Journal of Clinical Microbiology, May 2000, p. 1709-1712, Vol. 38, No. 5
Associated Regional and University
Pathologists Institute for Clinical and Experimental
Pathology1 and the University of
Utah Department of Pathology,2 Salt Lake City,
Utah
Received 9 November 1999/Returned for modification 13 January
2000/Accepted 21 February 2000
Legionella spp. are a common cause of
community-acquired respiratory tract infections and an occasional cause
of nosocomial pneumonia. A PCR method for the detection of legionellae
in respiratory samples was evaluated and was compared to culture. The
procedure can be performed in 6 to 8 h with a commercially
available DNA extraction kit (Qiagen, Valencia, Calif.) and by PCR with
gel detection. PCR is performed with primers previously determined to
amplify a 386-bp product within the 16S rRNA gene of Legionella pneumophila. We can specifically detect the clinically
significant Legionella species including L. pneumophila, L. micdadei, L. longbeachae, L. bozemanii, L. feeleii,
and L. dumoffii. The assay detects 10 fg (approximately two
organisms) of legionella DNA in each PCR. Of 212 clinical specimens
examined by culture, 100% of the culture-positive samples (31 of 31)
were positive by this assay. By gel detection of amplification
products, 12 of 181 culture-negative samples were positive for
Legionella species by PCR, resulting in 93% specificity.
Four of the 12 samples with discrepant results (culture negative, PCR
positive) were confirmed to be positive for Legionella species by sequencing of the amplicons. The legionella-specific PCR
assay that is described demonstrates high sensitivity and high
specificity for routine detection of legionellae in respiratory samples.
Legionella spp. are
gram-negative bacteria responsible for epidemic and sporadic cases of
pneumonia after inhalation of contaminated water droplets from a
variety of water sources (7). Currently, there are 42 species that comprise 64 serogroups of legionellae (3), most
of which have caused disease (6, 15).
An estimated 3 to 8% of all community-acquired pneumonias are caused
by Legionella spp. (3). The case-fatality rate
for patients with legionellosis is 5 to 30%, with the elderly and immunocompromised patients at greater risk of death (4).
Although nosocomial cases are not uncommon, with sources of infection
often being hospital showers and respiratory therapy equipment, most cases occur sporadically. Generally, the clinical presentation of
legionellosis includes systemic problems that involve kidney, liver,
gastrointestinal, and central nervous system dysfunction (5).
Currently, culture is regarded as the "gold standard" for detection
of legionellae. However, due to the fastidious nature of the organism
and problems inherent with specimen collection and transport, culture
may have a sensitivity as low as 50 to 60% (4). While
culture for legionellae will detect many different species, special
medium is required and the organisms take 3 to 5 days to form visible
colonies. Also, testing for legionella is often requested after the
patient has been treated with antibiotics, which decreases the chance
of isolation by culture. Serologic testing is useful only for the
retrospective diagnosis of infection, and as many as 25% of patients
with legionellosis may fail to demonstrate diagnostic titers
(8). Urinary antigen tests are convenient but only detect
antigens of Legionella pneumophila serogroup 1, which cause
more than 70% of cases of legionellosis in the United States
(4). A properly designed PCR assay could improve the speed,
accuracy, and sensitivity of diagnosis.
A PCR assay was developed in which the target for the test is the 16S
rRNA gene, which exists in multiple copies per genome, thus improving
the sensitivity of detection. Several medically relevant
Legionella species including L. pneumophila,
L. micdadei, L. bozemanii, L. longbeachae, L. feeleii, and L. dumoffii can be detected without the use of culture. This study compares PCR to
culture with respiratory samples from patients suspected of having
legionella infection.
Clinical specimens.
The PCR procedure described here is
designed for use with DNA isolated from respiratory specimens. One
hundred eighty-six respiratory samples from patients suspected of
having pneumonia caused by Legionella spp. were submitted to
the Associated Regional and University Pathologists Diagnostic
Infectious Diseases Laboratory for culture with or without direct
fluorescent-antibody assay (DFA). The number and type of each specimen
submitted were as follows: 31 sputum specimens, 66 bronchial washing
specimens, 74 bronchoalveolar lavage (BAL) specimens, 8 pleural fluid
specimens, and 7 lung tissue specimens. Portions of 26 of the
respiratory samples (13 BAL, 7 sputum, 4 bronchial washing, and 2 pleural fluid specimens) were spiked so that they contained a
final concentration of approximately 102 to 103
organisms of L. pneumophila (ATCC 33512) per ml.
