A quantitative baiting assay system for Phytophthora cinnamomi with greater sensitivity than the orthodox serial dilution end-point method was developed. A system of efficiently handling large numbers of subsamples using racks of tubes is described, and the factors determining baiting efficiency with blue lupin (Lupinus angustifolius) radicles were studied. Significantly greater baiting efficiency was obtained when the subsample size was decreased. The optimum temperature was found to be ≈ 25°C for baiting and 20–25°C for bait incubation. Air-drying of soil for two days killed P. cinnamomi propagules. Escape of zoospores from test soil was reduced by the presence of overlying material. Reduction was greatest when the overlying material was of fine particle size and of greater depth. This confirmed the need in assay work to keep soil subsample size small and avoid destruction of the soil structure in preparation. The peak of zoospore release from naturally infested soil occurred on the first day. Infection on radicles decreased as a function of distance from radicle tip.
As part of a larger programme of research to develop biological control of Phytophthora spp., an avocado orchard site that was infested with Phytophthora cinnamomi, and which was eventually to be replanted with avocado, was used to test biocontrol treatments. In order to assess these treatments, it was necessary to quantify the level of P. cinnamomi present in the soil over time. The orthodox serial dilution end-point technique ( Tsao, 1960) was considered appropriate as an initial guide to infestation levels, but neither well suited to quantifying low levels of inoculum nor sufficiently accurate. This work aimed to develop a more sensitive and statistically more robust technique.
A variety of techniques have been used to estimate numbers of propagules of a pathogen in soil. Historically, the most significant and most widely used have been the soil dilution plate method and the serial dilution end-point method. The principles and application of the soil dilution plate method are described by Johnson & Curl (1972). Tsao (1983), in discussing the method, stated that it is effective when the Phytophthora inoculum density is relatively high (> 10 propagules per g of dried soil). Naturally infested soils often have low levels of Phytophthora and hence require low dilutions, and it was not deemed an efficient method unless the antimicrobial agents in the selective medium were highly effective ( Tsao, 1983). Lower dilutions normally result in much greater interference by commensal organisms.
The serial dilution end-point method of Tsao (1960) is commonly used to estimate the disease potential of Phytophthora spp. in soil. The soil is diluted with sterilized soil in a series. Soil samples are saturated with distilled water and a susceptible host is incubated at a suitable temperature in the water above the soil as a bait. Typically, three seedlings in three replicate pots are used for each weight, in a series from 50 g down to 1/256 of this. The disease potential index is the reciprocal of the highest dilution giving a positive result. Weste & Ruppin (1977), in isolating P. cinnamomi from forest soils in Australia, used a modified method, omitting the sterile dilutant soil.
A variety of seedlings, plant parts and fruit have been used as baits for Phytophthora spp. ( Tsao, 1983; Erwin & Ribeiro, 1996). For P. cinnamomi, Eucalyptus seedlings, conifer needles and lupins are among the baits commonly used.
Surface disinfection of plant material with ethanol or sodium hypochlorite to reduce surface contamination prior to plating is normal practice. However, disinfection may reduce isolation success ( Montgomerie & Kennedy, 1983). The bacterial and fungal antibiotics necessary to produce a suitably selective medium have been well documented by Jeffers & Martin (1986).
More modern techniques of assay use immuno- and DNA-based testing ( Cahill & Hardham, 1994; Judelson & Messenger-Routh, 1996). The advantages and shortcomings of these are reviewed by Erwin & Ribeiro (1996). These methods are often compared with the orthodox baiting and plating isolation methods in order to give some gauge of their sensitivity and confirm their validity. They are likely to play an increasingly important role in detection and quantification of Phytophthora spp. The purpose of the research described in this paper was to identify and quantify factors influencing the efficiency of baiting techniques with particular focus on P. cinnamomi, and to develop a more efficient means of handling multiple samples for such assays.
