The dominant modes of action of macrocidins, bioherbicidal metabolites of Phoma macrostoma, differ between susceptible plant species

Environmental and Experimental Botany 132 (2016) 80–91

Contents lists available at ScienceDirect

Environmental and Experimental Botany journal homepage: www.elsevier.com/locate/envexpbot

The dominant modes of action of macrocidins, bioherbicidal metabolites of Phoma macrostoma, differ between susceptible plant species M. Hubbarda , W.G. Taylorb , K.L. Baileyc , R.K. Hynesa,* a b c

Agriculture and Agri-Food Canada, 107 Science Place, Saskatoon, SK, S7N 0X2, Canada 111 Staigh Crescent, Saskatoon, SK, S7N 3T2, Canada Box 143 Heriot Bay, BC, V0P 1H0, Canada

A R T I C L E I N F O

Article history: Received 26 May 2016 Received in revised form 23 August 2016 Accepted 24 August 2016 Available online 28 August 2016 Keywords: Bioherbicide Phytotoxin Carotenoids Chlorophyll fluorescence Modes of action

A B S T R A C T

The bioherbicidal fungus Phoma macrostoma produces macrocidins, which induce chlorosis in susceptible plant species by interfering with the carotenoid biosynthesis. Macrocidins inhibit the enzyme phytoene desaturase (PDS) and likely act on other components of the carotenoid biosynthetic pathway. It was hypothesized that macrocidins’ mode(s) of action differ between susceptible plant species that respond differently to these phytotoxins. This idea was tested by exploring the impact of macrocidins on the carotenoid and carotenoid precursor profiles over time, as well as chlorophyll fluorescence parameters, symptom severity and biomass, in dandelions, groundsel and chickpea. While PDS was inhibited in all three plants, this impact was strongest in dandelion and weakest in groundsel. However, groundsel showed the most severe macrocidin-induced symptoms and biomass reduction. In solution, macrocidin A bound iron and magnesium cations. Macrocidin-induced changes in OJIP chlorophyll fluorescence parameters are consistent with inhibited electron transfer from QA to QB (potentially due to iron-binding), and uncoupling the light-harvesting complex (LHC) of photosystem II (PSII) from the reaction centre, leading to an increase in photoprotective xanthophylls. This latter impact was stronger in groundsel than chickpea or dandelion. The decrease in total carotenoid and carotenoid precursor content in macrocidin-treated plants could be explained by macrocidins’ magnesium-binding activity reducing the efficiency of 1-deoxy-D-xylulose (DXP) reductoisomerase (DXR) and/or phytoene synthase (PSY). The putatively metal-binding-related modes of action of macrocidins occurred more rapidly than inhibition of PDS. These inter-specific variations in, and diversity of, modes of action suggest that the risk of resistance developing to macrocidins is very low. Crown Copyright ã 2016 Published by Elsevier B.V. All rights reserved.

1. Introduction Phoma macrostoma isolate 94–44B is registered as a bioherbicide in Canada and United States for the suppression of broadleaf

Abbreviations: ABA, abscisic acid; a.i., active ingredient; ANOVA, analysis of variance; DF, diflufenican; DXP, 1-deoxy-D-xylulose; DXR, 1-deoxy-D-xylulose reductoisomerase; ESI–MS, electrospray ionization mass spectrometry; HPLC, high-performance liquid chromatography; LC, liquid chromatography; LCY-b, lycopene b-cyclase; LHCII, light harvesting complex of photosystem II; LSD, least significant difference; MU, macrocidin units; NIC, nicotine; NIC+DF, nicotine and diflufenican; PDS, phytoene desaturase; PSII, photosystem II; PSY, phytoene synthase. * Corresponding author. E-mail addresses: [email protected] (M. Hubbard), [email protected] (W.G. Taylor), [email protected] (K.L. Bailey), [email protected] (R.K. Hynes). http://dx.doi.org/10.1016/j.envexpbot.2016.08.009 0098-8472/Crown Copyright ã 2016 Published by Elsevier B.V. All rights reserved.

weeds in turfgrass (Bailey and Derby, 2001). Macrocidins, polyketide secondary metabolites of P. macrostoma (Graupner et al., 2003), are able to induce photobleaching in susceptible plants, even in the absence of the living fungus (Hubbard et al., 2014) by inhibiting carotenoid biosynthesis. Hubbard et al. (2015) determined that macrocidins inhibit the enzyme phytoene desaturase (PDS) in the carotenoid biosynthesis pathway in the susceptible weeds (dandelion and thistle) but not in the resistant plants (pumpkin and wheat). However, Hubbard et al. (2015) also observed that macrocidin-treatment led to a drop in the b-carotene to lutein ratio and total carotenoid and carotenoid precursor content not seen in plants treated with diflufenican (DF), a PDS inhibitor (Wightman and Hayes, 1985). These findings suggest that macrocidins have additional modes of action. By examining the changes in carotenoid profiles over time, rather than at a single post-treatment time point, in plants treated with

