May-Ling Yap, Zamberi Sekawi, Hui-Yee Chee, Han Kiat Alan Ong, Vasantha Kumari Neela

1Department of Medical Microbiology and Parasitology, Faculty of Medicine and Health Sciences, Universiti Putra Malaysia, 43400 Serdang, Selangor,Malaysia

2Department of Pre-Clinical Sciences, Faculty of Medicine and Health Sciences, Universiti Tunku Abdul Rahman, Bandar Sungai Long, 43000 Kajang,Selangor, Malaysia

ABSTRACT Objective: To compare the efficiency of routine diagnostic PCR assays in detecting pathogenic Leptospira isolated from water and soils.Methods: Seven routine assays targeting six genes (lipL32, flaB, gyrB, lfb1, secY and ligB)were evaluated and compared on the cultures of two groups of pathogenic Leptospira from different sources. One group included 19 described reference strains recovered from infected human or animals, and another group included 22 environmental isolates from recreational and residential sites in Malaysia. The latter have been confirmed for presence of pathogenic Leptospira DNA. PCR positivity or detection sensitivity of each assay was determined and compared between the two groups.Results: Validation on reference strains showed 100.0% PCR sensitivity for all assays except ligB-PCR (95.0%) that failed to amplify Leptospira interrogans serovar Pomona. In marked contrast, there was a notable decline in sensitivity in the environmental isolates (lipL32-PCR,95.5%； flaB-PCR, 90.9%； gyrB-PCR, 77.3%； lfb1-PCR, 59.1%； secY-PCRs, 40.9% G1/G2-PCR, 36.4%； ligB-PCR, 13.6%), implying a large genetic distance between the two groups, as well as nucleotide polymorphism among environmental isolates.Conclusions: High proportion of false-negative PCR results suggests a need of prudent selection of primers in detecting environmental pathogenic Leptospira. These findings offer valuable insights on the extensive biodiversity of genus Leptospira and its impact on the efficacy and development of molecular detection tool.

Keywords:Leptospira Pathogenic species Environmental samples PCR Sensitivity

1. Introduction

Leptospirosis is a globally widespread zoonotic infectious disease of public health concern[1] with high annual morbidity in tropical low-income countries[2]. The aetiology is highly motile spirochaete bacteria of genus Leptospira. Rats, rodents and domestic animals such as pigs, cattle, and dogs are common reservoirs that maintain leptospires in renal tubules and chronically excrete the bacteria in urine into environment. Humans may get infected by contact of abraded skin or mucosa membrane with the urine directly, or most often, via exposure to contaminated surface water or soils[3].

Although most human leptospirosis are mild and self-resolving, the disease may causes serious damage to multiple organs including kidney, liver, lung, and sometimes fatal in severe cases.

The members of genus Leptospira are diverse. More than 300 serovars have been described based on antigenic relatedness as defined by the structural heterogeneity of lipopolysaccharide[4].With the emergence of sequence-based typing method, to date Leptospira strains have been classified into 35 species according to whole-genome DNA homology[4,5]. The genus is also divided into three clusters based on pathogenicity and phylogeny: the parasitic pathogens, harmless free-living saprophytes, and intermediates or opportunistic pathogens.

A proper validation of analytical specificity and sensitivity of a laboratory-developed molecular diagnostics is crucial prior to its use in specimen screening practically. Numerous PCR assays targeting a number of leptospiral genes and sequences have been described for leptospirosis diagnosis and typing of Leptospira[6,7]. They may have been validated on a global collection of reference Leptospira strains or spiked clinical specimens, but PCR has yet to be developed to an extent where it is universally applied in environmental screening[8].An accurate detection is of great importance especially during outbreak investigation where a public health investigator would trace environmental point source of transmission[9]. Recent discovery of extensive Leptospira biodiversity in the soils[10] and the low sensitivity of two diagnostic quantitative PCRs targeting leptospiral lipL32 and lfb1 genes[5] reinforces the need for re-assessment of reliability of the current assays. In view of this, this study pursues to evaluate and compare the sensitivity of selected routine PCR assays in detecting pathogenic and environmental Leptospira isolates.

2. Materials and methods

2.1. PCR primers

Seven primer sets targeting six different leptospiral genes, which were described in previous studies[11-17], were used in this study.The target genes include the lipL32 gene coding for major outer membrane lipoprotein[11], flaB gene for flagellin[12], gyrB gene for DNA gyrase B[13], lfb1 gene for putative leptospiral fibronectinbinding protein[14], ligB for leptospiral immunoglobulin-like protein[15], and secY gene for protein translocase subunit[16,17].To differentiate the two secY-based PCRs, the assay described by Gravecamp and co-workers[16] was named thereafter as G1/G2-PCR.

