|Home | About | Journals | Submit | Contact Us | Français|
Staphylococcus lugdunensis is an aggressive, virulent member of the coagulase-negative staphylococci (CoNS) that is responsible for severe, rapidly progressive skin and soft tissue infections and native valve endocarditis. To facilitate prompt identification and appropriate therapy, we describe here a rapid and robust multiplex real-time PCR assay that is able to definitively distinguish S. lugdunensis from other staphylococci. Using melting curve analysis, the assay also identifies Staphylococcus aureus and CoNS other than S. lugdunensis and determines MecA-dependent resistance to methicillin (meticillin). When applied to a panel of well-characterized staphylococcal reference strains, as well as 165 clinical isolates previously identified by conventional methods, the assay was both sensitive and specific for S. lugdunensis, correctly identifying the reference strain and all 47 S. lugdunensis isolates without inappropriate amplification of other staphylococci. Furthermore, rapid biochemical identification using the WEE-TAB system to detect ornithine decarboxylase activity was found to be unsuitable as an alternative to PCR identification, displaying just 31% sensitivity and 77% specificity when tested on a subset (90 isolates) of the clinical strains. We therefore propose that this simple, accurate PCR approach will allow for the routine and timely identification of S. lugdunensis in the clinical microbiology laboratory.
Staphylococcus lugdunensis is a member of the coagulase-negative staphylococci (CoNS), a group of organisms classified on the basis of their inability to clot plasma due to the lack of the secreted enzyme coagulase. Though the CoNS are normal skin flora, they are a significant cause of bloodstream infections and endocarditis (19). Whereas most CoNS infections are subacute with rare mortality, S. lugdunensis causes severe, aggressive disease similar to that caused by the well-characterized pathogen Staphylococcus aureus (6). For this reason, the clinical importance of S. lugdunensis is increasingly being recognized (2).
Currently, screening for S. lugdunensis in the clinical microbiology laboratory is done by traditional biochemical methods (9, 16). S. lugdunensis can be initially confused with S. aureus, as it often expresses clumping factor/bound coagulase and is therefore slide coagulase positive. However, unlike for S. aureus, the slide coagulase test typically shows a rough or grainy appearance without complete clearing. Furthermore, the tube coagulase test for free coagulase is by definition negative. S. lugdunensis is classified among the CoNS first by positive pyrrolidonyl-arylamidase activity (PYR reaction), which places it within a group of pathogens including Staphylococcus haemolyticus, Staphylococcus schleiferi, Staphylococcus xylosus, Staphylococcus simulans, Staphylococcus intermedius, Staphylococcus pseudintermedius, and Staphylococcus caprae. Subsequently, S. lugdunensis can be distinguished from these PYR-positive CoNS by the presence of ornithine decarboxylase (ODC) activity. In cases of weakly positive ODC, testing for mannose utilization may also be useful (13).
Accurate identification of S. lugdunensis is crucial for correct interpretation of oxacillin MIC and for avoiding misclassification as methicillin (meticillin) resistant-CoNS (3). While the traditional microbiologic identification of S. lugdunensis is effective, it requires at least 18 to 24 additional hours from the time a staphylococcus-like colony is recognized from primary culture. In addition, it is not possible to directly apply biochemical tests to positive blood cultures. Because infection with S. lugdunensis can be rapidly progressive and the optimal treatment regimen is distinct from those for other CoNS, prompt and accurate identification is critical to the timely initiation of appropriate antibiotic therapy (6). To address this diagnostic deficiency, we have developed a rapid and simple, real-time PCR assay on the Cepheid SmartCycler that is able to distinguish S. lugdunensis from other staphylococci using melting curve analysis.
A list of ATCC strains used in this study is shown in Table Table1.1. The 165 clinical isolates utilized in this study were recovered from patient specimens at the Stanford Hospital Clinical Microbiology Laboratory (Stanford, CA) and included 124 wound, 19 respiratory, 9 tissue, 5 blood, 4 fluid, 3 catheter tip, and 2 urine isolates.
Bacteria were cultured on 5% sheep blood agar and incubated at 35°C for 18 to 24 h. One loopful (1-μl loop) of bacteria was resuspended in 500 μl of molecular-grade water (Sigma, St. Louis, MO) and then boiled for 10 min. Cellular debris was sedimented by centrifugation for 1 min at 11,000 × g, and PCR was performed on the supernatant. Three of 166 clinical isolates did not amplify after the initial extraction. Two isolates were reextracted and subsequently amplified. A repeat Gram stain of the third isolate revealed gram-positive rods, and further evaluation was not pursued.
