A dual approach employing MALDI-TOF MS and real-time PCR for fast species identification within the Enterobacter cloacae complex

Abstract

A real-time PCR procedure targeting the gene of the molecular cochaperon DnaJ (dnaJ) was developed for specific detection of strains belonging to the Enterobacter cloacae group. The inclusivity and exclusivity of the real-time PCR assay were assessed with seven reference strains of E. cloacae, 12 other Enterobacter species and 41 non-Enterobacter strains. Inclusivity as well as exclusivity of the duplex real-time PCR was 100%. In contrast, resolution of matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) was inadequate for delineation of Enterobacter asburiae, Enterobacter hormaechei, Enterobacter kobei and Enterobacter ludwigii from E. cloacae. Eleven of 56 (20%) clinical isolates of the E. cloacae group could not be clearly identified as a certain species using MALDI-TOF MS. In summary, the combination of MALDI-TOF MS with the E. cloacae-specific duplex real-time PCR is an appropriate method for identification of the six species of the E. cloacae complex.

Introduction

Enterobacter cloacae are rod-shaped, gram-negative bacteria from the Enterobacteriaceae family. They can be found on plants, particulary fruits and vegetables, as well as on human skin and tissues, insects or water reservoirs (Hoffmann & Roggenkamp, 2003; Neto et al., 2003). Besides Enterobacter aerogenes, E. cloacae is by far the most frequent nosocomial pathogen among Enterobacter species (Sanders & Sanders, 1997). It is responsible for various infections, including bacteremia or lower respiratory tract infections (Sanders & Sanders, 1997). The widespread application of antibiotics results in an increased resistance of E. cloacae to antibiotics like ampicillin or narrow-spectrum cephalosporins (Seeberg et al., 1983; Tzelepi et al., 2000). Resistant bacteria may be released directly to the environment, particularly from clinical wastewater systems. Once present in the environment, resistance genes may spread across taxons and habitats via horizontal gene transfer. Here, E. cloacae acts as an indicator organism for a critical antibiotic resistance status among microbial communities in water systems.

Currently, six species have been assigned to the E. cloacae complex including Enterobacter asburiae, E. cloacae, Enterobacter hormaechei, Enterobacter kobei, Enterobacter ludwigii and Enterobacter nimipressuralis (Hoffmann et al., 2005a; Paauw et al., 2008). Discrimination of these species by phenotypic methods as well as 16S rDNA sequencing is difficult. Indeed, single-locus-based molecular methods like sequence analysis of oriC, gyrB, rpoB or hsp60 resulted in distinct genetic clusters, but not all clusters could be assigned to a specific species. Other molecular methods described for accurate identification of these species like comparative genomic hybridization analysis (CGH), and especially combination of CGH with multilocus sequence analysis (MLSA), worked well (Hoffmann & Roggenkamp, 2003; Paauw et al., 2008) but are too expensive and labour-intensive for routine analysis. Correct species identification is clinically relevant as the different clusters of the E. cloacae nomenspecies result in different virulence outcomes. Here, we describe a method combining matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF-MS) and real-time PCR for rapid and accurate identification of E. cloacae.

Materials and methods

Bacterial strains

The following E. cloacae reference strains were used in this study: DSM 3264, DSM 6234, DSM 16657, DSM 30054, DSM 30060, DSM 30062 and DSM 46348. Additionally, 56 reference strains representing 37 species of the Enterobacteriaceae family and five species of other taxonomic families were chosen for determination of the selectivity of the real-time PCR assay (Tables 1 and 2). Fifty-six biochemically characterized clinical isolates of E. cloacae were obtained from four different sources, synlab (Dachau, Germany), Klinikum Bogenhausen (Munich, Germany), Labor Becker, Olgemöller & Kollegen (Munich, Germany) and the Bavarian Health and Food Safety Authority (LGL) routine diagnostic laboratory. All isolates were subjected to both MALDI-TOF MS and the newly developed real-time PCR. All reference strains and clinical isolates were subcultured at 37 °C on Columbia sheep blood agar.

Extraction of DNA

DNA was extracted from bacterial strains following either the instructions of the High Pure Template Preparation kit (Roche Applied Science, Mannheim, Germany) or via heat lysis. For heat lysis, bacteria grown on appropriate media (Endo-Agar or Columbia sheep blood agar; Oxoid, Wesel, Germany) were resuspended in 1.5 mL physiological saline solution (0.9%). Twenty microlitres of this solution were added to 400 μL sterile water and heated at 95 °C for 15 min. After centrifugation, the supernatant was used for amplification. Purity and concentration of the DNA were analysed with the Nanodrop 1000 Spectrophotometer (Peqlab Biotechnologie GmbH, Erlangen, Germany).