Culture.
Each specimen was aliquoted into two samples, one
was frozen at Extraction of DNA from respiratory samples. (i) Sample
preparation.
Samples over 1 ml in volume were centrifuged in a
Beckman model TJ-6 centrifuge at 2,000 × g for 5 min,
the supernatant was discarded, and the pellet volume was brought up to
1 ml with phosphate-buffered saline or sterile water before DNA
extraction. Thick samples were diluted with phosphate-buffered saline
and vortexed until they were homogeneous, and 1 ml was aliquoted for analysis.
(ii) DNA extraction.
DNA extraction was performed with the
QIAamp Tissue Kit (Qiagen, Valencia, Calif.) according to the
manufacturer's instructions with the following modifications. The
columns were washed twice with the wash buffer provided with the kit
and were then washed once with absolute alcohol, followed by elution of
the DNA with 100 µl of the elution buffer (provided with the kit).
The extracted DNA was either amplified immediately or stored at
(iii) Control material.
Positive and negative controls,
previously prepared and stored frozen, were included in the extraction
procedure. The positive control consisted of a dilution in water of
approximately 10 to 100 L. pneumophila ATCC 33152 organisms.
The negative control was 20 µl of a BAL specimen previously
determined to be negative for Legionella species by culture
and PCR.
DNA amplification. (i) Primers.
The primers for the PCR
assay amplified a portion of the 16S rRNA gene from approximately base
451 to base 837 of L. pneumophila ATCC 33152. Forward primer
JFP (5'-AGGGTTGATAGGTTAAGAGC) and reverse primer JRP
(5'-CCAACAGCTAGTTGACATCG) were previously used by Jonas et
al. (9).
(ii) PCR conditions.
Five microliters of extracted template
DNA was used in a 50-µl reaction mixture that included 1× PCR buffer
(50 mM KCl, 10 mM Tris-HCl [pH 8.3]), 3 mM MgCl2, 200 µM dATP, 200 µM dCTP, 200 µM dGTP, 400 µM dUTP, 1 µM primer
JFP, 1 µM primer JRP, 2.8 U of AmpliTaq Gold per µl, and 1 U of
uracil DNA glycosylase (UDG) (Life Technologies, Gaithersburg, Md.) per
µl. Thermal cycling was performed with a Perkin-Elmer GeneAmp PCR
System 2400 (PE Applied Biosystems, Foster City, Calif.). Cycling
conditions began with an initial incubation at 37°C for 10 min to
allow UDG degradation of uracil residues to prevent carryover amplicon
contamination. After a 20-min hold at 95°C, 38 cycles consisting of
94°C for 45 s, 57°C for 45 s, and 72°C for 45 s
were followed by a final extension at 72°C for 60 min, and the
mixture was held at this temperature until gel analysis.
(iii) Amplicon detection.
Ten microliters of the PCR
amplicon was mixed with 1 µl of gel loading buffer (0.5 M Tris, 2.5 mg of bovine serum albumin per ml, 20% sucrose, 1 mM cresol red, 20 mM
MgCl2), and the mixture was electrophoresed on 1.5%
agarose gels in TBE buffer (5 mM Tris, 5 mM boric acid, 0.1 mM EDTA)
with ethidium bromide (0.5 mg/dl) at 80 V for 60 min. The gels were
viewed under UV light. The migration distance was compared to that of
BioMarker Low (BioVentures, Inc., Murfreesboro, Tenn.) to determine the
approximate number of base pairs of the amplification product.
Confirmation of positive results by sequencing. (i) Primers.
The primers used for sequencing analysis of DNA samples with discrepant
results (culture negative, PCR positive) were Leg F2
(5'-GAGGCAGCAGTGGGGAAT) and Leg R2
(5'-CCCAGGCGGTCAACTTAT).