Materials and methods
Preparation and storage of soil
Soil (a yellow-brown loam) was collected from five locations in a 10- to 20-year-old avocado orchard severely infected with P. cinnamomi. Samples to a depth of 10 cm were taken from the root zone and consolidated. The soil was thoroughly mixed, sieved through an 850-µm sieve (when the soil samples from the field were more moist, a 1·5-mm sieve was used), mixed thoroughly again, and stored until use at ambient laboratory temperature (20–25°C) in plastic bags to minimize dehydration. Soil could be kept in the laboratory in this way for 6–8 weeks without major loss of inoculum viability. (Four weeks' storage resulted in a reduction in infectivity to 92% at ambient temperature (≈ 20–25°C) and to 30% at 6°C.)
Standard method of baiting, plating and assessment
Blue lupin (Lupinus angustifolius) was chosen as the most suitable bait since large numbers could be prepared within two to three days and, in preliminary work, suffered less contamination on the plates of selective medium than did pine or cedar needles. Lupin seed was soaked for 6 h in cold water, sown in fine grade vermiculite, and kept moist until use. A shorter presoaking time was found to result in poorer germination and variable radicle length. Prior to use, the vermiculite was washed from the 1- to 5-cm-long radicles of the germinated seedlings.
Soil was apportioned to tubes (33 mL) using the multiple partitioning device described below at a specified volume per unit (normally 0·785 mL, 0·5 g). The tubes were completely filled with reverse osmosis (RO)-treated water, and lupin radicles were suspended (one per tube) through holes in aluminium foil for three days. In all cases, baits were suspended in the water above the saturated soil. Unless otherwise specified, baiting was carried out at ambient laboratory temperature (20–25°C).
Handling of baited radicles
After exposure for a specified period, lupin radicles were removed and washed in running tap water. Radicles were handled in bulk, normally 40 per replicate. The distal 2–5 cm was cut off, immersed in 1% sodium hypochlorite solution for 15 s, washed immediately in running sterile RO water for 15 s, blotted dry on filter paper, and plated onto a selective medium.
This consisted of half-strength prune extract agar, with 10 µg g−1 carbendazim, 25 µg g−1 nystatin, 25 µg g−1 pentachloronitrobenzene (Terrachlor 75%WP), 10 µg g−1 rifampicin, 500 µg g−1 ampicillin, 50 µg g−1 hymexazol. This is a modification of the medium used by Massago et al. (1977) .
Incubation and assessment of radicles
Unless otherwise specified, plated radicles were incubated at ambient laboratory temperature in darkness. Plates were examined after 6–10 days for the presence around each test lupin radicle of mycelial growth typical of P. cinnamomi. While some colonies could be observed as early as 3 days, final assessment was not made for at least 6 days because some were slow to develop. Confirmation of P. cinnamomi was made in early experiments by examining radicle squashes under the microscope for the presence of typical clusters of chlamydospores. Sporangia typical of P. cinnamomi could be induced with nonsterile pond water for checking, but not using the aseptic technique of Chen & Zentmyer (1970), the fungus in the soil under study not responding to this method.
Mass handling techniques
Simple devices were contrived to facilitate the handling of large numbers of subsamples.
1 Multiple partitioning device: This was a set of two perspex plates, with holes drilled to correspond to the grid pattern of the tubes in standard plastic racks (4 × 10 units) ( Fig. 1). The base plate was made with small blocks underneath that held the plate with its holes aligned to those of the tubes in the rack. On top of this base plate there were side and end blocks that aligned the holes in a second (upper) plate to those of the base plate. With the two plates together and with holes out of alignment, soil was spread evenly over the top, allowing it to fill the wells formed by the holes of the upper plate and out-of-line top of the base plate underneath. Even filling was achieved with gentle shaking and tapping, and by smoothing off the excess with a ruler. When the two plates were brought into alignment, the soil in the wells of the upper plate fell through the holes of the lower plate into the tubes. Stubborn subsamples were dislodged easily by gentle tapping. To prevent soil dropping down the outsides of the loosely held tubes, it was necessary to stabilize them. This was done using two sets of thin PVC fingers slipped between the tubes in the rack, at right angles to each other.
