A LAMP at the end of the tunnel: A rapid, field deployable assay for the kauri dieback pathogen, Phytophthora agathidicida

Summary

The root rot causing oomycete, Phytophthora agathidicida, threatens the long-term survival of the iconic New Zealand kauri. Currently, testing for this pathogen involves an extended soil bioassay that takes 14–20 days and requires specialised staff, consumables, and infrastructure. Here we describe a loop-mediated isothermal amplification (LAMP) assay for the detection of P. agathidicida that targets a portion of the mitochondrial apocytochrome b coding sequence. This assay has high specificity and sensitivity; it did not cross react with a range of other Phytophthora isolates and detected as little as 1 fg of total P. agathidicida DNA or 116 copies of the target locus. Assay performance was further investigated by testing plant tissue baits from flooded soil samples using both the extended soil bioassay and LAMP testing of DNA extracted from baits. In these comparisons, P. agathidicida was detected more frequently using the LAMP test. In addition to greater sensitivity, by removing the need for culturing, the hybrid baiting plus LAMP approach is more cost effective than the extended soil bioassay and, importantly, does not require a centralised laboratory facility with specialised staff, consumables, and equipment. Such testing will allow us to address outstanding questions about P. agathidicida. For example, the hybrid approach could enable monitoring of the pathogen beyond areas with visible disease symptoms, allow direct evaluation of rates and patterns of spread, and allow the effectiveness of disease control to be evaluated. The hybrid LAMP bioassay also has the potential to empower local communities to evaluate the pathogen status of local kauri stands, providing information for disease management and conservation initiatives.

Keywords:

Polymerase chain reaction – Forests – mitochondria – Sequence databases – DNA extraction – DNA isolation – Phytophthora – Oomycetes

Introduction

The long-term survival of kauri, Agathis australis (D.Don) Loudon (Araucariaceae), is threatened by the oomycete Phytophthora agathidicida B.S. Weir, Beever, Pennycook & Bellgard (Peronosporaceae) [1]. This soil-borne pathogen causes a root rot that results in yellowing of the foliage, bleeding cankers on the lower trunk, thinning of the canopy and eventually in tree death. Initially reported from Great Barrier Island in the early 1970’s [2], since the late 1990’s the disease has spread rapidly across northern New Zealand [3].

Broad phylogenies for the Peronosporaceae [e.g., 4–7] have identified strongly supported sub-clades. For example, Bourret et al. [7] described total of 16 sub-clades, the majority of which contain at least one Phytophthora species. The phylogenetic analysis of Weir et al. [8] placed P. agathidicida, along with three other formally recognized species, within Clade 5. Although P. agathidicida is currently the only clade 5 species reported from New Zealand, it is not the only Phytophthora species associated with kauri forests. A further five Phytophthora species have been reported from kauri forest soils. Specifically, P. chlamydospora Brasier and Hansen (Clade 6), P. cinnamomi Rands (Clade 7), P. cryptogea Pethybr. & Laff. (Clade 8), P. kernoviae Brasier (Clade 10), and P. nicotianae Breda de Haan (Clade 1) [1, 3]. Other oomycetes, including several Pythium (Pythiaceae) species, have also been reported from kauri forest soils [3].

Methods for the isolation and identification of Phytophthora typically involve recovery from infected host root material [9, 10] or from soil samples using plant tissue fragments (e.g., detached leaves) as bait for zoospores [11, 12]. Currently, the preferred method of testing for P. agathidicida is an extended soil bioassay [13]. Briefly, soil samples are first air dried (2–8 days), then moist incubated (4 days), and finally flooded with water and baited (2 days). At this point the baits are surface sterilised, plated onto Phytophthora-selective media, and incubated at 18°C (6 days). Phytophthora agathidicida is identified from the resulting cultures using morphology or molecular diagnostics both of which require trained laboratory staff. The former is typically based upon the size and reflection of the antheridium [14] while an RT-PCR assay targeting the internal transcribed spacer (ITS) regions of the ribosomal DNA [15] is routinely used for identification of cultured P. agathidicida isolates.