M. Hubbard et al. / Environmental and Experimental Botany 132 (2016) 80–91

macrocidins or synthetic herbicides with known modes of action, it was hoped these additional modes of action could be elucidated. Nicotine (NIC), a known inhibitor of lycopene b-cyclase (LCY-b) (Howes, 1974), was used as a control to determine if macrocidins inhibit the activity of this enzyme, testing the hypothesis that LCYb inhibition contributes to the change in b-carotene to lutein ratio observed by Hubbard et al. (2015). NIC-treatment was also combined with the application of DF (the combined treatment will be referred to as NIC+DF), which is known to inhibit PDS (Wightman and Hayes, 1985), to assess if the combination of these two modes of action could mimic that of macrocidins. Macrocidins are structurally similar to other natural products with metal-chelating capacity (Schobert and Schlenk, 2008). Hence, it is possible that metal binding could play a role in the biological activity of macrocidins. Iron plays critical roles in the photosynthetic apparatus (see review by Krohling et al. (2016)), particularly in the passing of electrons from plastoquinone A (QA) to plastoquinone B (QB) and in the Rieske FeS protein of the cytochrome b6/f complex (see Fig. 2 in Stirbet et al. (2014)). Plants grown with insufficient iron display chlorosis in new growth, decreased total chlorophyll and carotenoid content, as well as changes in carotenoid composition (Abadia et al., 1999; Larbi et al., 2004; Morales et al., 1990), reminiscent of the symptoms induced by macrocidin treatment. While magnesium deficiency can also lead to chlorosis, these symptoms tend to appear in older growth first, unlike macrocidin-induced yellowing. Magnesium forms the centre of chlorophyll, meaning that a lack of availability of this element can lead to breakdown of chlorophyll in older leaves in order to make it available for new growth (see Verbruggen and Hermans (2013) for a review). Chlorophyll fluorescence parameters, reviewed by Rohacek (2002), are useful for measuring the type and severity of stress experienced by plants. Fv/Fm is a measure of the maximum darkadapted quantum efficiency of photosystem II (PSII) (Kitajima and Butler, 1975). Fv is a measure of variable fluorescence, which is equal to Fm (maximum fluorescence) minus FO (initial fluorescence). Lower Fv/Fm values indicate a greater degree of damaged to PSII (Farquhar et al., 1989). F'v/F'm, or effective photosynthetic yield, is the same as Fv/Fm, except that is it measured in lightadapted, or steady-steady, photosynthetic organisms (reviewed by Rohacek (2002)). Decreases in Fv/Fm and/or F'v/F'm have been measured in iron-deficient plants (Abadia et al., 1999; Bertamini et al., 2001; Larbi et al., 2004). The transient fluorescence of chlorophyll, measured by the OJIP transient (Strasser et al., 2000), has been used to evaluate the level of damaged to electron transport chains in plants under adverse conditions, including iron deficiency (Jiang et al., 2007), herbicide, heat, salt (Percival, 2005) and potential bioherbicide (Chen et al., 2015) stress. Because they include points in the fluorescent transient other than the initial fluorescence (O, which is equivalent to FO) and maximum fluorescence (P, or FP, equal to Fm in Fv/Fm) (see Stirbet et al. (2014) for a review), the OJIP parameters can also yield information about the type of inhibition experienced by the electron transport chain (Chen et al., 2005; Chen et al., 2007; Chen et al., 2008; Hirakil et al., 2003). The symptom severity and percent mortality vary between plant species treated with P. macrostoma (Bailey et al., 2011a). Dandelions exhibited both chlorosis and death as a result of treatment with P. macrostoma, while groundsel displayed only mortality and chickpea only chlorosis. The hypothesis that macrocidins act by different modes of action on dandelion, groundsel and chickpea, which respond differently to treatment with macrocidins was formulated. The impact and mode(s) of action of macrocidin-treatment were assessed in terms of chlorotic symptoms, biomass, chlorophyll fluorescence parameters and carotenoid profiles over time.

81

2. Materials and methods 2.1. Experimental design The experiments were carried out twice, in 2014 and 2015, with four replicates in each year. Plants were arranged according to a randomized block design. Three plants species, dandelion, groundsel and chickpea, were used. The treatments applied to dandelions were untreated control, 32 macrocidin units (MU) per pot, 128 MU, 256 MU, nicotine alone (NIC), diflufenican (DF) alone and NIC and DF in combination (NIC+DF). Groundsel and chickpeas were subjected to the following four treatments: control (zero), 32, 128 and 256 MU. The macrocidin doses used were equivalent to 1/4, 1 and 2 the label rate of 64 g/m2 (based on a 200 MU/g granular product) for post-emergent weed control. The 1/4 post-emergent rate is also the pre-emergent recommended dose (Bailey et al., 2011a). 2.2. Plant material Chickpeas (Cicer arietinum, Sanford kabuli, a gift from Sabine Banniza (Professor, University of Saskatchewan, Saskatoon, SK, Canada)) were planted at a rate of five (first experiment) or seven (second experiment) seeds per pot. Twenty dandelion seeds were planted in each pot (Richters, Goodwood, ON, Canada). Groundsel seeds, collected in Saskatoon, SK, in 2011, were planted 25 seeds to a pot. Prior to treatment, dandelions, groundsel and chickpeas were thinned to five to six, five and three to five plants per pot, respectively. Treatments were applied to four-week-old groundsel and chickpeas and five-week-old dandelions. All plants were grown in soil-less mix in 10 cm  10 cm pots in a greenhouse at the AAFC Saskatoon Research Centre as described by (Hubbard et al., 2015). A track sprayer (Halltech Ag GPS, Guelph, ON, Canada) with a TeeJet 8002 XR8002 nozzle (Spraying Systems Co., Wheaton, IL, USA) 40 cm (2014) or 51 cm (2015) above the surface of the soil-less mix was used to apply NIC and DF. NIC was applied at 100 mmol/ pot, which is the equivalent of 16.2 kg active ingredient (a.i.)/ hectare in a total volume of approximately 0.7 mL/pot in the first experiment. In the second experiment, 90 mmol/pot (14.6 kg a.i./ hectare) was applied in 1.1 mL. In the first and second experiments, 50 and 45 mg of DF was applied per pot (the equivalent of 100 and 90 mL/ha or 50 and 45 g a.i./hectare) in a total volume of approximately 0.7 mL and 1.05 mL. Plants treated with both NIC +DF received herbicide in a total volume of roughly 1.4 mL/pot (first experiment) and 2.1 mL/pot (second experiment). A 1-touch plant mister (Cepia LLC, St. Louis Mo. 63124) was used to spray macrocidins onto leaves and soil surface. Groundsel and chickpea plants were enclosed in plastic sleeves during this process and for several weeks thereafter in order to retain run-off. 2.3. Macrocidins and chemicals Macrocidins were produced by P. macrostoma 94-44B grown in a fermenter, prepared and quantified as described by Hubbard et al. (2015). Dandelions treated with 32, 128 and 256 MU per pot in 2014 received 1.8, 7.2 and 14.4 mL/pot, respectively, of the concentrated solution of batch SCO-11-F12 (17.8 MU/mL), diluted to 15 mL/pot. Batch 2013-01-22 contained 7.6 MU/mL and was applied to groundsel and chickpeas in 2014 at 4.18 mL (32 MU), 16.75 mL (128 MU) and 33.50 mL (256 MU) per pot. Batch 2014-05-12 was applied to dandelion, groundsel and chickpeas in 2015 at 2.1 mL/ pot for 32 MU, 8.2 mL/pot for 128 MU and 16.4 mL/pot for 256 MU of the 15.6 MU/mL solution, diluted to 17 mL/pot.