(ii) PCR conditions.
Five microliters of template DNA was
used in a 50-µl reaction mixture that included 1× PCR buffer (50 mM
KCl, 10 mM Tris-HCl [pH 8.3]), 2.5 mM MgCl2, 200 µM
dATP, 200 µM dCTP, 200 µM dGTP, 400 µM dUTP, 1 µM primer Leg
F2, 1 µM primer Leg R2, 2.8 U of AmpliTaq Gold per µl, and 1 U of
UDG per µl. Cycling conditions began with an initial incubation at
37°C for 10 min to allow UDG degradation of uracil residues in
carryover amplicon contamination. After a 15-min hold at 95°C, 36 cycles consisting of 94°C for 30 s, 59°C for 40 s, and
72°C for 40 s were followed by a final extension at 72°C for
20 min, and the mixture was held at this temperature until purification.
(iii) Amplicon purification.
PCR products were purified with
the QIAquick PCR purification kit (Qiagen) according to the
manufacturer's instructions and were held at (iv) Sequencing.
Sequencing was performed by use of BigDye
chemistry with the ABI Prism 377 Sequencer (PE Applied Biosystems).
Sequence analysis was performed by a BLAST search of the GenBank database.
Dilution studies.
The detection limit of the current PCR
assay was determined by using limiting dilutions of DNA. DNAs from
several separate Legionella species were quantified
spectrophotometrically and were brought to 2 pg/µl with water.
Tenfold dilutions were made, and 5 µl of each dilution was analyzed
in a 50-µl PCR. Detection limits were determined for six separate
species of legionellae consisting of L. pneumophila ATCC
33152, L. micdadei ATCC 33218, L. bozemanii ATCC
33217, L. longbeachae ATCC 33462, L. feeleii ATCC
33849, and L. dumoffii ATCC 33279.
0095-1137/00/$04.00+0
Copyright © 2000, American Society for Microbiology. All rights reserved.
Detection of Legionella Species in
Respiratory Specimens Using PCR with Sequencing Confirmation
![]()
ABSTRACT
Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References
![]()
INTRODUCTION
Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References
![]()
MATERIALS AND METHODS
Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References
70°C, and the other was applied to buffered charcoal
yeast extract (BCYE) agar and BCYE agar with polymyxin B, anisomycin, and vancomycin (Remel, Lenexa, Kans.). For culture analysis, sputum specimens and specimens obtained by bronchoscopy were plated undiluted, pleural fluid specimens were concentrated by centrifugation, and lung
tissue was homogenized in a Stomacher 80 Lab Blender (CMS/Fisher, Houston, Tex.) with approximately 2.0 ml of Trypticase soy broth. Specimens were processed immediately upon receipt. The plates were
incubated at 35°C with 70 to 80% relative humidity. All cultures were examined daily and were held for 5 days before final reporting of
results. Visible colonies were subcultured onto a BCYE agar plate and a
blood agar plate with 5% sheep blood (Remel). Bacteria that grew on
BCYE agar but not on a blood agar plate were considered possible
Legionella spp. and were tested by DFA (Genetic Systems Corp., Redmond, Wash.) for L. pneumophila serogroups 1 to 6. If the DFA with the isolate was negative, the organism was sent to the
Utah State Health Laboratory for further identification.
20°C for up to 6 months until analysis.
20°C until sequencing.
4 dilution and a 10
5 dilution in water,
from which 10 µl of each dilution was plated onto BCYE agar plates
for colony counts. Two fivefold dilutions were further obtained, and 10 µl of each solution was used for DNA extraction.
| |
RESULTS |
|---|
|
|
|---|
Detection of a 386-bp amplification product is a positive result by the present PCR assay for legionellae. The amount of DNA consistently detected in each PCR is 10 fg by use of limiting dilutions for each of L. pneumophila, L. micdadei, L. bozemanii, L. longbeachae, L. feelii, and L. dumoffii. Occasionally, 1 fg of DNA was detected, suggesting that the exact detection limit is probably between 1 and 10 fg of DNA. L. pneumophila has a genome consisting of approximately 3.9 × 106 bp (about 4.3 fg of DNA) (2), indicating that 2.3 organisms can be detected by this assay. Since 5 of the 100 µl of DNA extracted from a sample was used in each PCR, 5 to 46 organisms per specimen corresponds to a detection limit of 1 to 10 fg of DNA per PCR.