Increasing the immediacy of results from diagnostic testing can substantially improve outcomes in terms of disease control. Historically, shortening the time needed to diagnose a disease has involved the introduction of new laboratory protocols or tools. For example, culture-based methods have in some cases been replaced by molecular diagnostics. More recently, diagnostic approaches that can be conducted onsite, thereby reducing reliance on centralised laboratory facilities, have become an increasingly important means of enhancing the immediacy of results [16, 17]. The emergence of new amplification technologies is enabling the development of rapid, field-deployable approaches to genetic diagnostics [18]. In particular, loop-mediated isothermal amplification (LAMP), an approach that combines rapid amplification with high specificity and sensitivity, is rapidly becoming an important diagnostic tool, especially for point of care applications. There are now numerous examples of LAMP tests for applications in human medicine [e.g., 19, 20], agriculture [e.g., 21, 22], and animal health [e.g., 23, 24]. LAMP has several features that make it particularly well suited to non-laboratory applications. First, the DNA polymerases used to catalyse the LAMP reaction have strand-displacement activity. Therefore, unlike PCR-based diagnostics, LAMP assays may be carried out using relatively simple equipment [25]. Second, the LAMP polymerases are less sensitive to inhibitors than those used for PCR reactions allowing simpler methods of DNA isolation to be used [26]. Finally, turbidity or colorimetry can be used for end-point detection of LAMP products, again reducing reliance on sophisticated equipment [26].

The cost of the extended soil bioassay together with the requirement for specialised staff and infrastructure means that, typically, this assay is conducted only after the appearance of physical disease symptoms. Such restricted testing severely limits our basic understanding of P. agathidicida and our ability to make informed management decisions at local, regional and national scales. For example, we are yet to characterise the distribution of the pathogen, how fast it is spreading, or the efficacy of interventions (e.g., track closures) aimed at disease control. Addressing these knowledge gaps requires ongoing, active monitoring of both diseased and healthy sites across the distribution of kauri. This cannot be achieved using the existing test. Instead, a reliable and rapid assay for P. agathidicida that is both cost effective and robust enough to be deployed outside of a laboratory is needed. Beyond increasing existing capacity, such testing has the potential to enable individual landowners and community groups to evaluate pathogen status in their area and thereby engage in an informed way with regional and national initiatives.

LAMP assays have already been reported for the detection of several Phytophthora species. These include tests for P. capsici Leonian [27], P. cinnamomi [28], P. infestans (Mont.) de Bary [29, 30], P. melonis Katsura [31], P. nicotianae [32], P. ramorum Werres, De Cock & Man in 't Veld [22] and P. sojae Kaufm. & Gerd. [33]. The genetic targets of these tests include the Ras-related protein (ypt1) gene [i.e., 30–33] and the ITS regions [i.e., 22, 27]. As might be expected, LAMP tests for Phytophthora species differ in terms of both their absolute detection limits and performance relative to PCR-based assays. For example, Hansen et al. [29] and Khan et al. [30] reported P. infestans LAMP tests with detection limits of between 128 fg and 200 pg. The most sensitive of these, that of Khan et al. [30], was ten times more sensitive than a test based on nested PCR and at least 100 times more sensitive than either RT-PCR or conventional PCR tests for the corresponding locus.

Here we describe a hybrid bioassay for the detection of P. agathidicida. This combines conventional soil baiting with a highly specific and sensitive LAMP assay to directly test the plant bait tissues for the presence of the pathogen. By reducing assay cost, the time needed for pathogen detection, and reliance on centralised laboratories this approach has the potential both to overcome key limitations of the currently used extended soil bioassay and provide data that will inform our basic understanding of the disease and its management.

Materials and methods

Target region identification

To identify potential targets for genetic testing, publically available Phytophthora and Pythium mitochondrial genome sequences were obtained from the NCBI RefSeq database (https://www.ncbi.nlm.nih.gov). These were combined with mitochondrial genome sequences from Phytophthora and related taxa (e.g., Pythium, Plasmopara) assembled at Massey University. The combined collection comprised mitochondrial genomes from 25 species representing 12 of the 16 clades of Phytophthora and downy mildews reported by Bourret et al. [7]. This sample included all five Phytophthora species reported from kauri forests [1, 3], all currently recognized representatives of Phytophthora clade 5 [8], and two accessions of P. agathidicida [8]. A list of species and accession numbers for publicly available sequence data are provided in S1 Table.

To identify potential targets for LAMP assays, mitochondrial genome sequences from five Phytophthora clade 5 species were aligned using the MUSCLE [34] alignment tool as implemented in Geneious V9 (Biomatters, Auckland, New Zealand). To evaluate the utility of the identified loci, 0.5–1 kb sections of DNA sequence containing these regions were extracted from all 25 mitochondrial genomes and multiple sequence alignments for each constructed as before.