82

M. Hubbard et al. / Environmental and Experimental Botany 132 (2016) 80–91

DF was formulated and obtained as described by (Hubbard et al., 2015). NIC was obtained as a liquid formulation of ( )-nicotine from Sigma-Aldrich (product number N3876). 2.4. Metal binding by macrocidin A Macrocidin A was purified in two main steps from a pooled sample of P. macrostoma, batch 2003-STD-44B, grown on barley (Hordeum vulgare, Tercel) grain. Typically, 80 g of barley grain was soaked in water and autoclaved before being inoculated with a 15 mL sub-sample of one or two 9 cm Petri plates of P. macrostoma homogenized in 300 mL of sterile water and plated in 250 mL glass Bernard canning jars for 14-21 days. Diaion HP-20 resin was used with ethanol as solvent, yielding 0.77 g of enriched macrocidins in the third ethanol wash from 50 g of grain culture. This preliminary purification step was followed by medium pressure liquid chromatography (LC) (AKTAExplorer 100 LC system from Amersham Biosciences Inc.) with a Sephasil Peptide C18 steel column (4.6  250 mm, 12 mm particle size, also from Amersham). The mobile phase was 0.1% formic acid in water (eluent A) and a 9:1 (v:v) mixture of acetonitrile in eluent A (eluent B). The gradient was 5% B to 58% B in 1.5 column volumes (4.11 min), 58% B to 62.5% B in 4.5 column volumes (12.33 min) and 62.5% B to 100% B in 1.5 column volumes. A total of 155 mg of the Diaion HP-20 ethanol (third) fraction was injected over 8 runs with the AKTA LC. This yielded 2.98 mg of macrocidin A. In theory, 50 g of grains culture (with a high-performance liquid chromatography (HPLC) peak area ratio of macrocidin A to decanophenone of approximately 10) will yield 14.8 mg of macrocidin A by these techniques. Purified macrocidin A (1.58 mg) was dissolved in 60% ethanol (5 mL) to obtain a stock solution at a concentration of 316 mg/mL (0.89 mmol/mL). For electrospray ionization mass spectrometry (ESI–MS), this solution was diluted to 100 mg/mL with 60% ethanol before the addition of an equal volume of an aqueous solution of formic acid (1 mg/mL). Stock solutions of ferric chloride [iron(III) chloride hexahydrate, 97%, Sigma-Aldrich] and magnesium chloride (hexahydrate, 99.995%, Sigma-Aldrich) were prepared in 60% ethanol at 9.94 and 10.74 mg/mL, respectively. Both stock solutions were diluted 100-fold with 60% ethanol to give working solutions of 99.4 mg/mL (0.37 mmol/mL) and 107.4 mg/mL (0.53 mmol/mL). Experimental solutions for ESI–MS were prepared by adding the working solution of ferric chloride (600 mL) or magnesium chloride (640 mL) to the stock solution of macrocidin A (840 or 890 mL), diluting to 2.5 mL with 60% ethanol and mixing with 2.5 mL of aqueous formic acid (1 mg/mL). Solutions were delivered to a benchtop triple quadrupole mass spectrometer (Quattro LC from Micromass (Manchester, UK) equipped with an atmospheric pressure electrospray ionization source and MassLynx ver. 4) with a syringe attached to an infusion pump (KD Scientific, model 100) at a flow of 0.02 mL/min. Nitrogen gas was employed for nebulization and desolvation. During infusion, the capillary voltage was set at 3.17 kV, the cone voltage at 40–41 V, the source temperature at 100  C and the desolvation temperature at 200  C. Mass range was 100–1200 Da and scan duration 1 s.

The fresh aerial biomass of groundsel and chickpea was measured at 21 days post-treatment, when the plants were 49 days old, while that of dandelions was measured 20 days after treatment, when plants were 55 days old, on a Mettler Toledo PG603-S or ML3002E balance. 2.6. Chlorophyll fluorescence Chlorophyll fluorescence measurements were taken with an OS5p chlorophyll fluorometer (Opti-Sciences, Inc., 8 Winn Avenue Hudson, NH 03051, USA). Plants were dark-adapted for at least 20 min prior to collection of Fv/Fm and OJIP parameter data. Chickpea and groundsel plants were 38 days old (10 days posttreatment) when chlorophyll fluorescent data was collected, while dandelions were 38 and 48 days old (3 and 13 days posttreatment). Dandelion OJIP data was gathered at 13, but not 3, days after treatment. F'v/F'm and Fv/Fm information was recorded for dandelions at both time points. One symptomatic leaf per plant and three plants per pot were selected for measurement where possible. The youngest leaf that was large enough to facilitate fluorometer clip placement was chosen. Only the O, J, I, P and T fluorescent parameters, rather than the entire OJIP transient, were recorded. Initial fluorescence, FO, is equivalent to O. Similarly, FJ, FI, FP and FT represent the fluorescence at J, I, P and T. Maximum fluorescence (Fm) is equivalent to FP. Relative variable fluorescence at the J step (VJ), was calculated according to the formula VJ = (FJ Fo)/(Fm Fo). An increase in VJ data indicates that electrons are not being passed beyond QA. VJ was calculated both for all available data points and excluding data from leaves in which the OJIP transient was flat (Fm FO < 20), indicating near complete breakdown of the electron transport chain. VJ data was separated into these two categories because information on the type of damage to PSII could be lost, due to extreme variability of VJ values obtained by examining severely damages leaves. Quantum yield of electron transport through PSII, fEo, was calculated using the formula fEo = (1 – FO/Fm) (1 – VJ). 2.7. Carotenoid extraction and quantification Leaf samples for carotenoid extraction were taken from symptomatic leaves, if any were present, from at least two plants per pot at one, three, six, 10, 13 and 17 days after treatment (dandelion) and one, three, nine and 17 days post-treatment (groundsel and chickpeas). Leaves were placed in a pre-weighed cryovial, weighed and flash-frozen in liquid nitrogen. Samples were stored at 80  C for future carotenoid and carotenoid precursor extraction. Carotenoids were extracted according to the procedure described by Hubbard et al. (2015). All the carotenoid standards employed by Hubbard et al. (2015) were used in this study. In addition, a lycopene standard (CaroteNature, Gerbestrasse 12, 3072 Ostermundigen, Switzerland) was quantified at 500 nm and used to generate a standard curve. Lycopene was particularly challenging to work with because it was prone to degradation as evident from detection of minor HPLC peaks. 2.8. Statistical analysis

2.5. Symptoms and biomass Symptoms severity was rated on a scale of 0–5, with 0 corresponding to the absence of any photobleaching-related symptoms and a rating of 5 indicating that all plants were dead. Ratings 1, 2, 3 and 4 indicate some chlorotic (yellow) plants, some yellow and some white plants, many white plants with some necrosis and many necrotic plants, with some white plants, respectively.

