A Multiplex TaqMan qPCR Assay for Detection and Quantification of Clade 1 and Clade 2 Isolates of Pseudoperonospora cubensis and Pseudoperonospora humuli

Downy mildew caused by oomycetes within the Pseudoperonospora genus causes a destructive foliar disease of hop (Humulus lupulus; Cannabaceae) and several Cucurbitaceae spp. These pathogens are dispersed by airborne sporangia that develop on sporangiophores from the abaxial surface of infected leaves, resulting in the characteristic “downy-like” appearance (Savory et al. 2011). Pseudoperonospora humuli, the causal agent of downy mildew on hop, occurs worldwide and is of economic importance for growers in the United States, where it negatively impacts cone quality and yield (Gent et al. 2010; Higgins et al. 2020; Royle and Kremheller 1981). Pseudoperonospora cubensis has a broad host range on cucurbits worldwide, including wild species and agronomically important crops such as cucumber, pumpkin, watermelon, and squash (Thomas et al. 2017; Wallace et al. 2020). In 2004, P. cubensis re-emerged in the United States, causing an outbreak of cucurbit downy mildew (CDM) on cucumbers after previously being controlled with resistant cultivars (Holmes et al. 2015). The destructive potential of this pathogen and the development of fungicide resistance has created an urgent need for improved and integrated management strategies through forecasting, detection, and crop resistance (Crandall et al. 2018; Holmes et al. 2015).

Morphological similarities between P. cubensis and P. humuli occasionally have resulted in these organisms being considered taxonomically synonymous (Choi et al. 2005), but some work differentiates isolates within P. cubensis and between P. cubensis and P. humuli. The accurate detection of each species is key for the assessment of disease risk in the field and for regulatory purposes (Runge et al. 2011; Thomas et al. 2017). Previously, Thomas et al. (1987) separated P. cubensis isolates into pathotypes based on a set of differential hosts. Molecular phylogenetic analysis using three loci (nrITS, COX-2, and YPT-1) divided P. cubensis into two clades or cryptic species (Runge et al. 2011). Isolates in clade 1 were associated with isolates that existed in the United States before the 2004 resurgence of CDM, whereas isolates in clade 2 were associated with U.S. isolates after 2004 and those from Europe and East Asia (Japan and Korea) (Runge et al. 2011). In addition, Runge et al. (2011) demonstrated that P. humuli formed a distinct but closely related clade. P. humuli and P. cubensis show limited ability to infect the primary host(s) of the other pathogen, but overall rates of infection by P. humuli on cucurbits and by P. cubensis on hop are low and do not support conspecificity (Mitchell et al. 2011; Runge et al. 2011).

Additional research combining isolate phenotype (host) and whole genome sequencing of nine isolates has confirmed the division of P. cubensis into two distinct lineages, with isolates recovered from Cucurbita pepo, Cucurbita moschata, and Citrullus lanatus tending to be lineage 1 and isolates recovered from Cucumis spp. and C. maxima, including those recovered after the re-emergence of CDM in 2004, tending to be lineage 2 (Thomas et al. 2017). Simple sequence repeat analysis of P. cubensis populations from commercial and wild cucurbits in North Carolina also confirmed the delineation of isolates into two distinct clades and detected evidence of random mating (recombination) only within isolates of clade 1 (Wallace et al. 2020).

Extensive molecular characterization provides differentiation among P. humuli and P. cubensis clades 1 and 2, and corresponding host range data indicate host preference for these isolate groups. Accurate diagnostic tests are needed to distinguish these phenotypically and genetically similar species. Whereas some morphological differences may be observed among Pseudoperonospora spp. on different hosts in a controlled laboratory setting, these differences are imperceptible in the field because of complicating variables, including temperature and relative humidity (Runge and Thines 2012). Specific and rapid molecular tests such as polymerase chain reaction (PCR) and quantitative PCR (qPCR), which target genetic loci or variants specific to each pathogen group, are the best way to distinguish accurately between P. humuli and the two clades of P. cubensis. Previous PCR assays have been used to monitor airborne P. humuli sporangia near hop yards to inform the timing of the first fungicide application, but because this assay also detects P. cubensis, its accuracy decreases in areas in which hops and cucurbits are grown in close proximity (Gent et al. 2009). Specificity was improved with the development of a real-time PCR and high-resolution melt curve assay using a single nucleotide polymorphism in the mitochondrially encoded COX-2 gene to distinguish between P. humuli and P. cubensis (Summers et al. 2015). However, this assay does not separate the two P. cubensis clades, which is needed to track pathogen spread and risk to specific cucurbit hosts.

An adapted version of this real-time PCR assay was used to differentiate between species in spore trap samples, but the specific detection and quantification of P. cubensis and P. humuli were shown to be compromised in samples that contained DNA from both species (Bello et al. 2021). Unique nuclear loci capable of differentiating P. cubensis or P. humuli were identified and suggested as possible candidates for development of a specific diagnostic assay for these taxa (Rahman et al. 2019; Withers et al. 2016). Using a marker that differentiates between P. cubensis clade 1 and 2 isolates, Rahman et al. (2020) described a qPCR assay that can detect as few as 10 sporangia on an impaction spore trap. Having a diagnostic assay capable of detecting all three pathogens simultaneously would improve the ability to manage the diseases they cause, especially if the assay were sensitive enough to detect low levels of the pathogens. An approach that has been used successfully to develop sensitive diagnostic tools for other oomycetes has been to target polymorphisms in the mitochondrial genome (Bilodeau et al. 2014; Crandall et al. 2018; Kunjeti et al. 2016; LeBlanc et al. 2021; Miles et al. 2015, 2017).