LAMP primer design

LAMP primer sets were generated for potential target regions using PrimerExplorer v5 (https://primerexplorer.jp/e/). For each region, an initial search for regular primers used the P. agathidicida sequence along with default parameter values; the GC content threshold was progressively lowered in subsequent searches until at least one regular primer set was recovered. Candidate primer sets were then compared to multiple sequence alignments for the corresponding target locus; primer sets where annealing sites did not distinguish P. agathidicida were not considered further. For the remaining primer sets, a search was then made for loop primers using PrimerExplorer v5 and the same GC content threshold as generated the regular primer set.

Finally, each primer set was queried against the NCBI nucleotide database (https://www.ncbi.nlm.nih.gov) using BLAST [35] to investigate non-specific annealing.

LAMP assay optimisation

For reaction optimisation, all LAMP assays were conducted in 25 μL volumes consisting of 15 μL OptiGene Isothermal Master Mix (OptiGene Ltd., Horsham, West Sussex, England) plus primer cocktail, extracted DNA and milliQ water (H₂O). Volumes of the three latter components were varied depending on the reaction conditions and template concentration. Reaction sets typically included both positive (i.e., 2 ng total DNA from cultured P. agathidicida isolates NZFS 3128 or ICMP 18244; see S1 Table for accession details) and negative (i.e., no DNA) controls. LAMP assays were performed using a BioRanger LAMP device (Diagenetix, Inc., Honolulu, HI).

Initial optimisation of the LAMP assay evaluated three parameters. First, we investigated the impact of varying the ratio of F3/B3 to FIP/BIP primer pairs. Ratios of F3/B3 to FIP/BIP primers of 1:3, 1:4, 1:6, and 1:8 were trialled; in each case the final concentration of the F3 and B3 primers was maintained at 0.2 μM with the concentrations of FIP and BIP primers being 0.6 μM, 0.8 μM, 1.2 μM, and 1.6 μM, respectively. Second, we examined the effect of amplification temperature. LAMP assays were performed at amplification temperatures of 60°C, 63°C, and 65°C. In all cases, DNA amplification was followed by enzyme denaturation at 80°C for 5 min. Finally, the amplification time was varied from 30–90 minutes.

LAMP assay specificity

The specificity of the P. agathidicida LAMP assay was evaluated using optimised reaction mixes and conditions. These tests were conducted using 2 pg total DNA from six P. agathidicida isolates as well as from isolates of 11 other Phytophthora species representing nine of the 16 clades reported by Bourret et al. [7]. All Phytophthora species reported from kauri forests and all currently recognized representatives of Phytophthora clade 5 were included in the test set (S1 Table). Individual reaction sets also included both positive (i.e., 2 ng total DNA from cultured P. agathidicida isolates NZFS 3128 or ICMP 18244) and negative (i.e., no DNA) controls.

LAMP assay sensitivity

We first evaluated LAMP assay sensitivity in terms of total P. agathidicida DNA. For these tests, a ten-fold dilution series with between 1 ng and 10 ag of total DNA from cultured P. agathidicida isolate ICMP 18244 together with optimised reaction mixes and conditions were used. Testing of the dilution series was conducted both with and without the addition of DNA from a plant tissue commonly used when baiting for P. agathidicida; specifically, 2 ng total Cedrus deodara (Roxb.) G.Don (Himalayan cedar) DNA. Reaction sets also included both positive (i.e., 2 ng total DNA from cultured P. agathidicida isolates NZFS 3128 or ICMP 18244) and negative (i.e., no DNA) controls.

We also evaluated assay sensitivity in terms of target copy number. As a template for these tests, we produced a PCR fragment 799 base pairs (bp) in length. Amplifications were typically performed in 20 μL reaction volumes containing 1× EmeraldAmp GT PCR Master Mix (Takara Bio Inc., Kusatsu City, Shiga Prefecture, Japan) and 0.5 μM of each amplification primer (PTA_pcrF, 5’-CCAAACATAGCTATAACCCCACCA-3’; PTA_pcrR, 5’-GGTTTCGGTTCGTTAGCCG-3’). Thermocycling was performed using a T1 Thermocycler (Biometra GmbH, Göttingen, Germany) and with standard cycling conditions including an initial 4 min denaturation at 94°C, then 35 cycles of 94°C for 30 secs, 58°C for 30 secs and 72°C for 30 secs, with a final 5 min extension at 72°C. Amplification products were prepared for DNA sequencing using shrimp alkaline phosphatase (ThermoFisher Scientific, Waltham, MA) and exonuclease 1 (ThermoFisher Scientific), following a manufacturer recommended protocol. Sensitivity tests using the PCR fragment as template were conducted as for total DNA.

