Introduction

Ustilago maydis is the causative agent of common smut of corn, which is considered to be one of the major prevalent diseases of maize plants in China. The average annual incidence of corn smut is between 5% and 10%. In regions where the disease is prevalent, it may occur in up to 20% of plants1. The hemibasidiomycete U. maydis displays a complex life cycle. To achieve successful infection, two haploid sporidia that form dikaryotic hyphae must have different alleles at both the a and b loci2. The most obvious symptom of infection by U. maydis is the formation of tumor-like galls that are found most frequently on ears, tassels, or sometimes on other meristematic tissues3. The tumors provide the environment where the hyphae differentiate, culminating in the formation of masses of black teliospores. Teliospores are probably the major source of the dispersal of inoculum in the field. Under the appropriate conditions, teliospores can germinate, undergo meiosis, and produce more haploid progeny4,5. Host resistance is the only practical method of controlling common smut in areas where U. maydis is prevalent3. To the best of our knowledge, there are currently no cultivated varieties or lines of maize available that have been shown to be immune to U. maydis once plants have become infected3. Maize common smut is still a major threat to maize production and hence a method of early, efficient, and specific detection of U. maydis could help improve the efficiency of management and resistance breeding with respect to this disease.

Many of the molecular assays currently employed for detecting U. maydis are based on the conventional PCR assay. Primers for PCR have been designed based on the insert terminal regions of a U. maydis plasmid library or the specific region downstream of the homeodomain of the bE gene at the b locus, respectively6,7. For the detection of another Ustilago species, U. scitaminea (also referred to as S. scitamineum), PCR-based methods were reported in which primers were designed based either on the bE mating-type gene sequence of U. maydis 8 or on the internal transcribed spacer (ITS) sequence of S. scitamineum 9. PCR assays using ITS sequence have also been reported for detecting the pathogen of barley smut, U. hordei 10. A Taq-man real-time PCR-based assay was developed for the early detection and identification of S. scitamineum using primers specific to the bE gene and a Taqman probe11. These methods may be specific for detecting U. maydis and other pathogenic fungi which share a similar life cycle or induce symptoms similar to smut disease. However, they fail to provide a rapid and quantitative tool capable of diagnosing the pathogen in crop and soil samples directly. Thus, a rapid, sensitive, simple, and economical detection method for practical applications would be preferred. Recently, two research groups successfully developed LAMP (loop-mediated isothermal amplification)-based assays for the detection of S. scitamineum in which internal transcribed spacer sequences and the genomic sequence of the pep1 gene were selected as targets for LAMP primer design, respectively12,13. These two LAMP protocols provided a specific, sensitive, and rapid test for the detection of S. scitamineum infection in sugarcane.

In addition to being an important agricultural pathogen of global significance, U. maydis has served as an excellent model for the study of the molecular interactions between biotrophic basidiomycete fungal plant pathogens and their host plants14,15. U. maydis deploys many effector proteins during establishment of infection and development of symptoms in maize plants16. Pep1 (protein essential for penetration1) is involved in the successful penetration of fungal hyphae into plant tissue and the establishment of initial compatibility by targeting and inhibiting the activity of the plant peroxidase POX1217,18. Another effector, Pit2 (protein involved in tumor2), is essential for the maintenance of biotrophy and the induction of tumors It has also been shown to be involved in the suppression of host defense response by its inhibition of maize apoplastic cysteine proteases19,20. The recently identified effector See1 (seedling efficient effector1) directly and specifically affects the formation of leaf tumors in maize by interfering with the MAPK-triggered phosphorylation of a maize homolog of SGT1 (suppressor of allele of skp1)16.

The LAMP assay is a novel nucleic acid amplification method that is performed under isothermal conditions utilizing Bst DNA polymerase, which has strand-displacement activity. A set of four or six custom-designed primers, which recognize at least six distinct sequences on the target DNA, are used to generate amplification products that contain single-stranded loops thereby allowing primers to bind to these sequences without the need for repeated cycles of thermal denaturation21,22,23. The detection of LAMP products by way of a fluorescent dye can be quantified using a real-time PCR thermal cycler or a portable fluorescent reader called ESE-Quant Tube Scanner, which is equipped with a battery pack. The latter is advantageous in that it offers a convenient method for rapid onsite detection24,25,26,27. Real-time fluorescence LAMP (RealAmp) is currently employed for quantitative assays of soil-borne pathogens such as Fusarium oxysporum f. sp. cubense and Pythium inflatum 26,27.

The objectives of this study were to develop a U. maydis-specific RealAmp assay and to evaluate this assay for the rapid and direct quantitative detection of U. maydis in naturally infested soil samples or maize plants prior to planting or before symptoms of smut disease appear. Genomic sequences of U. maydis effector genes Pep1, Pit2 and See1 were used to design candidate primers for screening to determine the optimal primer set. A single-factor experiment combined with an orthogonal experimental design was conducted in order to optimize three important reaction components (quantity of Bst DNA polymerase, primer ratio, and Mg2+ concentration) of the LAMP reaction system. The specificity and sensitivity of the method were evaluated and the feasibility of the LAMP-based quantitative detection assay was corroborated by testing with artificially and naturally infested samples. The optimized RealAmp method that was established from this study might lay a foundation for research on U. maydis-maize plant molecular interactions and on common smut disease management and resistance breeding.