In northern regions of the United States, where P. cubensis is not known to overwinter in the field, volumetric spore traps along with light microscopy have been used to detect airborne Pseudoperonospora spp. sporangia to inform the timing of fungicide applications (Granke et al. 2014; Holmes et al. 2015). In addition to this type of analysis being time consuming, it is not possible to quantify each taxon accurately because of similarities in sporangial morphology. Having a rapid, multiplex qPCR assay to detect and quantify concentrations of airborne sporangia for each taxon at specific times would improve the ability to generate these quantitative data. However, when working with environmental samples, it is important to ensure that PCR inhibitors from contaminated soil and other environmental debris do not alter the amplification efficiency of the assay and result in underreporting of pathogen inoculum. To identify whether PCR inhibitors are present, an internal control (IC) often is added to the master mix (Bilodeau et al. 2012; Haudenshield and Hartman 2011; Kunjeti et al. 2016).

Our goal is to provide a means for hop and cucurbit growers to make informed decisions to improve the management of downy mildew. The objective of this research is to use observed polymorphisms in the mitochondrial genomes as targets to develop a multiplexed diagnostic assay to detect and quantify sporangia of P. humuli and P. cubensis clades 1 and 2.

Materials and Methods

Pseudoperonospora isolates.

Purified DNA of P. humuli and P. cubensis was obtained from sporangial suspensions washed from naturally and artificially inoculated infected leaves (Table 1). The isolates were processed in the F. Martin laboratory at the U.S. Department of Agriculture, Agricultural Research Service (USDA-ARS) in Salinas, California. Sporangial samples obtained from North Carolina were processed following the procedures of Withers et al. (2016), whereas sporangia of P. humuli were multiplied on and harvested from the leaves of the susceptible cultivar Pacific Gem as described previously (Mitchell et al. 2011).

Table 1. Isolates of Pseudoperonospora cubensis (clades 1 and 2) and Pseudoperonospora humuli used in this experimentation to validate the multiplexed TaqMan qPCR assay using DNA extracted from recovered sporangia

Isolate	Host	Location
P. cubensis sensu stricto (clade 1)
1928	Cucurbita pepo	Italy
1933	Cucurbita pepo	Puerto Rico
SCD3	Cucurbita moschata	South Carolina
SC1982	Cucumis melo	South Carolina
ew043	Cucurbita maxima	North Carolina
P. cubensis (clade 2)
Way2-2A-1s	Cucumis sativus	North Carolina
Way2-2B-2s	Cucumis sativus	North Carolina
MSU-1	Cucumis sativus	Michigan
ar069	Cucumis sativus	North Carolina
1155	Squash	Italy
1922	Cucumis sativus	Germany
1925	Cucumis sativus	Spain
1926	Cucumis sativus	Spain
P. humuli
Zeus-HDM	Humulus lupulus	North Carolina
Sant2-HDM	Humulus lupulus	North Carolina
OR502AA	Humulus lupulus	Oregon
HDM-269	Humulus lupulus	Oregon
HDM-270	Humulus lupulus	Oregon
HDM-342	Humulus lupulus	Oregon
HDM-444	Humulus lupulus	Oregon
HDM-481	Humulus lupulus	New York
HDM-482	Humulus lupulus	New York
HDM-483	Humulus lupulus	New York
HDM-491	Humulus lupulus	Japan
HDM-494	Humulus lupulus	Japan
HDM-497	Humulus lupulus	Japan
HDM-504	Humulus lupulus	Vermont
HDM-507	Humulus lupulus	New York
HDM-508	Humulus lupulus	New York
HDM-509	Humulus lupulus	Washington
HDM-510	Humulus lupulus	Wisconsin

View as image HTML

To develop a dilution series of sporangial concentrations and enable correlating cycle threshold (C_t) values from the multiplexed TaqMan assay with sporangial numbers, intact sporangia from P. cubensis clades 1 (isolate 043) and 2 (isolate 069) as defined by Wallace et al. (2020) from North Carolina and samples representing a field population of P. humuli from Oregon were shipped in 70% ethanol to the USDA-ARS laboratory in Salinas, CA. P. humuli sporangia were derived from a heterogeneous field population collected from an experimental breeding plot in Oregon and processed as described previously (Mitchell et al. 2011). Before DNA extraction, ethanol was removed before adding the DNA extraction buffer. First, the sporangial slurry was gently vortexed and 20 µl of the solution were added to a separate 2-ml centrifuge tube and diluted with 200 µl of ddH₂O. After mixing, the tubes were spun down for 1 min at 21,000 relative centrifugal force in a microfuge, the supernatant was removed, and the pellet was saved. This step was repeated if the pellet was not properly formed. The pellet was resuspended in 400 µl of PL1 extraction buffer using the NucleoSpin Plant II Kit for Genomic DNA (Macherey-Nagel, Bethlehem, PA) with DNA extraction, as described in a later section.