For conventional testing of ODC activity, one loopful (1-μl loop) of bacteria was resuspended in 1 ml of Moeller's ornithine broth (Becton-Dickinson, Franklin Lakes, NJ), overlaid with mineral oil, and incubated at 35°C. The color of the broth (positive, purple; negative, yellow or gray) was evaluated at 4 h and after overnight incubation. For ODC testing via WEE-TAB (Key Scientific, Stamford, TX), a heavy suspension of the organism was made in 500 μl of distilled water, the ODC WEE-TAB tablet was added, and the mix was incubated at 37°C for 4 h. The development of a red/pink color indicated a positive test, while the development of a yellow color indicated a negative test. If the tablet remained white, the test was considered indeterminate.
Organisms with a colony morphology resembling that of CoNS (unpigmented, smooth, slightly raised to convex, opaque) on blood agar were initially screened for catalase-positive, gram-positive cocci in clusters and then tested for bound coagulase/clumping factor by the slide coagulase test. Whether the organism was slide coagulase positive or negative, the tube coagulase test was set up for overnight incubation and pure or predominant cultures were tested for PYR activity. If PYR positive, the organism was subcultured to fresh 5% sheep blood agar and incubated overnight for possible full identification and susceptibility testing on the MicroScan Walkaway (Siemens, Deerfield, IL). In addition, PYR-positive isolates were tested for ODC activity. PYR-positive, ODC-positive, tube coagulase-negative isolates were reported as S. lugdunensis. PYR-positive, ODC-negative, tube coagulase-negative isolates were reported based on the MicroScan identification (if the identification was >85%). If the identification was less than 85%, the API Staph strip (bioMérieux, Durham, NC) was used for confirmation. Tube coagulase-negative, PYR-negative isolates were reported as CoNS.
Organisms with a colony morphology resembling that of S. aureus (pigmented, smooth, slightly raised, hemolytic) on blood agar were initially screened for catalase-positive, gram-positive cocci in clusters and then evaluated for coagulase activity. Tube and/or slide coagulase-positive strains were reported as S. aureus. Methicillin resistance was determined based on growth on oxacillin-screening agar (Mueller-Hinton agar with 6% NaCl and 6 μg/ml oxacillin) and MicroScan evaluation of oxacillin resistance.
Each 25-μl reaction mixture on the SmartCycler (Cepheid, Mountain View, CA) contained 12.5 μl of 2× FastStart SYBR green Master (Roche Diagnostics, Indianapolis, IN), 7.5 μl of extracted DNA, and 5 μl of the appropriate primer mix. The mixture for reaction 1 contained the following primers (final concentration in parentheses): fbl forward (1,600 nM), fbl reverse (1,600 nM), Staph16S forward (100 nM), and Staph16S reverse (100 nM). The mixture for reaction 2 contained the primers ftsZ forward (800 nM), ftsZ reverse (800 nM), mecA forward (600 nM), and mecA reverse (600 nM), and SYBR green fluorescence was monitored in real time. After initial denaturation at 95°C for 5 min, the reaction mixtures underwent 40 cycles of 95°C for 15 s, 60°C for 30 s, and 72°C for 30 s. The final step involved a 60°C to 95°C temperature ramp at a rate of 0.1°C/second to generate the melting curve. The product sizes and melting temperatures are shown in Table Table22.
Each 10-μl reaction mixture on the Rotor-Gene 6000 (Corbett Life Science, Sydney, Australia) contained 5 μl of 2× FastStart SYBR green Master, 3 μl of extracted DNA, and 2 μl of the appropriate primer mix. The mixture for reaction 1 contained the following primers (final concentration in parentheses): fbl forward (600 nM), fbl reverse (600 nM), Staph16S forward (100 nM), and Staph16S reverse (100 nM). The final primer concentrations and the PCR program for reaction 2 were unchanged from those described above.