Results

LAMP primer screening

In order to screen for an optimal target gene for use in the LAMP assay, a total of 12 sets of LAMP primers were designed that targeted three U. maydis effector genes, which included UmPep1, UmPit2 and UmSee1. The DNA of another maize pathogen, S. reiliana was used as negative control template.

For the LAMP primer sets Pep-1, Pep-3, Pit-2, Pit-3, See-2, See-3 and See-4, pseudo-positive amplification was observed (i.e. isothermal amplification occurred in both reaction mixtures containing either SG200 or S. reiliana genomic DNA), indicating that they are not specific to U. maydis DNA (Fig. 1). For the primer sets Pep-4 and See-1, the LAMP amplification curve indicated that the corresponding target segments were amplified only from the SG200 DNA (Fig. 1). However, melt peak data from the same reaction mixtures indicated that some amplification was occurring in the negative control DNA samples (Supplementary Fig. S1). Similar melt peaks were also observed in the negative control LAMP reactions for primer sets Pep-1, Pep-3, Pit-2, Pit-3, See-2, See-3 and See-4 (Supplementary Fig. S1).

Figure 1
figure 1

Screening of Ustilago maydis LAMP primers. The LAMP fluorescence signal (RFU) vs. Cq [threshold time (min)] amplification curve graph was plotted automatically by the Bio-Rad CFX96 Real-Time PCR system. 12 sets of primers that targeted one of three U. maydis effector genes (See1, Pit2 and Pep1) were designed for the primers optimization experiment. 1–4 indicates one of the four sets of primers designed for each gene; NC indicates the reaction using negative control templates (the DNA of another maize pathogens, Sporisorium reiliana, which causes symptoms similar to corn common smut).

Full size image

Based on the above data, we speculated that Pit-1, Pit-4 and Pep-2 might be suitable for the LAMP assay specific to U. maydis. Further analysis indicated that the Pep-2 primer set is the most optimal of these three candidates for use in the LAMP assay as it yields a lower Cq value (reaction threshold time, 10.37 min), and has better repeatability (Fig. 1). Therefore, Pep-2 was selected for the further optimization of the LAMP assay of U. maydis in this study. The sequences of the final set of primers used in the LAMP protocol and their respective target sites are indicated in Supplementary Fig. S2.

Optimization of LAMP reaction

The screening of candidate LAMP primers outlined above was based on our previous work27. In order to test whether the LAMP reaction system for U. maydis could be improved, modifications to three parameters of LAMP reaction mixture (quantity of Bst DNA polymerase, the ratio of the concentration of inner primer pair FIB/BIP to outer primer pair F3/B3, and concentration of Mg2+) were screened (Fig. 2). In evaluating the effect of quantity of Bst enzyme, it was shown that the Cq value of the amplification decreased as the amount of Bst enzyme was increased in the reaction mixture while keeping the other parameters constant (concentrations of Mg2+, FIP/BIP, and F3/B3 were fixed to 8 mM, 1.6 μM and 0.2 μM, respectively). The reaction time differed by 7 min between samples containing either 2 U or 8 U Bst DNA polymerase (Fig. 2a). In testing the concentration ratio of inner primers to outer primers, non-specific amplification was observed when the ratio was set at 4:1 and 6:1. The reaction reached its highest amplification efficiency when the ratio was adjusted to 8:1 (Fig. 2b). For the optimization of Mg2+, the average threshold time at a concentration of 5 mM was 16.3 min. The amplification efficiency exhibited a sharp increase when the Mg2+ was adjusted to 6 mM or 7 mM and remained high up to 8 mM MgSO4 (Fig. 2c).

Figure 2
figure 2

Single factor experiment of RealAmp. The influence of each variable on the LAMP reaction was analyzed by the amplification curve (a) and the melt peak (b). For the Bst DNA polymerase optimization, Bst polymerase quantities were adjusted to 2 U, 4 U, 6 U and 8 U, respectively. Ratios of inner to outer primers were set at 2:1, 4:1, 6:1 and 8:1 with the outer primer concentration fixed to 0.2 μM to optimize the primer ratios. Mg2+ concentrations in the LAMP reactions were varied from 5 mM, to 8 mM for the optimization of Mg2+.