A dilution series and DNA extraction from suspended spores was performed at the USDA-ARS in Salinas (the laboratory of F. Martin), and additional dilution samples were created and sent to two other collaborating laboratories at Michigan State University (the laboratory of M. Hausbeck) and the USDA-ARS in Corvallis, OR (the laboratory of D. Gent) to test the assay. Specifically, the dilution series of sporangia was made by transferring approximately 20 µl of sporangia from a concentrated slurry in 70% ethanol to a 2-ml centrifuge tube, adding 200 µl of ddH₂O to dilute the ethanol, briefly spinning down in a microfuge at high speed, and removing the resulting supernatant. The sporangia were resuspended in a separate tube with ddH₂O. To visualize and isolate sporangia, a single drop of 15 µl of molecular grade ddH₂O was applied to two microscope slides, and 5 µl of the sporangial suspension was added to the first slide. Individual sporangia were isolated under a dissection scope using a fine-pointed glass pipette. First, the end of a disposable glass Pasteur pipette was heated using a Bunsen burner and then pulled apart to produce a fine tip, which was gently broken with tweezers to leave the smallest opening possible. This glass pipette was then attached to a rubber hose connected to a Bel-Art pipette pump to provide the suction needed to extract the desired number of sporangia from the first slide. To confirm the number of sporangia that were extracted, they were added to the droplet on the second slide. Once the desired number of sporangia were confirmed, the droplet from the second slide was added to a 2-ml centrifuge tube containing 25 µl of ddH₂O. The tube was briefly vortexed to homogenize the sample and then immediately flash frozen using liquid nitrogen for approximately 30 s and placed into a −80°C freezer. For each of the three Pseudoperonospora taxa, 69 sporangial-count samples were assembled, for a total of 207 samples. There were 15 replicates each for the sporangial counts of 1, 3, and 5 and eight replicates each for sporangial counts of 10, 25, and 50. DNA was extracted from these samples using the NucleoSpin Plant II Kit for Genomic DNA and modified protocol noted in a later section.

DNA extraction from purified sporangial suspensions.

The DNA extraction protocol for this study was based on spore extraction methods developed for Bremia lactucae, the causal agent of downy mildew of lettuce (Kunjeti et al. 2016). The protocol was further modified to include a bead beating and chloroform extraction step. After adding the extraction buffer to the spore sample, a capful of glass beads from Millipore Sigma (St. Louis, MO; size 425 to 600 µm, cap from a 1.5-ml microcentrifuge tube) was added to each sample tube. The samples were taped sideways on a platform vortexer and vortexed for 20 min (speed 1, lowest setting). Next, 10 µl of RNAse A (0.01 mg/µl) was added, and the samples were placed in a water bath at 65°C for 30 min to facilitate lysis of the sporangial cell wall. Samples were vortexed for 10 min and 100 µl of chloroform was added, and then samples were vortexed briefly and spun down for 5 min at maximum speed (∼14,000 rpm). Finally, the NucleoSpin Plant II Kit for Genomic DNA was used following the manufacturer’s instructions. DNA concentrations were determined with a model 3 Qubit fluorometer (Invitrogen, Carlsbad, CA).

Pseudoperonospora spp. specific probe and design of primers.

Comparative genomics of assembled mitochondrial genomes of P. cubensis clades 1 and 2 and P. humuli (NC042478) identified a region containing a putative open reading frame in all taxa that exhibited sequence polymorphisms and length differences; ORF374 for P. cubensis clade 1, ORF367 for clade 2, and ORF329 for P. humuli (F. Martin, unpublished data) (Supplementary Fig. S1). BLAST analysis against GenBank and a database of assembled oomycete mitochondrial genomes representing a wide range of taxa (F. Martin, unpublished data) confirmed that these sequences were unique to their respective taxa. Amplification primers were designed that bracketed a 21-bp indel in P. cubensis clades 1 and 2 and generated amplicons variable in length among all three taxa; 298 bp for P. cubensis clade 1, 277 bp for clade 2, and 256 bp for P. humuli. The same forward primer (PC_RFLP_2F; Table 2) was used for amplification of P. cubensis clade 1 and P. humuli; reverse primer PC_RFLP_3R was added for amplification of P. cubensis clade 1; and a second reverse primer (PH_RFLP_4R) was used for the amplification of P. humuli. TaqMan probes were designed spanning the regions variable in length (Table 2). Because of the repeat nature of the polymorphism between P. cubensis clades 1 and 2, the clade 2 TaqMan probe designed for this locus occasionally had background amplification of clade 1 DNAs, so amplification primers (PC-4F and PC-4R) and a probe (Pcub2) were designed for a 135 bp polymorphic region in ORF168 upstream of ORF367 as a target for detection of P. cubensis clade 2.