Three microliters of genomic DNA was amplified in a 50-μl reaction mixture consisting of 0.5 μM of TGGAGAGTTTGATCCTGGCTCAG and AAGGAGGTGATCCATCCGCA universal primers and 1× HotStarTaq Plus Master Mix (Qiagen Inc., Valencia, CA). The PCR was performed on a PE GeneAmp 9700 thermal cycler (Applied Biosystems, Foster City, CA). The PCR conditions included a 95°C activation step for 5 min, followed by 35 cycles of 95°C for 40 s, 60°C for 30 s, and 72°C for 2 min and a final 72°C elongation step for 10 min. Eight microliters of the PCR mix was visualized on a 1% agarose gel for the presence of a 1.5-kb amplicon. Cycle sequencing was performed in separate reaction volumes consisting of 2 pM of GTTTGATYMTGGCTCAG, TGCCAGCAGCCGCGGTAA, GGACTACCAGGGTATCTAAT, and TACCGCGGCTGCTGGCAC primers plus 2 μl of Big Dye Terminator mix, 3 μl of 5× Big Dye Terminator buffer (Perkin-Elmer Applied Biosystems, Foster City, CA), and 10 μl of diluted (1:6 in water) PCR mix. Cycling conditions included 25 cycles at 96°C for 10 s, 50°C for 5 s, and 60°C for 4 min. Sequencing products were purified with a Big Dye Xterminator purification kit and separated on an ABI 3730 Genetic Analyzer (Applied Biosystems). DNA sequences were assembled with Lasergene software (DNASTAR, Madison, WI) and compared to those in the LeBIBI database (http://umr5558-sud-str1.univ-lyon1.fr/lebibi/lebibi.cgi). A distance score of 0% to less than 1% was used as the criterion for species identification.
In order to rapidly identify S. lugdunensis and other clinically important staphylococci in the clinical microbiology laboratory, we designed a series of primers for use in a multiplex, real-time assay. The primer targets included the S. lugdunensis fbl (fibrinogen binding protein) gene (10); the S. aureus ftsZ (filamenting temperature-sensitive mutant Z) gene (12); and the mecA gene, encoding the penicillin binding protein PBP2a, which is responsible for methicillin resistance in both CoNS and S. aureus (4) (Table (Table22).
We next combined primer sets into two reactions to minimize the number of reactions needed to identify these staphylococcal species. The reaction 1 mixture contained the fbl primer set designed to detect S. lugdunensis. To this reaction mixture we also added primers targeting a region of the 16S rRNA gene that is conserved in all staphylococci. If nucleic acid from any staphylococcal species is present, these primers amplify a specific product, thus controlling for extraction efficiency and reaction setup. Critically, if the specimen is not S. lugdunensis, this genus control primer set distinguishes between a failed reaction and an alternative staphylococcal isolate. The second PCR mixture contained the mecA primer set to detect this molecular determinant of methicillin resistance and the ftsZ primer set to identify S. aureus. This reaction allowed us to distinguish methicillin-resistant S. aureus (MRSA) from methicillin-sensitive S. aureus (MSSA), as well as to rapidly identify S. lugdunensis and other CoNS that carry mecA.
Reactions 1 and 2 were then validated against the well-characterized staphylococcal reference strains S. lugdunensis ATCC 49576, S. epidermidis ATCC 12228, MSSA ATCC 29213, and MRSA ATCC 43300. Figure Figure1A1A shows a characteristic melting curve in reaction 1 for the identification of S. lugdunensis using template DNA from ATCC 49576. The peak at 72.7°C indicates that the isolate is S. lugdunensis, and the peak at 80.2°C confirms the staphylococcal genus. In contrast, when reaction 1 is performed on template DNA from S. epidermidis, only the genus-specific peak is present (Fig. (Fig.1B).1B). Similarly, only the genus-specific melting curve is present when reaction 1 is run on DNA from the MSSA or MRSA reference strain (data not shown). Figure 1C and D show characteristic melting curves in reaction 2 when performed on DNA from MSSA and MRSA, respectively. Whereas both strains show a specific peak at ~76°C indicating that they are indeed S. aureus (Fig. (Fig.1C),1C), only the methicillin-resistant strain carries mecA and has a peak at 73.3°C (Fig. (Fig.1D).1D). As expected, no peaks are present when reaction 2 is run on DNA from methicillin-sensitive S. lugdunensis and S. epidermidis (data not shown).
To further assess the specificity of these primer sets, we ran both multiplex reactions (1 and 2) using DNAs extracted from 12 additional staphylococcal reference strains (Table (Table1),1), including S. intermedius (ATCC 29663), a species that also expresses free coagulase and that may be confused with S. aureus by both phenotypic and molecular identification strategies (11). While the genus control product was amplified from all isolates, no false-positive peaks were observed and all strains were properly identified (data not shown).