Statistical analysis was performed using SAS (Version 9.3). Data from both experiments was pooled. Treatments were compared by analysis of variance (ANOVA), followed by least significant difference (LSD) analysis. For data, such as chlorophyll fluorescence parameters, where multiple measurements were taken per replicate, an average of all data from a given replicate was used for statistically analysis to reduce pseudo-replication, or overestimation of sample size. Percentage data (with the exception of

M. Hubbard et al. / Environmental and Experimental Botany 132 (2016) 80–91

83

Fig. 1. Ratings of symptom severity (0–5), where 0 indicates no symptoms and 5 means that all plants were dead, in dandelion, groundsel and chickpea. For a given plant species, treatments with the same lower-case letter are not significantly different (p  0.05). Upper-case letters compare groundsel and chickpea for a given treatment and time point; points with the same letter are not significantly different (p  0.05).

gamma-carotene data collected from dandelion 17 days posttreatment) was subjected to an arcsine transformation before statistical analysis. Ratios of b-carotene to lutein and total carotenoid and carotenoid precursor content were transformed according to the formulae y = log(x + 10) or y = log(x), respectively, prior to statistical analysis. Groundsel and chickpea were compared in terms of symptom severity, F'v/F'm, Fv/Fm, b-carotene to lutein ratios and total carotenoid and carotenoid precursor content. The latter two parameters were divided by the average value for that metric for control plants of that species for statistical analysis. Differences were considered statistically significant if the associated p-value was  0.05. 3. Results Throughout the results, comparisons of treatment that were not significantly different from one another have been omitted. 3.1. Symptoms and biomass At three days post-treatment, dandelions exhibited the more severe symptoms when treated with NIC+DF than when subjected to other treatments (Fig. 1). The symptoms of dandelions exposed to DF were less severe. Macrocidin- and NIC-treated plants displayed still milder symptoms. By both 13 and 17 days after treatment, dandelions treated with the higher two doses of macrocidins, DF or NIC+DF all showed similarly severe symptoms. Plants treated with 32 MU or NIC had less severe symptoms (Fig. 1). Groundsel symptoms were absent in control, significantly more severe for 32 MU-treated plants and higher still for plants exposed

to 128 and 256 MU. However, by 17 days after treatment, groundsel challenged with 256 MU showed more severe symptoms than their 128 MU-treated counterparts (Fig. 1). Chickpea symptom severity increased in a macrocidin-dose-dependent manner at both nine and 17 days post-treatment, with the exception that 32 MU- and 128 MU-treated plants did not differ from each other at 9 days. Treatment with 128 or 256 MU resulted in more severe symptoms in groundsel than chickpea at 17, but not nine, days post-treatment (Fig. 1). The aerial biomass of dandelions exposed to 32 MU, 128 MU, 256 MU, DF or NIC+DF was less than that of controls (Fig. 2). Groundsel exhibited a statistically significant drop in fresh biomass with each increase in macrocidin dose (Fig. 2). 256 MU-treated chickpeas had a biomass significantly less than control or 32 MUtreated plants (Fig. 2). 3.2. Chlorophyll fluorescence Three days after treatment, F'v/F'm values for DF-treated dandelions were significantly lower than those of untreated plants (Fig. 3). At 13 days post-treatment, F'v/F'm values were significantly lower in plants treated with DF or NIC+DF, than in their control or NIC-treated counterparts. At three days after treatment, the F'v/F'm values of dandelions treated with any dose of macrocidins or NIC+DF were significantly lower than the other three treatments at three days post-treatment. By 13 days after treatment, a macrocidin-dose-dependent drop in F'v/F'm was measured (Fig. 3). DF-treated dandelions had Fv/Fm values significantly lower than untreated plants at both three and 13 days post-treatment. Three days after treatment, Fv/Fm values of

Fig. 2. Fresh above-ground biomass. For a given plant species, treatments with the same letter are not significantly different (p  0.05).

84

M. Hubbard et al. / Environmental and Experimental Botany 132 (2016) 80–91

Fig. 3. Photosystem II (PSII) maximum quantum yield (Fv/Fm) and effective quantum yield (F'v/F'm) in dandelions. For a given time point and parameter, treatments with the same letter are not significantly different (p  0.05).

dandelions exposed to both NIC+DF were significantly lower than their DF-treated counterparts. Thirteen days after treatment, the Fv/Fm values of 32 MU-treated dandelions were less than those of NIC+DF-treated plants. Dandelions treated with 128 MU had Fv/Fm values significantly lower than their 32 MU-treated counterparts and higher than those treated with 256 MU (Fig. 3). Groundsel had lower F'v/F'm values 10 days post-treatment with 32 MU as compared to the untreated controls (Fig. 4). F'v/F'm values of plants treated with 128 or 256 MU were significantly lower than those of their 32 MU-treated counterparts. Ten days after treatment, Fv/Fm values were significantly decreased in groundsel treated with 32 MU relative to the control. Groundsel treated with 128 or 256 MU had Fv/Fm values that were below those of untreated or 32 MU-treated plants (Fig. 4). Both light and dark adapted photosynthetic yields (F'v/F'm and Fv/Fm) were significantly lower in chickpeas treated with 32 MU than in their untreated counterparts. Chickpeas exposed 128 or 256 MU were significantly lower than control or 32 MU-treated plants. In terms of both F'v/F'm and Fv/Fm, groundsel was more

severely impacted by the treatment with 32 or 128 MU, but not 256 MU, as compared to chickpea (Fig. 4). Treatment with macrocidins induced a flattening of the OJIP fluorescent transient in dandelions that became more pronounced with increasing dose (Table 1). Initial fluorescence values (O) were significantly higher in 32 MU- or 128 MU-treated dandelions than in plants exposed to any of the other treatments. P and T values of dandelions exposed to two lower doses of macrocidins were lower than in controls. Dandelions treated with DF or NIC+DF had OJIP transients that were somewhat flattened, but started at the same initial fluorescence as the control. Groundsel treated with macrocidins also displayed a flattened OJIP transient. The greatest increase in O values was observed in groundsel exposed to the lowest dose of macrocidins. In 32 MU-treated groundsel, J, I, P and T values were all higher than in controls. However, in groundsel exposed to 128 MU, only O and J values were above those of the control. For 256 MU-treated groundsel, P and T values were lower than the control (Table 1). O values were higher for chickpea treated with all three doses of macrocidins, relative to the control.

Fig. 4. Photosystem II (PSII) maximum quantum yield (Fv/Fm) and effective quantum yield (F'v/F'm) in groundsel and chickpea. For a given plant species and time point, treatments with the same lower-case letter are not significantly different (p  0.05). Upper-case letters compare Fv/Fm or F'v/F'm between groundsel and chickpea for a given treatment; bars with the same letter are not significantly different (p  0.05).