Table 2. Primer and probe sequences developed for the TaqMan assay to distinguish between Pseudoperonospora humuli and Pseudoperonospora cubensis clades 1 and 2

Name	DNA sequence
PC_RFLP_2F	5-CTGCTTTATCTTTTTCTTTTTG-3′
PC_RFLP_3R	5′-AGAGAAGATTTAGATTATAATTC-3′
PH_RFLP_4R	5′-AGAGACGATTTGGATTATAATT-3′
PC-4F	5′-CAAGACCACCATTTTTATGTC-3′
PC-4R	5′-TGGAAATTAAAAATTTTCTATTAC-3′
Vdf	5′-CGTTTCCCGTTACTCTTCT-3′
Vdr	5′-GGATTTCGGCCCAGAAACT-3′
Pcub_RFLP_qP1	5′-FAM AACAAACTCAAGTAGAACTTC AACAAA – BHQ-I-3′
Phum_RFLP_qP4	5′-CAL Fluor Red 610 CCAACAGTTATACTTGTAATAAACA TCAAG – BHQ-2-3′
Pcub2	5′-HEX AGGATTGATTTTCATTAATTCCTT TTTGTAATAGAA – BHQ1-3′
Vd quasar_probe	5′-Quasar 705 AAAGTAAGCTTATCGATACCGTCGACCT – BHQ2-3′

View as image HTML

Multiplex qPCR assay.

Assays were developed by testing each primer and probe combination independently to confirm specificity and optimize amplification at the same annealing temperature. To identify false-negative results caused by DNA extractions of field samples containing PCR inhibitors, the IC template, probe, and amplification primers reported in Bilodeau et al. (2012) were included in the amplification master mix (Table 2). The concentration of the IC template added to the master mix was diluted to give a final C_t between 32 and 34 (IC dilution experimentally verified) so its amplification would have limited interference with amplification of the pathogen targets. Subsequent analysis validating the assays and processing field samples was accomplished using a multiplex assay that contained components for detecting all three taxa and the IC because this was the intended combination for end use. The 25-µl amplification volume contained 1× PerfeCTa Multiplex qPCR ToughMix (Quantbio, Beverly, MA), 3 mM MgCl₂, 0.5 µM forward primer PC_RFLP_2F, 0.25 µM for the remaining primers for pathogen and IC amplification (Table 2), 50 nM for probe Pcub_RFLP_qP1, and 250 nM for the remaining probe concentration, diluted IC template to provide a C_t between 32 and 34 (empirically determined by testing a 1:10 dilution series of amplified template) and 1 ng DNA. Amplification was achieved in a BioRad CFX96 Realtime PCR Detection System (Hercules, CA) with the following amplification parameters: 95°C for 3 min; 45 cycles of 95°C for 15 s, and 58°C for 45 s, with a plate read at the end of each cycle. The baseline threshold cutoffs used for each fluorophore were the following: Texas Red for P. humuli at 50, HEX for P. cubensis clade 2 at 50, FAM for P. cubensis clade 1 at 25, and Quasar 705 for the IC at 50. Standard curves for each pathogen taxa were generated using singleplex reactions with 10-fold dilution series of DNA (1 ng to 1 fg). Standard curve generation was repeated in multiplex reactions containing all primers, probes, and IC in the amplification master mix to ensure that the IC did not interfere with the accuracy of results. This procedure was repeated in two different laboratories with the same DNA samples to evaluate the reproducibility of the assay. Specificity of amplification for each pathogen taxa was tested with all DNA samples in Table 1.

Correlating C_t with sporangial counts.

In an effort to correlate the number of sporangia with the C_t from the multiplex assay, a sporangial dilution series was prepared as previously noted. For each of the three taxa, 2 µl of DNA extracted from the sporangial dilutions was amplified under the following conditions: (1) with sporangial DNA only; (2) with spiked-in DNA extracted from a greased rod to mimic the elevated presence of PCR inhibitors found in field-collected DNA samples (described in a later section); and (3) as a repeat of the prior condition with the addition of the IC. Samples were amplified on two replicate 96-well plates using the qPCR primers and probes listed in Table 2. Results were analyzed by conducting an Ordinary Least Squared regression model, in which the sum of the squared residuals was minimized and estimated by drawing a linear trendline through our data points (James et al. 2017). Using the data and chart functions in Microsoft Excel version 14.7.1, we generated graphs showing sporangial counts as the independent variable and as the dependent variable, the number of amplification cycles necessary for the signal from a fluorescent dye to reach a threshold amount, or the C_t value.

To evaluate the performance of the diagnostic assays with known DNA concentrations under conditions that would be used when assaying field samples, an impact spore trap (Klosterman et al. 2014) was deployed near the USDA-ARS in Salinas (36.6729° N, 121.6092° W). This was done in an area in which there was no known report of hop or CDMs so that only airborne particulates and sporangia from endemic downy mildews (Klosterman et al. 2014; Kunjeti et al. 2016) were expected to be collected. The rods were lightly coated with silicon grease and deployed within the impaction spore trap (Thiessen et al. 2016) for 72 h before collection and returned to the laboratory for DNA extraction. Each sampling rod was carefully placed into a 2-ml screw-cap tube using a sterilized tweezer. Extraction buffer (400 µl of PL1) was added to the tube and the was DNA extracted using the aforementioned NucleoSpin Plant II Kit for Genomic DNA protocol (Kunjeti et al. 2016). The extracted samples were then combined to ensure a homogenous mixture that was used to spike the pathogen sporangial dilution DNA extractions to evaluate the influence of DNA extraction from environmental samples on amplification efficiency.

Analysis of field samples.

