| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Departments of Medical Microbiology1, and Pathology and Laboratory Medicine2, Section Medical Biology, University of Groningen, Hanzeplein 1, 9713 GZ Groningen, The Netherlands
Correspondence Hermie J. M. Harmsen h.j.m.harmse{at}med.umcg.nl
Received January 20, 2005
Accepted July 6, 2005
Enterococcus faecalis and Enterococcus faecium are among the leading causes of hospital-acquired infections. Reliable and quick identification of E. faecalis and E. faecium is important for accurate treatment and understanding their role in the pathogenesis of infections. Fluorescent in situ hybridization (FISH) of whole bacterial cells with oligonucleotides targeted at the 16S rRNA molecule leads to a reduced time to identification. In clinical practice, FISH therefore can be used in situations in which quick identification is necessary for optimal treatment of the patient. Furthermore, the abundance, spatial distribution and bacterial cell morphology can be observed in situ. This report describes the design of two fluorescent-labelled oligonucleotides that, respectively, detect the 16S rRNA of E. faecalis and the 16S rRNA of E. faecium, Enterococcus hirae, Enterococcus mundtii, Enterococcus villorum and Enterococcus saccharolyticus. Different protocols for the application of these oligonucleotides with FISH in different clinical samples such as faeces or blood cultures are given. Enterococci in a biofilm attached to a biomaterial were also visualized. Embedding of the biomaterial preserved the morphology and therefore the architecture of the biofilm could be observed. The usefulness of other studies describing FISH for detection of enterococci is generally hampered by the fact that they have only focused on one material and one protocol to detect the enterococci. However, the results of this study show that the probes can be used both in the routine laboratory to detect and determine the enterococcal species in different clinical samples and in a research setting to enumerate and detect the enterococci in their physical environment.
 |
INTRODUCTION |
|---|
 TOP INTRODUCTION METHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Reliable identification of enterococci, especially E. faecalis and E. faecium, is important for accurate diagnosis and treatment. Current methods of identification are based on phenotypic characteristics that may show variation on testing, and for which the tests are difficult to perform and often take up to several days for a final result (Facklam & Collins, 1989). In recent years, authors have described molecular methods for the detection and identification of Enterococcus species by the use of labelled oligonucleotide probes based on 16S and 23S rRNA genes (Behr et al., 2000; Beimfohr et al., 1993; Betzl et al., 1990; Monstein et al., 1998; Pryce et al., 1999). However, most of these studies use direct DNA-hybridization techniques for the identification of enterococci. Only a few studies describe whole cell in situ hybridization for identification and enumeration of enterococci (Beimfohr et al., 1993; Harmsen et al., 1999).
Fluorescent in situ hybridization (FISH) of whole bacterial cells with oligonucleotides that detect the 16S rRNA molecule has the advantage that bacteria do not need to be cultured before detection and this would lead to a reduced time to identification of the infecting organism (Jansen et al., 2000). In clinical practice, FISH can be used in situations in which quick identification of the infection organism has an advantage in the treatment of the patient. Also, different studies have shown that not all of the bacteria in a specimen can be cultured because of some of the bacteria having entered a non-culturable state, the culture conditions not being suitable or the patient having already been treated with antibiotics (Amann et al., 1995).
The studies using FISH for detection of enterococci that have been published so far have been hampered by the fact that the specificities of the probes were not tested in different mixed bacterial populations or that they were only focused on one specific type of material (Beimfohr et al., 1993; Harmsen et al., 1999; Jansen et al., 2000). In this study, we describe the design of two fluorescent-labelled oligonucleotides that specifically detect the 16S rRNA of E. faecalis or the combination of E. faecium, Enterococcus hirae, Enterococcus mundtii, Enterococcus villorum and Enterococcus saccharolyticus. Furthermore, we describe protocols for the application of these oligonucleotides in FISH on different clinical samples.
|
METHODS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
|
Design and testing of DNA probes.
Oligonucleotide probes were designed with the ARB software package (Ludwig et al., 1998), and rRNA sequences were obtained in an aligned form from the Ribosomal Database Project (RDP) (Maidak et al., 1997) supplemented with newly deposited rRNA sequences from GenBank. Fluorescein-labelled oligonucleotides against selected specific target sequences of E. faecalis and E. faecium were synthesized commercially (Eurogentec) and tested for specificity against the set of reference organisms listed in Table 1.