2.2. Leptospira strains

Nineteen reference strains representing four pathogenic species of genus Leptospira (Table 1), obtained from WHO/FAO/OIE Collaborating Centre for Reference and Research on Leptospirosis(the Netherlands) were used to determine the analytical sensitivity of selected PCR primers. To determine the diagnostic sensitivity in environmental isolates, twenty-two water and soil samples collected from eight amenity forests and three wet markets located in six districts in Perak, Malaysia, were used (Table 2). The presence of pathogenic Leptospira in culture of these samples was confirmed by PCR targeting lipL32 gene using primers designed by Stoddard and his co-workers[18].

2.3. Collection and culture of environmental samples

Sample collection was carried out according to recommended protocols[19,20] with some modifications. Fifty millilitre of water sample was taken from waterfalls, rivers, streams, puddles, or drains by dipping a sterile tube about one foot below the water surface. Approximate 40 cm3of damp soil sample was taken within a 15 cm×15 cm area and 3 cm underneath the ground. Soil washing was prepared by suspending the soil sample in 10 mL sterile phosphate buffered saline and allowed to sediment for one hour.

A 5 mL aliquot of each water or soil washing sample was filtered through a nitrocellulose syringe filter (0.45 μm pore size, Pall Corp,USA). Two mL of filtrate was inoculated into 5 mL of Ellinghausen-McCullough-Johnson-Harris liquid medium (Difco?, USA)supplemented with 10% (v/v) Leptospira enrichment medium(Difco?, USA) and selective agent 5-fluorouracil at 200 μg/mL. The culture was then incubated at 28 ℃ for 4 weeks in dark condition.

2.4. Genomic DNA extraction and PCR amplification

Bacterial genomic DNA was extracted from 1.5 mL culture of both reference strains and environmental isolates using the EZ-10 Spin Column Bacterial Genomic DNA Mini-Preps kit (BioBasic, Canada)following the recommendations given for Gram-negative bacteria.The culture was firstly spun at 15 000×g for 10 min at 4 ℃ and the cell pellet was resuspended in 200 μL cold sterile Tris-EDTA buffer.The DNA extract was suspended in 50 μL of elution buffer and stored at -20 ℃.

Although the previous validation of selected PCR assays reported a flawless discrimination between the pathogenic and nonpathogenic species (intermediate and saprophytic), their analytical sensitivity was verified again using reference Leptospira strains before application in the environmental isolates. PCR amplification was performed in a total volume of 25 μL containing 12.5 μL of MyTaq? mix (Bioline, UK), 400 nM of each forward and reverse primers (Integrated DNA Technologies, Singapore), and 1 μL of DNA extract from culture of reference strains, or 2 μL from sample cultures.

All amplifications were performed in a thermal cycler (MyCycler?,Bio-Rad, USA) with a programme comprising one cycle of predenaturation at 95 ℃ for 3 min followed by 35 cycles of amplification at 95 ℃ for 15 s, 55 ℃ for 15 s and 72 ℃ for 10 s. The amplificationwas completed with a final extension at 72 ℃ for 10 min. A nontemplate control, in which DNA template was replaced by sterile distilled water, was included in each reaction. Reactions were repeated in duplicates. Five μL of amplicon was visualised using 1.8% (w/v) Tris-Acetate-EDTA (TAE)-agarose gel electrophoresis at 150 mA for 35 min. All gels were prepared with HealthView?Nucleic Acid Stain (Genomics, Taiwan) according to manufacturer's recommendation.

Table 1. Nineteen reference Leptospira strains used in this study.

Table 2. Source of pathogenic Leptospira isolates from environmental samples.

2.5. Sequencing

The selected PCRs were also used to amplify Leptospira (L.)interrogans serovar Icterohaemorrhagiae to confirm if target gene has been successfully amplified. Amplicons were purified and sequenced by a DNA sequencing service provider (1st BASE, Selangor,Malaysia) which uses Applied Biosystems genetic analyzer platform and BigDye? Terminator v3.1 cycle sequencing kit chemistry.Chromatograms were assembled and analysed in Sequence Scanner software version 2.0 (Applied Biosystems).

2.6. Determination of lowest limit of detection (LLOD)

The LLOD of each PCR primer set was also determined using serial dilutions of DNA of L. interrogans serovar Canicola ranging from 10 000 to 10 pg per reaction. The concentration of DNA template was firstly adjusted to 10 ng/μL, and then the 10-fold serial dilutions were prepared. One μL of each DNA dilution was then added into corresponding reaction. The lowest concentration of DNA that showed a visible band in PCR was taken as the LLOD.