Primer and probe design

The sequences for the E. cloacae-specific oligonucleotide primers (dnaJ_f1 and dnaJ_r2) and the E. cloacae target probe (dnaJ_p3) were designed based on a multiple alignment of dnaJ sequences of species belonging to the Enterobacteriaceae family which were deposited in GenBank. Primer sets and probes for the internal amplification control (IAC) ntb2, a 125-bp sequence of Nicotiana tabacum, were adapted from the study by Anderson et al. (2011). Sequences of all primers and probes used in the multiplex PCR are listed in Table 3.

Conventional PCR

Conventional PCR was performed in 25 μL reactions. The reaction mixtures contained 2.5 mM MgCl₂, 0.2 mM dNTP, 0.4 pM primer (Table 3) and 0.06 U μL⁻¹ HotStar-Taq-DNA-polymerase (Qiagen, Hilden, Germany). The PCR program consisted of an initial activation step for 5 min at 94 °C followed by 32 cycles of denaturation for 60 s at 94 °C, annealing for 30 s at 56 °C and extension for 60 s at 72 °C.

Duplex real-time PCR

Real-time PCR was performed in 20 μL reactions in a LightCycler^® 480 multiwell plate 96 (Roche Applied Science). A quantity of 10× primer–probe mixes were prepared for each individual primer–probe set (Table 3). Each primer–probe mix contained the respective primers and probes at a final concentration of 2 μM.

Each reaction mix contained 10 μL 2× QuantiTect Multiplex RT-PCR NoRox Mastermix (Qiagen); 2.0 μL primer–probe mix from each of the 10× primer–probe mix for detection of dnaJ and ntb2; 1 μL of 25 copies of IPC-ntb2 plasmid DNA (Anderson et al., 2011); and 4 μL template DNA. A quantity of 0.5 μL sterile PCR grade water was then added to bring the final volume to 20 μL. The duplex real-time PCR was performed in a LightCycler^® 480 instrument. The PCR program consisted of an initial activation step for 15 min at 95 °C followed by 40 cycles of denaturation for 60 s at 95 °C and annealing/extension for 60 s at the optimized temperature of 59 °C. Owing to the careful selection of the two detection channels employed (FAM/TexasRed), a colour compensation experiment was not necessary.

Determination of the detection limit

To determine the detection limit of the duplex real-time PCR and the corresponding singleplex PCR, a DNA dilution series ranging from 50 ng μL⁻¹ to 0.5 fg μL⁻¹ was measured. Each measurement was repeated three times with DNA of E. cloacae ssp. cloacae DSM 30054^T. The same dilution series was used for calculating PCR linearity and efficiencies from the formula E = 10^−1/slope (Pfaffl, 2001).

MALDI-TOF-MS analysis

MALDI-TOF MS sample preparation

All isolates were grown for 20 h on Columbia sheep blood agar plates at 37 °C. Single colonies were picked and resuspended in 300 μl of sterile water. Nine hundred microlitres of ethanol abs. was added. The mixture was centrifuged at 10 000 g for 2 min. After the supernatant was discarded, the pellet was centrifuged again. Residual ethanol was completely removed by pipetting, and the pellet was allowed to dry at room temperature. Subsequently 30 μL of formic acid (70%) was added and mixed with the pellet by vortexting. Next, 30 μL of acetonitrile was added and mixed thoroughly. The solution was centrifuged at maximum speed for 2 min again and 1.5 μL of the supernatant was spotted on the MALDI target plate (Bruker Daltonics, Bremen, Germany) in two replicates. Immediately after drying, 1.5 μL of the Matrix solution was added to each spot and allowed to air dry. The matrix used was a saturated solution of α-cyano-4-hydroxycinnamic acid (Bruker Daltonics) dissolved in 50% acetonitrile (v/v), with 0.025% trifluoroacetic acid (v/v). Brukers Bacterial Test Standard (Bruker Daltonik GmbH, Bremen, Germany) was used as mass calibration standard. Samples were then processed in the MALDI-TOF MS spectrometer (Microflex LT; Bruker Daltonics) with flex control software (Bruker Daltonics). Each spectrum was obtained by averaging 500 laser shots acquired in the automatic mode at the minimum laser power necessary for ionization of the samples. The spectra have been analysed in an m/z range of 2–20 kDa. Data analysis was performed using BioTyper™ 1.1 software (Bruker Daltonics). MALDI-TOF identifications were classified using score values proposed by the manufacturer: a score ≥ 2 indicated species identification; a score between 1.7 and 1.9 indicated genus identification; and a score < 1.7 indicated no reliable identification. According to Mellmann et al. (2009), a score value distance of at least 0.15 between the two best-scored species was defined as necessary for a precise species identification.