The detection range of 5 to 46 organisms per sample was further confirmed by determining the number of organisms detected when performing PCR with the same dilutions that had been plated for CFU determination. It remains difficult to find the ultimate endpoint; however, after four separate dilution series with different strains of legionellae, we found a lower endpoint of 13.4 organisms for L. bozemanii. We showed the limit of detection to be as high as 55 organisms (L. pneumophila). Dilutions of L. micdadei and L. longbeachae resulted in the detection of 30.8 and 23.2 organisms, respectively.
Twenty separate species of organisms commonly occurring in the respiratory tract were analyzed by the current PCR assay. No cross-reactivity was found with any of these organisms, which included Candida albicans, Streptococcus pneumoniae, Arcanobacterium haemolyticum, Staphylococcus aureus, Actinobacillus actinomycetemcomitans, Capnocytophaga granulosa, Eikenella corrodens, Pseudomonas aeruginosa, Moraxella catarrhalis, Pasteurella multocida, Haemophilus influenzae, Escherichia coli, Klebsiella pneumoniae, Neisseria sicca, Neisseria meningitidis, Bordetella bronchiseptica, Acinetobacter baumannii, Acinetobacter junii, Chlamydia pneumoniae, and Mycoplasma pneumoniae. Limiting dilutions of A. baumannii ATCC 19606 and A. junii ATCC 17908 were further analyzed and were not found to amplify a 386-bp product.
Of 212 specimens examined by both culture and PCR for detection of Legionella spp., 31 samples positive by culture were also positive by PCR. Sequencing was performed with DNA amplicons from 12 samples that were PCR positive and culture negative. The results for 4 of 181 culture-negative samples remained discrepant after an attempt to confirm the results by sequencing. A GenBank BLAST search (1) indicated that all samples with discrepant results did, in fact, contain Legionella spp. and that the remaining eight PCR-positive samples were perhaps contaminated. Addition of a sequencing step to confirm all PCR-positive results for culture-negative samples improved the specificity from 93 to 98%.
We designed sequencing primers whose sequences were upstream and
downstream of the original PCR amplicon so that the legionella primer
sequences (primers JFP and JRP) were included in our analysis (Table
1). Comparison of the sequences of the
JFP and JRP primer regions among different Legionella
species showed very few, if any, base differences from the L. pneumophila sequence. For routine analysis, we have also validated
that sequencing of the amplicon obtained in the original PCR with
primers JFP and JRP provides the same results obtained with the
amplicon generated by using the less specific outside primers.
|
A BLAST search of the 386-bp PCR product from the first eight samples with discrepant results showed that they contained sequences with homology to the 16S rRNA gene of Acinetobacter species or an unidentified Proteobacterium 16S rRNA gene and does not give any indication that they are Legionella species. The sequence in each of the remaining samples with discrepant results (samples 9 to 12 in Table 1) was homologous to those of Legionella species and had fewer base pair differences from L. pneumophila than did samples 1 to 8, which had discrepant results. The sequences of the entire 386-bp amplicon of samples 9 and 10, which had discrepant results, were identical to that of L. micdadei ATCC 33218, while the sequence in sample 11, which had a discrepant result, was identical to that of L. longbeachae ATCC 33462.
| |
DISCUSSION |
|---|
|
|
|---|
The legionella-specific PCR assay described in our study targets the 16S rRNA gene and is very sensitive, detecting 100% of all culture-positive respiratory samples. Among those culture-negative samples that were PCR positive, we were able to confirm the results for samples with true-positive results and distinguish them from samples with false-positive results with good confidence using sequence analysis. By sequencing confirmation, 98% specificity was accomplished compared to the results of culture.