Full size image

Based on the results of the single-parameter experiments, further LAMP analysis was performed with 16 combinations using an orthogonal experimental design (Table 1). Amongst the 16 combinations, the negative control reactions for combinations 8 and 12 showed false positive amplification, combinations 1, 9, and 10 resulted in poor reproducibility according to the amplification curves or melt peaks (Fig. 3), and combinations 2, 3, 4, 5, 7, 13, and 14 yielded sub-optimal results in terms of reaction threshold time ( > 10 min) (Fig. 3). The threshold times of combinations 6, 11, 15 were between 9 and10 min, whereas that of combination 16 was approximately 8 min; plus, this reaction combination exhibited good repeatability (Fig. 3). Taking into consideration the previously performed single parameter experiments along with the orthogonal experimental setup, combination 16 was considered to represent the optimal conditions for a LAMP system used to detect U. maydis. This optimized reaction mixture (total, 25 μl) contained 12.5 μl 2 × Bst DNA polymerase (NEB, Ipswich, MA, USA) reaction buffer [1 × containing 1.6 mM dNTPs, 1 M betaine, 4 mM MgSO4, 20 mM Tris-HCl (pH 8.8), 10 mM KCl, 10 mM (NH4)2SO4, and 0.1% Triton X-20 (Sigma-Aldrich Inc., Saint Louis, USA), Double Helix Tech. Inc., Guangzhou, China], 0.2 μM SYTO-9 fluorescent dye (Invitrogen, Carlsbad, CA, USA), 2 μl of the template DNA, 4 mM MgSO4, 1.6 μM each of FIP and BIP, 0.2 μM each of F3 and B3, 0.8 μM each of loop primers LF3 and LB3, and 8 U of Bst DNA polymerase (NEB).

Table 1 [L16(45)] orthogonal design for Ustilago maydis LAMP system
Full size table
Figure 3
figure 3

Reaction results based on different combinations of reaction components according to an orthogonal test. (a) Reactions of 16 combinations monitored by the amplification curves. (b) Corresponding melt peaks of the above reactions. 1–16 represent each combination performed in the orthogonal test shown in Table 1. NC indicates reactions that share the same conditions with each combination except for the substitution of a negative control template (genomic DNA of Sporisorium reiliana).

Full size image

Specificity test of LAMP reaction system

To investigate the specificity of LAMP primer set Pep-2 and the optimized LAMP reaction for U. maydis, six other pathogens that infect maize plants and one peanut pathogen Colletotrichum truncatum (which has an overwintering pattern similar to U. maydis) were tested using this same system.

The reaction curves indicated that the Pep-2 primer set was able to specifically amplify the target DNA sequence of U. maydis strains SG200 and A1, but not the DNA of the other plant pathogens that were evaluated (Fig. 4a; Supplementary Fig. S3). By visual assessment, the color of the LAMP reaction products changed from orange to green when U. maydis and the positive control were detected with SYBR Green I, whereas the color remained orange for the negative control, for the other maize pathogens, and for the peanut pathogen (Fig. 4b).

Figure 4
figure 4

Specificity test of RealAmp assay for the detection of Ustilago maydis. (a) The RealAmp fluorescence (RFU) vs. Cq [threshold time (min)] amplification curve graph was plotted automatically by the Bio-Rad CFX96 Real-Time PCR system. (b) Visual inspection of the RealAmp amplification products using SYBR Green I. The original orange color of SYBR green turned to green in a reaction mixture showing a positive result. The templates consisted of genomic DNA isolated from the following controls and other pathogens: PC, positive control (Staphylococcus aureus); NC-1, negative control 1 (Salmonella choleraesusi); NC-2, negative control 2 (S. bongori); SG200, Ustilago maydis strain SG-200; 1, Curvularia lunata; 2, Bipolaris maydis; 3, Colletotrichum truncatum; 4, Sporisorium reiliana; 5, Fusarium moniliforme; 6, F. graminearum; 7, Pythium inflatum.

Full size image

To evaluate the stability of our optimized U. maydis LAMP reaction system, a 100 × diluted template DNA was also evaluated in parallel with the specificity test experiment. The data showed that amplification curves of LAMP products were similar to those of using the undiluted templates (Supplementary Fig. S3).

Sensitivity of LAMP

For the sensitivity tests, 10-fold serial dilutions of the pEasy-Pep plasmid DNA mixed with either DNA isolated from soil or maize genomic DNA were used as a reference. The preliminary experiment showed that amplification of LAMP products with exponential growth occurred within the range from 0.44 ng/μl to 4.4 × 10−5 ng/μl of plasmid DNA (Supplementary Fig. S4). To further determine the detection limit of U. maydis using RealAmp, an experiment was conducted using serial dilutions of the pEasy-Pep plasmid ranging from 0.44 ng/μl to 4.4 × 10−8 ng/μl. Amplification results indicated that the RealAmp assay could detect concentrations as low as approximately 44 fg/μl of plasmid DNA when mixed with total DNA extracted from soil or maize plants (Fig. 5a), which is in agreement with the preliminary experiment. Consequently, a standard curve between the Cq-value and the logarithm concentrations of template plasmid DNA was constructed (Fig. 5a).

