Detection of pathogenic leptospires by real-time quantitative PCR

Leptospirosis is a zoonosis of worldwide distribution (World Health Organization, 1999), caused by infection with pathogenic spirochaetes of the genus Leptospira. The disease is maintained in nature by chronic renal infection of carrier mammals, which excrete the organism in their urine (Faine et al., 1999). Humans become infected through direct exposure to infected animals or their urine, or through indirect contact via contaminated water or soil (Levett, 2001). Leptospirosis has been recognized as an emerging infectious disease, in part because of recent large-scale outbreaks associated with recreational activities (Morgan et al., 2002; Sejvar et al., 2003). Investigation of large outbreaks would be greatly enhanced by the availability of rapid and sensitive diagnostic assays which can confirm the diagnosis early in the clinical illness.

Diagnosis of leptospirosis is usually accomplished retrospectively by serology, because culture requires both special media and incubation for several weeks (Levett, 2003). Serologic diagnosis by microscopic agglutination test invariably requires testing of acute and convalescent sera, since agglutinating antibodies often are not detectable during the acute illness. IgM antibodies become detectable 57 days after the onset of symptoms, and the use of IgM-ELISA assays for presumptive diagnosis has been evaluated in numerous populations (Adler et al., 1980; Bajani et al., 2003; Smits et al., 1999; Terpstra et al., 1985).

A number of PCR assays for leptospiral DNA have been described, but only two have been evaluated in clinical studies (Brown et al., 1995; Merien et al., 1995) and used extensively for diagnosis. Despite their wide use these assays have limitations: that described by Merien et al. (1995) is a genus-specific assay which amplifies both pathogenic and non-pathogenic serovars, while the approach described by Gravekamp et al. (1993) and evaluated by Brown et al. (1995) requires amplification of two distinct targets in order to detect all species containing pathogens. More recently, a real-time PCR was developed using TaqMan chemistry (Smythe et al., 2002) which targeted an 87 bp section of the 16S rRNA gene of Leptospira spp. In this study, we report the development of a real-time assay using SYBR green chemistry, in which the target is the gene for the major outer-membrane lipoprotein LipL32 (Haake et al., 2000), which appears to be an important virulence factor (Yang et al., 2002), confined to pathogenic strains of all Leptospira spp.

Bacterial strains and DNA extraction.
Leptospiral strains (Table 1) were maintained in semi-solid PLM-5 medium (Serologicals) containing 1.5 % agar (Difco) at room temperature. Subcultures in liquid PLM-5 medium were incubated at 30 °C for 7 days. Cultures of other clinically encountered bacterial species, including Acinetobacter calcoaceticus, Bacillus subtilis, Bacillus thuringiensis, Bifidobacterium longum, Bordetella bronchiseptica, Bordetella pertussis, Borrelia burgdorferi, Brucella melitensis, Burkholderia cepacia, Campylobacter jejuni (two strains), Corynebacterium diphtheriae, Corynebacterium xerosis, Enterobacter aerogenes, Enterococcus faecalis (four strains), Enterococcus faecium (three strains), Escherichia coli, Helicobacter pylori (two strains), Klebsiella pneumoniae, Lactobacillus plantarum, Listeria monocytogenes, Neisseria canis, Pasteurella multocida, Proteus mirabilis, Pseudomonas aeruginosa, Staphylococcus aureus, Streptococcus pneumoniae, Streptococcus pyogenes and Streptococcus sanguis, on blood agar plates were incubated aerobically at 37 °C for 18 h. DNA was extracted from bacterial cultures using QIAamp DNA Mini kits (Qiagen). DNA from Candida albicans (two strains), Candida dublinensis (two strains), Candida glabrata, Candida krusei, Candida parapsilosis, Candida tropicalis, Treponema denticola, Treponema pallidum Nichols strain, Treponema phagedenis and Treponema refringens was received in extract form. To determine the analytical sensitivity of the assay, Leptospira interrogans serovar Icterohaemorrhagiae strain RGA was cultured in 500 ml liquid PLM-5 medium at 30 °C for 7 days. DNA was extracted, purified and quantified as described previously (Brenner et al., 1982). The genome size of L. interrogans has been determined to be 4.659 Mb (Nascimento et al., 2004; Ren et al., 2003) and this value was used to calculate the number of genome equivalents in the assay.