M. Hubbard et al. / Environmental and Experimental Botany 132 (2016) 80–91

85

Table 1 OJIP parameters in dandelion, groundsel and chickpea. For a given parameter (O, J, I P, T, variable fluorescence at the J step (VJ) or quantum yield of electron transport through photosystem II (PSII), fEo), treatments with the same letter are not significantly different (p  0.05). Treatment

O

J

I

P

T

VJ

VJf

fEo

Dandelion Control 32 MU 128 MU 256 MU Nicotine Diflufenican Nicotine and diflufenican

4839a 898162b 966193b 472123a 52617a 47562a 38441a

84523abc 1200212c 1137220c 492126a 88635bc 63184ab 60396ab

125739b 1363242b 1199232b 501127a 121954b 746103a 771133a

213934e 1609283cd 1253242bc 489123a 207588de 939144ab 1054176b

204043e 1550280cd 1202237bc 462118a 194493de 895136ab 1004171b

0.2190.010a 0.5820.056a 0.6380.237a 1.1800.531a 0.2330.007a 0.3770.033a 0.3390.047a

0.2190.010a 0.5540.047cd 0.6250.043d 0.8660.146e 0.2330.007ab 0.4310.049cd 0.3630.035bc

0.6050.008a 0.1950.036d 0.0540.019e 0.0010.010e 0.5700.009a 0.2650.033c 0.3420.035b

Groundsel Control 32 MU 128 MU 256 MU

44489a 1557171c 1269188bc 865101ab

722143a 1861159c 1371193b 932114ab

984150a 1986132a 1399193a 943115a

1742104b 217571c 1385190b 926118a

1656108b 204362c 1315185b 838100a

0.2200.052a 0.8470.423a 0.6340.844a 0.3440.797a

0.2200.052a 0.6470.099b 1.2710.212c 1.2090.148c

0.6010.056a 0.1450.084b 0.0080.032c 0.0150.009c

Chickpea Control 32 MU 128 MU 256 MU

3745a 801107b 940113b 771159b

5489a 1031134a 1106129a 899211a

73716a 1203150a 1188138a 938225a

177846b 1707211b 1385115ab 955238a

156449b 1572201b 1245117ab 881231a

0.1250.007a 0.4540.126ab 0.1880.312a 0.8900.166b

0.1250.007a 0.3700.077b 0.5160.067b 0.7700.078c

0.6900.008a 0.3430.087b 0.1530.045c 0.0350.019c

f

Excluding leaves for which Fm

FO < 20.

P and T values were lower for 256 MU-treated chickpea than for controls. VJ values were higher in macrocidin-treated plants relative to the control, when leaves with Fm FO < 20 were excluded. When VJ values were calculated from all leaves 256 MUtreated chickpeas had higher VJ values than untreated plants (Table 1). Quantum yield of electron transport through PSII, fEo, was lower in macrocidin-treated plants than their untreated counterparts. Dandelions treated with DF or NIC+DF, also had lower fEo values than controls (Table 1). 3.3. Iron and magnesium binding properties of macrocidin A A purified sample of macrocidin A in acidic aqueous ethanol solution showed a strong molecular ion (macrocidin A + H)+ at m/z 358 under positive electrospray conditions (Fig. 5A). That ion corresponded to the known molecular weight for macrocidin A of 357 mass units (Daltons). The ions of lower mass, for example at m/z 340 and 192, probably arose from a process of cone voltage fragmentation, a common occurrence during ESI–MS of organic natural products. These low mass ions were not considered further. The ion at m/z 380 corresponded to a sodium adduct ion (macrocidin A + Na)+, again a common occurrence. Ions of higher mass, such as those at m/z 739, 768, 1096 and 1125 in the purified macrocidin A solution (Fig. 5A), were of interest because ESI–MS has been shown to be capable of detecting dissolved metal ion complexes of various organic compounds

(Capon et al., 1999; Kaufmann et al., 2005). In that regard, numerous applications of ESI–MS have been investigated, from determination of metal-binding stoichiometry of proteins to investigation of dissolved metal speciation in natural waters (Ross et al., 1998). Mixing of ferric chloride and the acidic aqueous ethanol solution of macrocidin A gave the spectrum shown in Fig. 5B. The high mass ions at m/z 768 and 1125 had increased in abundance compared to Fig. 5A. These observations could be rationalized by recognizing that iron chelates were formed, one with two and the other with three macrocidin A molecules complexed with a ferric ion (mass of 56) (Fig. 5B). Both iron chelates occurred at low levels in the purified solution of macrocidin A (m/z 768 and 1125 in Fig. 5A). The dissolved iron may have originated from the grains culture, forming low concentrations of the chelates. The metal chelate complexes apparently carried through the purification steps intact. Fig. 5C depicts the mass spectrum of the solution of purified macrocidin A mixed with magnesium chloride (mass of Mg is 24). The ion at m/z 737 corresponds in mass to the binding of Mg2+ with two molecules of macrocidin A. This ion was also observed in an experiment under neutral conditions (without the addition of formic acid). The abundance of this ion was slightly greater than the abundance observed under acidic conditions. The ions at m/z 1096 (Fig. 5A–C) and m/z 739 (Fig. 5A and B) were probably associated with the aluminum (27 Da) chelated

Fig. 5. Electrospray mass spectrometry (ESI–MS) of macrocidin A (A), macrocidin A and ferric chloride (B) and macrocidin A and magnesium chloride (C), all in aqueous ethanol containing formic acid.

86

with macrocidin A as (2A respectively.