Consistent weights were obtained over the set of 40 chambers in the partitioning device. For example, a 10-mm-thick plate with 10-mm-diameter holes (volume 0·785 mL) partitioned subsamples of ≈ 0·5 g. Mean weights for two replicates of 40 weights were 0·513 g and 0·482 g, and standard deviations were 0·0164 and 0·016, respectively.
2 Tubes were cleaned within their racks for reuse using a stainless steel mesh container. In addition, simple, circular, stainless steel gauze containers (45 × 38 mm diameter) facilitated handling of 40 radicles at a time through the disinfection and wash operations.
Determination of factors affecting efficiency of baiting of P. cinnamomi
Soil volume and weight
The effect of volume or weight of soil subsamples on baiting efficiency was determined using a series of weights between 0·25 and 4 g. The 0·25 g treatment had four replicate racks of 40 tubes, all others having two replicates. Numbers of infected radicles per gram of soil were calculated for each replicate. The standard baiting and assessment procedure was used.
Blockage of zoospore escape by overlying soil material
To determine the extent to which zoospore escape was blocked by overlying soil, autoclaved sand (300–850 µm sieved particle size) was used to cover the test soil in the tubes. A standard 0·5 g test soil per tube was used for all treatments. After flooding, sand was apportioned to the tubes and allowed to cover the now settled soil. Sand weight ranged in a series from 0 to 4 g (depth 0–23 mm). The experiment was repeated using autoclaved sieved soil instead of graded sand, and with a range of 0–2 g of added soil (depth 0–9·5 mm). Four and five replicates of 40 tubes per treatment were used, respectively. A further experiment was set up to examine the effect of sand particle size on zoospore escape. Uniform layers of sand (0·5 g, 1·5 mm in depth) were laid over 0·5 g test soil in tubes after flooding, and baited as normal. Sand particle size ranges were 0–150, 150–210, 210–300, 300–500, 500–850 µm. Three replicate racks of 40 tubes for each treatment were used.
Location of infection
The portion of radicle that was most likely to be infected by P. cinnamomi was determined by baiting three racks of 40 tubes, each with 0·5 g soil, with radicles 2–8 cm long. The standard procedure was modified as follows: after two days of exposure, radicles were removed, surface-sterilized as before, washed, dried, then cut into 10-mm sections and plated sequentially, so that the location of infection in every radicle could be determined.
One-day infection profile
The time-course of lupin radicle infection during the first 4 h and the first 24 h was determined by the standard procedure, except that the dip in 1% NaOCl was omitted to avoid loss of infections still restricted to the surface tissue of radicles. Four and three replicate racks of 40 tubes, respectively, were used. For the 4-h profile, fresh radicles were placed in the tubes at one-hourly intervals and the exposed radicles were plated onto selective agar. For the 24-h profile, lupins were similarly replaced and plated at intervals of 4 h. The incidence of infection over time is referred to as the ‘infection profile’.
Three-day infection profile and effect of premoistening and air-drying soil
The effect of premoistening on the time course of infection was examined by moistening one set of tubes, closing them with foil to minimize dehydration, and holding them for three days prior to flooding and baiting. These tubes were compared with another set, the soil in which had not been premoistened. The standard procedure was modified: radicles were removed daily, washed, dipped in 1% NaOCl for 5 s, blotted dry and plated as normal. Fresh lupins replaced the plated ones at the time of removal. The infection of radicles in these tubes was compared with that in a further set with standard three-day baiting. In a further experiment using standard procedure, the effect of air-drying soil for 0, 2 and 4 days was tested. Five replicate racks of 40 tubes for each treatment were used in all cases.
The effect of temperature on the infection of bait radicles was examined in a series of experiments in which all other factors were standard, using a subsample weight of 0·5 g. Due to limited incubator availability, experiments were carried out with sets of temperature ranging from 13 to 25°C, then from 20 to 35°C. The effect of temperature during incubation of plated baits on subsequent growth of P. cinnamomi colonies was similarly examined. In this case, baiting was carried out at ambient laboratory temperature. Each treatment block of 40 tubes was replicated five times for each experiment.