3. Results

3.1. Amplicon and detection limit of PCR assays

Each assay has successfully yielded single amplicon corresponding to the expected size and a correct gene target as confirmed by sequencing. In determining the LLOD, the amplicon was detectable for all PCRs using the least DNA template (10 pg) containing an average of 2 000 genome equivalents per reaction [based on the genome size of L. interrogans strain Fiocruz L1-130 (4.6 Mb)； 1 genome=5 fg gDNA], except the G1/G2-PCR and ligB-PCR of which the yields were 10- and 100-fold lesser, respectively (Figure 1). Two μL of 50 mM MgCl2was added for ligB-PCR per reaction to increase the amplicon yields.

3.2. Validation of PCR assays on reference Leptospira strains

Results showed a 100% analytical sensitivity or true positive rates for all PCR assays but a slightly lower rate (95%) for ligBPCR (Table 3). Obtaining negative result consistently despite the increased amount of DNA template (up to 8.3 ng) has confirmed the unsuccessful amplification of L. interrogans serovar Pomona by ligBPCR in this present study.

3.3. Use of PCR assays on environmental Leptospira isolates

In contrast, the PCR amplification of environmental isolates revealed a range of lower and varied sensitivities (13.6% to 95.5%)(Table 3). lipL32-PCR showed the highest sensitivity (95.5%),followed by flaB-, gyrB- and lfb1-PCR, whereas secY-PCR, G1/G2-PCR and ligB-PCR detected less than 50.0% of samples.Comparison between the two distinct groups of pathogenic Leptospira showed a notably decrease in sensitivity (4.5% to 86.4%) and high percentage of false-negative PCR reactions in the environmental isolates (40.9%, 63/154) in relation to the host-associated strains(0.8%, 1/133) (Table 3). This finding thus raised a concern on the accuracy of current PCR assays in detecting environmental isolates.

Table 3. PCR sensitivity determined in reference strains and environmental isolates of pathogenic Leptospira species.

4. Discussion

Although the original studies[11-17] reported a faultless detection of all described pathogenic species of genus Leptospira, the reproducibility of the selected PCR primers was assessed. The 100% analytical sensitivity determined in the present study was well founded since similar reference strains were tested. However, a falsenegative reaction occurred unexpectedly where L. interrogans serovar Pomona could not be detected by ligB-PCR as opposed to the previous study[15]. The template DNA amount used in amplification(8 300 pg) should not be an issue with regards to the LLOD of ligBPCR determined in the present study (1 000 pg per reaction). The discrepancy raises a concern with the result inaccuracy concerning this assay to be used for diagnosis and detection of leptospires. As a result, there is a need to fully verify and improve the robustness and efficacy of ligB-PCR in detecting all pathogenic species.

Following the validation, similar primer sets were used to amplify pathogenic Leptospira of unknown type isolated from environmental water and soils. All PCRs are supposedly positive regardless of target gene given the presence of DNA of pathogenic Leptospira in these samples. Nonetheless, none of the primer sets could detect all environmental isolates, even the one targeting lipL32. The large total number of false-negative reactions is a matter of great concern,implying the shortcomings of current PCR assays particularly in detecting environmental isolates. Also, the PCR sensitivities determined in environmental isolates were much lower and widely varied to that in the host-sourced strains, suggesting genetic distance and divergence between the two groups.

Figure 1. Comparison of lowest detection limit among PCRs. Lanes 1-5: varied DNA template amount added per reaction from 20 000 pg (lane 1), 10 000 pg (lane 2), 1 000 pg (lane 3), 100 pg (lane 4) to 10 pg (lane 5)； M, MassRuler? DNA ladder mix； NT, non-template control.

Absence of amplification was likely attributed to mismatches between template and primer as a result of sequence variation.Impact of single-nucleotide polymorphisms on PCR detection of Leptospira was firstly reported by Bourhy and his co-workers[21]who failed to amplify soil-derived L. kmetyi using secY-PCR. The sequencing reveals a total of five mismatches, two in the forward primer and three in the reverse primer, among which one is located two bases from the 3' end of forward primer[21]. Having such a frequency and position of mismatches would block a reaction completely for lacking primer extension[22]. Overall, our findings suggest a large genomic plasticity and unexplored biodiversity in genus Leptospira and are in agreement with Thibeaux and his coworkers[5] who also reported low detection sensitivity of quantitative PCRs as a result of polymorphisms in lipL32 and lfb1 genes[10].