To determine the clinical utility of this two-reaction approach for the identification of these clinically important staphylococci and to monitor the performance of our novel multiplex, real-time assay, we analyzed 165 clinical isolates that had been previously identified as S. lugdunensis (40 isolates), MRSA (48 isolates), MSSA (42 isolates), or CoNS (35 isolates) by traditional methods. When we compared our PCR results to the identification provided in the final laboratory report, we found 12 discrepancies. Eight of the isolates were identified as S. lugdunensis by PCR. Six of these eight were initially identified as either S. haemolyticus (five isolates) or Staphylococcus auricularis (one isolate) by biochemical analysis on the MicroScan, and ODC testing was not performed. One of these eight discrepant isolates was identified from a mixed culture and at the time was not worked up further. However, subsequent testing revealed that the seven isolates were both PYR and ODC positive, most consistent with S. lugdunensis. The eighth discrepant isolate was identified as S. lugdunensis by the automated instrument but was ODC negative and so was reported as CoNS. Sequencing of the 16S rRNA gene confirmed that this strain was S. lugdunensis. Therefore, our assay identified eight additional S. lugdunensis strains that were not recognized during routine workup (Table (Table33).
The four remaining discrepant results were resolved in a similar manner. One isolate was reported as S. lugdunensis based on positive PYR and ODC tests, as well as automated biochemical identification. However, our real-time assay indicated that the strain was a CoNS other than S. lugdunensis. Upon repeat testing the isolate was ODC negative, and 16S rRNA gene sequencing confirmed that this strain was either Staphylococcus capitis or S. caprae, two closely related CoNS species that cannot be distinguished by their 16S sequences (7, 18). Two isolates were reported as MSSA based on a positive tube coagulase test, but the isolates were identified as CoNS by PCR. Interestingly, the biochemical profiles of these isolates were most consistent with S. intermedius. The final discrepant isolate was reported as MRSA, but PCR indicated it was a mecA-positive CoNS. Repeat testing confirmed that the isolate was tube coagulase positive, and automated biochemical identification/antibiotic susceptibility results were consistent with MRSA. When the reaction 2 primer sets were run individually, rather than in multiplex, the ftsZ and mecA peaks confirmed MRSA. Taking into account the resolution of discrepant results, this multiplex assay showed 100% sensitivity and specificity for S. lugdunensis and was also able to effectively identify MRSA (sensitivity, 98%; specificity, 100%), MSSA (sensitivity, 100%; specificity, 100%), and other CoNS (sensitivity, 100%; specificity, 99%).
Given the frequency with which staphylococci are found in clinical specimens, we also adapted this assay to the Corbett Rotor-Gene (Materials and Methods). This instrument has a greater reaction capacity (up to 100 reactions per run) and allows smaller reaction volumes (as low as 5 μl) than the Cepheid SmartCycler (16 reactions/run, 25 μl). When reaction 1 was repeated on the 165 isolates using the Rotor-Gene, no discrepancies with the SmartCycler results were found (data not shown).
One significant advantage of this molecular approach to the identification of S. lugdunensis is the rapid turnaround time (approximately 2 h). However, a rapid ODC assay may also be sufficient to adequately screen for S. lugdunensis. We therefore evaluated the ODC WEE-TAB, a 4-hour ODC assay. When tested against a subset of the isolates identified above (Table (Table4),4), the ODC WEE-TAB performed poorly. Only 31% (13/42) of the S. lugdunensis isolates were ODC WEE-TAB positive, compared to 100% (42/42) of those tested using conventional ODC broth medium incubated overnight. Similarly, 23% (11/48) of the strains identified as CoNS, MRSA, or MSSA were also ODC WEE-TAB positive. Though conventional testing was not performed on all of these isolates that were not S. lugdunensis to confirm the negative ODC, these results indicate that the ODC WEE-TAB is neither sensitive nor specific. Particularly concerning was the number of ODC WEE-TAB results that were simply uninterpretable (42%, 38/90).
