Disease risk prediction for Phytophthora root rot of chickpeas

¹NSW DPI, Tamworth, NSW, ²DAFQ Warwick, ³SARDI, Adelaide, SA

Take home message

• Phytophthora medicaginis (Pm) inoculum concentrations decline to low levels (within 6-12 months) of a diseased crop and the distribution becomes more uneven
• Resting populations (oospores) can be below detectable levels based on both soil DNA and isolate baiting methods
• These factors limit the ability of PreDicta B^® to identify paddocks which have a significant disease risk
• The Pm test is useful for disease diagnosis when the pathogen is active and inoculum decline has not taken place

Note: the SARDI PreDicta B test for Phytophthora medicaginis is under development and is not yet available commercially.

Phytophthora medicaginis detection in soil

Phytophthora medicaginis, the cause of chickpea Phytophthora root rot (PRR) is endemic and widespread in the northern grains region.  Under conducive conditions, PRR can cause 100% loss.  The pathogen survives from season to season on chickpea volunteers, lucerne, native medics, sulla and as resting structures (oospores) in roots and soil.  It is known that Pm inoculum concentrations is difficult to detect and quantify in paddocks when a susceptible host such as chickpeas is not present (Dale and Irwin, 1990). 

A PreDicta B soil DNA test has been developed by the South Australian Research and Development Institute (SARDI) to quantify the amount of Pm DNA in soil samples and so provide a measure of the amount of Pm inoculum (infected root tissue and oospores) in paddocks.  We report on three seasons of studies to assess the capability of this test to:

1. Detect Pm in soil from commercial paddocks
2. Predict the risk of PRR disease and potential yield losses in chickpea

Methods

Ability to detect Pm in paddock samples:  Soil samples were collected during winter-spring period from fields in northern NSW and southern QLD in 2013, and from central (16) and south-western Queensland (10), and Victoria (7) in 2014.  All paddocks included chickpeas in the rotation but not all had chickpeas in the previous year.

Eight sites were sampled per paddock using with 10 soil cores (15 long 1 cm wide cm AccuCore soil corer).  At each site 10 cores were collected every 20 – 25 paces along a ‘W’ collection pattern (total distance 200 – 250 m per sample site).  If soil conditions were too dry, a 15 cm long by 6 cm wide trowel tapering to 2 cm was used to sample.  Soil samples were stored in sealed plastic bags at 5°C.

After sieving (4 mm aperture), a 400 g sub sample was sent to SARDI for DNA analysis the remainder of each sample was used for baiting of the pathogen using a glasshouse based technique. Seedlings (cv. Sonali) were assessed for disease (chlorosis, stem cankers, death) three times a week.  Stem tissues were plated to isolate Pm.  Cultures with Phytophthora like growth on cornmeal agar were plated on low strength V8 agar and colony morphology, oospore production and oospore size used to identify Pm like cultures.  Isolation of Pm was attempted from all treatments that produced chlorosis followed by the appearance of Pm like stem cankers or seedlings with poor growth.  After six weeks the experiment was terminated.  To fulfil Koch’s postulates on putative Pm isolates, seedlings of the susceptible chickpea cv. Sonali were inoculated in controlled environment experiments.

Detection capability across a range of concentrations:  A Pm concentration:DNA yield series was prepared using a Pm isolate (943c-1) grown on plates of low strength V8 agar then prepared as a oospore-mycelium solution with distilled water at 1752 oospores/mL (average of five separate sample counts).  The required volume of solution containing 0, 100, 500, 1000, 2000, 4000, and 8000 oospores was then pipetted into 400 g of dry sand and distilled water added to bring the total volume of each solution to 5 mL, with three replicates prepared.  The samples were sent to SARDI for Pm DNA analysis. 

Disease and yield loss prediction trials:  In 2014, 2015 and 2016 disease and yield loss prediction trials were carried out. A range of Pm levels were established by applying, different rates of oospores (a mixture of 10 isolates) in-furrow at seedling to four row plots. At sowing or early in the season soil cores (150 mm depth cores, in row coring) were collected from the two middle rows of each plot and pooled and analysed for soil Pm concentration by SARDI.  During the each season PRR disease assessments (% infected plants, or row length of severely infected plants) and grain yields were measured from the two middle rows of each plot.  The trials were also soil sampled at the end of season as described previously.  The 2015 and 2016 trials had four row buffer plots around each plot to limit spread between plots.  Irrigation treatments were watered with dripper tape delivering between 0.6 to 0.7 mm/hr. 

Ability to monitor Pm concentrations in paddocks:  To develop sampling recommendations, Pm concentrations were assessed on three farms (Coonamble NSW, Moree NSW, Goondiwindi QLD) where PRR had been an ongoing problem, “Hotspots” were marked and GPS recorded.

Four samples were collected from each hotspot area following a W collection pattern (32 points along the pattern, each point 6 m apart, using a 150 mm AccuCorer). At each point cores were sampled from a single stubble row, with each core for each separate sample taken 2-3 cm apart. Low lying areas of paddocks where pooling following rainfall occurred (below contour banks, low areas of paddocks, dips) were also sampled and raised or uniform areas were also collected, these areas provided three 32 core ‘low area’ samples and three 32 core ‘high area’ samples from each field, their GPS positions were also marked.  Using this method 12 paddocks were sampled in April 2016, another four paddocks were also sampled with either hotspot or low-high samples collected.  In Nov and Dec 2016 all hotspot sites were resampled, all low-high sample sites were revisited and samples collected from paddocks in chickpeas in 2016 showing any disease problems. 