All strains were cultured in the appropriate fluid medium until exponential phase; this medium was diluted 1 : 50 in PBS (8 g NaCl l–1, 0.2 g KCl l–1, 1.44 g Na2HPO4 l–1 and 0.24 g KH2PO4 l–1). From this cell suspension, 10–20 µl was smeared on a glass slide and air-dried at room temperature (RT). The cells on the glass slides were fixed for 5 min in ethanol (96 %) and again air-dried. To allow permeabilization of their cell membrane, the Gram-positive strains were treated with enzymes prior to hybridization. The other strains were hybridized immediately after fixation. The permeabilization buffer consisted of a fresh mixture of 100 mM Tris/HCl pH 7.5 and 50 mM EDTA. For the permeabilization of Streptococcus species, Enterococcus species, Bifidobacterium species, Eubacterium species and Peptostreptococcus species, permeabilization buffer with 1 mg lysozyme ml–1 (Boehringer) was applied for 30 min at 37 °C. For the permeabilization of Clostridium species, Lactobacillus species and Lactococcus species, permeabilization buffer with 5 mg lysozyme ml–1 and 0.1 mg pancreatic lipase ml–1 (Sigma) was applied for 30 min at 37 °C. After the addition of 20 µl of the enzyme mixture, the cells smears were incubated for 30 min at 37 °C in a buffer-saturated chamber. After treatment the slides were briefly rinsed with milli-Q water and air-dried.
Subsequently, 10 µl hybridization buffer [0.9 M NaCl, 20 mM Tris/Cl (pH 7.5) and 0.1 % (w/v) SDS] containing FITC- or Cy3-labelled probe (10 ng µl–1) was added to the smears. The smears were incubated under a coverslip in a buffer-saturated chamber. The stringency of hybridization was optimized by trying hybridization temperatures ranging from 45 °C to 55 °C, at various duration times from 15 min to 16 h and with formamide concentrations in the hybridization buffer ranging from 0 % to 60 %. Finally, the slides were washed in 50 ml wash solution [0.9 M NaCl, 20 mM Tris-Cl (pH 7.5)], air-dried and mounted in Vectashield (Vector Labs). In all experiments, the Eub338 probe (3'-TGAG GATGCCCTCCGTCG) was used as a positive control (Amann et al., 1990). The slides were evaluated with an Olympus BH2 epifluorescence microscope.
Detection of E. faecalis and E. faecium in clinical samples of faeces and blood cultures.
The application of the probes on different faecal samples was tested by enumerating the enterococci in diluted faecal samples that yielded enterococci on routine culture. The faeces of 20 healthy adult volunteers and nine premature babies between 3 and 21 days old (mean 10 days) with a gestational time between 32 weeks 1 day and 35 weeks (mean 33 weeks 4 days) was tested. Portions (0.5 g) of each stool were diluted 1 : 80 and fixed with paraformaldehyde (PFA) as described previously (Franks et al., 1998). The diluted and fixed solution was mixed 1 : 1 with PBS and 10 µl was smeared on a glass slide and air-dried at RT. Further treatment was performed as described above with modifications as listed in Table 2. To determine the total number of bacteria present in the samples, 0.2 mg 4',6-diamidino-2-phenylindole (DAPI) ml–1 in hybridization buffer was added as a DNA stain to duplicate samples on glass slides, which were incubated and further processed in the same way as the slides with the specific probes.
|
Thirty positive blood cultures with Gram-positive cocci in chains on the Gram stain as described by Jansen et al. (2000) were used and also hybridized with the newly described probes for E. faecalis and E. faecium. A total of 15 µl from a positive blood culture was smeared onto a glass slide and air-dried at RT. Further treatment was performed as described above with modifications as listed in Table 2.
Detection of E. faecalis and E. faecium in bile drain biofilms.