Although the environmental samples have been previously detected positive for pathogenic Leptospira, a question has been in dispute if other genetic markers, besides lipL32, are also conserved among these isolates. Another question worth exploring is about the sensitivity of different primer sets targeting similar gene. The PCR primer sets of study amplify a broad range of major leptospiral genes, including three pathogen-restricted genes (lipL32, ligB, lfb1)and three housekeeping genes (secY, flaB, gyrB), and are routinely used in reference laboratories globally and surveillance studies[7]. It is postulated that the PCR sensitivity may correlate with the degree of sequence variability of the target gene. The higher polymorphism,the more mismatches and hence blocked amplification[22]. In this study, the sensitivities of PCRs targeting secY and ligB were notably lower than that of lipL32, suggesting a greater sequence diversity in the former. Further evidence supporting this was the higher percentage of variable nucleotide sites in secY than in rrs,lipL32 and lipL41 during an assessment of gene loci for genotyping by multilocus sequence typing method[23]. However, such a high diversity may be a major obstacle in development of sequence-based molecular diagnostics by the difficulty in identifying conserved sequences, as exemplified by the lig genes[24].

It is also important to highlight that G1/G2- and ligB-PCR need more DNA template for a detectable yield of amplicon given their LLOD was one to two log higher than other PCRs. Also, the detection sensitivity is unlikely dependent with genetic marker for the LLOD determined for two PCRs targeting similar gene (secY)were not similar. In the light of these findings, it is interesting to further explore the impact of LLOD on efficacy of a primer set.

In this present study, lipL32- and flaB-PCR showed the highest sensitivity. Pathogen-restricted lipL32 gene was most often targeted in development of molecular diagnostics[25] and used in environmental surveillance[26-28]. Findings of other studies in a variety of sample types support the higher sensitivity of PCR targeting lipL32 to other genetic markers. For example, lipL32-PCR was found to be more sensitive than ompL1-PCR in human and bovine serum samples[29] and in buffalo kidney samples as compared to G1/G2-primed secY-PCR[30]. On the other hand, findings of some studies rather support better performance of G1/G2-PCR, such as in detecting L. borgpetersenii Hardjo-bovis in bovine urine samples in comparison to rrs-PCR[31], in spiked human blood and urine samples than LP1/LP2 primers[32], and in environmental sample than lipL32-PCR[33]. Due to lacking of well-ground and consistent demonstration of high detection sensitivity of a particular assay, it is difficult to make a valid conclusion on which genetic marker or PCR primers is most sensitive for accurate detection. The underlying cause for the variability in different circumstances is unclear but it may be dependent on localised Leptospira biodiversity in a particular region or source.

Most of the current diagnostic PCRs take the important conserved sequences in the genome of L. interrogans, or among a panel of Leptospira strains representing global source, as the model target in designing primers. Its analytical specificity and sensitivity are validated experimentally using known relevant strains, followed by the spiked specimens for diagnostic applicability. The primers may not complement to all strains considering the overlooked biodiversity of the genus Leptospira[5]. The high specificity of PCR technique would compromise the sensitivity, especially in the case of rare or atypical strains isolated from the environment[10].

The G1/G2-PCR has been widely used in Malaysia for environmental epidemiological studies[33-36] besides the lipL32-PCR[26-28]. Because of the comparatively low sensitivity of G1/G2-PCR as determined in this present study, it is of great concern if the occurrence of leptospiral DNA reported (0.7%-7.5%)[33-36] was underrated. Moreover, there is one shortcoming of G1/G2 primers for its incapability in detecting L.kirschneri[16]. To ensure accurate detection of all pathogenic species,another primer set (bim-based B64-I/B64-Ⅱ) that specifically detect L. kirschneri[16] should be included. Overall, our findings implies that a comprehensive validation of a laboratory-developed PCR assay is important before its use in field setting in view of the large biodiversity of genus Leptospira. An earlier environmental epidemiological study reported a rich genotype diversity of Leptospira in Malaysia by typing the pathogenic Leptospira isolated from recreational lakes and wet markets using pulse-field gel electrophoresis[35].

In conclusion, recent works as well as our findings bring the reliability into question of most diagnostic PCRs, particularly secY-and ligB-PCRs, in the context of environmental Leptospira. Our findings also suggest a large genetic polymorphism and biodiversity in environmental Leptospira, which may be further verified using whole-genome sequences and a comprehensive identification tool. There is also a growing need to continue identify conserved PCR target sequence to improve the relevance of sequence-based molecular diagnostics.

Conflict of interest statement

We declare that we have no conflict of interest.

Authors' contribution

Y.M.L. developed the concepts, designed the experimental study,acquired the data and prepared the manuscript. Y.M.L., C.H.Y. and O.H.K.A. performed the data analysis. All authors reviewed and edited the manuscript.