Results

Optimization of the duplex real-time PCR

Performance, optimal annealing temperatures of the dnaJ primer pairs (Table 3) and expected product sizes were first determined with a conventional PCR followed by electrophoresis on agarose gels and sequencing of the PCR products with dnaJ_f1. Use of DNA from the E. cloacae reference strain DSM 30054^T resulted in amplification of an appropriate PCR product, while the PCR was negative for other members of the E. cloacae complex, E. asburiae, E. hormaechei, E. kobei, E. ludwigii and E. nimipressuralis (Table 1).

The duplex real-time PCR was optimized by varying the annealing temperature from 54 to 60 °C and the number of ntb2 copies. It was found that an annealing temperature of 59 °C was optimal for the reaction. Decreasing the annealing temperature resulted in the formation of false positive results for other Enterobacter species than E. cloacae. Furthermore, the concentration of ntb2-DNA was set to 25 copies per μL corresponding to a C_t of 35.00 cycles.

Selectivity

Selectivity of the duplex real-time PCR assay was examined using seven reference strains of E. cloacae, 12 other Enterobacter species and 41 non-Enterobacter strains. All strains used for selectivity testing were obtained directly from official culture collections (DSMZ), or were well-characterized strains from the LGL strain collection, or the Robert Koch Institut (Wernigerode, Germany). Tables 1 and 2 show the results of the inclusivity and exclusivity tests. As all seven E. cloacae reference strains tested were identified correctly, the inclusivity of the duplex real-time PCR was 100%.

All non-E. cloacae strains tested were positive for the IAC with C_t-values ranging from 34.43 to 35.00. Thus, presence of inhibitory substances could be excluded. No false positive results for the dnaJ gene were obtained for all strains used for exclusivity testing (Tables 1 and 2). In particular, none of the other members of the E. cloacae complex was misidentified as E. cloacae (Table 1). Therefore, exclusivity of the duplex real-time PCR was 100%.

Detection limit and PCR efficiency

Detection limit and PCR efficiency of the dnaJ system was determined by measuring DNA dilution series from E. cloacae ssp. cloacae DSM 30054^T ranging from 50 ng μL⁻¹ to 0.5 fg μL⁻¹. The detection limit of the dnaJ primer–probe system was 500 fg μL⁻¹ for both the singleplex and the duplex assay. The dnaJ system also showed good linearity across the range of detection with a slope of 3.49 and r² values of > 0.99, resulting in a PCR efficiency of 1.93 for the duplex real-time PCR (Table 4). The PCR efficiencies for the dnaJ and the ntb2 system are illustrated in Fig. 1.

Performance of the duplex real-time PCR compared with MALDI-TOF MS

MALDI-TOF MS spectra were obtained for seven reference strains of E. cloacae, one reference strain of each of the five other species of the E. cloacae complex and 56 clinical isolates of E. cloacae (Tables 1 and 2). Typical mass spectrometric fingerprints of reference strains are shown in Fig. 2. In addition, DNA of all clinical isolates was subjected to dnaJ duplex real-time PCR.

While application of the dnaJ duplex real-time PCR to reference strains allowed delineation of E. cloacae from the other members (Table 1) of the E. cloacae complex, MALDI-TOF MS did not (Table 6). A distinct identification with MALDI-TOF MS was only possible for E. cloacae and E. nimipressuralis. Analysis of E. asburiae, E. hormaechei, E. kobei and E. ludwigii resulted in log(score) values that did not allow for the definitive assignation of the analysed strains to E. cloacae or the respective species. For example, log(score) values for E. asburiae DSM 17506 were 2.26 ± 0.00 and 2.23 ± 0.07 for E. cloacae.

Analysis of clinical isolates

To test the performance of the duplex real-time PCR and MALDI-TOF MS compared with biochemical characterization, 56 clinical isolates previously characterized as E. cloacae with biochemical methods were obtained from different routine laboratories.

Only 45 clinical isolates (80%) were assigned to a certain species using MALDI-TOF MS (Table 6). All of them were identified as E. cloacae. No definite results were obtained for 11 strains (20%) as minor differences of log(score) values did not allow for a clear decision, whether the respective isolate was E. cloacae or belonged to another member of the E. cloacae complex. Fortunately, clear identification of these isolates was not hindered by species not belonging to the E. cloacae complex.

In contrast, 53 isolates (95%) could be identified as E. cloacae using the dnaJ duplex real-time-PCR. Only for three isolates, divergent results were obtained for biochemical characterization and the real-time PCR.

