Novel PCR-Probe Assay for Detection of and Discrimination between Legionella pneumophila and Other Legionella Species in Clinical Samples

In recent years, several PCR-based methods for detection of Legionella DNA have been described, but all PCR assays commonly lacked the ability to discriminate between Legionella pneumophila and other Legionella species (1-3, 5). A recent study indicated that Legionella species other than pneumophila may be important in the etiology of community-acquired pneumonia (4) and that their prevalence may have been underestimated due to inadequate diagnostic methods presently in use. To improve Legionella detection we designed a novel PCR-probe assay. The primer set, LEG1 (5′-TACCTACCCTTGACATACAGTG-3′) and LEG2 (5′-CTTCCTCCGGTTTGTCAC-3′), was derived from the 16S rRNA gene sequence and used to amplify a 200-bp DNA fragment specific for all Legionella species. Probes LSPP-B (5′-GGTTGCGCTCGTTACG-3′) and LPN-B (5′-GAGTCCCCACCATCACATG-3′) were selected to discriminate L. pneumophila from other Legionella species (with the exception of Legionella londonensis, which has a probe sequence identical to that of L. pneumophila).

The performance of the PCR-probe assay was investigated using a pool of negative clinical material spiked with 2.5 μl of positive control containing 2.5 or 1 L. pneumophila cell. As sensitivity and hybridization controls, 1-, 5, and 10-fg quantities of Legionella bozemanii, and Legionella tucsonensis chromosomal DNA were included. An inhibition control was included in each reaction. A negative control was included in each run. The PCR reaction mixture contained 10 pmol (each) of primers LEG1 and LEG2, 200 μM (each) deoxynucleoside triphosphate, 3 mM MgCl₂, and 0.625 U of AmpliTaq Gold polymerase in 1× PCR buffer (Perkin-Elmer), in a final reaction volume of 25 μl. Samples were preheated for 10 min at 95°C, followed by 40 cycles of 30 s at 94°C, 30 s at 60°C, and 30 s at 72°C, with a final extension of 5 min at 72°C. Hybridization of the PCR product was carried out by spotting 3 μl of the sample on Hybond-N membrane (Amersham, Little Chalfont, United Kingdom). and hybridizing with Legionella species- and L. pneumophila-specific 5′ biotinylated probes at 61°C (Fig. 1). PCR products were detected only when Legionella species were used as template and not with sequence-related organisms such as Pseudomonas aeruginosa, Lactobacillus casei, and Acinetobacter baumannii. The detection limit was at least 1 bacterial cell, 0.2 CFU, or 5 fg of chromosomal DNA per reaction.

FIG. 1.

Performance of PCR-probe assay in detection of Legionella species other than L. pneumophila. PCR was performed on the equivalent of 100 cells of two different clinical isolates of L. pneumophila (lanes 1 and 2), L. tucsoniensis (lane 3), L. birminghamiensis (lane 4), Legionella haekeliae (lane 5), Legionella brunensis (lane 6), Legionella jordanis (lane 7), Legionella jamestowniensis (lane 8), L. bozemanii (lane 9), Acinetobacter baumanii (lane 10), L. casei (lane 11), and P. aeruginosa (lane 12). Lane M, marker DNA (sizes given at right). The PCR products were hybridized to probes LSPP-B and LPN-B at 61°C, as shown at the left.

The clinical specificity of the PCR-probe assay was assessed using proteinase K-treated and boiled bronchoalveolar fluid or sputum samples from 100 hospitalized patients with pneumonia. Pneumonia was due to unknown (52%) or well-defined (48%) causes. Among the known causes was a diagnosis of infection with Aspergillus fumigatus, Bordetella pertussis, Chlamydia psittaci, Chlamydia pneumoniae, Haemophilus influenzae, Klebsiella oxytoca, Klebsiella pneumoniae, L. pneumophila, Moraxella catarrhalis, Mycobacterium tuberculosis, Mycoplasma pneumoniae, Pseudomonas aeruginosa, P. fluorescens, Serratia marcescens, Streptococcus pneumoniae, cytomegalovirus, of respiratory syncytial virus. Five patients were positive in the PCR-probe assay for Legionella. In three cases the PCR products also hybridized to the L. pneumophila-specific probe LPN-B. Those patients all had a positive culture for Legionella. The other 2 of 100 samples were positive for Legionella species other than pneumophila. The PCR products were sequenced, and after alignment to Genbank database sequences, these samples were identified as Legionella species with 98% homology to Legionella waltersii and 97% homologies to Legionella birminghamensis and Legionella maceachernii, respectively. No alternative etiologic agent was found in patients diagnosed with infection by Legionella species other than L. pneumophila. A collaborate prospective study is intended to present a future validation of the PCR-probe assay and to investigate the role of Legionella species other than L. pneumophila in causing pneumonia.