dias et al. 2010).
C. pseudotuberculosis is the aetiological agent of caseous lymphadenitis (CLA) in small ruminants, such as sheep and goats. This infection sometimes presents as pneumonia, hepatitis, pericarditis, mastitis, arthritis or subcutaneous abscesses. This pathogen is also associated with lymphadenitis in horses and with ulcerative lymphangitis in cattle (Foley et al. 2004, Perkins et al. 2004, Sharpe et al. 2010. Two biotypes of C. pseudotuberculosis, which are differentiated based on nitrate-reducing ability, have been reported: nitrate-negative strains are referred to as serotype I (biotype ovis) and nitrate-positive strains are classified as serotype II (biotype equi). Isolates from sheep and goats are usually nitrate negative, whereas strains from horses are typically nitrate positive; isolates from cattle are variable (Tejedor-Junco et al. 2008, Wagner et al. 2011). Although C. pseudotuberculosis is distributed worldwide, it has the most serious economic impact in Oceania, Africa and South America, including Argentina and Brazil (Estevao et al. 2007, Komala et al. 2008, Stefanska et al. 2008, Tarello & Theneyan 2008, Seyffert et al. 2010. Once established, CLA is difficult to eradicate because drug therapy is not effective and the clinical detection of infected animals is of limited efficiency (Dorella et al. 2006). Early microbiological diagnosis and long-term antimicrobial treatment are important for a successful outcome in horses with C. pseudotuberculosis infection (Pratt et al. 2005). Similar to diphtheria in humans (Kombarova et al. 2001, Saikya et al. 2010, C. pseudotuberculosis can cause clonally expanding epidemics in animals. The increase in the number of infections could be the result of a reporting bias, environmental factors that facilitate infection or host factors, such as greater herd susceptibility (Foley et al. 2004).
C. pseudotuberculosis can be spread among animals by fly vectors, such as Musca domestica and Hippobosca equina, as previously observed in the United States of America (USA), Israel and Egypt. This pathogen can be found in up to 20% of flies in the vicinity of diseased animals (Yeruham et al. 1996, Braverman et al. 1999, Selim 2001. Oedematous skin disease (OSD) is an endemic disease of buffaloes in Egypt. C. pseudotuberculosis serotype II (biotype equi; nitrate positive) is the primary cause of OSD. The pathogenesis of these strains is related to the secretion of toxigenic factor(s) and phospholipase D (PLD) and the lipid contents of their cell walls (Selim 2001). The appearance of OSD outbreaks during the summer months was found to be related to the fact that the summer months are the breeding season for H. equina, the primary vector of the causative agent. The control of OSD presents several problems because there is insufficient knowledge about the epidemiology and pathogenesis of this disease (Syame et al. 2008). The role of DT in CLA, OSD and other C. pseudotuberculosis diseases in animals has not been adequately studied.
Few studies have demonstrated the isolation of the causal agent of CLA and OSD from humans worldwide. Although rare, human infections caused by C. pseudotuberculosis are frequently similar to those observed in sheep and goats (CLA); these infections usually require the excision of infected lymph nodes accompanied by supplementary antimicrobial treatment, but do not involve toxaemic manifestations. C. pseudotuberculosis is usually acquired after close contact with an infected animal and no underlying diseases or predisposing conditions have been identified in infected patients (Peel et al. 1997, Romero-Perez et al. 2004, Join-Lambert et al. 2006, Hemond et al. 2009). Only one patient with toxaemic symptoms, an injecting drug user with endocarditis, has been reported. This patient had no history of animal contact and no possible source of the C. pseudotuberculosis infection was identified (Wagner et al. 2011). Additional studies are necessary to investigate the correlation between the prevalence of C. ulcerans and C. pseudotuberculosis infections in humans and the prevalence in local cattle populations, most notably in developing countries, as previously performed for bovine tuberculosis infection in animal and human populations in Ethiopia (Shitaye et al. 2007).
There has been an increase in the incidence of disease caused by non-toxigenic C. diphtheriae. Non-toxigenic strains generally cause persistent sore throats and severe pharyngitis/tonsillitis and invasive diseases, such as endocarditis, septic arthritis, splenic abscesses and osteomyelitis, are not uncommon (Galazka 2000, Dzupova et al. 2005, Farfour et al. 2012. In non-industrialised countries there is an overall lack of information on the prevalence of colonisation by and diseases caused by non-toxigenic C. diphtheriae in the population due to a reduction in the frequency of routine screening for this organism. Because of non-toxigenic C. diphtheriae infection may not be recognised by healthcare professionals and non-toxigenic C. diphtheriae is considered an emerging pathogen of increasing significance worldwide there is an urgent need for increased clinical awareness, especially for immunocompromised patients, in whom complications can arise (Gomes et al. 2009, Edwards et al. 2011, Mattos-Guaraldi et al. 2011. The data indicate that there is still a lack of information on the prevalence of disease caused by potentially toxigenic corynebacteria. Therefore, to better understand the mechanisms of carriage/transmission, determine the incidence rates and analyse the variation among non-toxigenic and toxigenic strains, it is imperative to use assays that can rapidly and accurately detect these three pathogens (Connor et al. 2007, Estevao et al. 2007, Kraeva et al. 2007, Baird & Malone 2010, Konrad et al. 2010, Sharpe et al. 2010). In the present study, a multiplex polymerase chain reaction (mPCR) assay was developed to identify C. diphtheriae, C. ulcerans and C. pseudotuberculosis and differentiate between toxigenic and non-toxigenic strains. This assay will serve as a complementary method for the timely laboratory diagnosis of infected animals and/or humans and for epidemiological studies.