DNA from field samples extracted from infected tissue (Table 3) or rod samples (Table 4) was tested with the multiplex qPCR assay. Infected tissue samples (1 cm²) and rod samples were collected in Michigan 20 m from a commercial cucumber field. These infected tissue samples were carefully placed into impact-resistant 2-ml tubes (Lysing Matrix H, MP Biomedicals, Burlington, NC) containing 100 µl of PL1 extraction buffer. The samples were homogenized using a TissueLyser II (Qiagen, Valencia, CA) for 4 min at 30 hertz, and DNA was extracted using the NucleoSpin Plant II Kit for Genomic DNA following the manufacturer’s instructions. Infected tissue and rod samples from North Carolina were collected and extracted as described in Wallace et al. (2020) and Rahman et al. (2020), respectively. Amplification conditions for analysis of samples was the same as described previously.

Table 3. Detection of Pseudoperonospora cubensis clades 1 and 2 and Pseudoperonospora humuli in multiplexed assay when tested on DNA extracted from infected host tissue from North Carolina and Michigan

Sample	Host	Location (state, county)	Date collected	P. cubensis clade 1	P. cubensis clade 2	P. humuli
5-3	Cucurbita maxima	North Carolina	2016	22.1a	36.5	–
SCD3	Cucurbita moschata	South Carolina	2012	18.4	34.3	–
Kin2-6-4	Cucurbita moschata	North Carolina	2013	21.0	38.0	–
KIN2-6-3	Cucurbita moschata	North Carolina	2013	19.1	33.7	–
KIN2-6-10	Cucurbita moschata	North Carolina	2013	22.2	–	34.2
CLAY2-6-3b	Cucurbita moschata	North Carolina	2015	20.9	–	–
CLAY4-4	Cucurbita pepo	North Carolina	2015	22.2	36.9	–
KIN-1-1	Cucumis sativus	North Carolina	2013	–	30.6	–
KIN-1-6	Cucumis sativus	North Carolina	2013	–	29.3	–
KIN2-3-4	Cucumis melo	North Carolina	2013	–	25.2	–
KIN2-3-3	Cucumis melo	North Carolina	2013	–	23.7	–
113c	Cucumis sativus	Michigan, Ingham	2018	–	22.1	–
109d	Cucumis sativus	Michigan, Ingham	2018	–	18.4	–
202b	Cucumis sativus	Michigan, Ingham	2018	–	26.3	–
217d	Cucumis sativus	Michigan, Ingham	2018	–	21.8	36.7
312a	Cucumis sativus	Michigan, Ingham	2018	–	20.0	–
310a	Cucumis sativus	Michigan, Ingham	2018	–	22.0	–
412	Cucumis sativus	Michigan, Ingham	2018	–	24.0	38.2
415	Cucumis sativus	Michigan, Ingham	2018	–	21.4	–
23	Cucurbita pepo	Michigan, Oceana	2015	–	26.9	31.5

^a C_t of sample when 2 µl of DNA was added to the multiplexed diagnostic assay for P. cubensis clade 1 and clade 2 and P. humuli. Values represent the average of two replicates. There was no difference in amplification of the internal control in samples with added environmental DNA and the control wells. The dashes indicate no detection.

View as image HTML

Table 4. Detection of Pseudoperonospora cubensis clade and clade 2 and Pseudoperonospora humuli in multiplexed assays using DNA samples extracted from impaction spore traps deployed in the field in North Carolina and Michigan

Sample	Location (state, county)	Year	P. cubensis clade 1	P. cubensis clade 2	P. humuli
17	North Carolina, Lenoir	2017	30.1a	23.8	–
18	North Carolina, Lenoir	2017	27.7	25.5	–
19	North Carolina, Lenoir	2017	28.6	30.6	–
20	North Carolina, Lenoir	2017	27.1	28.1	–
21	North Carolina, Lenoir	2017	36.5	38.9	–
22	North Carolina, Lenoir	2017	37.9	37.4	–
RR1-4	Michigan, Ingham	2019	–	–	29.9
RR1-5	Michigan, Ingham	2019	–	–	28.2
RR3-7	Michigan, Ingham	2019	–	–	23.2
RR3-11	Michigan, Ingham	2019	–	–	20.9
RR3-29	Michigan, Ingham	2019	–	34.5	21.1
RR4-55	Michigan, Muskegon	2019	–	34.7	–
RR4-58	Michigan, Muskegon	2019	–	25.4	–
RR4-61	Michigan, Muskegon	2019	–	21.1	–
RR6-58	Michigan, Allegan	2019	–	36.3	–
RR6-62	Michigan, Allegan	2019	–	33.7	–
RR7-66	Michigan, Saginaw	2019	–	29.4	–
RR7-68	Michigan, Saginaw	2019	–	29.7	–

^a C_t of sample when 2 µl of DNA was added to the multiplexed diagnostic assay for P. cubensis clade 1 and clade 2 and P. humuli. Values represent the average of two replicates. There was no difference in amplification of the internal control in samples with added environmental DNA and the control wells. The dashes indicate no detection.

View as image HTML

Results

Evaluation of marker specificity and sensitivity.