Figure 5
figure 5

Sensitivity comparison of Ustilago maydis RealAmp and real-time PCR using the SG200 strain plasmid DNA as template. (a) Sensitivity of the RealAmp assay conducted on the Bio-Rad CFX96 Real-Time PCR system. The RealAmp fluorescence (RFU) vs. Cq [threshold time (min)] amplification curve graph and standard curve in which threshold time was plotted against the logarithmic amount of initial template DNA were plotted automatically by the CFX96 Real-Time PCR system. (b) Sensitivity test of real-time PCR for detecting U. maydis. The real-time fluorescence units were plotted against the initial concentration of plasmid DNA ranging from 4.4 × 10−1 ng/μl to 4.4 × 10−7 ng/μl by the CFX96 Real-Time PCR system. The standard curve was generated using known Log10-transformed concentrations of a 10-fold serial dilution of the pEasy-Pep plasmid DNA and the corresponding threshold cycle (Ct) values. (c) RealAmp sensitivity test using ZYD-S1 instrument. The fluorescence vs. time amplification curves of RealAmp were plotted automatically by the ZYD-S1 Tube Scanner. The standard curve was calculated by ZYD-S1 in which threshold time (Tt) was plotted against the logarithmic amount of initial template plasmid DNA.

Full size image

To compare the sensitivity between the RealAmp assay and conventional real-time PCR, serially diluted template DNA samples were used to conduct a real-time PCR experiment. This experiment detected amplification of PCR products and, in light of the regression analysis, plotting Cq against initial template concentrations showed that the resulting standard curve was linear over a concentration range of at least six orders of magnitude, from 10−3 to 10−8. Thus, the detection limit of the real-time PCR was approximately 4.4 fg/μl plasmid DNA (Fig. 5b). The sensitivity of the RealAmp and real-time PCR assays for detecting U. maydis were further verified by a detection limit experiment, in which reactions using 10−7/10−8 and 10−8/10−9 concentrations of the original plasmid template were repeated 30 times and 20 times, respectively (Supplementary Fig. S5).

The sensitivity of RealAmp was also tested using the ZYD-S1TM detection system in this study. Curves representing the standard fluorescence amplification of LAMP products with exponential growth were obtained when employing pEasy-Pep template DNA ranging from 440 ng/μl to 44 fg/μl (Fig. 5c). These results indicated that the detection limit of the U. maydis RealAmp system is approximately equal to that of LAMP using the BioRad Real-Time system, thereby confirming that amplification was reliable and the RealAmp assay employing the ZYD-S1TM instrument could be used for pathogen DNA quantification through traditional standard curves.

RealAmp detection of artificially inoculated samples

Artificially infected soil and maize plant samples were prepared, and then DNA was extracted as described in the methods section. No U. maydis DNA was detected either in the uninoculated control soil samples with the RealAmp assay or with real-time PCR. The RealAmp assay could sensitively detect U. maydis contamination as low as 1.6 × 103 teliospores/g of soil in artificially infested soil samples (Fig. 6a).

Figure 6
figure 6

Quantitative detection of artificially inoculated soil and maize plant samples using RealAmp and real-time PCR. (a) Detection of Ustilago maydis from serial 10-fold dilutions ranging from 1.6 × 107 to 1.6 × 102 spores per gram of artificially infested soil using RealAmp. The quantitative detection of U. maydis in maize plants challenged with sporidia of U. maydis strain SG200 or compatible mating strains at 1 DPI (day post inoculation) and 6 DPI using RealAmp (b) and real-time PCR (c). Serially diluted template DNA (ranging from 100 to 10−3) of 1 DPI were used to test the sensitivity of RealAmp in detecting U. maydis at the early stage of infection in the artificially inoculated maize plants. Approximately 354 ng/μl genomic DNA of U. maydis strain SG200 was employed as a quantitative reference.

Full size image

For the artificially inoculated maize samples, genomic DNA was extracted from maize plants that were injected with 108 sporidia of U. maydis strain SG200 (or compatible mating strains) at 1 day and 6 days post inoculation (DPI), before the appearance of tumors. Approximately 354 ng/μl of SG200 genomic DNA was used as a quantitative reference. The data obtained indicated that the RealAmp can detect concentrations as low as 4.36 × 10−2 ng/μl U. maydis DNA at 1 DPI, and 34 ng/μl U. maydis was detected at 6 DPI by both RealAmp (Fig. 6b) and real-time PCR (Fig. 6c), respectively. The successful sensitivity test of the RealAmp assay in artificially inoculated samples demonstrated that it might be applicable for the detection of U. maydis in field samples either prior to or following infection.

RealAmp detection in field samples

The feasibility of RealAmp for detecting artificially infested samples led us to attempt to quantitatively assay U. maydis in field conditions using this method.

Initially, samples were collected from eight separate maize inbred lines at the big trumpet and tasseling stages, prior the appearance of tumor. These individuals were planted in a field infected with common smut for purposes of disease tolerance testing. The data from this experiment indicated that U. maydis could be detected in samples isolated from maize inbred lines A801, B73, B104 and 87-1, while the other four lines were indicated to be negative for the pathogen (Fig. 7). In addition, 150 soil and 72 maize plants samples were tested for the presence of U. maydis by RealAmp using the ZYD-S1 system. Results from the RealAmp assay showed that 190 of these samples tested positive, whereas 32 samples tested negative (Supplementary Table S2).