Comparison of standard bioassay and hybrid LAMP bioassay

A direct comparison of the extended bioassay and hybrid LAMP bioassay was performed for two sets of soil samples, one collected from sites in the Waitākere Ranges Regional Park and the other from the Waipoua Forest Sanctuary (S2 Table). These collections were made by the Healthy Trees Healthy Future (HTHF) programme following consultation with representatives of the mana whenua–Te Kawerau ā Maki (Waitākere Ranges) and Te Roroa (Waipoua)–and under permit from the New Zealand Department of Conservation (e.g., 69218-GEO). Samples from each site typically consisted of 1–2 kg of soil from the upper 15 cm of the mineral horizon in the vicinity of kauri trees displaying dieback symptoms. Subsamples of 500 g were first air dried in open plastic containers for two days, moist incubated for four and subsequently flooded with 500 mL of reverse osmosis H₂O. Fifteen detached Cedrus deodara needles were then floated on the water surface. Baits were removed after 48 h (Fig 1); ten were immediately used for the standard bioassay with the remainder frozen at -20°C prior to DNA extraction and LAMP testing.

For the standard bioassay, cedar baits were first rinsed with reverse osmosis (RO) H₂O, soaked in 70% ethanol for 30 s, then rinsed again with RO H₂O before being dried on clean paper towels. Surface sterilised baits were then placed on Phytophthora-selective media [36] using sterile technique and incubated at 18°C for 5–7 days. To identify the resulting cultures, asexual and sexual structures were examined using a Nikon ECLIPSE 80i compound light microscope (Nikon Corporation, Tokyo, Japan) with micrographs captured using a Nikon DS-Fi1 digital microscope camera head (Nikon Corporation) and processed using NIS-Elements BR (version 5.05, Nikon Corporation) [8, 37] (Fig 1).

For the hybrid LAMP bioassay total DNA was extracted from two or three frozen cedar baits using the Macherey-Nagel Plant Kit II (Macherey-Nagel GmbH & Co. KG, Düren, Germany) and the manufacturer’s recommended protocol for plant material. Following extraction, the concentration of total DNA was determined for each sample using a Qubit 2.0 Fluorometer (Thermo Fisher Scientific, Waltham, MA). LAMP assays were performed using optimised reaction mixes and conditions with up to 5 ng of total bait DNA added as template. Each reaction set included both positive (i.e., 2 ng total DNA from a cultured isolate of P. agathidicida) and negative (i.e., milliQ H₂O) controls (Fig 1).

We also used PCR and sequencing of the ITS region to assess the presence of P. agathidicida in total bait DNA samples. Amplifications were performed in 25 μL volumes containing 1 × Platinum SuperFi PCR Master Mix (Invitrogen, Carlsbad, California, USA), 12.5 pM of amplification primer ITS_PTA_F2 [15], 12.5 pM of amplification primer ITS4 [38], and 4 ng of total bait DNA. Thermocycling consisted of an initial 30 sec denaturation at 98°C, followed by 35 cycles of 98°C for 10 secs, 58°C for 10 secs, and 72°C for 30 secs with a final extension of 72°C for 5 mins. Amplification products were purified using the Macherey-Nagel NucleoSpin Gel and PCR Clean-up (Macherey-Nagel GmbH & Co. KG) following the manufacturer’s recommended protocol. Sequencing products were generated from each amplification primer using ABI PRISM BigDye Terminator Cycle Sequencing Ready Reaction Kits (Applied Biosystems, Foster City, California, USA) and run on an ABI 3730 DNA Analyzer (Applied Biosystems). Sequences for each amplification product were assembled using Geneious R9 (Biomatters) and queried against the NCBI nucleotide database (https://www.ncbi.nlm.nih.gov) using BLAST [35].