In singleplex amplifications with DNA recovered from isolated sporangia (Table 1), the diagnostic markers were specific for the taxa that they were designed to detect; background amplification of the other taxa was not observed. Multiplex reactions containing primers and probes for all taxa and the IC yielded the same results. All isolates of each taxon (Table 1) were detected by their respective diagnostic markers but not by markers for the other taxa.

The serial dilutions of DNA (1 ng to 1 fg) for all three taxa showed linearity when the linear regression analysis was conducted with the following mean C_t and R² values: P. humuli (35.73−20.91; R² = 0.994), P. cubensis clade 1 (36.97 − 25.03; R² = 0.976) and P. cubensis clade 2 (37.06 − 23.32; R² = 0.999) (Fig. 1). To confirm results from the laboratory of F. Martin, the serial dilution series for P. humuli and P. cubensis clades 1 and 2 were sent to the laboratory of M. Hausbeck at Michigan State University to test the assay under different laboratory conditions; similar results were obtained (Supplementary Fig. S2). The only difference observed was for amplification of P. humuli; the C_t in the laboratory of M. Hausbeck was approximately 2.5 lower than that observed in the laboratory of F. Martin.

Download as PowerPoint

Correlation between qPCR results and sporangial counts.

Experimentation to correlate sporangial counts with C_t under conditions consistent with processing field samples was conducted with multiplexed amplifications of the three assays using extracted DNA from a sporangial dilution series. The first multiplexed amplification was with the sporangial DNA alone, followed by a second multiplexed amplification in which the reaction was spiked with DNA extracted from a field-deployed impaction spore trap and a final multiplexed amplification in which the second amplification was repeated with the addition of an IC. (It should be noted that detection was not observed when DNA extracted from a field-deployed impaction spore trap was tested with the multiplexed assay.) The results demonstrated reliable pathogen detection with as few as three sporangia for each taxon (Figs. 2, 3, and 4). Amplification results of DNA from a single sporangium were inconsistent and displayed C_t values that either were outside of the linear relationship between sporangial counts or failed to detect DNA (failed amplification). Addition of DNA extracted from field-deployed rods lacking the pathogen did not impact amplification efficiency; the relationship between sporangial count and C_t remained unchanged. Addition of the IC to evaluate whether PCR inhibitors made it through the DNA extraction step did cause a slight increase in C_t for the lowest sporangial counts of P. humuli and P. cubensis clade 1. Independent verification of the assay in the laboratory of M. Hausbeck using the same samples yielded similar results (Supplementary Fig. S3). The laboratory of D. Gent assayed the P. humuli sporangial dilution samples using only the P. humuli assay without addition of an IC with similar linear results observed, albeit with lower C_t values than observed in the laboratory of F. Martin (Supplementary Fig. S4).

Download as PowerPoint

Download as PowerPoint

Download as PowerPoint

Evaluation of plant and field sampling.

Multiplexed detection of seven North Carolina samples from C. moschata and C. maxima collected in 2016 indicated the presence of high amounts of P. cubensis clade 1 DNA (C_t between 18.4 and 22.2), but lower concentrations of P. cubensis clade 2 DNA were also detected in five samples, and one sample detected P. humuli DNA (C_t 33.7 to 38; Table 3). Only P. cubensis clade 2 was detected in the remaining North Carolina samples from C. melo and C. sativus. P. cubensis clade 2 was also detected from C. sativus and C. pepo samples from Michigan collected in 2015 and 2018, with three of the samples also having lower amounts of P. humuli DNA detected (C_t 31.5 to 38.2). Assays of North Carolina impaction spore trap samples collected in 2017 detected both clades of P. cubensis; with the exception of sample 17, both taxa were detected in roughly similar amounts (Table 4). Analysis of rods collected in Michigan detected only P. cubensis clade 2 in seven of the samples and only P. humuli in four samples, with one sample having both taxa present.

Discussion

The economic impact of downy mildew of cucurbits (P. cubensis clades 1 and 2) and hop (P. humuli) is significant in yield loss and cost of control. Because the sporangia of these three Pseudoperonospora taxa cannot be differentiated reliably based on morphology, an accurate and sensitive molecular diagnostic assay is needed. The ability to accurately detect and simultaneously quantify the airborne inoculum of these pathogen(s) can contribute to improved scheduling of fungicide applications and other control measures. The assay described herein targeting unique sequences for each taxon within a putative open reading frame in the mitochondrial genome addressed this need. BLAST analysis against sequences deposited in GenBank and against a library of mitochondrial genomes representing >170 oomycete taxa (F. Martin, unpublished data), including the three taxa of interest in our study, indicated that the targets selected for each taxon were unique. Specificity, and the ability to detect other isolates of the respective taxa, was confirmed in laboratory-based assays using isolated sporangia from P. cubensis clades 1 and 2 and P. humuli; cross-reactivity was not observed for any of these assays even when run in multiplexed experiments. DNA was extracted from a field-deployed impaction spore trap deployed in Salinas, where other downy mildew taxa are endemic and gave negative results when tested with the multiplex assay, indicating nonspecific amplification against taxa commonly found in the area was not observed (e.g., Peronospora effusa and P. schachtii, Klosterman et al. 2014; B. lactucae, Kunjeti et al. 2016). Given that all three taxa, P. cubensis clades 1 and 2 and P. humuli, may be present at the same geographic location, the ability to multiplex the three diagnostic assays in a single amplification reaction enhances the use of the assay.