Discussion

In this study, a duplex real-time PCR was developed for delineation of E. cloacae from other species of the E. cloacae complex. The combination of this PCR with MALDI-TOF MS allowed the correct identification of the respective species of the E. cloacae complex (Tables 1 and 5).

Generally, identification of a specific species within the E. cloacae complex is difficult. The taxonomy of the E. cloacae complex is mainly based on whole-genome DNA–DNA hybridization and differentiation of phenotypic characteristics (Hoffmann & Roggenkamp, 2003). The taxonomic classification of the E. cloacae complex is still ongoing. In recent years, several descriptions for new species as well as reassignments took place (Brenner et al., 1986; O'Hara et al., 1989; Kosako et al., 1996; Hoffmann et al., 2005a, b, c). Hence, it is not surprising that sequencing of 16S rDNA and several other housekeeping genes like oriC, gyrB, rpoB or hsp60 alone is not suitable for the identification of a specific species within this complex. Combination of MLSA with array CGH seems to be most promising for this purpose (Hoffmann & Roggenkamp, 2003; Paauw et al., 2008). As more precise identification of E. cloacae complex is of particular interest for clinical diagnosis [different members of the complex are believed to be involved in pathogenesis in different ways (Morand et al., 2009)], an identification method suitable for routine diagnosis is needed. In this context, MLST and array CGH are by far too time-consuming and cost-intensive, as previously mentioned.

Recently, MALDI-TOF MS has become a rapid and reliable method for the identification of bacteria at genus, species and strain level (van Baar, 2000). Irrespective of the various advances of MALDI-TOF MS compared with biochemical methods, resolution of certain taxonomic groups still remain a daunting challenge. One of these difficult groups is the E. cloacae complex. Indeed, reference strains of E. cloacae itself and E. nimipressuralis were identified correctly and reliably using MALDI-TOF MS. In contrast, E. asburiae, E. hormaechei, E. kobei and E. ludwigii could not be delineated from E. cloacae (Table 6). Eleven of 56 (20%) clinical isolates precharacterized as E. cloacae by biochemical methods could only be assigned to the E. cloacae complex and not to a certain species (Table 5). This is not satisfying with regard to a reproducible and reliable method for species delineation within the E. cloacae complex.

Another method feasible for routine analysis with regard to time-efficiency and reliability are real-time PCRs. More recently, Pham et al. (2007) identified the gene of the molecular cochaperon DnaJ (dnaJ) as a gene with higher discriminatory power among Enterobacteriaceae than 16S rDNA, tuf and atpD genes. We generated alignments for the E. cloacae complex based on different genes like oriC, rpoB or gyrB. Again, dnaJ turned out to be the most powerful marker for delineation of E. cloacae from the other species of the complex. The selectivity of the primers and probe based on dnaj was tested both by homology searches of a nucleotide database (blastn) and by screening of seven E. cloacae and 56 non-E. cloacae strains including at least one representative of each species belonging to the E. cloacae complex. Neither false negatives nor false positives were recorded. The combination with ntb2 as IAC allowed exclusion of the presence of inhibitory substances and dysfunctions of the PCR in case of dnaJ-negative results.

Application of the duplex real-time PCR to clinical isolates, biochemically precharacterized as E. cloacae, resulted in the identification of 53 isolates as E. cloacae. Three isolates were identified as non-E. cloacae isolates. As MALDI-TOF MS identified these isolates as: (a) E. hormaechei (log score value 2.39) or E. cloacae (2.32); (b) E. kobei (2.24 ± 0.08) or E. cloacae (2.20 ± 0.07) and (c) E. asburiae (2.15 ± 0.08) or E. cloacae (2.14 ± 0.01), and E. cloacae was excluded by the duplex real-time PCR, we suggest that these isolates are indeed: (A) E. hormaechei; (B) E. kobei and (C) E. asburiae regarding the known difficulties of biochemical discrimination of these species.

Conclusions

In conclusion, the duplex real-time PCR described here has high selectivity and is suitable for reliable identification of E. cloacae. Exclusive use of MALDI-TOF MS does not have the discriminatory power for clear and reliable identification of certain species within the E. cloacae complex. However, supplementing MALDI-TOF MS with determination of the presence or absence of E. cloacae in a specific isolate using the duplex real-time PCR allows for correct identification of all reference strains used.

Acknowledgements

The authors would like to acknowledge the financial support of the Bavarian State Ministry of the Environment and Public Health. We are grateful to the Klinikum Bogenhausen (Munich, Germany) and the laboratories synlab (Dachau, Germany) and Becker, Olgemöller and Partner (Munich, Germany) for providing us with isolates of Enterobacter cloacae. We would like to thank Henrike Skala and Anika Luze for invaluable technical assistance.

References

Author notes