Melampsora medusae Thümen is one of the causal agents of poplar leaf rust, which is considered among the most damaging disease of poplars worldwide. This disease causes defoliation, growth reduction and in severe cases mortality, leading to important economic losses in commercial poplar cultivation. M. medusae is classified as a quarantine pest in Europe (Anonymous 2000) and its presence is thus strictly controlled. Two formae speciales have been described within M. medusae (Mm), M. medusae f. sp. deltoidae Shain (Mmd) and M. medusae f. sp. tremuloidae Shain (Mmt) on the basis of their telial host range: Mmd is pathogenic on Populus species of the sections Aigeiros and Tacamahaca, including cultivated poplars such as P. deltoides (Bartr.) Marsh. and P. x euramericana (Dode) Guinier and P. x interamericana Brockh., whereas Mmt infects Populus species of the section Populus, including wild poplars such as P. tremuloides Michx. A recent multilocus phylogenetic study showed that these two formae speciales should be considered as distinct species (Vialle et al. 2013), but the European phytosanitary regulations in force does not make the distinction between both formae speciales and the two taxa are still so far equally considered as quarantine fungi.

The first symptoms of infection by Melampsora species are small yellow pustules (uredinia, containing urediniospores), which appear within 2–3 weeks on the underside of the leaves, or on both sides in case of heavy infections. Different species of Melampsora can infect the same poplar leaf (Pinon and Frey 1997). Melampsora species have complex life cycles that generally require two unrelated host plants, in this case poplars (telial host) and conifers (aecial host), and exhibit five different spore stages. In order to comply with the current European phytosanitary regulations, EU members have to conduct surveys in order to detect any introduction of M. medusae and prevent its spread especially through the trade of infected poplar saplings produced by nurseries. During the surveys, the sampling procedure is restricted to the poplar trees, since the aecial stage of M. medusae was never reported on any conifer in the EPPO region (EPPO 2009). Infected leaves or uredinia (scraped from the leaves) are sent to the laboratory for analysis. To date, techniques available for the detection of M. medusae are based on either: (i) morphological identification (Pinon 1973; EPPO 2009; Vialle et al. 2011), or (ii) conventional PCR (Bourassa et al. 2005; Husson et al. 2013). On the one hand, morphological identification of M. medusae based on the microscopic observation of urediniospores and paraphyses is time consuming and, it may be difficult to visually identify a few urediniospores of M. medusae mixed in a large amount of urediniospores of other indigenous Melampsora species, resulting in a lack of sensitivity that may not be compatible with the nil tolerance considering quarantine organisms. On the other hand, the currently available methods based on conventional PCR (Bourassa et al. 2005; Husson et al. 2013) are sensitive and rapid but these techniques still lack specificity or are not fully inclusive (EPPO 2009). Real-time PCR using a hydrolysis probe has already been successfully used for the detection of numerous regulated plant pathogenic fungi, due to its enhanced specificity (Van Gent-Pelzer et al. 2007; Tan et al. 2010; Ioos et al. 2012). In the present study, a new real-time PCR assay was developed to detect both formae speciales of M. medusae, with improved specificity and sensitivity that overcome existing protocols.

Partial sequences of the internal transcribed spacer (ITS) rDNA region (23 sequences) and of the 28S rDNA gene (24 sequences) for all inventoried poplar Melampsora species (17 species considered as recognized taxa) with accurate identification were retrieved from GenBank or generated during a previous phylogenetic species recognition study (Vialle et al. 2013) and screened for regions showing interspecific polymorphisms. When relevant, sequences showing intraspecific polymorphisms were included in the study. For each marker, sequences were analyzed by multiple alignments using CLUSTALW (Thompson et al. 1994) available on the online PBiL platform (http://npsa-pbil.ibcp.fr/cgi-bin/npsa_automat.pl?page=/NPSA/npsa_clustalwan.html). The ITS showed regions with high levels of interspecific polymorphism and a series of forward and reverse primer and probe combinations specific to M. medusae (Mm-F/-R/-P), but conserved between both formae speciales, were designed using Primer 3 online software (http://frodo.wi.mit.edu/primer3/) (Rozen and Skaletsky 2000). Primers were designed and manually adjusted to amplify short PCR fragments (~ 150 bp) and meet the thermodynamic constraints for a primers/probe combination. Melting temperatures (Tm) and potential secondary structures were first evaluated in silico using Beacon DesignerTM software (Premier Biosoft). Following in vitro tests by SYBR-green real-time PCR, primers with the lowest tendency to form secondary structures and the best sensitivity were retained for the design of a M. medusae-specific probe. In parallel, regions of the 28S that are highly conserved in the genus Melampsora were selected and a set of forward and reverse primers and probe (Mel-F/-R/-P) was designed. The Mel-F/-R/-P combination, used in a separate real-time PCR assay, was used to check the quality of the DNA extracted from Melampsora urediniospores (e.g. inhibition effect, DNA losses during extraction). Primer and probe sequences and probe reporter/quencher dyes retained for the study are listed in Table 1. All primers and probes were custom synthesized by Eurogentec.