M. Hubbard et al. / Environmental and Experimental Botany 132 (2016) 80–91

2H + Al)+ and (3A

3H + Al + H)+,

3.4. Carotenoids and carotenoid precursors Phytoene made up a greater percent of carotenoids and carotenoid precursors in the DF or NIC+DF-treated dandelions at all time points as compared to the untreated controls (Fig. 6A). Macrocidin-treated dandelions took longer for the percent phytoene to increase above the level of controls (Fig. 6A). In groundsel, 17 days after treatment, the two higher doses of macrocidins contained elevated percentages of phytoene relative to the control or 32 MU-treated plants (Fig. 6B). Percent phytoene in macrocidin-treated chickpea began to increase compared to the control at three days post-treatment. The macrocidin-dosedependency of the increase in percent phytoene grew with increasing time post-treatment (Fig. 6C). In dandelion, percent lutein first increased (all three doses) and then decreased (128 MU and 256 MU, but not 32 MU) in response to macrocidin treatment (Fig. 6D). Treatment with NIC corresponded with a decrease in percent lutein at some time points and no change relative to the control at others. Dandelions exposed to DF or NIC+DF had lower percentages of lutein than the control at all time points (Fig. 6D). Lutein comprised a greater percentage of carotenoids and their precursors at three, nine and 17 days posttreatment with macrocidins in groundsel (Fig. 6E). At three days post-treatment, percent lutein was greater in 128 MU- and 256 MU-treated chickpea than in control plants (Fig. 6F). All three doses of macrocidins resulted in a drop in percent b-carotene relative to untreated dandelions at all time points except for one day post-treatment. At 10, 13 and 17 days after treatment, dandelions exposed to the two higher doses of macrocidins had lower percent b-carotene than dandelions treated with 32 MU. Treatment of dandelions with NIC led to a decrease in percent b-carotene at one, three and six days posttreatment, but not thereafter. At all time points, dandelions challenged with NIC+DF or DF had lower percent b-carotene values than their untreated counterparts (Fig. 6G). Three days posttreatment and later, percent b-carotene was lower in macrocidintreated groundsel than in controls, with the exception of 32 MUtreated plants nine days after treatment. This drop was more pronounced in groundsel treated with the higher two doses of macrocidins (Fig. 6H). Nine and 17 days following treatment with 128 or 256 MU, chickpeas had lower percent b-carotene than control and 32 MU-treated plants. By 17 days after treatment with 256 MU, chickpeas had lower percent b-carotene than those exposed to 128 MU (Fig. 6I). One day after treatment, DF or NIC+DF-treated dandelions had lower percent violaxanthin values than those in any other treatment. Violaxanthin made up a larger percent of total carotenoids and carotenoid precursors in dandelions treated with any of the three doses of macrocidins three and six days posttreatment (with the exception of 128 MU-treated plants six days after treatment), compared to the other treatments. By 17 days post-treatment, dandelions treated with 128 or 256 MU had lower percent violaxanthin than the control. Dandelions treated with NIC, DF or NIC+DF had lower percent violaxanthin at all time points, except for NIC-treated plants at one day post-treatment (Fig. 6J). Groundsel exposed to 128 or 256 MU had higher percent violaxanthin values at three, nine and 17 days post-treatment, as did plants treated with 32 MU three days after treatment (Fig. 6K). Three days after treatment with any dose of macrocidins, chickpeas had increased percent violaxanthin compared to the control (Fig. 6L). One day after treatment, lycopene made up the greatest percentage of carotenoids and their precursors in dandelions

treated with NIC (0.8 a standard error of 0.1%) as compared to all other treatments (all 0.0  0.0%; data not shown). By three days post-treatment, percent lycopene was greatest in NIC-treated dandelions (2.9  0.6%), with significantly less in plants treated with NIC+DF (0.2  0.1%) and less again in all remaining treatments (all 0.0  0.0%). Lycopene comprised the highest percent of carotenoids and precursors six days after treatment in NIC-treated dandelions (2.9  0.8%), with significantly lower values (0.1  0.0%) being detected in plants treated with both NIC+DF. No lycopene was found in dandelions belonging to any other treatment group at six days post-treatment. NIC-treated dandelions had higher percent lycopene values (0.9  0.3%) at ten days after treatment than any other treatment. All other treatments contained 0.0  0.0% lycopene. By 13 days post-treatment, 128 MU-treated dandelions contained the most lycopene (1.7  1.1%), significantly more than in any other treatment except NIC (0.3  0.1%). NICtreated dandelions had higher percent lycopene values than those in any other treatment (all 0.0  0.0%). When 17 days had elapsed after treatment, dandelions exposed to NIC contained 0.3  0.1% lycopene, which was greater than plants in any other treatment, all of which had 0.0  0.0% lycopene. No lycopene was detected in groundsel or chickpea in any of the treatments or time points tested (data not shown). The b-carotene to lutein ratio was lower in macrocidin-treated dandelions at three days post-treatment and later, but not earlier (Fig. 7). In contrast, dandelions treated with NIC had a reduced b-carotene to lutein ratio one day after treatment, but not at later time points. Dandelions exposed to NIC+DF had lower b-carotene to lutein ratios one, three, six and 10 days, but not 13 and 17 days, post-treatment. The b-carotene to lutein ratio differed between control and DF-treated dandelions only at 10 days post-treatment (Fig. 7). At three and 17 days post-treatment, groundsel exposed to any dose of macrocidins had lower b-carotene to lutein ratios than control plants. This drop was greater in plants challenged with 128 or 256 MU. Nine days after treatment, only groundsel exposed to the two higher doses of macrocidins showed a drop in this parameter (Fig. 7). The b-carotene to lutein was lower in 32 MUtreated chickpea, as compared to the control, at 17 days posttreatment. Chickpea exposed to 128 or 256 MU had b-carotene to lutein values below those of control plants at three, nine and 17 days post-treatment. At one day post-treatment, none of the treatments differed from one another in terms of b-carotene to lutein ratio. The drop in b-carotene to lutein ratio when treated with a given dose of macrocidins was greater in groundsel than in chickpea at three and 17 days post-treatment for all three doses (Fig. 7). Dandelions treated with macrocidins had decreased total carotenoid and carotenoid precursor content relative to the controls (Fig. 8). This decrease was significant one, three and six days after treatment for 256 MU-, 128 MU- and 32 MU-treated plants, respectively, and at all time points thereafter. Lower total carotenoid and carotenoid precursor values were measured in 256 MU-treated dandelions than in their 32 MU-treated counterparts. Dandelions exposed to DF or NIC+DF had higher total carotenoid and carotenoid precursor content than did controls (Fig. 8). Macrocidin-treated groundsel had lower levels of total carotenoids and their precursors than did untreated plants at three, nine and 17 days post-treatment. At the two latter time points, groundsel exposed to 128 and 256 MU had still lower values for this parameter than did their 32 MU-treated counterparts (Fig. 8). Total carotenoid and carotenoid precursor content was lower in macrocidin-treated chickpeas three, nine and 17 days following treatment. At nine days after treatment, there was a significant dose-dependent drop in total carotenoid and precursor levels. Total carotenoid and carotenoid precursor content (as a fraction of the mean for the appropriate control) did not vary

M. Hubbard et al. / Environmental and Experimental Botany 132 (2016) 80–91

87

Fig. 6. Percent phytoene (A, B, C), lutein (D, E, F), b-carotene (G, H, I) andviolaxanthin (J, K, L) in dandelion (A, D, G and J), groundsel (B, E, H and K) and chickpea (C, F, I and L) over time. Within a given carotenoid or carotenoid precursor and time point, treatments with the same letter are not significantly different (p  0.05).

88

M. Hubbard et al. / Environmental and Experimental Botany 132 (2016) 80–91

Fig. 7. Ratio of b-carotene to lutein in dandelion, groundsel and chickpea over time. At a given time point, treatments with the same lower-case letter are not significantly different (p  0.05). Upper-case letters compare the ratio of b-carotene to lutein between groundsel and chickpea for a given treatment and time point; points with the same letter are not significantly different (p  0.05).