Tang et al. (14) showed that identification of bacterial species by sequence analysis can successfully be performed with the MicroSeq system (PE Applied Biosystems Division, Foster City, Calif.), which includes a sequence database library to be used for identification of several bacteria to the species level. We cannot use the MicroSeq system primers to amplify DNA obtained from respiratory samples because the primers are not legionella specific and there may be several different types of bacteria in a respiratory sample. In our study, there appears to be enough variability within the 386-bp amplicon generated (approximate bases 451 to 837 of the 16S rRNA gene) to determine the species of several strains of legionellae.
Less stringent conditions were applied to the assay for detection of most, if not all, of the 64 or more species and serogroups of legionellae. Jonas et al. (9) used identical primers and similar conditions compared with those used for PCR in the current study in conjunction with a probe-capture technique to delineate samples with false-positive results. While we favored their choice of primers for detection of multiple species, we found their detection procedure to be undesirable since their probe did not detect L. micdadei, which is a common disease-causing strain. Our studies indicate that sequencing of all PCR-positive samples for confirmation of results provides a higher degree of specificity without a loss of sensitivity, as may occur with probe-capture detection systems.
Genetic targets other than the 16S rRNA gene, such as the intergenic 16S-23S ribosomal spacer region (11, 13) and the mip gene (12), have been used to determine the species of pure isolates of legionellae. For the detection of legionella DNA extracted directly from respiratory samples, our goal was to use genus-specific primers so that other non-Legionella organisms would not be detected. In the present study, the primers that target the 16S rRNA gene appear to be very specific for Legionella spp., while they amplify a variety of different species which cause legionellosis.
Although the sequences of the false-positive samples were very homologous to those of Acinetobacter spp., we could not get a positive PCR result with DNA extracted from A. baumannii or A. junii. There remains the possibility that the sequence with a false-positive result is not that of an Acinetobacter sp., although the BLAST search showed high scores (767 bits; 100% [387 of 387] identities for the A. junii 16S rRNA gene).
It is also interesting that samples 9, 10, and 11, which had discrepant results (they were not detected by culture), were non-L. pneumophila strains of legionellae. Without sufficient data with which to draw conclusions, a suspicion of the loss of sensitivity of culture for detection of legionellae that are not L. pneumophila arises. It may be possible that less common strains are not as easily detected in the microbiology laboratory by culture and DFA. There is a broad range (50 to 99%) of reported sensitivities of culture in the literature (4, 6). This discrepancy is most likely due to differences in the types and qualities of the specimens used in each study. Collection of specimens in saline (BAL specimens, bronchial washes) lowers the sensitivity of culture because saline inhibits the growth of legionellae (6). Sputum samples may include contamination with enough oropharyngeal organisms to overgrow and mask Legionella spp. Another contribution to a lowered sensitivity of culture is the use of samples collected from patients currently being treated with antibiotics. A PCR test is likely to overcome some of these issues.
The accurate diagnosis of legionella pneumonia has important implications for treatment of the infection. Many first-line antibiotics commonly used to treat typical bacterial pneumonias (i.e., beta-lactams) are ineffective against Legionella species (10). This is at least partially due to the fact that legionella bacteria are intracellular pathogens. There is a great need for a rapid diagnosis of legionella infection. While DFA is rapid, the test detects only L. pneumophila serogroups 1 to 6 and fails to detect other Legionella species and serogroups. Testing for legionella antigen in urine is a good test but also fails to detect multiple serotypes and species. Culture of respiratory samples can detect multiple species, but it may take up to 5 days for results. Our findings show that a legionella-specific PCR of respiratory samples that targets a 386-bp portion of the 16S rRNA gene can be performed in 6 to 8 h with high sensitivity and high specificity of results. The specificity is further improved by confirming positive results by sequencing analysis. If the PCR-positive samples confirmed to be positive by sequencing are accepted as true positives on the basis of a comparison with sequences in the GenBank database, then no specimens have false-positive results (100% specificity) by the combined PCR-sequencing assay.
| |
ACKNOWLEDGMENTS |
|---|
We are very grateful to Martha Bale for expert opinions. We also appreciate the assistance of the Associated Regional and University Pathologists Diagnostic Infectious Disease Laboratory in supplying bacterial strains and patient samples.
The work was supported by the Associated Regional and University Pathologists Institute for Clinical and Experimental Pathology.
| |
FOOTNOTES |
|---|
* Corresponding author. Mailing address: ARUP Institute for Clinical and Experimental Pathology, 500 Chipeta Way, Salt Lake City, UT 84108. Phone: (801) 583-2787, ext. 2439. Fax: (801) 584-5207. E-mail: cloudjl{at}arup-lab.com.
| |
REFERENCES |
|---|
|
|
|---|
| 1. |
Altschul, S. F.,
T. L. Madden,
A. A. Schaffer,
J. Zhang,
Z. Zhang,
W. Miller, and D. J. Lipman.
1997.
Gapped BLAST and PSI-BLAST: a new generation of protein database search programs.
Nucleic Acids Res.
25:3389-3402 |
| 2. | Bender, L., M. Ott, R. Marre, and J. Hacker. 1990. Genome analysis of Legionella spp. by orthogonal field alternation gel electrophoresis (OFAGE). FEMS Microbiol. Let. 60:253-257[Medline]. |
| 3. | Benson, R. F., and B. S. Fields. 1998. Classification of the genus Legionella. Semin. Respir. Infect. 13:90-99[Medline]. |
| 4. | Breiman, R. F., and J. C. Butler. 1998. Legionnaires' disease: clinical, epidemiological, and public health perspectives. Semin. Respir. Infect. 13:84-89[Medline]. |
| 5. | Cunha, B. A. 1998. Clinical features of Legionnaires' disease. Semin. Respir. Infect. 13:116-127[Medline]. |
| 6. | File, T. M., J. S. Tan, and J. F. Plouffe. 1998. The role of atypical pathogens: Mycoplasma pneumoniae, Chlamydia pneumoniae, and Legionella pneumophila in respiratory infection. Infect. Dis. Clin. N. Am. 12:569-592[Medline]. |
| 7. | Friedman, H., Y. Yamamoto, C. Newton, and T. Klein. 1998. Immunologic response and pathophysiology of legionella infection. Semin. Respir. Infect. 13:100-108[Medline]. |
| 8. | Harrison, T. G., and A. G. Taylor. 1988. Timing of seroconversion in Legionnaires' disease. Lancet ii:795. |
| 9. | Jonas, D., A. Rosenbaum, S. Weyrich, and B. Sucharit. 1995. Enzyme-linked immunoassay for detection of PCR-amplified DNA of legionellae in bronchoalveolar fluid. J. Clin. Microbiol. 33:1247-1252[Abstract]. |
| 10. | Klein, N. C., and B. A. Cunha. 1998. Treatment of Legionnaires' disease. Semin. Respir. Infect. 13:140-146[Medline]. |
| 11. |
Presti, F. L.,
S. Riffard,
F. Vandenesch, and J. Etienne.
1998.
Identification of Legionella species by random amplified polymorphic DNA profiles.
J. Clin. Microbiol.
36:3193-3197 |
| 12. |
Ratcliff, R. M.,
J. A. Lanser,
P. A. Manning, and M. W. Heuzenroeder.
1998.
Sequence-based classification scheme for the genus Legionella targeting the mip gene.
J. Clin. Microbiol.
36:1560-1567 |
| 13. |
Riffard, S.,
F. L. Presti,
P. Normand,
F. Forey,
M. Reyrolle,
J. Etienne, and F. Vandenesch.
1998.
Species identification of Legionella via intergenic 16S-23S ribosomal spacer PCR analysis.
Int. J. Syst. Bacteriol.
48:723-730 |
| 14. |
Tang, Y.,
N. M. Ellis,
M. K. Hopkins,
D. H. Smith,
D. E. Dodge, and D. H. Persing.
1998.
Comparison of phenotypic and genotypic techniques for identification of unusual aerobic pathogenic gram-negative bacilli.
J. Clin. Microbiol.
36:3674-3679 |
| 15. |
Winn, W. C.
1988.
Legionnaires' disease: historical perspective.
Clin. Microbiol. Rev.
1:60-81 |
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»