Table 1 List of primer pairs and dual labelled probes developed in this study
Full size table

For each sample, approximately 2 mg of uredinia were scrapped from rusted leaves using a sterile scalpel blade or a spatula and transferred into a 2-ml microtube. Total DNA was extracted using a commercial DNeasy® Plant Mini Kit (Qiagen). Four hundred μl of AP1 buffer (supplied with the kit), 4 μl RNase A (100 mg ml−1, supplied with the kit) and 1 mm glass beads (~8 mg) were added to the harvested urediniospores and the microtubes were shaken for 2 min at a frequency of 6.5 units using the FastPrep®-24 system (MP Biomedicals). The microtubes were then incubated for 20 min at 65 °C and the samples were processed according to the protocol supplied with the DNeasy® Plant Mini Kit. The DNA concentration was quantified using the Nanodrop 2000 Spectrophotometer (Thermo Scientific).

Real-time PCR reactions were performed on a Rotor-Gene 6500 (Corbett Research) set with an autogain optimization for each channel performed before the first fluorescence acquisition. No template controls (NTC) were systematically included in triplicate to check the absence of contamination in all reactions of real-time PCR. The Ct value for each reaction was determined using the Rotor-Gene software version 1.7.75, setting manually the threshold line at 0.02 for all the experiments. Due to inacceptable loss of sensitivity observed during preliminary experiments, Mm and Mel tests could not be combined in a duplex assay and were carried out in separate reactions. The Mm real-time PCR assay was carried out in a total volume of 20 μl using the qPCR core kit no ROX (Eurogentec) consisting of ultra pure water, 1× reaction buffer, 5 mM MgCl2, 4 × 0.2 mM dNTPs, 0.2 μM of the forward primer (Mm-F), 0.2 μM of the reverse primer (Mm-R), 0.05 μM of the probe (Mm-P), 0.5 U of Hot GoldStar Taq polymerase and 2 μl of template DNA (~ 1 ng). For the Mel real-time PCR assay, the mix contained 0.3 μM of the forward primer (Mel-F), 0.3 μM of the reverse primer (Mel-R), 0.1 μM of the probe (Mel-P) and all the other reagents at the same concentration than listed above for the Mm real-time PCR assay. Real-time PCR cycling conditions for both assays included an initial denaturation step at 95 °C for 10 min, followed by 40 cycles of denaturation at 95 °C for 15 s and annealing/elongation at 64 °C (for Mm) or 62 °C (for Mel) for 45 s.

The specificity of the Mm real-time PCR assay was verified on DNA extracts from a collection of 123 Melampsora samples collected on a wide range of host plants, from various geographic locations and collected at different years (Table 2). In total, 12 Melampsora species were represented. Melampsora samples mostly consisted in urediniospores, except a few consisting in dry specimens of infected leaf tissues, either from aecial or telial hosts, retrieved from herbaria. Melampsora samples were tested in triplicate. The 123 Melampsora samples showed acceptable Ct values for the Mel real-time PCR assay (mean Ct = 20.4, SD = 2.9) (Table 2), confirming that the DNA was amplifiable. Mean Ct values obtained with DNA from herbaria samples were similar to those yielded with DNA from urediniospores, suggesting that after normalization they contained similar quantities of Melampsora DNA. Herbaria samples could therefore be included in the specificity test. For the Mm real-time PCR assay, the DNA from all of the 41 M. medusae samples, including 39 Mmd and two Mmt samples, gave positive results (Table 2), regardless of the host, geographical origin and collection date, thus supporting the inclusivity of the assay. The Mm real-time PCR assay yielded negative results with DNA from all the other Melampsora species, except with the hybrid M. medusae-populina (3 isolates) and with M. occidentalis (3 isolates).