Eleven bile drains were collected from liver transplant patients. The drains were cut into 0.5 cm pieces, and fixed in 4 % (w/v) PFA for 48 h at 4 °C and subsequently stored in 50 % ethanol/PBS (v/v) at 4 °C. The drains were rinsed in distilled water, dehydrated in graded ethanol (50, 70, 96 and 100 %) and embedded in Technovit 7100 hydroxyethyl methacrylate (Kulzer Histo-Technik, Klinipath). Cross-sections of 10 µm were cut and hybridized as described above with modifications as listed in Table 2. Results of the hybridization were confirmed by routine culture.
|
RESULTS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
|
In Table 1 the results of the FISH with probes Enfl84 and Enfm93 on different clinical enterococcal isolates and reference organisms are shown. As reference organisms we chose streptococci because they look very similar to enterococci after Gram staining and other relevant organisms that are frequently found in the faeces to confirm that the enterococci could be distinguished in this complex environment. All the micro-organisms hybridized with the Eub338 probe. The Enfl84 probe only hybridized with E. faecalis isolates. The Enfm93 probe hybridized with E. faecium and also with E. hirae, E. mundtii and E. saccharolyticus. Table 3 shows the selected oligonucleotide sequences of Enfl84 and Enfm93, their target sites and the alignments of the corresponding sites of the enterococci that were used as reference organisms. Enfl84 has a minimum of eight mismatches with the reference enterococci used; this explains the high specificity of the probe. Enfm93 has a full match with E. hirae and E. villorum so it is also specific for these strains. In addition it has only one weak T-G mismatch with E. mundtii and two G-T mismatches at the extreme 5' end of the target with E. saccharolyticus. The hybridization conditions were not stringent enough to prevent hybridization with these targets. However, after increasing the stringency, the signal with E. faecium was too weak for positive identification.
Optimization of hybridization protocols of the clinical samples
Different human clinical sample types for microbiological detection are generally prepared and fixed for hybridization in different ways. For instance faeces is always a multi-species sample, for which a universal fixation protocol should be applied to maintain quantitative aspects, but in blood cultures fast-growing cells hybridize more easily and quantification is not an issue. Therefore material-specific optimized protocols were established. A summary of the specific conditions is given in Table 2. Only the fixation, preparation, permeabilization and duration of hybridization are different for the different materials. The specificity of the probes tested under the same conditions was not changed by these protocols since the stringency of the hybridization was maintained. Hybridization of the faecal samples was optimal without prior permeabilization; probably, the longer fixation in PFA had already permeabilized the bacteria. The blood cultures were permeabilized and hybridized for less time than the pure bacterial cultures; this was necessary for use in the routine laboratory, where quick detection was the priority. This protocol resulted in less clear signals; however, detection and identification were still possible.
Enumeration of enterococci in the faeces of healthy volunteers
In the faeces of the 20 adult volunteers, enterococci were only occasionally detected by FISH, which indicated that the enterococci were below the detection limit for enumeration. The calculated detection limit for our method is 107 cells (g faeces)–1. However, in the faecal samples of seven of the nine babies high numbers of enterococci were found; from the total number of DAPI-stained cells up to 41 % (mean 8.4 %) were E. faecalis and up to 23 % (mean 6.4 %) were E. faecium, with a total number of up to 2.6 x 1010 cells (g wet weight)–1 (mean 3.46 x 109) for E. faecalis and up to 1.1 x 1010 cells (g wet weight)–1 (mean 1.71 x 109) for E. faecium. Fig. 1(a) shows the fluorescence micrograph of E. faecium hybridized with the Enfm83 probe labelled with FITC in a faecal suspension from a premature baby.
|
Detection of E. faecalis and E. faecium in blood cultures
A total of 30 positive blood cultures with Gram-positive cocci in chains on the Gram stain were hybridized with the Enfl84 and Enfm93 probes. Of these blood cultures, 10 were positive for E. faecalis on routine culture. These 10 blood cultures were also positive for the Enfl84 probe after hybridization. E. faecium was not detected in any of the blood cultures. The remaining 20 blood cultures yielded other streptococci on routine culture and did not hybridize with either probe.
Detection of enterococci in biofilms on indwelling devices
After a liver transplantation the bile is often diverted outside the body of the patient for prolonged time by a bile drain inserted into the bile duct. These drains often get colonized with enterococci, and therefore we selected these materials for the detection of enterococci in biofilms attached to biomaterials. From five of the 11 bile drains that were collected from liver transplant patients, E. faecalis was both cultured and detected after embedding and hybridization with the Enfl84 probe. Micro-organisms that were co-cultured from these five drains but did not react with the probes were Enterobacter cloacae, Candida albicans, Staphylococcus aureus, coagulase-negative staphylococci, Enterococcus species (not E. faecalis or E. faecium), Klebsiella pneumoniae, Klebsiella oxytoca and Escherichia coli. From the remaining six bile drains, enterococci were neither cultured nor detected with FISH. Micro-organisms that were cultured from these Enterococcus-negative drains were S. aureus, coagulase-negative staphylococci, Gram-positive rods (not specified), Streptococcus species (not specified), C. albicans and ‘Streptococcus milleri'.