To further assess the presence of P. agathidicida on baits we also conducted whole genome sequencing of total bait DNA from each of the three sampling sites in the Waitākere Ranges Regional Park. Specifically, we sequenced HTHF 1014, HTHF 1020, and HTHF 1035 (S2 Table). Shotgun sequencing libraries were prepared for each DNA extraction using Illumina Nextera DNA library preparation kits (Illumina, Inc., San Diego, CA). The Massey Genome Service (Palmerston North, New Zealand) performed library preparation, paired-end DNA sequencing and quality assessment of the resulting reads. For each sample a preliminary de novo assembly was performed using idba_ud [39]. The resulting contigs were then compared to our collection of mitochondrial genome sequences (S1 Table) using BLAST [35]. Using the reference assembly tool implemented in Geneious R9 (Biomatters), contigs with high similarity to the reference set were then mapped to a complete mitochondrial genome sequence of P. agathidicida (S1 Table), P. cinnamomi (S1 Table), and Pythium ultimum Trow [40]. Assemblies were subsequently checked by eye; contigs were removed from an assembly if similarity to another reference genome was higher.

Results

Target region identification and primer design

A comparison of mitochondrial genome sequences for members of Phytophthora clade 5 suggested several potential targets for a LAMP assay specific to P. agathidicida. Seven sets of LAMP primers, each consisting of between four and six primers, were designed using PrimerExplorer v5. However, in initial trials all but one of these primer sets failed to discriminate P. agathidicida from other Clade 5 species. We assume that although there are sequence level differences between P. agathidicida and other Peronosporaceae at these loci, the overall number and/or distribution of these differences is not sufficient to prevent amplification in non-target species. The exception was a set of six primers targeting a 227 nucleotide long section of the apocytochrome b (cob) coding sequence spanning from nucleotide position 392 to position 617 (Table 1; Fig 2). Of the initial primer sets, only this one was tested further.

LAMP assay optimisation

Using 2 ng of total DNA from cultured P. agathidicida isolates NZFS 3128 or ICMP 18244 as template, the P. agathidicida LAMP assay gave similar results across a range of reaction conditions. Specifically, amplification was observed for all examined ratios of external to internal primers (i.e., 1:3, 1:4, 1:6 and 1:8), amplification temperatures (i.e., 60°C, 63°C, and 65°C), and amplification times (i.e., 30 min, 45 min, 60 min, and 90 min). Conversely, no amplification was observed under these same reaction conditions for controls containing no DNA.

For subsequent analyses, a 1:3 ratio of external to internal primers, an amplification temperature of 63°C, and an amplification time of 45 min were used.

LAMP assay specificity and sensitivity

Using optimised reaction mixes and conditions, we consistently recovered amplification products from tested P. agathidicida isolates (Table 2). In all cases, amplification was detected using real-time fluorescence (e.g., Fig 3, panel B curves B and C) and agarose gel electrophoresis (e.g., Fig 3, panel A lanes B and C). Conversely, amplification was not detected using either agarose gel electrophoresis or real-time fluorescence for reactions containing no DNA (e.g., Fig 3, panel A curve H and panel B lane H) or those containing total DNA from other representatives of Phytophthora (Table 2).

We initially examined assay sensitivity using total P. agathidicida DNA. Using optimised reaction mixes and conditions, we consistently detected as little as 1 fg total DNA from cultured P. agathidicida isolate ICMP 18244; this limit remained the same when 2 ng total Cedrus deodara DNA was also added to LAMP reactions. The detection limit when using a PCR amplification product containing the target locus as template was 100 ag. Again, this limit was unchanged by the addition of 2 ng total Cedrus deodara DNA. Given Avogadro's number (i.e., 6.022 × 10²³ molecules/mole), the predicted length of the amplification product (i.e., 799 bp), and average weight of a base pair (i.e., 650 Daltons) the observed detection limit of 100 ag corresponds to 116 copies of the target fragment.

Comparison of standard bioassay and hybrid LAMP bioassay

The extended and hybrid LAMP bioassays produced contrasting results for soil samples from sites in the Waitākere Ranges Regional Park and Waipoua Forest Sanctuary (Table 3). Using the extended soil bioassay, P. agathidicida was detected in two of six soil samples from the Waitākere Ranges Regional Park and none of the eight from the Waipoua Forest Sanctuary. Detections were five out of six from the Waitākere Ranges Regional Park and three of eight from the Waipoua Forest Sanctuary using the LAMP assay to test DNA extracted from cedar baits.