Fig. 8. Total carotenoid and carotenoid precursor content per gram of fresh weight in dandelion, groundsel and chickpea over time. At a given time point, treatments with the same lower-case letter are not significantly different (p  0.05). Upper-case letters compare groundsel and chickpea for a given treatment and time point; points with the same letter are not significantly different (p  0.05).

between groundsel and chickpeas exposed to the same treatment for the same length of time, except at nine days post-treatment, when groundsel had lower values for this parameter relative to the control (Fig. 8). Phytofluene content was initially elevated in dandelions treated with DF, relative to all other treatments (Fig. S1A). However, by 13 and 17 days post-treatment none of the dandelion treatments differed in terms of this parameter (Fig. S1A). At nine and 17 days post-treatment, groundsel exposed to all doses of macrocidins had lower percent phytofluene values than controls (Fig. S1B). At nine days post-treatment, control chickpeas had lower percent phytofluene values than plants exposed to 128 or 256 MU (Fig. S1C). Percent a-carotene was lower in macrocidin-treated dandelions, relative to controls, at three days post-treatment and later, with the exception of 32 MU-treated dandelion 10 days posttreatment (Fig. S1D). Dandelions exposed to DF or NIC+DF had percent a-carotene values below controls at all time points. At six and 13 days after treatment, NIC-treated dandelions had higher and lower percent a-carotene values, respectively, compared to their untreated counterparts (Fig. S1D). At all time points after one day percent a-carotene values were lower in macrocidin-treated

groundsel compared to controls (Fig. S1E). At three and nine days, macrocidin-treated chickpeas had lower percent a-carotene values than control plants (Fig. S1F). The percent g-carotene increased in NIC-treated dandelion at all time points compared to untreated controls (Fig. S1J). NIC+DF treatment resulted in a decreased in percent g-carotene 10, 13 and 17 days after treatment. Dandelions treated with DF had lower percent g-carotene values compared to their untreated counterparts at all time points. At six days post-treatment and later dandelions treated with 128 or 256 MU had lower percent g-carotene values than control plants (Fig. S1G). Nine days posttreatment, 32 and 128 MU-treated chickpeas had higher percent g-carotene values than plants challenged with 256 MU. Percent g-carotene was lower in macrocidin-treated chickpeas than control plants at 17 days post-treatment (Fig. S1K). 4. Discussion The carotenoid profiles of dandelion, groundsel and chickpea each responded differently to treatment with macrocidins. Most notable were differences in percentages of phytoene, lutein, and

M. Hubbard et al. / Environmental and Experimental Botany 132 (2016) 80–91

violaxanthin (Fig. 6A–E and J–L). Higher percent phytoene was taken to indicate more complete inhibition of phytoene desaturase (PDS). This interpretation is supported by the fact that treatment with DF, a known PDS inhibitor (Wightman and Hayes, 1985), led to a very high percent phytoene, both in the present study (Fig. 6A–C) and in Hubbard et al. (2015). PDS was most completely inhibited in dandelion, followed in decreasing order by chickpea and groundsel (Fig. 6A–C). Chickpea showed a greater inhibition of PDS, but less severe impacts on symptom severity (Fig. 1), biomass (Fig. 2), Fv/ Fm, F'v/F'm (Fig. 4) and, in Bailey et al. (2011a) reduced mortality, as compared to groundsel. Clearly plant mortality can occur in the absence of PDS inhibition, demonstrating the importance of other modes of action. An increase in relative xanthophyll (the hydroxyl groupcontaining carotenoids lutein and violaxanthin) content was observed in macrocidin-treated groundsel; percent lutein and violaxanthin increased while percent b-carotene decreased (Fig. 6E, H and K). This shift towards xanthophylls was also seen transiently in dandelions (at three and six days after treatment; Fig. 6D, G and J) and chickpea (at three days post-treatment; Fig. 6F, I and L), before the build-up of phytoene. The up-swing in percent phytoene in chickpea is less extreme than in dandelion (Fig. 6A and C), coinciding with a drop in percent lutein and violaxanthin in the latter (Fig. 6D and J) and a mere lack of change in the percentage of both xanthophylls in the former (Fig. 6I and L). These results demonstrate a trade-off between PDS inhibition and relative xanthophyll increase. Macrocidins do not appear to inhibit the enzyme LYC-b; NIC-treated, but not macrocidin-treated, dandelions exhibited a consistent increase in percent lycopene and g-carotene (Fig. S1G and J). Thus, the drop in b-carotene to lutein ratios seen in all three species (Fig. 7) can be understood as a manifestation of the macrocidin-induced rise in relative xanthophyll content, as opposed to a shift towards the lutein branch of the carotenoid biosynthetic pathway, away from the b-carotene branch, driven by LYC-b-inhibition and/or a b-carotene “drawdown” in order to potentially increase abscisic acid (ABA) production (discussed by Hubbard et al. (2015)). NIC and DF took effect on carotenoid profiles more quickly than did macrocidins. This speed of action and the fact that DF induced a dramatic increase in total carotenoid and carotenoid precursor content (Fig. 8; consistent with the findings of Hubbard et al. (2015)) and no increase in percent xanthophylls (Fig. 6G and J) or in the initial fluorescence, or FO or O values, measured as one of the OJIP parameters (Table 1), suggest that the modes of action of both these herbicides are distinct from that of macrocidins. Based on the finding that macrocidins can bind to iron (Fig. 5B) and the changes observed in the OJIP fluorescent transient of macrocidin-treated plants (Table 1), the hypothesis that macrocidins bind the Fe2+ involved in electron transfer between QA and QB in PSII (shown in Fig. 2 in Stirbet et al. (2014)) in susceptible, but not resistant, plants is put forward. This hypothesis predicts that QA remains reduced in dark-adapted macrocidins-treated susceptible plants because it is unable to pass electrons to QB. The increases in the FO values of macrocidin treated plants as compared to untreated controls (Table 1) are consistent with this prediction. Bertamini et al. (2001) also suggests that the increased FO values in iron-deficient plants are due to reduced QA. Consistently, elevated FO values are observed in iron-deficient plants (Msilini et al., 2011). In contrast, OJIP transients of plants exposed to other stressors have altered J, I and/or P values compared to controls, but exhibit no change in FO readings (Chen et al., 2015; Percival, 2005). In plants treated with the highest dose of macrocidins, the damage to carotenoids could be so great that the majority of the photosynthetic apparatus has been destroyed, explaining the total flattening of the OJIP transient. Abadia et al. (1999) and Larbi et al. (2004) demonstrated that iron deficiency can induce chlorotic symptoms