Figure 7
figure 7

Detection of Ustilago maydis in plants from various maize inbred lines using RealAmp for disease tolerance testing. (a) The amplification curve graph reporting the fluorescence in mV on the y-axis over time in min on the x-axis was plotted by the ZYD-S1 tube scanner. (b) Comparison of the RealAmp quantitative results for samples from different maize inbred lines between using the Bio-Rad CFX96 Real-time PCR system and the ZYD-S1 instrument for fluorescence detection. Data are presented as the mean ( ± SD) of three replicates. The SPSS (version 20, SPSS Inc., Chicago, IL, USA) independent t-test software package was employed to compare the RealAmp quantitative results of different samples between using the Bio-Rad and ZYD-S1 instruments.

Full size image

Discussion

Common smut of corn caused by U. maydis is considered to be a troublesome disease throughout most of the regions where maize is produced. Sensitive and effective detection of U. maydis prior to or at an early stage of pathogen infection is indispensable for disease management and resistance breeding practices. As a diagnostic tool, microscopy has been used to analyze the infection events of U. maydis in maize plants28,29, However, this technique suffers from low sensitivity and is technically challenging when asymptomatic plants or individuals at an early stage of infection are analyzed. PCR-based assays provided a sensitive and reliable tool for the detection of U. maydis as previously described6,7, but this method can be time-consuming and is not convenient for use in field conditions.

In this work, we described the development of a rapid, sensitive, and quantitative RealAmp method for U. maydis detection using species-specific primers. To screen for optimal LAMP primers, three U. maydis effector genes (Pep1, Pit2 and See1) were employed as targets for primer design. These virulence genes are involved in different events that occur during the U. maydis infection cycle16,17,18,19,20. Thus, the potential for successful infection of U. maydis on maize plants could be substantiated as the occurrence of the targeted gene was detected. Among the 12 sets of candidate LAMP primers, Pep-2, which presented the highest amplification efficiency and stability out of these sets, was selected for further development of our method (Fig. 1, Supplementary Fig. S1).

The Pep1 protein from U. maydis shows sequence identity ranging from 45.76% to 62.36% with its orthologs in six other smut fungi30. In order to enhance the specificity of the LAMP reaction, the variable region of this gene was targeted when designing candidate primers. In our study, the six LAMP primers of the Pep-2 set recognized the different variable regions of UmPep1 (Supplementary Fig. S2), which helps to sufficiently distinguish the sequence in U. maydis from those of other maize pathogens such as S. reilianum, the cause of symptoms of head smut in maize plants (Fig. 4). Previous research indicated that there exist two clades of U. maydis based on the sequence of Pep1 gene, most of the U. maydis belong to clade A, the Pep1 sequences of different strains or isolates in the same clade is conserved31. The coding sequence of Pep1 from clade B is identical to that of clade A except a 15 bp insertion at the position of nt 432 (Supplementary Fig. S7). The LAMP primer set Pep-2 used in this study were designed based on the Pep1 coding sequence from clade A (XM_011389599), the primer B1c of Pep-2 that targeting nt 422 to 439 of XM_011389599 was designed to separate the strains of clade A from the strains of clade B (Supplementary Figs S7 and S8). To detect the strains from clade B, seven sets of LAMP primers (mPep-1 to mPep-7) were also designed (Supplementary Table S1), in which primer set mPep-2 was thought to be capable of amplifying Pep1 segment specifically (Supplementary Figs S8 and S9). In a recent study, the Pep1 gene was also selected as an amplification target for detection of S. scitamineum using LAMP in sugarcane13. S. scitamineum Pep1 shares 63.7% and 73.33% nucleotide sequence identity with its orthologs in U. maydis and S. reilianum, respectively. Varying, non-conserved regions of the sequence were chosen for primer design in both studies; we selected more variable sequences in order to avoid the conserved central region13, (Supplementary Fig. S2). The use of loop primers further enhanced the specificity of the LAMP reaction as they targeted two separate variable regions of the target sequences (Supplementary Fig. S2). From what has been discussed above, we may speculate that the Pep-2 primer set is species-specific to U. maydis and to the exclusion of S. scitamineum and other smut pathogen as the specificity of the RealAmp primers was experimentally tested using several other pathogens.