The identification of corynebacteria was based on the results of both standard biochemical tests (Efstratiou & George 1999, Funke & Bernard 2007) and a commercial kit: the API Coryne System (bioMérieux, La-Balme-les-Grottes, France) (Soto et al. 1994). A DNase test was also performed to differentiate C. diphtheriae and C. ulcerans from C. pseudotuberculosis and other Corynebacterium spp (Pimenta et al. 2008b). A reverse CAMP reaction was performed using Staphylococcus aureus ATCC 25923. A nitrate test was used to differentiate the C. pseudotuberculosis biovars equi and ovis. The conventional phenotypic tests used to identify the potentially toxigenic species are listed in Table I.
Stock cultures were maintained as suspensions in 10% skim milk containing 25% glycerol at -20ºC. The microorganisms were cultured on Columbia agar base (CAB) (BBL, Sparks, USA) containing 5% sheep blood for 24-48 h at 37ºC for all the phenotypic and genotypic procedures described in this study. For DNA extraction, microorganisms were grown on CAB.
Toxigenicity tests -All C. diphtheriae, C. ulcerans and C. pseudotuberculosis strains were evaluated for the presence of the tox gene by PCR using the DT primer set (PCR-DT), which targets fragment A of diphtheria toxin. The primer pair (forward, ATCCACTTTTAGTGCGA-GAACCTTCGTCA and reverse, GAAAACTTTTCT-TCGTACCACGGGACTAA) was able to amplify a 248 bp fragment from both the control and clinical strains (Pallen et al. 1994, Efstratiou et al. 1998).
mPCR -The conditions for amplification using this mPCR assay were based on reactions described previously (Pacheco et al. 2007, Pimenta et al. 2008a. The nucleotide primers used for mPCR are listed in Table I. In this study, primers targeting the 16S rRNA sequences (816 bp) of both C. pseudotuberculosis and C. ulcerans strains were used along with a reverse primer for the amplification of a pld sequence specific for C. pseudotuberculosis (203 bp) to discriminate C. pseudotuberculosis from C. ulcerans. In addition, a primer pair, Dipht 4 (Nakao et al. 1996), that amplifies a region linking the A and B fragments of the DT gene was used to generate an  amplicon of 303 bp, which is between the size of the amplicons for rpoB (446 bp) and dtxR (258 bp) of C. diphtheriae strains. The amplification of the rpoB gene was used both as an internal control and to confirm that the isolate was a Corynebacterium sp. when no other amplicons were observed after the amplification reaction. Microbial DNA was extracted by boiling a suspension containing a loopful of freshly grown bacteria on CAB in 500 µL of sterile deionised MilliQ water for 10 min. The suspension was centrifuged at 13,000 rpm and 1 µL of the supernatant was used in the final multiplex reaction described in Table II, with primers diluted to 2 mM. The amplification was performed in a MyCycler thermal cycler (BioRad Laboratories).
The amplified products were observed after electrophoresis on 1% agarose gels and SYBR green staining.

Identification of C. diphtheriae, C. ulcerans and C. pseudotuberculosis isolates -The microorganisms
were identified by both biochemical assays and mPCR. All sucrose-fermenting and non-sucrose-fermenting C. diphtheriae and C. ulcerans strains exhibited DNase activity, whereas 93 of the C. pseudotuberculosis strains belonging to biotypes ovis and equi were negative for the expression of DNAse. The urease production assay showed positive results for C. ulcerans and C. pseudotuberculosis; both species also inhibited the beta-haemolysis of S. aureus ATCC 25923, indicating a positive reverse CAMP reaction.
The API profiles identified all microorganisms included in this study to the species level, with percentages greater than 98%. Only Corynebacterium strains that were identified with API at percentages higher than 98%, except for five strains belonging to the C. amycolatum complex, were used in this investigation. Using these confirmed strains and conventional biochemical tests was important for the accuracy of the mPCR assay.