Fig. 2 shows a micrograph of a bile drain covered with a thick layer of bacteria that hybridized with the FITC-labelled Eub338 probe and with a microcolony of E. faecalis on top of it that also hybridized with the Cy3-labelled Enfl84 probe. Routine culture of this drain contained E. faecalis, K. pneumoniae, K. oxytoca and Escherichia coli. The presence of high numbers of Enterobacteriaceae in this biofilm was confirmed by FISH with a DNA probe that hybridizes with Enterobacteriaceae (Poulsen et al., 1995).
|
|
DISCUSSION |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
The Enfl84 probe is specific only for E. faecalis, and the Enfm93 probe hybridized with E. faecium, E. hirae, E. mundtii, E. villorum and E. saccharolyticus. Although a specific rRNA-targeting FISH probe for E. faecium is desirable, a more specific target site could not be identified using a 16S rRNA database. There are at least 10 validated 16S rRNA sequences of E. faecium available in the latest version of the ARB database and sufficient other Enterococcus sequences for comparison. However, with our software no target site was found within the 16S rRNA sequence that was more specific. A positive result from using this probe on clinical samples will almost always indicate detection of E. faecium, because E. hirae and E. mundtii have been isolated only rarely from human sources and then usually as colonizers of humans and not from blood cultures or other clinically important materials (Facklam & Collins, 1989; Gilad et al., 1998; Kaufhold & Ferrieri, 1991; Ruoff et al., 1990). E. saccharolyticus and E. villorum have not been isolated from human sources (Facklam et al., 1999; Vancanneyt et al., 2001). When using FISH with the Enfm93 probe on very complex mixed bacterial populations, e.g. faeces, an organism giving a positive signal is most likely to be E. faecium but the possibility of one of the other three enterococcal species being present should be considered.
Different materials require varying FISH protocols. In this study we showed that only minor adjustments in the permeabilization procedure are necessary to make the protocol applicable for different materials, whereas the stringency and therefore the specificity is not changed in the different protocols. The method was evaluated on faeces because it is one of the most complex bacterial populations of the human body. The fact that enterococci could be detected in this environment implies that the probes also can be used on other clinical materials containing mixed bacterial populations such as sputum or pus from abdominal abscesses. Furthermore, human faeces is the natural habitat of enterococci. Although the numbers were below the detection limit in healthy volunteers, the FISH on faeces might be used in an epidemiological setting among patients highly colonized with enterococci.
Positive blood cultures often contain a single bacterial species that can be identified easily; however, this can take up to 24–48 h and antibiotic treatment is necessary immediately. A study by Munson et al. (2003) showed that the result of the Gram stain of a positive blood culture particularly influenced the antibiotic therapy, whereas fewer interventions were initiated after the antimicrobial susceptibility data became available. As E. faecium is generally more resistant to antibiotics compared to E. faecalis, early discrimination between these species can be important for the empirical therapy. Culture of the bacterium will still be necessary to determine the exact antimicrobial spectrum and further narrow down the antimicrobial therapy; however, we think that the advantage of FISH detection in clinical practice is the speed of determination and therefore quick implementation of correct empirical therapy.
FISH should therefore be used in situations in which the results of culture cannot be awaited because quick identification of the infection organism has an advantage in the treatment of the patient. The fact that FISH can detect non-culturable bacteria is an advantage. In a clinical setting where a patient is on antimicrobial therapy, enumeration of dead cells could be misleading, as the therapy may in fact be working against the enterococci. However, the advantage of FISH is that it does detect non-culturable bacteria that are in a ‘sleeping’ state but not dead and therefore cannot be cultured but possibly still play a role in infecting the patient, e.g. bacteria in a biofilm. Generally, in the treatment of a patient the early results of FISH have to be combined with the later results of culture and the general state of the patient and therefore we do not think that the detection of dead bacteria is a problem.