Different parameters of the LAMP protocol could be adjusted in order to optimize the reaction. The ratio between inner and outer primers and the quantity of Bst DNA polymerase used in reaction are two critical variables for optimization21. In this work, we showed that as the concentration of Bst DNA polymerase was increased from 4 U to 8 U, the amplification improved dramatically, however, the tendency of amplification efficiency improvement from 6 U to 8 U is far less than that of 4 U to 6 U (Figs 2 and 3). 8 U of Bst DNA polymerase produced the best LAMP results under the optimal ratio of inner to outer primers and MgSO4 concentration (Figs 2 and 3). We speculated that the improvement of efficiency reach a plateau when the Bst enzyme is set at 8 U. This is consistent with previous reports13,32. The initial study evaluating LAMP primer ratios demonstrated that the concentration of the outer primer should be 1/4-1/10 that of the inner primer21,32. In our study, an optimal ratio (8:1) of the FIP/BIP to F3/B3 primers successfully avoided false positive results and low-efficiency amplification (Fig. 2). Mg2+ ions are also known to affect both DNA polymerase activity and PCR specificity33. An extremely high concentration may lead to false positives because of nonspecific amplification. At least 4 mM Mg2+ is needed for a noticeable reaction to occur and 6 to 8 mM results in an optimal LAMP reaction32. The optimal concentration of Mg2+ that has been reported recently for several LAMP methods is 8 mM26,27,34,35,36, which is higher than that for conventional PCR. This is because SYBR Green I was used in these reaction system, as previous reports have demonstrated that Mg2+ concentration should be at least 0.5–3 mM higher when used for florescent PCR than when used for conventional PCR according to the SYBR Green I product manual34. In the orthogonal experiment outlined in this report, combinations 1–4 (which employed 5 mM Mg2+) showed low efficiency of amplification even when the Bst DNA polymerase and primers ratio were under optimized conditions (Fig. 3). As the concentration of MgSO4 in the buffer for Bst DNA polymerase was increased from 5 mM to 8 mM, the amplification efficiency also increased, reaching its highest level at 8 mM (Fig. 2). These patterns were also observed in previously described studies32,37. However, this differs from the recently reported LAMP method employing SYBR Green I for detecting the sugarcane smut pathogen S. scitamineum, in which 6.00 mM was claimed to be the optimal concentration of Mg2+ ions13. The difference might be caused by the differing amounts of dNTPs as well as the templates and the primers used in the reactions32,37. Furthermore, we cannot rule out the possibility that the use of loop primers influences the Mg2+ optimization. In most cases, the optimization of LAMP is accomplished using single variable experiments, by which different individual parameters of the reaction are altered to screen for approximately optimal concentration ranges of a single reaction component while all other variables are held constant. Nevertheless, it is arduous to obtain the optimal experimental results due to the inobservance of potential interactions between these factors13. To overcome this problem, single factor experiments were carried out in combination with an orthogonal experimental design, which allowed the effects of interactions between the different reaction parameters to be determined so that their optimization with respect to the LAMP assay for detection of U. maydis could be established13. In this study, the orthogonal arrangement was designed based upon the results of single factor experiments, and the latter were further authenticated by the orthogonal experiment (Table 1, Figs 2 and 3). Thus an optimized LAMP system that was suitable for the detecting of U. maydis was successfully developed by adopting a single factor experiment followed by enhanced evaluation through the orthogonal experiment schedule.

The proposed RealAmp assay could detect as low as 44 fg/μl U. maydis plasmid DNA among whole DNA extracted from soil (Fig. 5, Supplementary Fig. S5), which was almost 200 times higher than the previously described conventional PCR-based method6. The real-time PCR method that was evaluated showed 10 times more sensitivity than the RealAmp assay when tested with U. maydis genomic DNA (Fig. 5, Supplementary Fig. S5). However, when using the plasmid DNA mixed with maize genomic DNA the RealAmp assay had approximately the same detection limit as real-time PCR (Supplementary Fig. S4). The lower detection sensitivity of RealAmp utilizing a plasmid template that is contained in a mixed DNA isolated form soil sample might be caused by potential inhibitory substances in the soil that constrain the reaction efficiency. In some circumstances, infection by U. maydis does not lead to the development of any external or characteristic symptoms of common smut. Thus determination of the RealAmp detection limit using artificially infested samples, including the teliospore-infested soil and maize plants inoculated with sporidia should assist in future studies on plant-pathogen interactions, smut disease control, and resistance breeding. The detection limit of soil samples artificially infested with teliospores was shown to be 1.6 × 103 spores/g of soil (Fig. 6), which is equal to 8.0 × 10−4 ng/μl pEasy-Pep plasmid DNA according to the quantity of DNA extracted from the soil sample. This is 20 times lower than that calculated when using plasmid DNA diluted with DNA isolated from soil. This might be attributable to low yield encountered during preparation of teliospore nucleic acids from artificially inoculated soil owing to the thick wall of teliospores being somewhat resistant to lysis, and to co-purification inhibitors present in the sample. For the artificially inoculated maize plants samples, the RealAmp assay can detect as low as 4.36 × 10−2 ng/μl U. maydis genomic DNA at the earliest stage of infection prior to the appearance of symptoms. The sensitivity assessment using artificially infested samples suggested that RealAmp could be feasible for quantifying presence U. maydis in field samples. In this work, we were intrigued by the failure of RealAmp to amplify high-concentration template samples. In other words, no amplification was observed in reaction system using undiluted plasmid (440 ng/μl) in the preliminary experiment (Supplementary Fig. S4). We speculate that this is can be explained by the inability of the primers to anneal to their target sequences in the presence of large concentrations of template.

The portable fluorescent reader ZYD-S1TM, equipped with a battery pack, was used for detecting the presence of a plant pathogen for the first time in this study. This instrument has twice the tube holding capacity and twice the efficiency of detecting fluorescent signals than a similar tube scanner that has been previously described26. By means of the mobile DNA workstation, the ZYD-S1TM provides a major advancement towards “electricity-free” LAMP technology, from template preparation to amplification and product quantitative detection, for evaluating field samples (Fig. 7, Supplementary Table S1). The RealAmp assay we developed here could become the foundation for integrated disease management practices that could guide maize planting to avoid further dissemination of U. maydis. It also might be a useful tool for research on plant-pathogen interactions and disease resistance breeding.

Methods

Pathogens and plant materials preparation

U. maydis haploid isolates A1~A3 (clade A) and D1~D3 (clade B) were generously donated by Dr. Qingchang Meng at Jiangsu Academy of Agricultural Sciences. The solopathogenic strain SG200 was kindly provided by Professor Doehlemann at the Max Planck Institute for Terrestrial Microbiology, Marburg, Germany. U. maydis isolates and strains used in this study were maintained on potato dextrose agar (PDA) medium.

The artificial inoculation of U. maydis on maize plants was performed as previously described38 using compatible mating strains or SG200. The infected maize samples were harvested prior to tumor formation for further DNA preparation.

Soil samples preparation

Teliospores were prepared as previously described39. The harvested teliospores were adjusted to the desired concentration for soil inoculation (109 spores/ml) by counting in a hemocytometer. For the artificially infected soil samples, 1 ml titers of the U. maydis teliospore suspension (~1.6 × 108 spores) were inoculated onto 10 g of the twice-autoclaved soil substrate in 15 ml falcon tubes. The fungi were cultured at 25 °C for 10 d. Samples were then air-dried at ambient temperature (~3 d) and subsequently ground in liquid nitrogen to produce a fine powder, which was stored at −80 °C prior to DNA extraction.

DNA extraction

The haploid strains of U. maydis and S. reiliana were grown in YEPSL medium [containing 0.4% (w/v) yeast extracts (BD, Franklin Lakes, NJ, USA), 0.4% (w/v) peptone (BD) 2% (w/v) sucrose (Fishier Sci., Pittsburgh, PA, USA)] at 28 °C with shaking at 200 rpm to an OD600 of 0.8. Cultures were centrifuged at 900 x g for 5 min to pellet cells. Cells were then harvested for DNA extraction using the Mag-Bind® universal 96 Kit (Omega Bio-Tek, Norcross, GA, USA) according to the manufacturer’s instructions, or using a modified version of the CTAB method as described by Xu et al.6.

Mycelial DNA of fungal species other than U. maydis or S. reiliana was extracted with the E.Z.N.A.® SP Fungal DNA Kit (Omega Bio-Tek) following the manufacturer’s manual. The genomic DNA of maize plants was extracted with the E.Z.N.A.® SP Plant DNA Midi Kit (Omega Bio-Tek). Genomic DNA was extracted from the soil samples (0.5 g, with approximately 108 teliospores) as previously described40,41.

The soil and plants samples that isolated in field were treated and then subjected to DNA preparation in the mobile DH-WSWTM mini DNA workstation (Double Helix Tech. Inc.).

Screening of candidate LAMP primers

To screen for the optimal target gene for the U. maydis LAMP assay, three fungal effector genes [UmPep1 of clade A (XM_011389599), UmPep1 of clade B (HQ002865), UmPit2 (XM_752429) and UmSee1 (XM_011390277)] of U. maydis were employed as targets for LAMP primer design in this study.

For each gene, six LAMP primers [external primers F3 and B3; forward and backward internal primers (FIP and BIP, respectively); and loop primers, LoopF and LoopB] were designed according to the gene’s cDNA sequence using the PrimerExplorer V4 software (http://primerexplorer.jp/e/; Eiken Chemical Co., Ltd., Tokyo, Japan). Four to seven sets of LAMP primers were designed for each of the three genes being evaluated. The sequences of these primers are listed in the Supplementary Table S1.

All primers were synthesized by Sango BioTech (Shanghai, China) at HPLC purification grade. The specificity of each primer was checked using the basic local alignment search tool (BLAST) against human DNA and other fungal sequences in the non-redundant GenBank database.

Optimization of real-time LAMP reaction system

Real-time LAMP was conducted using methods described previously for screening U. maydis LAMP primers26,27. In order to optimize the LAMP reaction, a single factor experiment was conducted using three variables and four levels. For a fixed outer primer concentration of 0.2 μM, the concentration ratio of two primer pairs (outer primer F3/B3 and inner primer FIB/BIP) was adjusted to 1:2, 1:4, 1:6 and 1:8. The quantity of Bst DNA polymerase in the reaction mixture was set at 2.0 U, 4.0 U, 6.0 U and 8.0 U. Mg2+ concentrations were set to 5.0 mM, 6.0 mM, 7.0 mM and 8.0 mM, respectively (Fig. 2). All tests were repeated three times. In total, 16 combinations were studied sequentially in the LAMP optimization experiment employing L16 (45) orthogonal experimental design to avoid potential interference caused by interactions between different factors of the LAMP reaction42, (Table 1). Each combination was repeated twice.

The RealAmp reaction was carried out on the CFX96TM real-time PCR detection system (Bio-Rad, Hercules, CA, USA). The following reaction was repeated for 60 cycles: 63 °C for 30 s (stage 1), 63 °C for 15 s (stage 2), and 63 °C for 45 s (stage 3), followed by a melt curve (95 °C for 15 s, 65 °C to 95 °C increment 0.5 °C for 0.5 s). This reaction was coupled with detection of fluorescence signals after each cycle, which was analyzed by Bio-Rad CFX Manager 3.1. For all three reaction monitoring systems, a result was considered positive when an “S” amplification curve was obtained, while a linear or slightly oblique amplification curve indicated a negative result.

Alternatively, the RealAmp assay was performed at 63 °C for 60 min with pathogen detection instrument ZYD-S1TM (Double Helix Tech. Inc.), which is a small, easy-to-use fluorescence measurement system, that has heating blocks capable of holding sixteen tubes, adjustable temperature settings, and spectral devices to detect amplified products using a fluorescent dye. The tube scanner was set to collect fluorescence signals at 30 s intervals. The threshold validation test was used to verify that the signal had sufficiently increased to be considered indicative of a positive result. During real-time amplification, the fluorescence data was obtained on the 6-carboxyfluorescein channel (excitation at 470 nm and detection at 520 nm), and a fluorescence unit threshold value was used. The threshold time (Tt) was calculated as the time after initiation of the reaction when the fluorescence signal reached the threshold value. The threshold value was the sum of the mean fluorescence signal during the initial 5 min and 10 times the standard deviation values during the same time-frame. In the resulting plot, the y-axis denoted the fluorescence units in mV. The x-axis shows the time in min.

LAMP specificity

To confirm the specificity of the LAMP assay to U. maydis, DNA samples from other fungi including Sporisorium reiliana, Curvularia lunata, Bipolaris maydis, Fusarium graminearum, F. moniliforme, Pythium inflatum and Colletotrichum truncatum were employed in the assays. Additionally, the pathogenic bacteria Salmonella enterica and Staphylococcus aureus were used in the LAMP reactions as positive and negative controls, respectively.

LAMP sensitivity

To determine the sensitivity of the U. maydis LAMP assay, a DNA fragment (190 bp) containing the LAMP target region of the gene of the U. maydis genome was amplified by conventional PCR with the primers UmPepF (5′-TCGTGTACCAATGCCAAAG-3′) and UmPepR (5′-TACCGATTCCTCCTAGCAG-3′). The obtained PCR product was purified and cloned into the pGEM® -T Easy vector (Promega) according to the manufacturer’s instructions. The resulting recombinant plasmid, which was confirmed by sequencing (Supplementary Fig S6) and was designated as pEasy-Pep, was adjusted to a concentration of 440 ng/μl and diluted with a 10-fold serial dilution (1 × 10° to 1 × 107 copies) before mixing with extracted DNA from soil or maize plant genomic samples. This recombinant plasmid was used as a reference to assess the detection limit of the LAMP assay.

Real-time fluorescence PCR

The real-time PCR assay was performed thrice for each sample using the SYBR® Premix ExTaqTM kit (TaKaRa). Each PCR amplification procedure was performed with the CFX96TM system (Bio-Rad) at a total volume of 25 μl, following the manufacturer’s instructions. The U. maydis Pep1-specific primer set of Pep-qF (5′-ACAATTCGTACACACTGCCG-3′) and Pep-qR (5′-CCCTTCTTCTCCTGGTCGTT-3′) was used, which was designed with the Primer3 web version 4.0 (http://primer3.ut.ee/)43. The thermal cycling conditions were set up as previously described27.

Feasibility test of RealAmp

To evaluate the feasibility of the RealAmp method for the diagnosis of samples collected in the field, a total of 230 soil or maize plant samples were collected from established maize fields or fields that had never been planted with maize previously. The presence of U. maydis was detected with RealAmp using the ZY1-S1TM pathogen detection instrument.

The quantification of U. maydis DNA in field samples operating on ZYD-S1TM was analyzed statistically by calculating the corresponding threshold time value corrected with the standard curve formula that was determined in the RealAmp detection sensitivity experiment of artificially inoculated samples outlined above. To verify the RealAmp results of filed samples, the LAMP products were amplified using primers pair F2/B2 or mF2-2/mB2-2 and then were subjected to sequencing, in addition, the sporidia isolated from the above mentioned samples were propagated and which DNA templates were amplified using previously described primers Rok317/Rok31831 for sequencing validation.