Table 1. Leptospira and Leptonema strains studied

Effect of specimen matrix on amplification.
A 7-day-old culture of L. interrogans serovar Icterohaemorrhagiae strain RGA in PLM-5 was diluted 10-fold in sterile defibrinated horse blood (Lampire Biological Laboratories) and aliquots were added to sterile Vacutainer blood collection tubes (Becton Dickinson) containing anticoagulants or no coating, or to a paediatric Isolator tube (Wampole Laboratories). Aliquots (200 µl) were removed immediately for DNA extraction; further aliquots were removed after 2 and 5 days storage at ambient temperature.

A 7-day-old culture of L. interrogans serovar Icterohaemorrhagiae strain RGA in PLM-5 was diluted in 30 ml fresh human urine and added to a DNA/RNA Protect container (Sierra Diagnostics). Aliquots of urine were removed immediately for DNA extraction as described above; further aliquots were removed at intervals after storage at ambient temperature. DNA was extracted from blood and urine using QIAamp DNA Mini kits and resuspended in a total volume of 100 µl.

Diagnostic specimens and DNA extraction.
Serum (one) and urine (two) specimens from three patients were examined. DNA was extracted from 200 µl aliquots using QIAamp DNA Mini kits.

Primer design.
lipL32 sequences from the following serovars were aligned using ClustalW (DNASTAR): Leptospira borgpetersenii serovar Hardjo (GenBank accession no. AF181554), L. interrogans serovar Autumnalis (GenBank accession no. AF366366), L. interrogans serovar Copenhageni (GenBank accession no. AF245281), L. interrogans serovar Lai (GenBank accession no. LIU89708, L. interrogans serovar Pomona (GenBank accession no. AF181553), Leptospira kirschneri serovar Grippotyphosa (GenBank accession no. AF121192), Leptospira noguchii serovar Fortbragg (GenBank accession no. AF181556) and Leptospira santarosai serovar Tropica (GenBank accession no. AF181555). Primers were designed to conserved regions with no mixed bases using Oligo 6.0 (Molecular Biology Insights). After preliminary experiments, the modified primer set LipL32-270F (5'-CGCTGAAATGGGAGTTCG TATGATT-3') and LipL32-692R (5'-CCAACAGATGCAACGAAAG ATCCTTT-3') was selected, resulting in a 423 bp amplicon between positions 270 and 692 of the lipL32 coding region.

Real-time quantitative PCR.
Real-time quantitative PCR was performed on an MJ Opticon system (MJ Research), using SYBR green JumpStart Taq ReadyMix regents (Sigma) in a 20 µl volume containing 300 nM forward primer, 900 nM reverse primer, 25 µl ReadyMix reagent and 4 µl DNA template. The amplification protocol consisted of 20 s at 95 °C and 2 min at 94 °C, followed by 40 cycles of amplification (95 °C for 5 s, 67.2 °C for 15 s, 72 °C for 20 s), after which the reaction was stopped (96 °C for 2 min), cooled (20 °C for 1 min) and melted (7296 °C with plate readings every 0.4 °C). The entire program was completed in 73 min. All experiments were repeated at least twice for reproducibility.

Sensitivity and specificity of the real-time PCR assay
The concentration of DNA extracted from L. interrogans strain RGA was 550 µg ml¹, equivalent to approximately 1.13 x 10¹¹ genome equivalents ml¹. Serial dilutions of this extract were tested using the real-time assay, which was able to detect 3 genome equivalents per reaction (Fig. 1). The specificity of the assay was determined by using DNA extracted from 47 strains representing 38 different species of bacteria and fungi as the template. None of the non-leptospiral species gave positive results. The LipL32 target was amplified from pathogenic leptospires belonging to Leptospira alexanderi, L. borgpetersenii, L. interrogans, L. kirschneri, L. noguchii and L. santarosai, but was not amplified from any strains of the non-pathogenic species Leptospira biflexa, Leptospira meyeri and Leptospira parva, or from the intermediate species Leptospira inadai, Leptospira fainei and Leptonema illini (Table 1). The melting curve temperature for the leptospires tested ranged from 82.5 to 86 °C, depending on the species (Table 1).

(15K): Fig. 1. Dilution end point standard curve of log10 genome equivalents versus CT cycle number. The analytical sensitivity of this assay was approximately 3 genome equivalents per reaction.