Testing of total bait DNA samples using PCR amplification and Sanger sequencing of the nrITS region was consistent with the results of the LAMP assay. Specifically, PCR amplification products of appropriate size were detected for the same five Waitākere Ranges Regional Park and three Waipoua Forest Sanctuary samples as had tested positive using the LAMP assay. Moreover, in BLAST [35] searches of the NCBI nucleotide database the DNA sequences of all eight amplification products shared 100% identity with 14 publically available P. agathidicida nrITS sequences. However, as previously reported ITS sequences do not distinguish P. agathidicida from P. castaneae [8]. Consistent with this 13 publically available sequences of P. castaneae were also recovered with 100% identity.

We assembled mitochondrial genome sequences from whole genome sequencing of total bait DNA from three sites in the Waitākere Ranges Regional Park. Sequences assignable to the mitochondrial genomes of P. agathidicida, P. cinnamomi and a member of the genus Pythium were recovered in most cases. For P. agathidicida, the recovered contigs corresponded to 98.6–100% of the reference mitochondrial genome; smaller portions of the P. cinnamomi and Pythium mitochondrial genomes were recovered.

Discussion

Molecular assays have been reported for various Phytophthora species [e.g., 28, 41, 42]. In the present study, we have developed a LAMP assay for the detection of P. agathidicida, the causative agent of kauri dieback. When combined with soil baiting, this assay, which targets a region of the mitochondrial apocytochrome b gene, provides a powerful alternative to the currently used extended soil bioassay.

In specificity testing, our LAMP assay did not cross react with a range of other Phytophthora species, including all recognised members of Clade 5 and four of the five species (P. nicotianae was not tested) known to occur in kauri forest soils (Table 3). These results are consistent with pairwise comparisons of the assay target sequence from P. agathidicida and 64 other Oomycete taxa (S3 Table). These comparisons indicate pairwise sequence differences of 3.6–14.9% across the entire set; the remaining Clade 5 taxa (including P. sp. novaeguineae) differed by 3.5–4.4% and species from kauri forest soils (including P. nicotianae) by 6.1–8.3%. Together, results from specificity testing and sequence comparisons suggest our LAMP assay is diagnostic of P. agathidicida. In contrast to the currently available P. agathidicida RT-PCR test [15] our LAMP assay distinguishes between P. agathidicida and P. castaneae.

Our LAMP assay consistently detected 1 fg total P. agathidicida DNA. This detection limit is lower than that of several other Phytophthora-specific LAMP assays [e.g., 29] but similar to the limit reported by Than et al. [15] for their P. agathidicida RT-PCR assay (i.e., 2 fg). However, unlike the Than et al. [15] assay, which was ten-fold less sensitive in the presence of soil DNA, the sensitivity of our LAMP assay was found to be unchanged in the presence of background DNA. This result suggests that our LAMP assay is likely to outperform the Than et al. [15] assay for complex samples (e.g., soil or bait DNA). While it is common to report assay detection limits in terms of the total amount of DNA, these limits can be difficult to interpret and we therefore also estimated the detection limit based on target copy number. Using a PCR-amplified target fragment the observed detection limit was approximately 116 target copies. We are not aware of estimates for the numbers of mitochondrial genomes per cell in Oomycetes but for Saccharomyces cerevisae 20–200 mitochondrial genome copies per cell have been reported [43]. While we acknowledge there is considerable variation in mitochondrial genome counts between taxa and life stages, the S. cerevisae count implies that few P. agathidicida zoospores need to have colonised baits before the pathogen would be detected by our LAMP assay. Indeed, the testing conducted in this study suggests that the sensitivity of our LAMP assay is sufficient to detect P. agathidicida at the levels typically encountered on cedar baits used for soil baiting.

For samples from both the Waitakere Ranges Regional Park and Waipoua Forest Sanctuary, P. agathidicida was detected with higher frequency using the hybrid LAMP bioassay than the extended soil bioassay. Specifically, four times more samples tested positive using the hybrid LAMP bioassay (Table 3). That said, results from these two approaches are consistent; those samples that tested positive using the extended soil bioassay also tested positive using the hybrid LAMP bioassay. Given these markedly different results we used a PCR-based approach to further assess the presence of P. agathidicida. Amplification products that shared 100% identity with P. agathidicida ITS sequences were recovered for all eight samples that had tested positive using the hybrid LAMP bioassay. Although these ITS sequences are also consistent with the presence of P. castaneae [8], this latter species has not been reported from New Zealand suggesting P. agathidicida is the likely source. Moreover, from whole genome sequencing of total bait DNA we recovered complete, or nearly so (i.e., 98.6–100%), mitochondrial genome sequences from each of the three Waitakere Ranges Regional Park samples we examined. In all cases the recovered mitochondrial genome included an intact version of the LAMP assay target sequence. For two of these samples (i.e., HTHF 1020 and HTHF 1035) both the extended and hybrid LAMP bioassays had detected the presence of P. agathidicida whereas for the remaining sample (i.e., HTHF 1014) P. agathidicida had only been detected using the hybrid LAMP bioassay. Taken together these additional analyses strongly support the results of the hybrid LAMP bioassay; that is, that P. agathidicida was present in these samples. More generally our analyses also imply that although P. agathidicida zoospores may colonise plant tissue baits, this will not always result in visual detection of P. agathidicida following culturing.

At least partially the slow average in vitro growth rate of P. agathidicida (4.5 mm/day; [8]) may explain why culturing resulted in fewer P. agathidicida detections than did genetic testing. Specifically, for two thirds of the samples (i.e., four of the six) where the results of the extended and hybrid LAMP bioassay differed, oomycetes with faster in vitro growth rates–specifically, P. cinnamomi (8.3 mm/day [44]) and Pythium sp. (21–29 mm/day [45])–were found on the plates following culturing (Table 3). For these four samples our ability to visually detect P. agathidicida may have been compromised by the presence of faster growing species. If so, stochastic differences in patterns of competition across replicates may explain the inconsistent recovery of P. agathidicida from split replicate soil samples [46, 47]. That said, in these studies the choice of bait tissue was not standardised. Recently, Khaliq et al. [48] have shown that bait type and integrity (e.g., detached or intact) influences the diversity of Phytophthora recovered by traditional baiting and culturing. There are clearly multiple factors that impact upon our ability to detect P. agathidicida; at least some of these are either reduced or eliminated by the use of a genetic test to evaluate presence of the pathogen.

Our LAMP assay could be applied to DNA from other sources. One possibility would be to directly test soil DNA for the presence of P. agathidicida. However, low pathogen titre and soil heterogeneity pose considerable challenges for this approach; indeed, inconsistencies between RT-PCR testing of soil DNA and the extended bioassay have previously been reported [47]. Another possibility would be to test DNA from diseased tissue. In this case testing would be confirmatory and not provide information about the distribution of the pathogen beyond those sites where physical symptoms have already been recognised. Given these limitations we have instead focused on implementing a hybrid bioassay that combines baiting as a means of minimising the impact of soil heterogeneity and low pathogen titre with a LAMP assay that increases the sensitivity and reproducibility of detection from baits. Additionally by removing the need for culturing and morphological identification, as well as confirmatory sub-culturing and RT-PCR testing, this hybrid bioassay is both faster and more cost effective than the current extended soil bioassay. As a result the hybrid LAMP bioassay could dramatically enhance our ability to address the threat of kauri dieback. In particular, given the cost effectiveness of this diagnostic we could move from confirmatory testing at diseased sites to systematic monitoring of the pathogen across the distribution of kauri. The latter is necessary if we are to determine pathogen distribution, measure the rate and pattern of spread, and evaluate the efficacy of disease control measures.

Additionally, since culturing is not required, the hybrid LAMP bioassay can be performed without centralised laboratory facilities. Although we acknowledge that the approach is not equipment free, devices for DNA extraction and amplification are now available that enable testing to be carried out locally [e.g., 49, 50, 51]. Critically, the ability to implement testing outside a laboratory creates opportunities for landowners and community groups to engage directly with diagnostic technologies and hence with disease management and conservation programmes. The information provided by community-led testing could enhance the management of local kauri stands as well as contribute directly to regional and national initiatives.

Supporting information

S1 Table [docx]
Sources of the oomycete mitochondrial genome sequences used when designing the . loop-mediated isothermal amplification (LAMP) assay.

S2 Table [docx]
Location details for field collected soil samples used in comparisons of extended bioassay and . loop-mediated isothermal assay.

S3 Table [docx]
Comparison of the LAMP assay target sequence and the corresponding locus in other oomycetes.

S1 Fig [pdf]
Endpoint visualisation of PCR amplification products using SYBR safe (Invitrogen) following electrophoresis on a 1% TAE agarose gel.

S1 Raw images [pdf]
Original images supporting and .