Soil particle size
The effect of soil particle size on baiting efficiency was determined by sieving soil through a series of sieves: 1 500, 850 and 500 µm. This gave three grades of soil, those retained on the 850- and 500-µm sieves, and the residue of fines that passed through the 500-µm sieve. Standard procedure was used with five replicates per treatment, and the experiment was repeated.
Comparison with soil dilution end-point
The multiple partitioning system was compared with the orthodox dilution end-point system. A series of weights (see Table 2) of naturally infested sieved soil and corresponding weights of autoclaved soil were mixed and placed in 270-mL paper pots (four replicates), together with a similar series omitting the sterile dilutant soil; these were compared with 4 racks of 40 tubes, each with 0·5 g of the same soil. Paper pots in the dilution series were baited with four lupin radicles each and tubes with one each. Standard baiting and plating procedure was otherwise used.
Table 2. Comparison of dilution end-point and multiple partitioning systems, using lupin baits and soil naturally infested with Phytophthora cinnamomi
Dilution end-point: four lupin baits per test unit; multiple partitioning: one lupin bait per test unit.
Design and statistical analysis
Experiments were carried out using randomized complete block designs. Data were subjected to analysis of variance using Genstat 5 v. 3·2. Differences were taken as being significant at the 5% level. Graphs of lines and fitted curves together with 95% confidence bands were generated with Flexi software, which uses Bayesian smoothing techniques ( Upsdell, 1994).
Soil volume and weight
The number of lupins infected per gram of soil increased significantly with decreasing weight of subsample size ( Fig. 2).
Blockage of zoospore escape by overlying soil material
The number of infected baits was significantly reduced by the addition of 0·5 g of sand, and it was reduced to virtually zero by greater amounts ( Fig. 3a). There was a similar response to increasing weights of sterilized soil, but the reduction was less marked ( Fig. 3b). When test soil was capped with the same volume of increasingly fine grades of sand, there was a similar significant trend ( Fig. 3c).
Location of radicle infection
Infection of 10-mm radicle sections decreased significantly as a function of distance from the radicle apex ( Fig. 4).
One-day infection profile
There was a significant trend of increasing infection over the 24-h period, with up to 40% of the radicles infected after 3 h. Very little infection occurred in the first hour ( Fig. 5).
Three-day infection profile and effect of premoistening and air-drying soil
The time series showed that most infections occurred on the first day. There was no significant difference in the proportion of lupins infected between soil that was premoistened for three days and soil that was not ( Table 1). When soil was air-dried for either two or four days, there was no infection of lupins, the treatment without air-drying having 9% radicle infection (data not presented).
Table 1. Effect of premoistening of soil samples on a three-day profile of baiting efficiency, using lupin baits and soil naturally infested with Phytophthora cinnamomi
The proportion of lupins infected increased with temperature from 13 to 25°C but decreased at higher temperatures ( Fig. 6). The same trend in P. cinnamomi recovery was evident for incubation temperature of plated radicles ( Fig. 7). Baiting at 35°C prevented lupin infection and incubation of radicles at 35°C prevented growth of P. cinnamomi on the selective medium.
Soil particle size
There was significantly more lupin infection with soil of medium particle size range. The proportions of infected baits for each size range were: 0–500 µm, 0·15; 500–850 µm, 0·72; 850–1500 µm, 0·59 (LSD, 0·063). The results of a repeat experiment were not consistent with this pattern (see Discussion).
Comparison with soil dilution end-point
The experiment produced results typical of a dilution end-point series, with no major difference between the series with and without sterile diluting soil ( Table 2) . Each series produced a similar disease potential index, according to interpretation. The results were similar in an earlier experiment examining the sterile dilutant soil series only.