The mPCR was identified all C. pseudotuberculosis strains independently of the biotype (equi or ovis), yielding at least three amplicons: 816 bp, corresponding to 16S rRNA, 446 bp, corresponding to rpoB, and 203 bp, corresponding to pld. C. ulcerans had only two amplicons, which corresponded to 16S rRNA and rpoB. The C. ulcerans strains did not produce amplicons for the pld gene, although PLD was expressed by all strains, confirming the specificity of the primer (pld R) used in this study for C. pseudotuberculosis. The mPCR also identified all sucrose-fermenting and non-sucrose-fermenting C. diphtheriae strains independently of the biotype (gravis, mitis, intermedius and belfanti) or site of colonisation (respiratory tract, skin, bone and blood), yielding amplicons for at least two genes, rpoB (446 bp) and dtxR (258 bp).
In this investigation, all isolates of C. diphtheriae, C. pseudotuberculosis and C. ulcerans yielded the amplicons targeted in the mPCR assay, giving the amplification pattern presented in Figure. No isolates identified as the species above yielded unequivocally different amplification patterns using the mPCR conditions described in this study. The additional amplicon observed in DT gene- bearing strains (dipht 4-303 bp) was correlated with amplification by the DT primer set (PCR-DT), which targets fragment A of DT (248 bp). The primers targeting fragment A of DT gave a product with a molecular mass similar to the dtxR amplicon (258 bp) and therefore, these primers not included in this mPCR assay. All tested Corynebacterium spp that were not associated with C. diphtheriae, C. pseudotuberculosis or C. ulcerans, except for Corynebacterium pseudodiphtheriticum strains, only yielded amplicons for rpoB. The rpoB sequence of C. pseudodiphtheriticum could only be amplified using the cycling conditions described in this study after replacing C3130 R (rpoB) TCCATTTCGC-CGAAGCGCTG with the reverse primer 5'-TCCATCT-CACCGAAGCGCTC-3'.
The Gram-positive bacteria Nocardia sp, R. equi and Oerskovia sp. included in this investigation did not yield amplicons in this mPCR assay using the cycling conditions described.

DISCUSSION
The prevalence of C. diphtheriae, C. ulcerans and C. pseudotuberculosis infections is greater than previously recognised by the medical and veterinary communities in many regions worldwide due to the difficulty in accurately identifying these pathogens and their ability to produce DT.
The laboratory identification of Corynebacterium organisms is usually complex and expensive, requiring the use of conventional and/or miniaturised biochemical methods (Funke & Bernard 2007). The molecular typing methods used for the characterisation of C. diphtheriae, C. ulcerans and C. pseudotuberculosis are important tools for the rapid detection and identification of bacterial clones that harbour the tox gene and these methods have several advantages over the traditional methods used to determine the toxigenicity of corynebacterial isolates (Hauser et al. 1993, Pallen et al. 1994, Nakao & Popovic 1997, Cetinkaya et al. 2002, Mothershed et al. 2002, Sing et al. 2003, Pacheco et al. 2007, Pimenta et al. 2008a. In this study, a species-specific mPCR assay for the identification of potentially toxigenic human and animal isolates of C. diphtheriae, C. ulcerans and C. pseudotuberculosis was developed. The taxonomic position of most of the isolates could also be determined using this mPCR assay, which includes an rpoB-targeting genusspecific primer pair for the identification of Corynebacterium sp. (Khamis et al. 2004). All Corynebacterium strains used in this study, except for C. pseudodiphtheriticum isolates, yielded amplicons with the rpoB primers, as previously observed (Pacheco et al. 2007).
The mPCR method using primers for the detection of the dtxR gene was also able to identify C. diphtheriae isolates independent of DT production, including isolates of the biotypes gravis, mitis, intermedius and belfanti, as previously demonstrated (Pimenta et al. 2008a).
C. ulcerans and C. pseudotuberculosis yielded amplicons with the 16S rRNA primers, whereas only C. pseudotuberculosis, regardless of the biovar (ovis or equi), yielded amplicons for pld, as previously described (Pacheco et al. 2007). No diverse amplicons were observed for C. diphtheriae after amplification with primers targeting 16S rRNA or pld.
To detect toxin production by the three species included in this study, primers targeting Dipht 4 (Nakao et al. 1996), a sequence between the A and B fragments of DT, were chosen due to the adaptation of the primer sequences to the cycling conditions used for mPCR and the molecular mass of the generated amplicon, which allowed for an easier interpretation of the mPCR results after electrophoresis with agarose gels.
The mPCR assay yielded reliable results when performed by at least two different technicians who used the same strains at different times. For clinical purposes, C. pseudotuberculosis, C. ulcerans and toxigenic and non-toxigenic C. diphtheriae control strains should be included in the assay. The inclusion of these strains was shown to be essential to monitor the amplification profiles of the isolates. No non-specific amplicons were observed for other Corynebacterium spp with mPCR in this study. Amplicons were not observed for the coryneform strains tested.
The advantages of this mPCR assay over conventional biochemical procedures are its rapidity, ease of performance, the large number of strains that can be simultaneously tested and the easy interpretation of the mPCR results. This novel species-specific mPCR system may facilitate routine laboratory diagnosis and/or epidemiological research on this group of potentially zoonotic and toxigenic corynebacterial pathogens: C. pseudotuberculosis, C. ulcerans and C. diphtheriae.