Maximum sensitivity of a diagnostic assay is essential when its purpose is to quantify low levels of airborne inoculum, especially when the objective is to implement control strategies to prevent establishment of a disease within a field. Targeting multiple copy sequences improves this sensitivity relative to single copy sequences, which is one reason why diagnostic tests for oomycetes often target the internal transcribed spacer (ITS) of the rDNA or mitochondrial sequences (Crandall et al. 2018; Martin et al. 2012). Given the low level of polymorphisms in the ITS region among P. cubensis and P. humuli, it was not possible to design species-specific assays for these species (Gent et al. 2009), but identifying unique sequences in the mitochondrial genome of each of these three taxa was successful for developing specific and sensitive diagnostic assays. The sensitivity of the diagnostic assays when evaluating purified sporangial DNA was linear for concentrations ranging from 100 fg to 1 ng for all three pathogens tested (R² = 0.999) with a C_t of approximately 35 for 100 fg (Fig. 1). Given that the assay is intended to be used to quantify sporangia in environmental samples, further testing to determine the lower limit of detection for purified sporangial DNA was not done. These assays were not as sensitive as the B. lactucae assay that also targeted a unique putative open reading frame in the mitochondrial genome; this assay detected 20 fg DNA in the linear range of the regression analysis with a C_t of 35 (Kunjeti et al. 2016).

A sporangial dilution series was developed to facilitate correlating C_t with sporangial numbers. Although sporangial counts of less than three could be detected by all three assays, amplifications either were inconsistent or the results were outside of the linear relationship between C_t and sporangial number. The reliable limit of detection for each taxon was determined to be three sporangia and above, which is not as sensitive as that observed for B. lactucae (a single sporangium could be detected with a C_t around 32; Kunjeti et al. 2016). Whether this difference in sensitivity was the result of variation in the number of mitochondrial present in a sporangium or other aspects of the assay design is unknown. Because these assays were run under similar conditions as field-collected environmental samples (multiplexed assays with samples spiked with DNA extracted from field-deployed spore traps with an IC), our results provide a guide for estimating pathogen inoculum densities in environmental samples. However, given the different levels of sensitivity observed when the P. humuli assay was run in the laboratories of F. Martin, M. Hausbeck, and D. Gent, it will be important for individual laboratories to validate the correlation between C_t and sporangial counts before using qPCR results to predict pathogen inoculum density.

P. cubensis clades 1 and 2 and P. humuli were detected on both infected plant samples and field-deployed impaction spore trap samples. Analysis of most of the cucurbit plant samples from North Carolina indicated that two taxa were present, with P. cubensis clade 1 the primary taxa (lowest C_t) and lower amounts of P. cubensis clade 2 and, in one sample, P. humuli. Because both P. cubensis taxa are present at the same time in North Carolina (each taxa exhibits a host preference), both clade 1 and clade 2 isolates have been recovered from the same host species (Rahman et al. 2020; Wallace et al. 2020). For Michigan samples, P. cubensis clade 2 was detected, with three samples also having low amounts of P. humuli (higher C_t). P. humuli is normally not associated with cucurbits (Mitchell et al. 2011; Runge and Thines 2012), but this taxon is present on hop grown in the area; thus, it is likely that its sporangia had been deposited on the cucumber leaf surface. The airborne detection of P. humuli in spore trap samples collected in commercial cucumber fields of Michigan adds support to this hypothesis (Bello et al. 2021). In specificity tests with DNA extracted from purified sporangia, nonspecific background amplification with the other taxa in this study was not observed.

Field-deployed impaction spore traps often contain soil, insect parts, and other environmental debris, which results in PCR inhibitors in the final samples if the DNA extraction process is not performed properly. Thus, it is important to include the IC in the master mix so its amplification in sample wells can be compared with amplification in a control well in which environmental DNA has not been added. This would confirm whether the amplification efficiency of the sample has been reduced, thereby serving as a check against extrapolating a lower inoculum density than was actually present. To reduce competition for amplification between the pathogen target and IC, thereby increasing the C_t for the pathogen assay and affecting the accuracy of quantification, it is important the IC is diluted to provide a C_t between 32 and 34. As discussed in Bilodeau et al. (2012) and Kunjeti et al. (2016), to avoid having the IC amplification influence an accurate quantification of inoculum when the C_t for pathogen detection is above 30, if inhibition of amplification is not observed, the assay should be rerun without the IC to obtain an accurate quantification of the pathogen. If there is an increase in the IC C_t indicating a reduction in amplification efficiency, then the DNA must be diluted to reduce the concentration of PCR inhibitors.