The success of a baiting assay depends upon zoospores being released from inoculum sources in the test soil, and their subsequent movement to and infection of the root radicle of the bait plant. Contributing factors include the ability of zoospores to move up to 6 cm under their own power ( Newhook et al., 1981 ), negative geotaxis ( Cameron & Carlile, 1977) and the attraction of zoospores by root exudates ( Carlile, 1983). In addition, when test containers of soil are filled with water, turbulence can spread zoospores or soil particles. The combination of these factors probably allowed P. cinnamomi zoospores to cover the ≈ 13 cm distance from soil surface to bait radicle in the present study, and is consistent with the occurrence of infections within the first hour.
The experiments defining the role of soil weight/volume as a variable showed that zoospores can be detected more efficiently from a small volume of soil than from a larger volume. The water-to-soil ratio in baiting techniques should be high to optimize infection of baits ( Tsao, 1983). Many baiting procedures use a low ratio (≈1 : 1 or lower). The enhanced efficacy where the ratio is higher has been attributed to the dilution of chemical substances inhibitory to the germination of Phytophthora propagules; however, the physical blockage of zoospore movement ( Tsao, 1983) appeared to be a major factor in the present experiments. The water-to-soil ratio here was typically 26 : 1, and even thin layers (1–2 mm) of sterile sand or soil on the surface of the test soil clearly blocked zoospore movement. The failure of zoospores to escape from beneath greater thicknesses of soil material may result from physical blockage and the induction of premature encystment. Young et al. (1979) and Newhook et al. (1981) showed little effect of blockage of zoospores in ‘ideal soils’ (glass beads) where the pore size was greater than the size of the zoospores.
The portion of the lupin radicle with the highest incidence of infection was the distal 1 cm. This concords with the pattern most usually found ( Hickman, 1970). In the present study, incidence of infection decreased with distance from radicle tip. Shortening radicles prior to plating by discarding more of the proximal end will therefore result in reduced infection counts. This needs to be taken into account, or the problem avoided by baiting only with very short radicles to allow for growth (to 3–4 cm) before plating.
The level of infection that occurred in the first 24 h of baiting, and particularly in the first few hours, suggests the presence of sporangia or mycelium in the soil. The effect of drying for 2 or 4 days in reducing infection to zero shows that the fungus was in a form that was highly susceptible to desiccation. It is most likely that it was present as mycelium, sporangia and chlamydospores, since P. cinnamomi oospores, which are more resistant to desiccation, are not known to occur in New Zealand soils. Chee (1973) reported the death of chlamydospores of P. palmivora after only 4·5 h of dehydration at 26°C and 94% relative humidity. It is possible that chlamydospores of P. cinnamomi are similarly susceptible. Propagules of other Phytophthora spp. respond differently. Soil naturally infested with P. cactorum oospores, for example, lost no significant infectivity over 10 days of air-dryingm and also responded to premoistening ( Horner & Wilcox, 1995).
The optimum temperature for baiting with lupin radicles was found to be ≈ 25°C, and that for incubation of plated radicles was between 20 and 25°C. Tsao (1983) concluded that lower temperature (15–20°C) enhances Phytophthora recovery, even though for many Phytophthora spp. this is a suboptimal temperature. This was attributed to the greater proliferation of many bacteria at higher temperatures. In the present work, bacterial growth was not observed to be more prevalent at higher temperatures; surface disinfection and blotting dry of radicles prior to plating, together with the effect of the antibiotics in the medium, reduced this problem. The results indicate that the temperature at which plated lupin radicles are incubated has a direct effect on survival and growth of infections initiated during baiting.
Although some Pythium spp. are tolerant to hymexazol ( Jung et al., 1996 ; Tsao & Guy, 1977), they were not important contaminants in the present tests. The most usual contaminant was a Mortierella sp. The optimum temperature for this fungus was found to be 22°C for baiting and 19°C for plate incubation. Growth of this contaminant was worse when surface disinfection was omitted; this also made assessment more difficult.