The application of FISH on positive blood cultures was previously reported and the enterococcal DNA probe used only identified E. faecalis (Jansen et al., 2000). However, analysis of the oligonucleotide sequence with the ARB software package revealed that the probe used in that study detected many other bacterial species. These species are not commonly found in blood cultures but could be present in mixed bacterial populations, and applying the probe to these populations may cause problems. Another study, by Beimfohr et al. (1993), described two probes specific for the detection of E. faecalis and E. faecium in milk. These probes were based on the 23S rRNA sequence. The 23S rRNA database is not as extensive as the 16S rRNA database. Therefore it might be difficult to use these probes in clinical samples containing a complex mixture of bacteria as in faeces because the specificity of the probes can not be compared in silico to many other bacteria.
Enumeration of the enterococci in faeces revealed that the adult volunteers carried numbers under the detection limit for enumeration, but in a group of premature babies high numbers were found. This showed that FISH is a suitable tool for studying the enterococcal content of the faeces and the development of early bacterial gastro-intestinal colonization in this age group. Furthermore, many hospitalized patients use antibiotics with little or no anti-enterococcal activity, which might increase the number of enterococci in their faeces, and therefore this method will also be useful for samples from this group of patients.
Interestingly, we were able to detect enterococci by FISH in a biofilm attached to an explanted biomaterial. Embedding the material in Technovit 7100 preserved the morphology very well and we were able to observe the architecture of the biofilm. These biofilms are a frequent source of persistent infection and many studies have been conducted on preventing the formation of these biofilms (Costerton et al., 1999; Pascual, 2002). Recently it was reported that bacterial biofilms play a role in orthopaedic implant failure even if the routine culture does not reveal a pathogenic bacterium, and it was suggested that microscopic analysis of the surface of the explanted prosthesis could improve the detection of implant infections (Neut et al., 2003). With this protocol we provide a strong tool for detecting bacteria in the biofilm and observing the in vivo composition of the biofilm. Direct observation of the distribution of enterococci in biofilms might lead to a better insight into their development and the discovery of ways to prevent their formation on indwelling biomaterials.
In this study we used the culture of enterococci as a gold standard. The FISH technique is often considered more sensitive than culture because FISH can also show non-culturable or dead bacteria. However, to test the specificity of the FISH test it was necessary to have a gold standard and for this culture is the most applicable. To increase the sensitivity of the routine culture to a maximum, enrichment in liquid media was used prior to subculturing on selective media. However, this method is very elaborative, and quantification of the bacteria is not possible.
In other studies the 16S rRNA-based probe technique has been used to identify different enterococcal species (Behr et al., 2000; Monstein et al., 1998; Patel et al., 1998; Pryce et al., 1999). Only a few, however, have combined the advantage of direct visualization of the enterococci by FISH and the identification by the 16S or 23S rRNA sequence (Beimfohr et al., 1993; Harmsen et al., 1999; Jansen et al., 2000). The usefulness of these studies describing FISH for detection of enterococci is hampered by the fact that they have only focused on one material and therefore one protocol to detect the enterococci. In this study, we describe two DNA probes that can be used for the clinically important enterococci E. faecalis and E. faecium and hybridization protocols for three different materials. Therefore it can be concluded that these probes are ideal for the detection of enterococci in various clinical materials in a microbial laboratory setting where research and routine diagnostics are combined.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
|
REFERENCES |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  N. Wellinghausen, M. Bartel, A. Essig, and S. Poppert Rapid Identification of Clinically Relevant Enterococcus Species by Fluorescence In Situ Hybridization J. Clin. Microbiol., October 1, 2007; 45(10): 3424 - 3426. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. R. Pingle, K. Granger, P. Feinberg, R. Shatsky, B. Sterling, M. Rundell, E. Spitzer, D. Larone, L. Golightly, and F. Barany Multiplexed Identification of Blood-Borne Bacterial Pathogens by Use of a Novel 16S rRNA Gene PCR-Ligase Detection Reaction-Capillary Electrophoresis Assay J. Clin. Microbiol., June 1, 2007; 45(6): 1927 - 1935. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| INT J SYST EVOL MICROBIOL | J MED MICROBIOL | MICROBIOLOGY | J GEN VIROL | ALL SGM JOURNALS |