Table 2 Melampsora samples used in this study
Full size table

The sensitivity of the Mm real-time PCR assay was assessed with DNA extracts prepared from samples artificially spiked with Mmd urediniospores. Urediniospores of Mmd were handled individually under a stereomicroscope using a human eyelash fixed on a needle stick. Each sample finally contained 0, 1, 2, 5, 10, 20 or 50 urediniospores of Mmd mixed into 2 mg of urediniospores (ca 800 000 spores) of M. larici-populina (Mlp), which is the most commonly encountered rust species in France. Six replicates were prepared for each spore mixture. Total DNA was extracted from each sample and analyzed using the Mel and Mm real-time PCR assays described above. The samples showed expected low Ct values (mean Ct = 9.8, SD = 0.1) for the Mel real-time PCR assay (Table 3) due to the high concentration of target DNA (extracted from 2 mg of Mlp urediniospores) in the samples. The specific Mm real-time PCR assay was very sensitive; it was able to detect down to two Mmd urediniospores in a mixture of 2 mg of Mlp urediniospores in all replicates, and as little as a single Mmd urediniospore in four out of six replicates (Table 3). The DNA from samples containing one urediniospore of Mmd yielded mean Ct values of 29.2 ± 0.2, which can be considered as the cut-off value for detection in our experimental conditions (i.e. starting quantity of urediniospores ≤2 mg). In routine analysis, mean Ct values >30 should be cautiously interpreted and would require additional investigation.

Table 3 Mel and Mm detection using the real-time PCR assays optimized in this study
Full size table

This test detected all M. medusae samples tested, regardless of the forma specialis. This test did not cross-react with M. pinitorqua and M. rostrupii DNA, one of the drawbacks of the currently used conventional PCR method (Husson et al. 2013). Nevertheless, positive results were observed with M. medusae-populina and M. occidentalis DNA. On the one hand, this is not entirely surprising since M. medusae-populina is actually a natural interspecific hybrid between Mlp and Mmd and contains both types of ITS sequences in its genome (Frey et al. 2005). To confirm the hybrid status, cloning and sequencing the ITS region could demonstrate the presence of both ITS sequences. Alternatively, morphological urediniospore observations could be easily used to distinct the hybrid M. medusae-populina from Mmd.

On the other hand, despite three and two single nucleotide polymorphisms (SNP) in the forward and the reverse primers respectively, M. occidentalis DNA yielded positive results with the Mm real-time PCR assay, even with higher hybridization/polymerisation temperature (data not shown). Such results suggest: i) a lack of specificity of the present assay resulting in a cross amplification of M. occidentalis ITS sequence, or ii) the presence of ITS copies specific to Mmd within the DNA of the M. occidentalis specimens tested. Such ITS copies heterogeneity in M. occidentalis DNA sample is possible as this species is able to hybridize with Mmd in northwestern America (Newcombe et al. 2000, 2001), giving rise to the hybrid M. x Columbiana. The presence of hybrids involving M. medusae as a parent, i.e. M. medusae-populina and M. x columbiana, could therefore lead to false positive results. Nevertheless, M. medusae-populina has only been reported in New Zealand and South Africa so far (Spiers and Hopcroft 1994; Frey et al. 2005) and M. occidentalis and its hybrid M. x columbiana in western North America (Vialle et al. 2011). As for M. medusae, these species and hybrids are presently absent in Europe. Thus, positive results with the Mm real-time PCR assay would also give a chance to monitor any new introduction of M. occidentalis, M. x columbiana or M. medusae-populina in Europe. Morphological urediniospore observation could then be used to confirm the identity of the taxa detected.

In this study, a new real-time PCR test was developed in order to detect minute amounts of the European quarantine species M. medusae in a massive mixture of indigenous Melampsora urediniospores. This test offers improved specificity and sensitivity over currently existing tests and does not require specific taxonomic skills, which makes it a valid and useful detection tool for surveys in European nurseries. Due to its ability to target indistinctively both formae speciales of M. medusae, this tool can be used with samples collected from both cultivated and wild poplars.