Specific cations (Ca, K, Mg and Fe) at appropriate concentrations stimulate sporangium production ( Chen & Zentmyer, 1970). Baiting tests comparing commercially available mineral waters with RO-treated water showed significantly higher bait infection with one of the mineral waters. The results are not presented, as a detailed study of ion effects necessary for practical application was beyond the scope of the described work. However, the results emphasize the need for consistency in the mineral content of the water used for baiting.
Particle size effects were found to be inconsistent. Several factors could be involved. For example, smaller particle size may result in more P. cinnamomi propagules being closer to the surface of the particles and zoospores may as a result more easily escape into the free water of the test container. Once particle size drops below a critical level ( Young et al., 1979 ), blockage of zoospore movement can occur, and this could counter the effects of increasing ease of escape. In addition, particle size depends on the adhesion of various mineral materials with organic matter and soil moisture, these being variable in natural soil. While variation can be reduced with careful storage and handling of soil, the results highlight particle size as being a potential source of difficulty in quantification.
Both dilution end-point and multiple partitioning systems are semiquantitative and have flaws in quantifying Phytophthora spp., both in technique and in their interpretability.
The low replicate number and high volume of soil used in the dilution end-point method render it less suitable as a quantitative method. For example, calculations of infections per gram of test soil (figures that the dilution end-point technique was not intended to generate) ( Table 2) produced a trend for only the first three or four in the series. At higher dilutions of test soil, the statistical basis was insufficient for a trend to be apparent. Although Tsao (1960) did not refer to zeroes occurring in the sequence of reducing infections, they occurred in the data in this study, and probably resulted from the variation inherent in low replicate numbers and the uneven distribution of inoculum in the mass of soil tested. In the dilution end-point method, the depth of 25 g of soil is considerable. The blockage of zoospore escape demonstrated by the results means that it cannot be determined just how much soil is being tested for activity.
Inoculum levels in soil are often expressed as ‘propagules per gram of soil’. A single propagule of a Phytophthora spp., which germinates to produce many zoospores, could infect several baits in a test unit (as in the dilution end-point method). Conversely, infection of a single bait plant from two or more inoculum sources in the soil is possible (as in the multiple partitioning system). This means that the expression ‘propagules per gram of soil’ is not a precise measure for Phytophthora spp., and that direct comparison between different systems of baiting is difficult.
It is concluded that the dilution end-point method is suitable for initial establishment of the level of Phytophthora inoculum in soil, but that the multiple partitioning system described provides a more accurate quantification guide. The only limit to detecting low levels of inoculum in soil is the practical one of how many replicate tubes are used. A worthwhile refinement of the technique used here would be the use of a bait that gives a visible characteristic reaction such as Eucalyptus cotyledon colour change ( Marks & Kassaby, 1974), as it would obviate the need to plate and incubate baits.
The work described here permits the following recommendations to be made for estimating the inoculum level of P. cinnamomi in soil. To maximize the sensitivity of baiting, collect soil that is moist. Test the soil as soon after collection as possible, but if necessary store at ≈ 20°C in closed plastic bags to prevent dehydration. Mix the soil thoroughly, sieve gently (0·5–1·5 mm) to avoid excessive destruction of the soil structure, mix again and divide into 40 or more aliquots of 0·5 g or less. Flood with clean water (consistent in its mineral content) in narrow-necked vessels such as test tubes and bait with lupin radicles ((1 cm long, one per tube) for 2–3 days at 25°C. Detach the portion of radicle between the water meniscus and the tip. After washing, surface-disinfect in 1% sodium hypochlorite for a maximum of 15 s, rewash immediately in sterile water, blot dry and plate the radicles on a Phytophthora-selective medium. Incubate at 20–25°C in darkness. Depending on the rate of emergence of P. cinnamomi and the presence of contaminants, assess visually after 6 days for growth of mycelium from the radicles onto the medium. Early assessment (after 3+ days) may be necessary if fast-growing contaminants are present, but late-emerging colonies may be missed.
Statistical analysis and advice by Barbara Dow is gratefully acknowledged, as is the provision of incubators by Landcare Research NZ for the temperature work.