For the April samples three soil samples from each hotspot and all low-high samples were sent for analysis to SARDI.  The fourth hotspot sample was assessed for Pm in a glasshouse baiting experiment (5 reps, cv. Sonali grown in a soil:sand mix) as described previously.  At the end of the baiting experiment the soil-sand media in each cup was sent to SARDI for analysis. 

Hotspot sites in Coonamble and Goondiwindi were resampled in Nov to compare with April results and for paddocks in chickpeas, the low and high areas were also inspected and soil samples collected if any disease symptoms were observed. 

Results and discussion

Ability to detect Pm in paddock samples:  The 2013 and 2014 collected soils showed a similar pattern with most samples (which had positive DNA results also yielding Pm cultures 2014: 9/11, 82%; 2015: 8/9, 89%) and with most  samples which had negative DNA results not yielding Pm cultures (2014: 36/37, 97%; 2015: 33/34, 97%).  There were two false negatives, whereby the DNA method did not detect Pm, but the soil sample yielded Pm isolates.  If within sample variability is high, the lack of DNA detection may be due to low concentrations in the sub-sample submitted for DNA analysis.  Both sets of samples also had some cases of false positives, possible explanations for this is that more time may be required for symptoms to develop during baiting, or that the pathogen had died but some DNA had been detected.

Detection capability across a range of concentrations:  The experiment showed increasing DNA yields from an increasing pathogen concentration, although the spread among replicates was quite large for the three highest concentrations (Figure 1).  It was also notable that for the lowest concentration 0.25 oospores/g (100 oospore solution in 400 g sand) that yields were variable, values for the three replicates were 0, 261 and 858 copies/g sand.  It is possible that the small amounts of DNA present in low concentration samples, such as these, may have degraded during transportation to SARDI for analyses and lead to the nil detection in one replicate.  Further work is required in this area but this single set of results suggests that the detection of Pm DNA at low concentrations (≤ 0.25 oospores/g sample) may be variable and could deliver nil detection values.

Figure 1. Plot of three replicate sample of Pm oospore-mycelium in sand at a range of concentrations against the yield of Pm copies/g of sample.

Disease and yield loss prediction trials: ability to predict average disease and average yield loss: The relationships between Pm DNA values and disease or yield correlations for the three trials are summarised in Table 1. For 2014 with cv. Sonali there were relatively weak correlations between the early season soil Pm DNA concentrations and PRR disease (r = 0.46) and chickpea yields (r = - 0.37), but for the 2015 with var. Yorker trial Pm DNA values provided a common central relationship across data from both irrigation treatments for both disease and yield.  The 2016 trial with cv. Yorker provided a poor relationship between Pm DNA values and both disease and yield.  The high r values for PRR disease measurements and yield in each trial supported the assumption that the yield losses were principally due to PRR disease.

Table 1. Correlations r values for early season or sowing Pm DNA values, PRR disease and yield of three experiments (DNA14 (2014), HMDNA15 (2015) and HMDNA16 (2016).

Ability to consistently detect Pm DNA in yield loss trials on a plot by plot basis:  In each of the three seasons there were a number of plots where no Pm DNA was detected by the qPCR method either at sowing or early in the season or at the post-harvest sample (Table 2).  The nil DNA plots included both inoculated and non-inoculated treatments each season.  Of the nil Pm DNA early season or at sowing samples, across each of the three trials a proportion (2014 0.79, 2015 0.43 and 2016 0.94) of these then had PRR symptoms, in the majority of cases yield losses were high in these plots (data not presented).  The 2016 trial results were unusual with the very large number of nil DNA plots, Pm control samples included for analyses with these samples gave expected DNA values.  It is not known why so many 2016 samples gave negative Pm DNA results yet PRR symptoms occurred in the plots.  The 2015 trial was the most successful with only 3 of 7 nil DNA plots having PRR symptoms.  However, for the postharvest DNA results from 2015 trial 7 plots with PRR had nil DNA values.

Table 2. No of plots in each of three experiments (DNA14 32 plots (2014), HMDNA15 40 plots (2015) and HMDNA16 40 plots (2016)) that had nil Pm DNA results early in the season or at the post-harvest sample and the number of these plots that had PRR symptoms or not.

Time	post sowing or early season Pm DNA values			post harvest Pm DNA values
Trial date	Total nil DNA values	nil DNA values & nil PRR symptoms	nil DNA values & PRR symptoms	Total nil DNA values	nil DNA values & nil PRR symptoms	nil DNA values & PRR symptoms
2014	14	3	11	2	2	0
2015	7	4	3	11	4	7
2016	33	2	31	-	-	-

Of individual plots with nil DNA results, based on relatively high soil sampling intensities (2.5-3.2 cores/m row length), each season a differing proportion of these then had PRR later in the season.  If these findings are extrapolated to a paddock scenario where a single sample-single result from the paddock will be used to assess PRR disease risk, then the probability for nil DNA results but PRR later being observed (a false negative) will range from 0.18 to 0.83.  However, as the sampling intensity per unit area of paddocks will be much lower than those of plots in field trials, it may be expected that the probability of a false negative for paddock samples may be higher than those for field trials.