To evaluate whether a diagnostic assay and the procedures for its use can be duplicated, it is important that other laboratories evaluate the same samples to determine whether similar results are obtained (Crandall et al. 2018; Martin et al. 2012). The DNA dilution series presented in Figure 1 was assayed in the laboratory of M. Hausbeck with similar results; the only difference observed was the P. humuli assay, which had a C_t approximately 2.5 lower than the C_t observed by the laboratory of F. Martin. Likewise, similar results were obtained when DNA from the sporangial dilution series used in Figures 2, 3, and 4 was assayed by the laboratory of M. Hausbeck; the assay for P. humuli exhibited a similar reduction in C_t for the sporangial dilution series as observed for amplification of purified DNA. The laboratory of D. Gent also evaluated the P. humuli sporangial dilution series with the P. humuli diagnostic markers only and observed a similar linearity albeit with a lower C_t than the laboratory of F. Martin. The DNA samples extracted from spore trap samples in Michigan presented in Table 4 were provided by the laboratory of M. Hausbeck; the results obtained in the laboratory of F. Martin were similar to those observed in the laboratory of M. Hausbeck. Another benefit of testing assays in multiple laboratories, each independently ordering their own supplies to run the assay, is identifying possible problem areas that may prevent the assay from working properly. One of the laboratories in this project encountered problems with some wells (including the water control) having low levels of background fluorescence with the P. cubensis clade 2 assay late in the amplification cycles; reordering the probe for this taxa from a different supplier resolved this nonspecific detection problem. While the multiplexed assay described in this submission was developed using PerfeCTa Multiplex qPCR ToughMix, Bello et al. (unpublished data) observed similar results when using Prime-Time Gene Expression Master Mix (IDT, Skokie, IL).

Targeting mitochondrial sequences that are either unique for the taxa or have a unique gene order for development of diagnostic assays has been useful for detection of other downy mildew pathogens and oomycetes. Using this approach compared with relying on polymorphisms in primer annealing sites for specificity is advantageous as it reduces the importance of annealing temperature for obtaining accurate results, thereby improving the potential for consistency when the assays are run in different laboratories. The target for the diagnostic assay for detection and quantification of B. lactucae was a unique mitochondrially encoded putative open reading frame (orf286; Kunjeti et al. 2016) while the target for Plasmopara destructor (LeBlanc et al. 2021) was based on the gene order of COX-1–atp1 that is unique to the genus Plasmopara (F. Martin, unpublished data). Conserved gene order differences have also been used for developing a systematic approach for designing a multiplexed approach for genus and species-specific detection of Phytophthora. Targeting the atp9-nad9 locus a genus specific TaqMan probe in a highly conserved region confirms if a Phytophthora spp. Is present while the adjacent spacer region separating the two genes was used for designing species specific TaqMan probes; 50 species specific diagnostic assays have been validated with the data, suggesting species specific markers can be designed for 89% of the 146 taxa sequenced (Bilodeau et al. 2014; Miles et al. 2017). A second genus specific locus for Phytophthora based on the mitochondrial gene order of trnM-trnP-trnM, was also developed (Bilodeau et al. 2014). Both the atp9-nad9 and trnM-trnP-trnM loci were used for developing a genus and species specific isothermal recombinant polymerase amplification detection system for Phytophthora (Miles et al. 2015).

Comments have been made about the accuracy of using mitochondrial loci for quantification of P. cubensis by qPCR because of the varying numbers of mitochondria that may be present in a cell (Rahman et al. 2020). While investigating the use of a nuclear versus mitochondrial locus to quantify an individual species representing four arbuscular mycorrhizal genera by qPCR, Voříšková et al. (2017) observed different numbers of mitochondria in each taxa as well as different numbers of mitochondria present at different stages of growth (with actively growing regions having more than senescent regions). If this is also observed in oomycetes there may be limitations in using mitochondrial loci for quantification of multiple species in infected tissue, so the validity of targeting mitochondrial loci for this type of analysis should be experimentally verified prior to being used. However, the impact of any differences should be much less when using mitochondrial loci for quantification of airborne sporangia. The number of mitochondria in a Pseudoperonospora sporangium is not known, but the error bars in the calibration curves (Figs. 2, 3, and 4) suggest some variation in number among sporangia (differences in DNA extraction efficiencies when low amounts of DNA are present may also contribute to this variation). Because estimation of sporangial counts using the calibration plots is performed for each taxa separately, if there is variation in the numbers of mitochondria in sporangia among taxa, it would not affect the accuracy of results. It is also important to note that when using qPCR to monitor the presence of an airborne pathogen and timing of sprays to prevent the development of an epidemic, maximum sensitivity and the ability to detect when the pathogen arrives so a spray program can prevent disease establishment are more important than absolute quantification of the number of sporangia.

The multiplexed assay for simultaneous detection of P. cubensis clade 1 and 2 and P. humuli described herein is highly specific and sensitive enough to reliably detect sporangial counts down to three sporangia. Coupled with an IC to confirm whether PCR inhibitors are affecting amplification efficiency, the diagnostic assay provides a means for diagnosticians to accurately determine which taxa are present in the air and other environmental samples. The ability to reliably detect inoculum densities as low as three sporangia and amplify template from impaction spore traps should provide information to growers when the pathogens are present in low levels so control measures can be initiated earlier in the epidemic to limit disease development in a field.

Acknowledgments

The authors thank Savithri Purayannur from the laboratory of L. Quesada-Ocampo for providing sporangial suspensions of P. cubensis clade 1 and 2 isolates, Emma Wallace for sending DNA samples, and Alamgir Rahman for sending DNA from impaction spore trap samples. The technical support provided by Uma Crouch, Leah Martin, and Marc Jacquez from the laboratory of F. Martin is gratefully acknowledged. Mention of trade names or commercial products in this article is solely for providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture.