Quantification of root phosphite concentrations for evaluating the potential of foliar phosphonate sprays for the management of avocado root rot

Crop Protection 103 (2018) 87e97

Contents lists available at ScienceDirect

Crop Protection journal homepage: www.elsevier.com/locate/cropro

Quantification of root phosphite concentrations for evaluating the potential of foliar phosphonate sprays for the management of avocado root rot le McLeod a, *, Siyethemba L. Masikane a, Precious Novela b, Jing Ma a, Ade Philemon Mohale b, Makomborero Nyoni a, Marietjie Stander c, J.P.B. Wessels d, Pieter Pieterse b a

Department of Plant Pathology, Stellenbosch University, Private Bag X1, Matieland, 7600, South Africa Bertie van Zyl (Edms) Bpk, P.O. Box 19, Mooketsi, 0825, South Africa c Central Analytical Facility, Stellenbosch University, Private Bag X1, Matieland, 7600, South Africa d ProCrop Trust Consult, Wellington, South Africa b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 7 July 2017 Received in revised form 21 September 2017 Accepted 22 September 2017

In South Africa, phosphonate trunk injections are widely used in a preventative management strategy against avocado root rot caused by Phytophthora cinnamomi. Due to increasing costs, alternative application methods must be investigated. The efficacy of different phosphonate foliar spray treatments was evaluated in two trials that were each situated in a climatically different region. Efficacy was evaluated through quantification of root phosphite (breakdown product of phosphonates) concentrations at different time points, following fall and summer applications. Since no high-throughput cost-effective analytical methods are available for phosphite quantification from avocado roots, a phosphite extraction and purification method was first developed, from which phosphite was quantified using a publically available liquid chromatography-mass spectrometry (LC-MS/MS) method. Foliar potassium phosphonate sprays, applied as three weekly sprays (full- and ¾ volume sprays) in fall, did not result in significantly lower root phosphite concentrations (8, 12 and 23 weeks after application) than the trunk injection. This was also true for two potassium phosphonate foliar sprays applied in summer (8 and 14 weeks after application) in the one trial. However, in the other trial, the summer applied potassium phosphonate foliar sprays had significantly lower root phosphite concentrations than the trunk injection. Ammonium phosphonate foliar sprays, three sprays applied in fall and two in summer, consistently yielded higher or similar root phosphite concentrations than the trunk injection. The ammonium phosphonate foliar sprays furthermore yielded significantly higher root phosphite concentrations than the corresponding potassium phosphonate foliar spray treatment. This was true for almost all time points, except 8-weeks after the summer application in one trial. Phosphite fruit residues were significantly higher for the foliar spray treatments than for the trunk injection in the one trial, but in the other trial it was similar or lower. © 2017 Elsevier Ltd. All rights reserved.

Keywords: Phosphonates Phytophthora Avocado Phosphite Phosphonic acid Phosphorous acid

1. Introduction Avocado root rot caused by Phytophthora cinnamomi is effectively managed using phosphonate fungicides (salts and esters of phosphite [syn. phosphonic acid]) world-wide, including South Africa (Darvas et al., 1984; Pegg et al., 1987). In South Africa, the pathogen previously caused wide-spread destruction in orchards.

* Corresponding author. E-mail address: [email protected] (A. McLeod). https://doi.org/10.1016/j.cropro.2017.09.013 0261-2194/© 2017 Elsevier Ltd. All rights reserved.

This changed when Darvas et al. (1984) discovered fosetylaluminium (alkyl phosphonate) trunk injections, which was subsequently also registered in South Africa. In addition to fosetylaluminium trunk injections, potassium phosphonate has also been registered in South Africa as a trunk injection for preventative- and curative root rot management. This product is currently widely used due to its cost-effectiveness compared to fosetylaluminium. In addition to potassium phosphonates, ammonium phosphonate is also available in South Africa as a registered fungicide on crops other than avocado.

88

A. McLeod et al. / Crop Protection 103 (2018) 87e97

Initially, phosphonate management of avocado root rot was focused on the curative treatment of declining trees. However, as the health of declining trees improved, the focus has since moved to preventative phosphonate management strategies (Whiley et al., 1995). In South Africa, a preventative management strategy consists of two potassium phosphonate trunk injections applied annually, one in fall (after the summer flush hardened off) and another in summer (after the spring flush hardened off). These application windows are very effective due to root flushing occurring during these time points, and the source sink translocation of phosphonates (Whiley et al., 1995). The highly mobile nature of phosphonates in plants unfortunately also results in translocation to fruits, and exceedances of fruit residue limits. This was not problematic during the first registrations of phosphonates in the 1980's, since there were no set maximum residue levels. Consequently, residue data were not a requirement for product registration. However, this situation changed in 2014, when the European Union started to enforce residue limits (50 mg/kg) for phosphonate products on avocado and several other crops. The European Union is the largest avocado export market for South Africa. Due to increasing labor costs, and trunk injections possibly causing damage to tree trunks, alternative application methods are required for replacing phosphonate trunk injections that are widely used in South Africa. In South Africa, labour cost in the agriculture sector increased drastically in 2013 by approximately 50% due to labour unrest (Pahle, 2015). This has resulted in alternative application methods such as foliar sprays becoming more cost effective. Currently, approximately four 0.5% (a.i. phosphorous acid) foliar sprays are almost comparable in cost to two trunk injections in new high density orchards (unpublished data). In contrast, in 2001 only two foliar sprays were similar in cost to two trunk injections (Duvenhage, 2001). In addition to increasing costs, another negative aspect of trunk injections is that it can cause damage to the trunk wood with prolonged use (Robbertse and Duvenhage, 1999). Trunk sprays containing penetrants are an alternative application method that is effective on young avocado trees with green stems (Giblin et al., 2007), and on some threatened native plant species in some countries (Crane and Shearer, 2014; Dunstan and Hardy, 2005; Garbelotto et al., 2007). However, this application method is not effective for older bearing avocado trees (Giblin et al., 2007). A better alternative application method is foliar phosphonate sprays that were first registered in Australia as 0.1% (a.i. phosphorous acid) foliar potassium phosphonate sprays. However, these were later found ineffective by growers. Therefore, emergency use permits were obtained for 0.5% (a.i. phosphorous acid) foliar sprays, which are still being used in Australia (Whiley et al., 2001; personal communication, Elizabeth Dann, University of Queensland, Brisbane, Australia). The number of sprays required is not well defined, with the emergency use registration stating a limit of no more than five sprays. Whiley et al. (2001), in a non-peer reviewed article recommended applications of between three to eight 0.5% foliar sprays. The variable number of recommended foliar sprays might be due to differences in season, location and crop load (Whiley et al., 2001). Therefore, in Australia, it is recommended that growers monitor their root phosphite (breakdown product of phosphonates in plants) levels through a commercial laboratory to determine the number of sprays required (Whiley et al., 2001). In South Africa, limited work has been conducted on the efficacy of foliar phosphonate sprays. The studies have all only been published in non-peer reviewed journals. Duvenhage (2001) evaluated the efficacy of two 0.75% (a.i. phosphorous acid) potassium phosphonate foliar sprays (one after summer flush completion and the other after spring flush completion) in one orchard, and reported that these were effective. McLeod et al. (2015) evaluated the efficacy of three to four foliar potassium phosphonate sprays applied at

different concentrations (0.5%, 0.75% and 1% a.i. phosphorous acid) in two orchard trials. None of the foliar sprays were effective in comparison to the registered trunk injection. This was most likely due to the fact that foliar sprays were applied with a knapsack sprayer, which resulted in too low spray volumes being applied. The efficacy of phosphonate foliar sprays is known to be influenced by spray volume, but limited information is available. In native threatened plant communities in Australia, high volume aerial foliar sprays were shown to be more effective than low volume sprays (Shearer et al., 2012). A similar finding has been reported in avocado in Australia where high spray volumes were more effective. This is most likely due to the fact that more active ingredient is applied to trees (Whiley et al., 2001). The spray volume specified by the emergency use label for potassium phosphonates in Australia states “apply spray volume of 2000e3000 L/ ha for matures trees (depending on tree size)”. This is a rather wide range to select from, which may lead to suboptimal or inconsistent results. In deciduous fruit crops in South Africa, spray volume is determined using the Unrath tree-row-volume (TRV) model. This model could be useful for determining spray volumes for the application of foliar phosphonate sprays to avocado trees. The Unrath model calculates a high spray volume using the formula: canopy diameter  1200 Spray volume ¼ tree height  tree . The constant in row width formula (1200) can vary according to tree crop type (Unrath et al., 1986). In plants, phosphonate dissociates into anions, hereafter referred to as phosphite, which is important in plant tissue for pathogen suppression. This, however, may vary in different Phytophthora host plant systems. In general, a negative linear relationship has been reported for lesion length development and phosphite plant tissue concentration for threatened native Australian plant species that are effectively controlled with phosphonates. However, this is not true for species where phosphonates are less effective (Shearer and Crane, 2009; Shearer et al., 2012; Wilkinson et al., 2001a). El-Hamalawi et al. (1995) also found a negative linear relationship between bark phosphite concentration and inhibition of P. citricola in avocado. Smillie et al. (1989) furthermore reported a close correlation between the concentration of phosphite present at the site of inoculation and the extent of protection against P. cinnamomi, P. nicotianae and P. palmivora inoculated onto phosphonate treated lupine, tobacco and pawpaw plants respectively. Knowledge on phosphite concentrations in plants can be useful for optimizing phosphonate application methods and dosages. In peer reviewed literature, this has been done for the management of P. cinnamomi in threatened native plant species in Australia (Crane and Shearer, 2014; Fairbanks et al., 2000; Shearer and Crane, 2009; Shearer et al., 2012), P. citrophthora in citrus (Schutte et al., 1991), P. citricola canker development in avocado (El-Hamalawi et al., 1995) and P. cinnamomi avocado root rot (Ouimette and Coffey, 1989). A few non-peer reviewed articles have evaluated foliar phosphonate sprays based on root phosphite concentrations in avocado in Australia (Whiley et al., 2001) and South Africa (Duvenhage, 2001; McLeod et al., 2015). Quantification of root phosphite is especially useful for optimizing phosphonate applications for a preventative management strategy, since the pathogen is absent, or present at very low levels and does not cause enough damage, or do so inconsistently within orchards. Foliar sprays, in general, cannot be evaluated effectively on diseased trees since declining trees do not have sufficient foliage for the uptake of foliar applied phosphonates (Darvas, 1983). A number of analytical methods have been published for quantifying phosphite in different plant tissues including radiolabelling, gas chromatography, gas chromatography e mass

A. McLeod et al. / Crop Protection 103 (2018) 87e97

spectrometry (GC-MS), high-performance ion chromatography (IC) and liquid chromatography-mass spectrometry (LC-MS/MS). Highperformance ion chromatography has been used by several research groups for phosphite quantification in plants (Barrett et al., 2003; Borza et al., 2014; Dalio et al., 2014; Jackson et al., 2000; Nartvaranant et al., 2004; Orbovic et al., 2008; Ouimette and Coffey, 1989; Roos et al., 1999; Thao et al., 2008; Whiley et al., 2001; Wilkinson et al., 2001a). A draw-back of IC analyses is that sulfate, which is often applied to agricultural crops, can result in high root sulfate concentrations that have a peak that overlaps with that of phosphite, precluding phosphite quantification (Ma, 2016). GC-MS analyses have also been used by several studies (Barrett et al., 2003; Bezuidenhout et al., 1985; McKay et al., 1992; Schutte et al., 1988; Shearer and Crane, 2009; Shearer et al., 2012; Smillie et al., 1988; Torres Elguera et al., 2013; Van der Merwe and Kotze, 1994), but some of these were conducted by a commercial laboratory (Spadek Western Australian Chemistry center) using an undisclosed method (Crane and Shearer, 2014; Shearer and Crane, 2009, 2012; Shearer et al., 2012). GC-MS analyses require an additional derivatization step, which is time consuming (Alves et al., 2016). In Australia, SGS Australia (Toowoomba Queensland, Australia) is a commercial laboratory that provides a service for quantifying phosphite from avocado roots (Smith et al., 2010; Thomas, 2008). Their assay method is unknown and the service has recently been discontinued (personal communication, Elizabeth Dann, Queensland University, Australia). In New Zealand, Hill Laboratories (Hamilton, New Zeland) is still providing a service for phosphite quantification from roots using an undisclosed liquid chromatography-mass spectrometry (LC-MS/MS) method. LC-MS/MS is a powerful analytical tool that combines the scope and utility of liquid chromatography with the sensitivity and specificity inherent to mass spectrometry (Black and Read, 1998; Hern andez et al., 2003). This method is becoming increasingly popular for testing the presence of highly polar pesticide residues in foods of plant origin, due to the robust selectivity and sensitivity ndez et al., 2003; Hogenboom et al., 2000). LC-MS/MS has (Herna not been reported widely in literature as an analytical method for ndez et al. (2003) quantification of phosphite in plant tissue. Herna and Alves et al. (2016) published LC-MS/MS methods for determining fosetyl-aluminium (Al) residues in lettuce and soy nutraceuticals respectively. LC-MS/MS is furthermore the method that is most commonly used by commercial laboratories for determining phosphite residues in fruit. The aim of this study was to first develop a cost-effective and high-throughput phosphite extraction and purification method for the quantification of phosphite in avocado roots. The extracted phosphite was quantified using a publically available LC-MS/MS method, which is used for analyses of fruit residues. Subsequently, the method was used for evaluating preventative foliar phosphonate sprays (ammonium and potassium) in the 2015 production season in two trials that were each situated in a different climatic region in South Africa. The foliar spray volumes were calculated using the Unrath formula. Potassium phosphonate sprays were evaluated at two different spray volumes. The efficacy of treatments was evaluated through quantification of root phosphite concentrations at several time points after summer and fall applications. The effect of treatments on fruit residues were also investigated, due to the commercial importance to growers. 2. Materials and methods 2.1. Phosphite extraction from avocado roots Avocado root samples from two orchard trials (Section 2.3) were used for evaluating the LC/MS-MS method (Section 2.2). Feeder

89

roots collected from trees were washed using tap water, and the fresh and dry weights were determined. The washed roots were placed in brown paper bags and dried in an oven at 60  C for 2 days. The dried roots were ground into a fine powder using an IKA basic analytical mill® (IKA® - Werke GmbH and Co.KG, Staufen, Germany). Samples were sieved with a tea-strainer to remove large particles that remained after grinding. Particles sizes in sieved samples were between 75 and 150 mm. Phosphite was extracted from the sieved roots by combining 500 mg of roots with 10 ml distilled water in a 15 ml falcon tube. The tubes were shaken on a rotary shaker incubator (3082U, Labcon, Midrand, South Africa) at 100 rpm overnight at room temperature (~25  C). Subsequently, the tubes were centrifuged at 4000 g in a swing bucket centrifuge (Eppendorf 5810R) for 10 min at 20  C. Five milliliters of the supernatant was transferred to a new 15 ml tube. The tubes were briefly vortexed and 2 ml was subsequently passed through a 0.22 mm PALL acrodisc® syringe filter containing a Supor® membrane (Pall Corporation, Midrand, South Africa). Four hundred microliters of the filtrate were added to a 10 K Nanosep® centrifugal device (Pall corporation), and centrifuged at 14 000 g for 20 min. Two hundred microliters of the filtrate from the collection tubes were used for subsequent LC/MS-MS analyses. 2.2. LC-MS/MS phosphite quantification 2.2.1. Standard curves Phosphite standard curve solutions were prepared from a 200 g/ L phosphite stock solution by accurately weighing 20 g of phosphorous acid crystals (Sigma-Aldrich-Aldrich, Oakville, ON) and dissolving it in 80 ml deionized water. The pH was adjusted to 6.5 with 10N KOH and the solution was made up to 100 ml. The solution was diluted to 10 000 mg/ml, and was subsequently serially diluted for the standard curve. The standard curve included concentrations of 0.05, 0.1, 0.5, 1, 2.5, 5, 10 and 20 mg/ml. Internal reference control samples (low, medium and high concentration) of 0.98, 1.5 and 7 mg/ml were also included in runs along with the standard curve. 2.2.2. LC-MS/MS sample analyses All sample analyses were conducted by the Central Analytical Facility (CAF) at Stellenbosch University. The LC-MS/MS analysis method was based on the European Commission Reference Laboratories for residues of pesticides Single Residue Methods (EURLSRM): Quick method for the analysis of numerous highly polar pesticides in foods of plant origin via LC-MS/MS involving simultaneous extraction with methanol (QuPPe-Method). Method 1.3 “Glyphosate and Co. AS 11-HC” (http://www.crl-pesticides.eu/ library/docs/srm/meth_QuPPe.pdf) within this document was used. The analyses were conducted on a Waters Acquity Ultra Performance liquid chromatography system (UPLC) (Waters Corporation) connected to a Waters Xevo TQ mass spectrometer with electrospray ionization probe (Manchester, UK). The column used in LC separation was a Thermo Hypercarb (100  2.1 mm, 5 mM particle size) (Thermo Fisher Scientific, Waltham, USA) at a flow rate of 0.4 ml/min. The mobile phase was a gradient mixture of HPLC-grade water plus 1% acetic acid (Associated Chemical Enterprises, South Africa) (solvent A) and HPLC-grade methanol (Merck, Darmstadt, Germany) plus 1% acetic acid (solvent B) in which the percentages of solvent A and solvent B were changed linearly as follows: 0 min, 98% A and 2% B; 0.5 min, 98% A and 2% B; 5 min, 93% A and 7% B; 5.1 min, 10% A and 90% B; 5.2 min, 98% A and 2% B; 10 min, 98% A and 2% B. The column temperature was held at 40  C. For operation of MS, the settings on the instrument were optimized for maximum ion sensitivity: capillary voltage was 3.50 kV, cone

90

A. McLeod et al. / Crop Protection 103 (2018) 87e97

voltage 20 V, source temperature 140  C, and desolvation temperature 400  C. Desolvation gas flow was 800 L/Hr and cone gas flow 50 L/Hr. Nitrogen gas was supplied by a nitrogen generator and bottled argon was used as collision gas. Phosphite was detected using multiple reaction monitoring (MRM) mode with the 80.9 > 63 transition at a collision energy of 15 eV. Phosphite concentrations were determined by injection a 2 ml aliquot into the column. Masslynx and Targetlynx software (Ver.4.1) was used to process the quantitative data obtained from the calibration standards and from root samples. 2.2.3. Recovery rate, matrix effect and limit of detection Recovery rates were determined by spiking roots from the control treatments from two orchard trials (Section 2.3) with phosphite at the start of the phosphite extraction process (prespiking). Root samples were spiked to final concentrations of 0.5, 1.5, 2.5, 4 and 10 mg/ml (equivalent to 9.52, 28.57, 47.62, 76.19 and 190.48 mg/gFW), by adding the required phosphite concentrations to tubes containing 500 mg of dried roots and distilled water to a final volume of 10 ml. The samples were processed further as described in Section 2.1. Each spiking concentration was evaluated in three to five independent experiments on different days. The matrix effect (post spiking) was investigated by adding phosphite to cleaned root extracts (Section 2.1) from the control trees (Section 2.3). The samples went through the clean-up process ending with elution from the 10 K Nanosep device (Section 2.1). The eluted extracts were spiked to final concentrations of 1.5, 2.5, 4 and 10 mg/ml (equivalent 28.57, 47.62, 76.19 and 190.48 mg/gFW) in at least three independent experiments on different days. 2.2.4. Incurred sample reanalysis (ISR) The reproducibility and precision of the phosphite quantification method were determined using incurred sample reanalyses, as recommended by the AAPS workshop on Current topics in Good Laboratory Practices bioanalysis in 2008 (Yadav and Shrivastav, 2011). ISR consists of repeated analysis of naturally occurring test samples containing the molecule of interest, and is used to determine the reproducibility of a method within the biological matrix of samples. This is important, since assay outcomes can be substantially influenced by biological matrices as compared to when the molecule of interest is analyzed in water or after sample extraction and purification (FDA, 2013; Fluhler et al., 2014; Subramaniam et al., 2015; Yadav and Shrivastav, 2011). ISR was determined on an inter-day (different days) and intraday (within-run) basis. The samples were selected to represent a range of concentrations recorded within the orchard trial root samples across all sampling points (Section 2.3). The intra-day (20 samples) and inter-day (20 samples) precision was assessed by calculating the (i) coefficient of variation percentage   deviation *100) (Reed et al., 2002) and (ii) per(CV% ¼ standard mean    original *100) (Yadav centage difference (DF% ¼ mean repeat of repeat and original and Shrivastav, 2011). According to the FDA, precision for pharmaceuticals analyzed using LC-MS/MS analyses is acceptable if the CV% is below 15% and the DF% below 20% (FDA, 2013; Subramaniam et al., 2015; Yadav and Shrivastav, 2011). 2.2.5. Phosphite root quantification by international commercial laboratory A subset of eight root samples (5 g dried, milled and sieved) ranging in phosphite concentration from 14.22 to 87.11 mg/gFW, as determined using the LC/MS-MS method from the current study, was sent to Hill Laboratories (Hamilton, New Zealand) in New

Zealand. The laboratory uses LC-MS/MS analyses to test avocado roots for phosphite using an undisclosed method. 2.3. Orchard trial 2.3.1. Trial layout and treatments The evaluation of different phosphonate application methods was conducted in two orchard trials in the 2015 season. The trials were situated on the farms Boschoek (low rainfall) and Morgenson (high rainfall) that represent two climatically different regions. Both orchards contained Hass on Dusa tree varieties that had a 6 m canopy diameter, 3 m height and 10 m row width. The orchard trial included five treatments, replicated three times. The trial design was a completely randomized block design. Each replicate consisted of three full-length orchard rows, each containing between 24 and 43 trees. This resulted in the trials containing approximately 1575 trees at Boschoek and 1440 trees at Morgenson. The treatments are shown in Table 1. Phosphonate treatments included foliar ammonium- and potassium foliar sprays and the registered trunk injection. The foliar spray volume required for a full volume spray was calculated using the formula: tree canopy diameter ð6mÞ  1200 Spray volume ¼ tree height ð3mÞ row , which width ð10mÞ

resulted in a spray volume of 2160 L/ha. This volume yielded adequate coverage of the foliage and branches, without run-off occurring. Calculation of the ¾ spray volume treatment used a constant of 900. All foliar sprays were applied with a commercial axial fan sprayer. Trunk injection was conducted using the standard procedure (Darvas et al., 1984), at the South African registration rate for curative treatment. The active ingredient (phosphorous acid) of the potassium phosphonate foliar spray and trunk injection products was 500 g a.i./L, and the ammonium phosphonate foliar spray product 300 g a.i./L. All foliar spray solutions were adjusted to a pH of 7.2 using potassium hydroxide to prevent foliar burn. The dosage of the foliar sprays was reduced from 0.6% a.i. for the fall applications to 0.5% a.i. in summer, due to slight foliar burn occurring with the 0.6% sprays in one of the trials. The treatments were applied in fall after the summer flush hardened off (May 2015), and again in summer after the spring flush hardened off (December 2015). The first fall sprays were applied on 21 May 2015, followed by another two applications at weekly intervals, i.e. total of three sprays. The fall trunk injection was also applied on 21 May 2015. All treatments were re-applied in summer (Table 1). The first summer spray was applied on 1 December 2015, followed by a second spray one week later, i.e. total of two sprays. The summer trunk injection was also applied on 1 December 2015. 2.3.2. Root phosphite sampling and phosphite quantification The trees used in the two orchard trials were not treated with phosphonates for 1 year, at which time residual root phosphite concentrations are very low (unpublished data). Therefore, root phosphite concentrations were only quantified from the control treatment trees in each orchard, prior to the first phosphonate treatment applications. Subsequently, root phosphite concentrations were determined for all treatments at several time points after phosphonate applications were made in fall and summer. Samplings in fall were conducted 4, 8, 12 and 23 weeks after the first foliar applications and trunk injection were applied. In summer, samplings were made 4, 8 and 14 weeks after the first summer applied foliar spray and trunk injection. A total of six root samples were taken per treatment. Since each treatment only consisted of three replicates (three rows of trees/ replicate), the six root samples per treatment were obtained by

A. McLeod et al. / Crop Protection 103 (2018) 87e97

91

Table 1 Phosphonate treatments applied in avocado orchard trial. Treatment

Fall applicationa

Summer applicationb

Untreated control None Foliar potassium phosphonate 3  0.6% a.i. potassium phosphonate foliar sprays at full rate Unrath spray volume (2160 L/ha) Foliar potassium phosphonate 3  0.6% a.i. potassium phosphonate foliar sprays at the ¾ rate Unrath spray ¾ vol. volume (1628 L/ha) Foliar ammonium 3  0.6% a.i. ammonium phosphonate foliar sprays at full rate Unrath spray volume phosphonate (2160 L/ha) Trunk injection 1 x potassium phosphonate trunk injection at 0.5 g a.i./m2

None 2  0.5% a.i. potassium phosphonate foliar spray at 2160 L/ha 2  0.5% a.i. potassium phosphonate foliar sprays at 1628 L/ha. 2  0.5% a.i. ammonium phosphonate foliar spray at 2160 L/ha 1 x trunk potassium phosphonate injection at 0.5 g a.i./m2

a First applications were made on 21 May 2015. Foliar sprays were applied at 1-week intervals. Phosphorous acid is the active ingredient (a.i.) for all the phosphonate products. b First applications were made December 2015. Foliar sprays were applied at 1-week intervals. Phosphorous acid is the active ingredient (a.i.) for all phosphonate products.

taking two spatially separated sub-samples within the center row of each replicate. For each sub-sample, roots were sampled from the same five trees at each time point, which was pooled into one sample, usually resulting in approximately 30e60 g of roots (fresh weight). Samples consisted of a mixture of mature roots and root tips, with the latter depending on the root flush status of trees at the time of sampling. The roots were processed and phosphite extracted as described in Section 2.1, and phosphite was quantified using the LC/MS-MS method as described in Section 2.2. The root phosphite quantifications were conducted so that all treatments from one time point and from both trials were conducted on the same day (i.e. within the same run). This was possible since 200 samples can be run within the same LC-MS/MS run over a 2-day period. The LC/MS-MS values were multiplied by 20 to adjust for the dilution of samples (500 mg dried roots in 10 ml water). Subsequently, values were also adjusted for recovery rate (estimated at an average of 42%, Section 3.1.2) by multiplying each sample with 100÷ 42. Finally, the phosphite concentration was calculated for the fresh weight of roots, assuming a moisture content of 60%. This average moisture content was decided on since a random subset of avocado roots from the orchard trials had an average moisture content of 60% (range from 46% to 65%). The rate of phosphite change in roots per week was calculated for the treatments and seasons (fall and summer). For the fall application, the rate of change from week-4 to week-12 after fall application was calculated:

Rate of change ¼

2.3.3. Phosphite fruit residues Fruit samples were collected for residue analyses within the same six replicates (two sub-samples within each of three replicates) used for root sampling. A total of 1e2 kg of fruit was collected per replicate, from the same five trees that were also sampled for roots. The fruits were sent for phosphonic acid (synonym of phosphite) quantification to a commercial laboratory, Hearshaw and Kinnes Analytical laboratory (Pty) Ltd (Cape Town, South Africa). 2.3.4. Statistical data analyses Analyses of variance (ANOVA) was performed on the root phosphite concentrations, fruit phosphite residues and weekly rate of phosphite change using the GLM (General Linear Models) Procedure of SAS statistical software (Version 9.4; SAS Institute Inc, Cary, USA). The Shapiro-Wilk test was performed to test for deviation from normality (Shapiro and Francia, 1972). The root phosphite data were not normally distributed and therefore a Ln (xþ1) transformation was conducted to stabilize the variance and improve normality (Snedecor and Cochran, 1980). Fisher's least significant difference (LSD) test was calculated at the 5% level to compare means for significant effects (Ott, 1998). A probability of 5% was considered significant for all significance tests. Pearson's correlation analyses and the significance of correlations were conducted on the root phosphite concentrations versus fruit phosphite concentrations using XLStat (Version 2014, Addinsoft, New York, USA).

week 12 root phosphite concentration  week 4 root phosphite concentration 8 weeks

and for the summer applications:

Rate of change ¼

week 14 root phosphite concentration  week 4 root phosphite concentration 10 weeks

92

A. McLeod et al. / Crop Protection 103 (2018) 87e97

3. Results

3.2. Orchard trials

3.1. LC-MS/MS phosphite quantification

3.2.1. Root phosphite quantification Prior to the first phosphonate applications, root phosphite concentrations were determined for the six control replicates within each trial. At Morgenson, the average root phosphite concentration for control trees was 17.61 mg/gFW (10.77e25.49 mg/gFW), and at Boschoek the average was 8.85 mg/gFW (6.15e12.61 mg/gFW). ANOVA analyses showed that there were significant trial x treatment interactions for the week-4 and week-12 (P < 0.002) time points after the fall applications, and for weeks 4, 18 and 14 (P < 0.007) after the summer applications. The only time points for which there were no trial x treatment interactions were the week-8 and week-23 time points after the fall applications. Therefore, the data of the two replicate trials were not combined but analyzed separately. A significant treatment x time point interaction (P < 0.001) was recorded for both trials, and therefore time points were also analyzed separately. For both trials, there were significant differences (P < 0.001) in root phosphite concentrations between treatments for each of the seven investigated time points (weeks 4, 8, 12, 23 after fall application, and weeks 4, 8 and 14 after summer application). The root phosphite concentrations of the untreated controls in both trials were significantly lower than all the phosphonate treatments at all time points after phosphonates were applied (Table 2). The exception was for the Boschoek trial at the last sampling date (14-weeks after the summer application), where the potassium phosphonate full volume spray did not differ significantly from the control. The foliar ammonium- and potassium phosphonate sprays yielded root phosphite concentrations that were mostly comparable, or higher than that of the trunk injection following fall applications. However, for the summer applications this was not always the case for all treatments (Table 2). In both trials, the fall applied full volume and ¾ volume 0.6% a.i. potassium phosphonate foliar spray treatments, yielded root phosphite concentrations that were not significantly lower than the trunk injection treatment at weeks 8, 12 and 23 after application. This was also true for the two summer applied 0.5% a.i. potassium phosphonate foliar sprays at weeks 8 and 14 after application in the Morgenson trial. However, at the Boschoek trial all the summer foliar potassium phosphonate treatments had significantly less root phosphite concentrations than the trunk injection at weeks 8 and 14 after application. In both trials, there were no significant differences between the full volume and ¾ volume potassium phosphonate sprays for all the time points. The full volume summer and fall applied ammonium phosphonate foliar sprays in both trials yielded significantly higher or similar root phosphite concentrations than the trunk injection for

3.1.1. Standard curves Regression analysis of the analyte peak area response versus phosphite concentration exhibited a good quadratic relationship (R2 ¼ 0.9993) within the concentration range of 0.01e20 mg/ml. The average retention time of phosphite for avocado root samples was 1.7 min.

3.1.2. Recovery rates, matrix effect and limit of detection The recovery rates of the pre-spiked avocado root samples at all four phosphite concentrations (0.05, 1.5, 2.5, 4 and 10 mg/ml) were similar. The average recovery rate was 42% ± 8%. Samples were also spiked post clean-up, to determine if the observed recovery rates were due to a matrix effect. The post clean-up recovery rates were similar than the pre-spiked samples. This indicated that the observed recovery rates were, in fact, a matrix effect.

3.1.3. Incurred sample reanalysis (ISR) The intra-day ISR (within the same day) precision analyses of 20 root samples were within acceptable levels. The CV% (percentage coefficient of variation) and DF% (percentage difference) for all samples were below 15% and 20% respectively as specified by the FDA standards (FDA, 2013). The precision of the inter-day ISR (on two different days) analyses was higher than the intra-day ISR. The inter-day ISR of the 20 samples was within the acceptable FDA standards for 70% of samples. The remaining samples were not within acceptable ranges (CV % 18e32%; DF% 25e45%). The average inter-day ISR for all 20 root samples was CV% 13.58 ± 9.46 and DF% 10.70.14 ± 21.45. The internal reference standards were also analyzed for intraday and inter-day ISR. These standards were all below the FDA standards of CV 15% and DF 20%.

3.1.4. Phosphite root quantification by international commercial laboratory The eight root samples analyzed with the LC/MS-MS method (after adjustment for a recovery rate of 42%) from the current study had root phosphite concentrations that were comparable to those conducted by Hill's laboratory. The CV % and DF% values were below the accepted FDA values (CV% <15%; DF% < 20%) for six of the eight samples. Only two of the samples (64.35 and 87.11 mg/gFW as determined in the current study) had CV% (26 and 33%) and DF% (38 and 47%) values above the acceptable levels.

Table 2 Root phosphite concentrations (mg/gFW) in avocado trees receiving different phosphonate treatments at two trial sites (Boschoek and Morgenson), following fall and summer phosphonate applications. Treatments

Boschoek

Morgenson

Fall applicationa 4w June Untreated control 10.58 c Foliar potassium phosphonate 37.46 b Foliar potassium phosphonate ¾ vol. 27.37 b Foliar ammonium phosphonate 66.31 a Trunk injection 57.86 a

Summer applicationa

Fall applicationa

Summer applicationa

8w July

12 w 23 w August Oct.

4w Dec.

8w Jan.

14 w March

4w June

8w July

12 w August

23 w Oct.

4w Dec.

8w Jan.

4.32 c 26.91b 30.44 b 68.55 a 31.39 b

14.68 41.57 34.39 63.67 42.59

5.29 c 14.35 b 12.67 b 31.27 a 37.50 a

6.66 d 20.67 bc 18.15 c 31.54 ab 36.13 a

4.77 d 7.53 cd 8.35 bc 15.26 ab 25.85 a

16.59 d 64.49 bc 81.79 b 151.79 a 52.76c

24.69 c 113.6 b 109.05 b 233.69 a 77.95 b

19.98 d 108.12 b 123.81 b 204.34 a 73.83 c

9.47 c 40.10 b 30.15 b 64.42 a 30.27 b

18.78 c 83.54 b 127.35 ab 173.28 a 102.82 ab

11.44 34.10 40.50 54.21 42.21

c b b a b

3.15 c 9.10 b 9.57 b 16.73a 13.63 ab

14 w March b a a a a

7.99 c 17.32 b 13.14 b 32.03 a 17.00 b

a Roots samples were taken at several time points (weeks) after phosphonate application in fall and summer. Values within columns followed by the same letter do not differ significantly (P < 0.05), according to Fisher's least significant difference (LSD) test.

A. McLeod et al. / Crop Protection 103 (2018) 87e97

93

Table 3 Weekly rate in change of root phosphite concentrations (mg/gFW) in avocado trees receiving different phosphonate treatments at two trial sites (Morgenson and Boschoek) in fall and summer. Treatment

Untreated control Foliar potassium phosphonate Foliar potassium phosphonate ¾ vol. Foliar ammonium phosphonate Trunk injection

Weekly rate of change in phosphite (mg/gFW) concentrationa Fall Boschoek

Summer Boschoek

Fall Morgenson

Summer Morgenson

0.51 cd 0.51 cd 0.88 cd 0.33 cd 1.91 d

0.08 0.81 0.43 1.60 1.61

0.43 5.46 5.25 7.87 2.64

0.90 6.62 9.99 13.4 9.10

cd cd cd d d

cd ab ab a bc

cd e ef f e

a The rate of change was calculated between weeks 12 and 4 after the fall phosphonate applications, and between weeks 14 and 4 after the summer applications. Values in columns and rows followed by the same letter do not differ significantly (P < 0.05), according to Fisher's least significant difference (LSD) test.

Fig. 1. The effect of different phosphonate treatments on avocado fruit phosphite residues at two trial sites (Boschoek and Morgenson). Bars of the same colour followed by the same letter do not differ significantly (P < 0.05) from each other, according to Fisher's least significant difference (LSD) test. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

all the time points (Table 2). The ammonium phosphonate foliar sprays were the best phosphonate treatment with regards to yielding root phosphite concentrations. This treatment was not only often significantly better than the trunk injection, but also yielded significantly higher root phosphite concentrations than the corresponding potassium phosphonate foliar spray. This was true for both trials for almost all the measured time points. The only exception was 8-weeks after the summer applications, where the two treatments did not differ significantly from each other (Table 2). Important observations were made regarding the rate of phosphite change between fall and summer, which differed for the trials. ANOVA analyses of the weekly change in phosphite

concentration showed that there was a significant trial x treatment x season interaction (P < 0.0002). At the Morgenson trial, there was a significant difference between the weekly change in phosphite concentration between fall and summer for all treatments, except the untreated control. In fall, the phosphite concentrations increased from weeks 4e12, whereas in summer it decreased. However, at the Boschoek trial there were no significant differences in the weekly rate in phosphite decline for the fall and summer applications for all treatments (Table 3). Treatments were also investigated for differences in the rate of phosphite change. At Boschoek, there were no significant differences between treatments in the rate in phosphite change. This was also true at Morgenson, except for the ammonium phosphonate

Fig. 2. The effect of a phosphonate trunk injection treatment on six replicates of the treatment on avocado fruit phosphite residues at two trial sites (Morgenson and Boschoek).

94

A. McLeod et al. / Crop Protection 103 (2018) 87e97

foliar spray. This treatment had a significantly higher rate of decline compared to the trunk injection and the full volume potassium phosphonate spray in summer (Table 3). 3.2.2. Phosphite fruit residues There was a significant difference (P < 0.001) in fruit phosphite residues between treatments in both trials. At the Morgenson trial, all the foliar sprays had significantly higher fruit residues than the trunk injection treatment. However, at the Boschoek trial the foliar sprays all had similar, or significantly lower fruit residues than the trunk injection (Fig. 1). Considering only the registered trunk injection treatment, which is indicative of residues that growers will experience, there was considerable variation in fruit residues between replicates in both trials (Fig. 2). Pearson's correlation analyses showed that there were significant correlations between the root phosphite concentrations for all of the root sampling time points and fruit residues considering all treatments. At Boschoek the fruit phosphite concentrations were significantly correlated (P < 0.001) with the root phosphite concentrations for all the time points. Pearson correlation values ranged from 0.595 to 0.743. Correlations between root and fruit phosphite concentrations were also high (0.529e0.774) and significant (P < 0.005) at the Morgenson trial. In both trials, correlation values were comparable between the root phosphite concentrations for the fall and summer time points and fruit residues. 4. Discussion and conclusion The study was able to develop a phosphite extraction- and purification method from avocado roots. The extracted phosphite could be successfully quantified using a LC/MS-MS method that is publically available for quantification of phosphite from avocado fruit. The root phosphite quantification method was used for evaluating the efficacy of foliar-versus trunk injected phosphonate applications under orchard conditions. Phosphite was quantified at several time points following fall and summer applications. This was feasible since the method is high-throughput and relatively cost-effective. Potassium- and ammonium phosphonate foliar sprays were shown to have potential for replacing potassium phosphonate trunk injections based on root phosphite concentrations. Fruit residues, however, may be problematic with foliar sprays when applied in summer and fall. The developed LC/MS-MS method had a better intra-day (within run) ISR precision than inter-day (between different days) precision (reproducibility). The intra-day ISR precision was acceptable since the CV% and DF% values were all below the recommended 15 and 20% respectively. However, the inter-day precision was not within acceptable ranges for 30% of the samples. The precision of the method could potentially be improved in future, through the inclusion of an analogue internal standard (diethyl phosphate). ndez et al. (2003), who also had problems with unacceptable Herna precision for LC/MS-MS phosphite quantification from lettuce leaves, successfully used this approach to solve the problem. In the current study, the high inter-day ISR precision for some samples would not have affected conclusions regarding the efficacy of different application methods, since all treatments from the same time point from both trials were analyzed together within a single run (intra-day analysis). This is feasible due to the high-throughput nature of the LC/MS-MS method. Of all the studies that have reported on plant phosphite quantifications, only two studies reported the inter-day ISR precision of their phosphite quantification method. Hern andez et al. (2003) reported acceptable CV% values of <10% for LC/MS-MS quantifications from lettuce leaves, whereas Alves et al. (2016) reported that their relative standard deviation (synonym of CV%) values were

always equal to or lower than 17% for the LC/MS-MS analyses of soy nutraceuticals. The lack in reporting on precision in other studies, could be due to the fact that several published studies have made use of commercial laboratories for phosphite quantification (Barrett et al., 2004; Groves et al., 2015; Shearer and Crane, 2009, 2012; Shearer et al., 2012). In these commercial laboratories, precision assessments most likely form part of the quality control process. The recovery rate of the developed LC/MS-MS method was 42% across a range of phosphite concentrations. Therefore, phosphite values had to be adjusted for recovery rate. Following recovery rate adjustments, root phosphite concentrations from the current study were comparable to those generated by a commercial laboratory in New Zealand (Hill's laboratory). The low recovery rate of the LC-MS/MS method from the current study was due to a matrix effect, since samples that were spiked post clean-up yielded similar recovery rates than pre-spiked samples. The matrix effect is likely due to the nature of avocado roots, since roots contain compounds such as starch and phenolics that ndez et al., 2003). In may result in matrix ion suppression (Herna contrast to the current study, a high recovery rate (98%) has been reported for the LC/MS-MS analyses of lettuce leaves. However, for lettuce the initial recovery rate was much lower (21%), which was subsequently improved by using a 5-fold sample dilution, and a ndez et al., matrix-matched external standard calibration (Herna 2003). Our recovery rate might, therefore, be improved for avocado roots, if higher dilutions of the samples are conducted. The sensitivity of the assay, however, should not be reduced. The use of matrix-matched external standard calibrations is not deemed feasible for avocado roots, since this approach requires homogendez neity between the blank matrix and analyzed samples (Herna et al., 2003). The matrix of different avocado root samples will likely differ since the samples vary in the ratio of old suberized roots and newly flushing root tips. This mixture of root ages is inevitable to avoid during sampling from orchards at different time points, since root flushes differ throughout the year. Only a few other studies that have quantified phosphite from plant tissues have reported phosphite recovery rates. Ouimette and Coffey (1988) and Borza et al. (2014), using high-performance ion chromatography, reported recovery rates of 70% and 95% for quantification from pepper roots, and potato (leaves and tubers) respectively. For gas chromatography, a recovery rate of 80% has been reported for avocado root and shoot tissues (Nartvaranant et al., 2004; Whiley et al., 1995). Currently, in published literature, there is a general lack of reporting on recovery rates, matrix effects and precision when phosphite is quantified from plant tissues. Furthermore, most articles report phosphite concentrations based on fresh weight, but do not state what moisture content was used in calculations. Since moisture content will vary between samples according to soil moisture levels and transport conditions, it is more accurate to first determine the phosphite concentration based on the dry weight of samples. In the current study, our dry weight phosphite concentrations were converted to fresh weight concentrations using a 60% moisture content. However, a 75% moisture content is used by Hills laboratory for reporting fresh weight phosphite concentrations (personal communication Jill Rumney, Hills laboratory, Hamilton, New Zealand). This will result in lower fresh weight values, although the dry weight values will be similar. All the aforementioned information is important to report, since it can preclude the comparison of phosphite concentrations between studies. Other valuable information that should also be reported on, is whether bulk root systems were analyzed or only root tips. This is important, since it has been reported that phosphite concentrations in root tips are higher than in mature roots (Fairbanks et al., 2000). The limit of detection of the developed LC/MS-MS method in

A. McLeod et al. / Crop Protection 103 (2018) 87e97

spiked root samples was 9.52 mg/gFW. This is higher than the limit of detection of 2 mg/gFW reported by a commercial laboratory (Hill's laboratory) that also uses LC/MS-MS quantification for the analyses of avocado roots. Peer reviewed publications have also reported lower limits of detection for high-performance ion chromatography ranging from 0.5 to 3 mg/gFW (Borza et al., 2014; Ouimette and Coffey, 1988; Roos et al., 1999) and for gas chromatography values were <0.01 mg/gFW to 1 mg/mgDW (Barrett et al., 2003; Pilbeam et al., ndez 2000; Shearer and Crane, 2009; Shearer et al., 2012). Herna et al. (2003) reported a very low limit of quantification of 0.2 mg/ gFW for lettuce leaves using LC/MS-MS analyses. The 9.52 mg/gFW limit of detection for the developed LC/MS-MS method is considered acceptable, based on the reported phosphite threshold levels required for Phytophthora suppression. Van der Merwe and Kotze (1994) reported that suppression of P. cinnamomi in an excised root bioassay required phosphite concentrations of 9.5 mg/gFW or higher. In non-peer reviewed articles in Australia, the threshold root phosphite concentration required for suppression of P. cinnamomi are often referenced as 25e40 mg/gFW. Quantifications were usually conducted by the SGS Australian laboratory (Giblin et al., 2007). However, no scientific publication or information is provided indicating how these threshold values were determined. El-Hamalawi et al. (1995) found that in avocado, 21 mg/gFW or higher bark phosphite concentrations were required for suppression of Phytophthora citricola cankers. Shearer et al. (2012) reported that the amount of phosphite required for 50% inhibition of P. cinnamomi varied widely in plant species and were 26 mg/gDW (6.5 mg/gFW), 149 mg/gDW (37.25 mg/gFW) and 265 mg/gDW (66.25 mg/gFW) for Banksia coccinea, Hakea ferruginea and Allocasuarina humilis respectively. Variation in the reported phosphite threshold concentrations in different host pathogen systems is likely due to the complex mode of action of phosphite. The mode of action can include a direct toxic effect towards the pathogen, and/ or induction of host plant defense responses (Coffey and Bower, 1984; Daniel and Guest, 2006; Fenn and Coffey, 1984, 1985; Guest and Grant, 1991; Grant et al., 1990; Jackson et al., 2000). Although a few studies have reported threshold levels of phosphite required for disease suppression, there is still insufficient data available to assess whether these values are absolute and correct. It will thus be important to determine these values for suppression of P. cinnamomi in avocado roots in future studies. Another important aspect to investigate is whether phosphite concentrations above the threshold level are more effective in pathogen suppression, as suggested by the study of Massoud et al. (2012). Until this information is available, knowledge on the root phosphite concentrations in agricultural crops achieved with registered phosphonate fungicides, is useful for evaluating the efficacy of phosphonates as was done in the current study. The maintenance of threshold levels of phosphite in roots, is likely important for control. Phosphite is known to only inhibit the growth of Phytophthora spp., i.e. is fungistatic, but does not kill the pathogen. For example, it has been shown that the pathogen can still produce zoospores from plant tissue treated with phosphonates (Wilkinson et al., 2001a,b). Therefore, one can expect that the pathogen will become active once phosphite levels decrease below threshold levels (Wilkinson et al., 2001a). These assumptions, however, will require more information on the mode of action of phosphonates, and whether the longevity of host plant defense induction systems (priming) are important. For example, for other plant resistance inducers such as acibenzolar-S-methyl, the compound and its acid derivatives are rapidly degraded in plants. Yet plants subsequently remain protected from pathogens due to the activation of the plant's defense system (Scarponi et al., 2001). Based on the results of the orchard trial from the current study, foliar phosphonate sprays have potential for replacing trunk

95

injections. For the potassium phosphonate sprays, at least five sprays will be required (three in fall, followed by two in summer). The number of required ammonium phosphonate sprays might be less, since this treatment often had significantly higher root phosphite concentrations than the trunk injection treatment. This is the first report on a comparison of root phosphite concentrations obtained with ammonium phosphonate versus potassium phosphonate. In contrast to the current study, Duvenhage (2001) previously reported, based on root phosphite concentrations from one trial, that only two 0.75% phosphonate foliar sprays (one in fall and one in summer) applied at 943 L/ha are required for 8-year-old trees (unspecified size). The number of required foliar sprays from the current study is in agreement with the maximum of five sprays recommended by the Australian emergency use permit, where application volumes of 2000e3000 L/ha are advised for mature trees. The mature trees referred to in the permit is likely for old low density orchards, as compared to the smaller trees from the current study. For smaller trees (3 m height, 6 m canopy dia. and 10 m row width) we found that 1628 L/ha was as effective as 2160 L/ha. Therefore, spray volume can be calculated using the Unrath forcanopy diameter  900 mula: Spray volume ¼ tree height  tree . A reduced row width spray volume will assist in reducing the cost of foliar sprays. The weekly rate of decline in root phosphite concentrations after the summer and fall applications differed for the two orchards. At the Morgenson trial, root phosphite concentrations increased from week 4 to week 12, but decreased significantly after the summer applications. The decline following summer applications can be expected since rapid tree growth, but also fruit set occurs during this period, resulting in a dilution effect in the growing root system. At Boschoek, the root phosphite concentrations mostly showed a decline after summer and fall applications, which did not differ significantly from each other. The reason for the difference between the two trials is unclear, but it might be due to differences in climate and peaks of root flushing. Nonetheless, it is clear that for both trials re-applications are required in fall following the summer applications. Similarly, summer applications are required to supplement the declining root phosphite concentrations from the fall applications. In the current study, there were mostly no significant differences in the rate of phosphite change between the foliar sprays and trunk injection. In contrast, in native Australian species, foliar applications have been reported as not being as long lasting as injections (Hardy et al., 2001) as well as in citrus (Schutte et al., 1991). This variation in persistence of phosphite likely depends on plant species and the rate of application (Hardy et al., 2001). The fruit residues recorded in the two trials were almost all above the MRL value set by the European Union (50 mg/kg). The exception was the trunk injection treatment (32.87 mg/kg), but only for one of the two trials. Although the average fruit residue for this treatment was below the MRL, one of the replicates of this treatment did exceed 50 mg/kg. This would still place a grower at risk for the rejection of fruit consignments in Europe. Furthermore, some German supermarkets have an even lower MRL requirement of 13 mg/kg. The high fruit residues are likely due to the fact that small fruits were present during the summer application in November. Small fruit is a strong sink for not only photosynthesis products, but also phosphonates. Future studies should determine if applications within this time window can be avoided, while still being able to yield the required root phosphite concentrations for a period of 10e12 months. There will likely be a problem in consolidating effective pathogen control through high root phosphite concentrations with low fruit residues, since a significant correlation was observed between root phosphite concentrations and fruit residues. The fact that even the control fruit had phosphite residues in

96

A. McLeod et al. / Crop Protection 103 (2018) 87e97

both trials is difficult to explain. To the best of our knowledge, no phosphonate containing products were applied by growers in the trials. The only explanation could be that growers are unknowingly applying some organic products used for fertilization that may contain phosphite. In Italy, Malusa and Tosi (2005) reported the occurrence of phosphite in a few organic and fertilizer products that subsequently contributed to fruit residues in apple fruit. In conclusion, progress has been made in optimizing foliar sprays for replacing trunk injections in South Africa. Five foliar sprays of ammonium- or potassium phosphonate yielded root phosphite concentrations that were comparable to those of the currently registered trunk injection. However, the fruit residues of foliar sprays, as well as the trunk injection, are unacceptable. The main focus of future research should thus be to try and reduce fruit residue levels, while maintaining acceptable tree health. Determining the threshold root phosphite levels for suppression of P. cinnamomi in avocado will be important, and whether this differs in tolerant and susceptible rootstocks. The developed LC/MS-MS method from this study, along with gene expression studies, will be very useful in these studies. Acknowledgements We would like to thank the South African Avocado Growers' Association, Bertie van Zyl (Edms) Bpk. and the Technology and Human Resources for Industry Programme (THRIPTP13080425506) for financial support of the project. References Alves, R.D., Romero-Gonzalez, R., Lopez-Ruiz, R., Jimenez-Medina, M.L., Frenich, A.G., 2016. Fast determination of four polar contaminants in soy natraceutical products by liquid chromatography coupled to tandem mass spectrometry. Anal. Bioanal. Chem. 408, 8089e8098. Barrett, S.R., Shearer, B.L., Hardy, G.S.J., 2003. The efficacy of phosphite applied after inoculation on the colonisation of Banksia brownii stems by Phytophthora cinnamomi. Australas. Plant Pathol. 32, 1e7. Barrett, S.R., Shearer, B.L., Hardy, G.E., 2004. Phytotoxicity in relation to in planta concentration of the fungicide phosphite in nine Western Australian native species. Australas. Plant Pathol. 33, 521e528. Bezuidenhout, J.J., Korsten, L.I.S.E., Kotze, J.M., 1985. Monitoring phosphorus compounds in avocado tissues. South Afr. Avocado Grow. Assoc. Yearb. 8, 100e102. Black, R.M., Read, R.W., 1998. Analysis of degradation products of organophosphorus chemical warfare agents and related compounds by liquid chromatographyemass spectrometry using electrospray and atmospheric pressure chemical ionisation. J. Chromatogr. A 794, 233e244. Borza, T., Schofield, A., Sakthivel, G., Bergese, J., Gao, X., Rand, J., Wang-Pruski, G., 2014. Ion chromatography analysis of phosphite uptake and translocation by potato plants: dose-dependent uptake and inhibition of Phytophthora infestans development. Crop Prot. 56, 74e81. Coffey, M.D., Bower, L.A., 1984. In vitro variability among isolates of eight Phytophthora species in response to phosphorous acid. Phytopathology 74, 738e742. Crane, C.E., Shearer, B.L., 2014. Comparison of phosphite application methods for control of Phytophthora cinnamomi in threatened communities. Australas. Plant Pathol. 43, 143e149. Dalio, R.J., Fleischmann, F., Humez, M., Osswald, W., 2014. Phosphite protects Fagus sylvatica seedlings towards Phytophthora plurivora via local toxicity, priming and facilitation of pathogen recognition. PLoS One 9. Daniel, R., Guest, D., 2006. Defence responses induced by potassium phosphonate in Phytophthora palmivora-challenged Arabidopsis thaliana. Physiol. Mol. Plant Pathol. 67, 194e201. Darvas, J.M., 1983. Five years of continued chemical control of Phytophthora root rot of avocados. South Afr. Avocado Growers’ Assoc. Yearb. 6, 72e73. Darvas, J.M., Toerien, J.C., Milne, D.L., 1984. Control of avocado root rot by trunk injection with phosetyl-Al. Plant Dis. 68, 691e693. Dunstan, B., Hardy, G., 2005. Control of Phytophthora cinnamomi with phosphite: some recent developments in application methods. Australas. Plant Conserv. 13, 10e11. Duvenhage, J.A., 2001. Efficacy of H3PO3 leaf sprays and resistance of Phytophthora cinnamomi to H3PO3. South Afr. Avocado Growers’ Assoc. Yearb. 24, 13e15. El-Hamalawi, Z.A., Menge, J.A., Adams, C.J., 1995. Methods of fosetyl-al application and phosphonate levels in avocado tissue needed to control stem canker caused by Phytophthora citricola. Plant Dis. 79, 770e778. Fairbanks, M.M., Hardy, G.S.J., McComb, J.A., 2000. Comparisons of phosphite

concentrations in Corymbia (Eucalyptus) calophylla tissues after spray, mist or soil drench applications with the fungicide phosphite. Australas. Plant Pathol. 29, 96e101. FDA, 2013. Guidance for Industry: Bioanalytical Method Validation. Draft Guidance. US Department of Health and Human Services, Food and Drug Administration (CDER and CVM). September 2013. Available from: http://www.fda.gov/ downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ ucm368107.pdf. (Accessed 20 December 2015). Fenn, M.E., Coffey, M.D., 1984. Studies on the In vitro and in vivo antifungal activity of fosetyl-Al and phosphorous acid. Phytopathology 74, 606e611. Fenn, M.E., Coffey, M.D., 1985. Further evidence for the direct mode of action of fosetyl-Al and phosphorous acid. Phytopathology 75, 1064e1068. Fluhler, E., Vazvaei, F., Singhal, P., Vinck, P., Li, W., Bhatt, J., de Boer, T., Chaudhary, A., Masahiro, M., Rezende, V., Zhong, D., 2014. Repeat analysis and incurred sample reanalysis: recommendation for best practices and harmonization from the global bioanalysis consortium harmonization team. AAPS J. 16, 1167e1174. Garbelotto, M., Schmidt, D.J., Harnik, T.Y., 2007. Phosphite injections and bark application of phosphate þ PetrabarkTM control sudden oak death in coast live oak. Arboric. Urban For. 33, 309e317. Giblin, F., Pegg, K., Thomas, G., Whiley, A., Anderson, J., Smith, L., 2007. Phosphonate trunk injections and bark sprays. In: Proceedings VI World Avocado Congress, Vina Del Mar, Chile, ISBN 978-956-17-0413-8, pp. 12e16. Nov. 2007. Grant, B., Dunstan, R., Griffith, J., Niere, J., Smillie, R., 1990. The mechanism of phosphonic (phosphorous) acid action in Phytophthora. Australas. Plant Pathol. 19, 115e121. Groves, E., Howard, K., Hardy, G., Burgess, T., 2015. Role of salicylic acid in phosphite-induced protection against Oomycetes; a Phytophthora cinnamomi e Lupinus augustifolius model system. Eur. J. Plant Patholo 141, 559e569. Guest, D.I., Grant, B.R., 1991. The complex action of phosphonates as antifungal agents. Biol. Rev. 66, 159e187. Hardy, G.E. St. J., Barrett, S., Shrearer, B.L., 2001. The future of phosphite as a fungicide to control the soilborne palnt pathogen Phytophthora cinnamomi in natural ecosystems. Australas. Plant Pathol. 30, 133e139. ndez, F., Sancho, J.V., Pozo, O.J., Villaplana, C., Ib Herna anez, M., Grimalt, S., 2003. Rapid determination of fosetyl-aluminium residues in lettuce by liquid chromatography/electrospray tandem mass spectrometry. J. AOAC Int. 86, 832e838. Hogenboom, A.C., Hofman, M.P., Kok, S.J., Niessen, W.M.A., Brinkman, U.T., 2000. Determination of pesticides in vegetables using large-volume injection column liquid chromatography-electrospray tandem mass spectrometry. J. Chromatogr. A 892, 379e390. Jackson, T.J., Burgess, T., Colquhoun, I., Hardy, G.E.S., Jackson, T.J., 2000. Action of the fungicide phosphite on Eucalyptus marginata inoculated with Phytophthora cinnamomi. Plant Pathol. 49, 147e154. Ma, J., 2016. Phosphite Sensitivity of Phytophthora Cinnamomi and Methods for Quantifying Phosphite from Avocado Roots. MSc thesis. Stellenbosch University, South Africa. Massoud, K., Barchietto, T., Le Rudulier, T., Pallandre, L., Didierlaurent, L., Garmier, M., Saindrenan, P., 2012. Dissecting phosphite-induced priming in Arabidopsis infected with Hyaloperonospora arabidopsidis. Plant Physiol. 159, 286e298. Malusa, E., Tosi, L., 2005. Phosphorous acid residues in apples after foliar fertilization: results of field trials. Food Addit. Contam. 22, 541e548. McKay, A.G., Floyd, R.M., Boyd, C., 1992. Phosphonic acid controls downy mildew (Peronospora parasitica) in cauliflower curds. Animal Prod. Sci. 32, 127e129. McLeod, A., Ma, J., Novela, P., Nyoni, M., Nzanza, B., Stander, M., Pieterse, P., 2015. Investigating root phosphite quantification methods for optimizing foliar spray application in avocado. South Afr. Avocado Grow. Assoc. Yearb. 80, 80e87. Nartvaranant, P., Hamill, S., LeonardI, J., Whiley, A.W., Subhadrabandhu, S., 2004. Seasonal effects of foliar application of phosphonate on phosphonate translocation, In vitro pollen viability and pollen germination in ‘Hass’ avocado (Persea Americana Mill.). J. Hortic. Sci. Biotechnol. 79, 91e96. Orbovic, V., Syvertsen, J.P., Bright, D., Van Clief, D.L., Graham, J.H., 2008. Citrus seedling growth and susceptibility to root rot as affected by phosphite and phosphate. J. Plant Nutr. 31, 774e787. Ott, R.L., 1998. An Introduction to Statistical Methods and Data Analysis. California: Duxbury Press, Belmont, pp. 807e837 (pp. 1e1051). Ouimette, D.G., Coffey, D., 1988. Quantitative analysis of organic phosphonates, phosphonate, and other inorganic anions in plants and soil by using highperformance ion chromatography. Phytopathology 78, 1150e1155. Ouimette, D.G., Coffey, M.D., 1989. Phosphonate levels in avocado (Persea Americana) seedlings and soil following treatment with fosetyl-Al or potassium phosphonate. Plant Dis. 73, 212e215. Pahle, S., 2015. Stepchildren of liberation: South African farm worker's elusive rights to organise and bargain collectively. J. South. Afr. Stud. 41, 121e140. Pegg, K.G., Whiley, A.W., Langdon, P.W., Saranah, J.B., 1987. Comparison of fosetylaluminium, phosphorous acid and metalaxyl for the long-term control of Phytophthora root rot of avocado. Aust. J. Exp. Agric. 27, 471e474. Pilbeam, R.A., Colquhoun, I.J., Shearer, B., Hardy, G.E.St. J., 2000. Phosphite concentration: its effect on phytotoxicity symptoms and colonisation by Phytophthora cinnamomi in three understorey species of Eucalyptus marginata forest. Australas. Plant Pathol I29, 86e95. Reed, G.F., Lyn, F., Meade, B.D., 2002. Use of coefficient of variation in assessing variability of quantitative assays. Clin. Diagn. Lab. Immunol. 9, 1235e1239. Robbertse, P.J., Duvenhage, J.a, 1999. Reaction of avocado wood on injections into the trunk. South Afr. Avocado Grower’s Assoc. Yearb. 22, 39e47.

A. McLeod et al. / Crop Protection 103 (2018) 87e97 Roos, G.H.P., loane, C., Dell, B., Hardy, J.H.G.S., 1999. Facile high performance ion chromatographic analysis of phosphite and phosphate in plant samples. Commun. Soil Sci. Plant Analysis 30, 22332e23229. Scarponi, L., Buonaurio, R., Martinetti, L., 2001. Persistence and translocation of a benzothiadiazole derivative in tomato plants in relation to systemic acquired resistance against Pseudomonas syringae pv. tomato. Pest Manag. Sci. 57, 262e268. Schutte, G.C., Bezuidenhout, J.J., Kotze, J.M., 1991. Timing of application of phosphate fungicides using different application methods as determined by means of gasliquid-chromatography for Phytophthora root rot control of citrus. Phytophylactica 23, 69e71. , J.M., 1988. Distribution of phosSchutte, G.C., Botha, T., Bezuidenhout, J.J., Kotze phite in avocado trees after trunk injection with phosphorous acid and its possible response to Phytophthora cinnamomi. South Afr. Avocado Growers' Assoc. Yearb. 12, 32e34. Shapiro, S.S., Francia, R.S., 1972. An approximate analysis of variance test for normality. J. Am. Stat. Assoc. 67, 215e216. Shearer, B.L., Crane, C.E., 2009. Influence of site and rate of low-volume aerial phosphite spray on lesion development of Phytophthora cinnamomi and phosphite persistence in Lambertia inermis var. inermis and Banksia grandis. Australas. Plant Pathol. 38, 288e304. Shearer, B.L., Crane, C.E., Scott, P.M., Hardy, G.S., 2012. Variation between plant species of in-planta concentration and effectiveness of low-volume phosphite spray on Phytophthora cinnamomi lesion development. Australas. Plant Pathol. 41, 505e517. Shearer, B.L., Crane, C.E., 2012. Variation within the genus Lambertia in efficacy of low-volume aerial phosphite spray for control of Phytophthora cinnamomi. Australas. Plant Pathol. 41, 47e57. Smillie, R.H., Grant, B., Cribbes, R.L., 1988. Determination of phosphate and phosphite in plant material by gas chromatography mass spectrometry and ion chromatography. J. Chromatogr. 455, 253e261. Smillie, R., Grant, B.R., Guest, D., 1989. The mode of action of phosphite e Evidence for both direct and indirect modes of action on three Phytophthora species in plants. Phytopathology 79, 921e926. Smith, L.A., Dann, E.K., Pegg, K.G., Whiley, A.W., 2010. Management of Phytophthora cinnamomi in Australian avocado orchards. In: 6th Australasian Soilborne

97

Disease Symposium, pp. 9e10. Twin Waters, Queensland, Australia, August. Snedecor, G.W., Cochran, W.G., 1980. Statistical Methods, seventh ed. The Iowa State University Press, Ames. Subramaniam, S., Patel, D., Davit, B.M., Conner, D.P., 2015. Analysis of imprecision in incurred sample reanalysis for small molecules. AAPS J. 17, 206e215. Thao, H.T.B., Yamakawa, T., Shibata, K., Sarr, P.S., Myint, A.K., 2008. Growth response of komatsuna (Brassica rapa var. peruvirdiso) to root and foliar applications of phosphite. Plant Soil 308, 1e10. Thomas, F.M., 2008. Recent advances in cause-effect research on oak decline in Europe. CAB Rev. Perspect. Agric. Veterinary Sci. Nutr. Nat. Resour. 3, 1e12. Torres Elguera, J.C., Yan~ez Barrientos, E., Wrobel, K., Wrobel, K., 2013. Monitoring of phosphorus oxide ion for analytical speciation of phosphite and phosphate in transgenic plants by high-performance liquid chromatographyeinductively coupled plasma mass spectrometry. J. Agric. Food Chem. 61, 6622e6628. Unrath, C.R., Sutton, T.B., Obermiller, J.D., Williams, K.M., 1986. A Tree-row-volume Model for Determining Proper Spray Rates in Apple Orchards. NC Apple Grow Assoc Publ001, 8pp. Van der Merwe, M., Kotze, J.M., 1994. Fungicidal action of phosphite in avocado root tips on Phytophthora cinnamomi. South Afr. Growers' Assoc. Yearb. 17, 38e45. Whiley, A.W., Hargreaves, P.A., Pegg, K.G., Doogan, V.J., Ruddle, L.J., Saranah, J.B., Langdon, P.W., 1995. Changing sink strengths influence translocation of phosphonate in avocado trees (Persea Americana Mill.) trees. Aust. J. Agric. Res. 46, 1079e1090. Whiley, A.W., Leonardi, J., Pegg, K.G., Langdon, P.W., 2001. Use of foliar applications of phosphonate fungicides to control Phytophthora root rot in avocado. In: Proceedings of the Australian and New Zealand Avocado Growers' Conference ‘Vision 2002 (Brisbane, New Zealand). Wilkinson, C.J., Holmes, J.M., Tynan, K.M., Colquhoun, I.J., McComb, J.A., Hardy, G.S.J., Dell, B., 2001a. Ability of phosphite applied in a glasshouse trial to control Phytophthora cinnamomi in five plant species native to Western Australia. Australas. Plant Pathol. 30, 343e351. Wilkinson, C.J., Holmes, J.M., Dell, B., Tynan, K.M., McComb, J.A., Shearer, B.L., Colquhoun, I.J., Hardy, G.E. St. J., 2001b. Effect of phosphite on in planta zoospore production of Phytophthora cinnamomi. Plant Pathol. 50, 587e593. Yadav, M., Shrivastav, P.S., 2011. Incurred sample reanalysis (ISR): a decisive tool in bioanalytical research. Bioanalysis 3, 1007e1024.