Effect of specimen matrix on sensitivity
Leptospiral DNA was amplified successfully from the spiked blood samples added to Vacutainer tubes (Becton Dickinson) containing a range of chemical treatments. Tubes containing citrate or EDTA allowed the greatest amplification immediately after inoculation and after 2 and 5 days storage. Plain tubes and those containing ACD solution B or potassium oxalate/sodium fluoride also allowed amplification up to 5 days after inoculation with the spiked blood. In contrast, tubes containing lithium or sodium heparin, sodium polyanetholesulfonate or saponin were inhibitory to the PCR assay at all time points. Leptospiral DNA was amplified from spiked urine samples stored at ambient temperature up to 16 days after inoculation. The sensitivity of the assay when performed on urine extracts was approximately 10 genome equivalents per reaction.

Amplification from diagnostic specimens
Leptospiral DNA was amplified from two urine samples and one serum sample from patients with serologically confirmed leptospirosis (data not shown).

We developed and evaluated a highly sensitive and specific real-time PCR assay for the detection of pathogenic leptospires. The 423 bp target was amplified from pathogenic strains of Leptospira spp., but not from non-pathogenic species, and not from a wide range of other clinically significant bacteria and yeasts. The LipL32 primers can also be used in a conventional PCR assay (data not shown). The analytical sensitivity of the assay was 3 genome copies per reaction in blood and approximately 10 genome equivalents per reaction in urine, comparable to a real-time assay which uses a 16S rRNA gene target (Smythe et al., 2002).

Traditionally, leptospires are quantified using counting chambers viewed under dark-field microscopy (Ellinghausen et al., 1981). However, this is an imprecise method which yields results accurate only to within an order of magnitude. Since the genome size of L. interrogans is now known, we prepared a large quantity of genomic DNA and calculated the number of genome equivalents ml¹. This novel approach to leptospiral quantification allows more accurate determination of the analytical sensitivity of the assay.

The assay described above was developed for use as a rapid diagnostic test for use in public health investigations, such as outbreaks of undifferentiated febrile illness. Recent outbreaks of dengue and leptospirosis have highlighted the importance of a specific diagnosis, particularly as earlier diagnosis of leptospirosis facilitates prompt antibiotic treatment (Flannery et al., 2001; Levett et al., 2000; Sanders et al., 1999). In many outbreak investigations, specimens are shipped long distances under less than optimal conditions. Because the diagnosis of leptospirosis is often not considered initially, appropriate specimens may not be collected. Thus it was important to demonstrate the stability of leptospiral DNA added to a range of blood collection tubes and to urine. Vacutainer tubes containing citrate or EDTA gave optimal results up to 5 days after addition of blood containing viable leptospires; tubes containing heparin were inhibitory, as reported previously (Smythe et al., 2002), and also those containing sodium polyanetholesulfonate or saponin. Leptospiral DNA was detectable in human urine samples stored at ambient temperature for prolonged periods in DNA/RNA Protect containers. The assay was developed using SYBR green technology, and specificity derived from melting point analysis. The omission of a probe from this assay helped to limit the costs of testing large numbers of samples in an outbreak investigation.

The sequence detected in this study is located as a single copy on chromosome I of L. interrogans (BLAST search, 14 April 2004). LipL32 is a probable virulence factor and is restricted to pathogenic leptospires (Haake et al., 2000). This is important because assays described previously (Gravekamp et al., 1993; Kawabata et al., 2001; Kee et al., 1994; Merien et al., 1992) are genus-specific and detect all leptospiral serovars, both pathogenic and non-pathogenic. Although it is possible to develop species-specific assays (Gravekamp et al., 1993), several Leptospira species contain both pathogenic and non-pathogenic serovars (Levett, 2001). An assay which targets a gene encoding a surface lipoprotein which is expressed in vivo but not in vitro (Haake et al., 2000) has potentially greater specificity. The real-time PCR assay described will complement other diagnostic methods such as rapid IgM-detection assays (Bajani et al., 2003; Smits et al., 2000).

D. E. T. was the recipient of a Dr James A. Turner Emerging Infectious Diseases Fellowship from the Minority Health Professions Foundation. We thank the following colleagues who supplied cultures of bacteria or DNA from culture collections within the CDC: Jean Jordan, Caroline Mohr, Patti Fields, Lynne Shoemaker, Mary Brandt, Allan Pillay and Kathy Wilson.