We describe here a simple multiplex real-time PCR assay for the identification of S. lugdunensis that provides a rapid and effective alternative to conventional identification strategies and that is superior to rapid identification using the ODC WEE-TAB. Whereas biochemical identification typically requires an additional 18 to 24 h, this assay takes only ~2 h to identify S. lugdunensis from culture. Furthermore, this assay also identifies S. aureus and CoNS other than S. lugdunensis and detects the presence of mecA, thereby providing comprehensive coverage of the clinically important species of staphylococci and their sensitivity to methicillin. In the absence of mecA, this allows the timely initiation of optimal antistaphylococcal therapy with penicillinase-resistant penicillins (8, 15) and facilitates the rapid discontinuation of vancomycin, thus reducing adverse medication effects and preventing the potential emergence of drug resistance (1).
When we applied this assay to 165 clinical isolates, we identified 8 additional S. lugdunensis strains that had not been recognized during the initial workup. For seven of these isolates, ODC activity was assayed only retrospectively to resolve discrepant results. Taking this into account, we found that all PYR-positive, ODC-positive strains were accurately identified as S. lugdunensis by our real-time assay. Interestingly, we identified one S. lugdunensis isolate by PCR that was PYR positive and ODC negative, an identification that was confirmed by 16S rRNA gene sequencing. To our knowledge, this is the first report of an ODC-negative S. lugdunensis strain in the literature (9). Because ODC activity is the foundation of conventional S. lugdunensis identification, this finding raises the possibility of underidentification. Future work using molecular strategies such as the assay described here will be required to determine the worldwide prevalence of these ODC-negative strains and will facilitate our understanding of their role in human disease.
Despite our finding of an ODC-negative isolate, ODC testing remains a major criterion for S. lugdunensis identification. We hypothesized that a rapid ODC test would provide many of the advantages of our real-time assay without the technical requirements of molecular testing. However, we found that the rapid ODC WEE-TAB reagent was not a suitable substitute for conventional ODC testing and was outperformed by our real-time assay. Additional experiments with other rapid ODC tests, including the next-generation ODC WEE-TAB reagent, will be needed to further assess the utility of rapid biochemicals for the identification of S. lugdunensis.
We believe that the assay described here is the first molecular test specifically designed to identify S. lugdunensis. Though genotyping of the CoNS by real-time PCR and melting curve analysis has been proposed previously (5, 14), only one of these assays can definitively identify S. lugdunensis (5). These assays attempt to identify up to 15 CoNS species and in so doing lack the simplicity and ease of interpretation that our assay provides. The rationale for rapid species-level identification of CoNS other than S. lugdunensis is also unclear, as these organisms have equivalent antibiotic susceptibility interpretive criteria (3) and only rarely cause severe disease.
A competing molecular methodology for the identification of S. lugdunensis is the Staphplex system (Genaco Biomedical Products, Huntsville, AL), which involves multiplex PCR amplification followed by detection via a Luminex 100 flow cytometric suspension array (17). The multiple steps required for this protocol allow the introduction of error and add to the overall turnaround time, estimated at ~5 h. Although this assay may be adequate for the resolution of discrepant or ambiguous results in a reference laboratory, its complexity and turnaround time preclude its use in a routine laboratory. Furthermore, only one S. lugdunensis isolate was included in their sample set, suggesting that the assay requires additional validation to determine its performance characteristics for S. lugdunensis identification.
In summary, we have designed a simple and robust real-time PCR assay that appears well suited to provide the rapid and accurate staphylococcal identification required for critical clinical decision making. A flowchart describing the integration of this assay into the clinical microbiology laboratory is shown in Fig. Fig.2.2. We run this assay at least once daily, providing S. lugdunensis identification ~18 h before conventional ODC testing. While this approach requires that a laboratory have the instrumentation and personnel to perform nucleic acid testing, our real-time PCR reagent costs are comparable to those for conventional ODC media. We anticipate that in addition to evaluating cultures from solid media, direct identification of S. lugdunensis, S. aureus, CoNS other than S. lugdunensis, and mecA from blood and other body fluid culture broths will further expand the utility of this assay. Taken together, our work suggests that this staphylococcal identification strategy is appropriate for routine use in the clinical microbiology laboratory.
We thank the staff of the Stanford Clinical Microbiology Laboratory, in particular, Manjula Mudambi and Nancy Moore, for their hard work, support, and technical expertise. In addition, we are grateful to Laszlo Juhos for his tremendous help with data entry. Finally, we thank Thuy Doan for critical reading of the manuscript.
Published ahead of print on 9 September 2009.