89

and a drop in total leaf carotenoid content. All of these findings are consistent with macrocidins binding iron in the leaves of susceptible plants. Another cause of the increase in initial fluorescence could be an inability of the energy harvested from incoming light by the light harvesting complex of PSII (LHCII) to be passed on to QA, paralleling the impact of senesce observed by Prakash et al. (2003). Inhibitions in electron flow, both before QA and between QA and QB, could contribute to the observed drop in fEo and increase in VJ obtained from leaves with Fm FO  20 (Table 1). Consistently, Chen et al. (2014) found that the phytotoxic fungal metabolite and tetramic acid, tenuazonic acid, induced an increase in VJ, which the authors attributed to a build-up of QA-. Tenuazonic acid also induced a rise in FO. However, Chen et al. (2005) found that tenuazonic acid did not change Fv/Fm, concluding that the LHCII to PSII reaction centre connection remained intact. Thus, the macrocidin-induced drop in Fv/Fm implies that macrocidins do indeed uncouple LHCII from the PSII reactions centres. The fact that tenuazonic acid binds the D1 protein in PSII, thereby inhibiting electron transfer from QA to QB (Chen et al., 2007) suggests that exploring the potential binding of macrocidins to component(s) of the PSII could be worthwhile. Despite the similarly of their impacts on OJIP parameters, tenuazonic acid and macrocidins induce distinct symptoms: brown lesions (Chen et al., 2015; Qiang et al., 2010) and chlorosis (Hubbard et al., 2015), respectively. Hence, it is likely that the two tetramic acids have overlapping, but not identical, modes of action. Blocked electron flow in the photosystem II (PSII) reaction centre, which contains QA, QB and a total of 22 b-carotenes (Umena et al., 2011), would cause high-energy electrons to accumulate. These excess electrons likely contribute to reactive oxygen species formation, damaging the b-carotenes in the reaction centre. Because the reaction centre would have a reduced capacity to absorb incoming light energy, the LHCII would be forced to deal with more excess energy. Xanthophylls, including lutein, neoxanthin and another xanthophyll are found in LHCII (Liu et al., 2004), offering photoprotection (reviewed by Domonkos et al. (2013), Jahns and Holzwarth (2012) and Mozzo et al. (2008)). This explains the increased relative xanthophyll content in both macrocidin-treated and iron-deficient plants (Abadia et al., 1999; Larbi et al., 2004), as well as those exposed to excess light (Contin et al., 2014). The rise in percent xanthophyll (Fig. 6D, E, J and K) and drop in total carotenoid and carotenoid precursor content (Fig. 8) occurred more rapidly after macrocidin-treatment than the inhibition of PDS (Fig. 6A–C). So, perhaps macrocidins first bind iron, thus blocking electron transfer from QA to QB, which leads to the photoprotective increase in relative xanthophyll content, while also binding magnesium. The latter would reduce the amount of Mg2+ available to 1-deoxy-D-xylulose (DXP) reductoisomerase (DXR), an enzyme involved in carotenoid precursor synthesis that is activated by Mn2+, Co2+ or Mg2+ (reviewed by Murkin et al. (2014)), or to stabilize phytoene synthase (PSY) (Schofield and Paliyath, 2005), the enzyme required to synthesize phytoene (reviewed by Cunningham and Gantt (1998)). Diminished DXR and/or PSY activity could reasonably be predicted to lower total carotenoid and carotenoid precursor content. Carotenoids can improve plants’ ability to take up iron (Garcia-Casal and Leets, 2014). Thus, macrocidin-induced reductions in carotenoid concentrations could in turn lead to an iron deficiency, exacerbating macrocidin-induced chlorosis. Consistent with the hypothesis that metal-binding related impacts of macrocidins take place prior to PDS-inhibition, the impact of macrocidin-treatment on Fv/Fm and F'v/F'm values of dandelions was dramatic at three days (Fig. 3), suggesting damage to PSII, but symptom severity was still quite mild (Fig. 1). At three days, the xanthophyll build-up (Fig. 6J) and decreased carotenoid

90

M. Hubbard et al. / Environmental and Experimental Botany 132 (2016) 80–91

and carotenoid precursor content was observed (Fig. 7), while the inhibition of PDS was not (Fig. 6A). As macrocidins are phloem-mobile (Graupner et al., 2003) and hence transported to new growth, macrocidin A likely first binds magnesium in the youngest leaves. When chlorophyll starts to break down, due to inhibiting the biosynthesis of carotenoid precursors and disrupted electron transport between QA and QB, the absence of free magnesium prevents chlorophyll from being formed in new leaves. This could explain the absence of chlorosis in older leaves, a symptom commonly seen in magnesium-deficient plants (Yeh et al., 2000). Once no new chlorophyll is being produced, the OJIP parameters would flatten out, as observed in plants treated with 256 MU (Table 1). As different species react differently to macrocidins, the relative strength of the interaction between macrocidins and the iron in PSII, magnesium in chlorophyll and/or involved in promoting DXR and/or PSY efficiency and the inhibition of PDS vary depending on speciesspecific factors, which could include the structure of PDS and/or differences in plant uptake and/or transport of macrocidins. Macrocidin A only binding magnesium in already-damaged plants might explain the absence of magnesium deficiency symptoms in resistant plants. Alternatively, resistant plants may not take up macrocidins. This latter possibility is supported by the fact that, while resistant plants are colonized by P. macrostoma, the fungus does not penetrate the vascular cylinder (Bailey et al., 2011b). Future investigation on the presence and concentration of macrocidins in the sap of plants species with divergent responses to macrocidintreatment could help clarify this point. It is also possible that macrocidins are broken down by resistance plants and/or excluded from chloroplasts, where their modes of action occur. The hypothesis that being bound to iron increases macrocidin bioactivity seems unlikely in light of Bailey et al. (2013)’s results. Bailey et al. (2013) found that the application of Scotts Turf Builder Pro fertilizer, that contains 2% iron and would thus be expected to increase the ratio of macrocidin bound to iron to iron-free macrocidins, increased percent reduction of dandelion (Table 2 of Bailey et al. (2013)). However, both Scotts Turf Builder Pro fertilizer and Scotts Lawn Pro (with no iron) led to increased percent dandelion reduction (Bailey et al., 2013), so it cannot be concluded that iron is the reason for increased dandelion control. Increased nitrate- or ammonium-nitrogen as the cause is better supported by the data. In conclusion, the modes of action of macrocidins include 1) hampering of electron transfer from QA to QB via in-planta ironbinding, leading to a photoprotective increase in relative abundance of xanthophylls and 2) binding magnesium, potentially leading to decreased functionality of DXR and/or PSY and reduced ability to synthesize new chlorophyll, 3) uncoupling of LHCII from the reaction centres of PSII and 4) inhibition of PDS. The first three modes of action occur more rapidly than the latter. The relative importance of mechanisms one and three are negatively related; one is more pronounced in groundsel than dandelion or chickpeas and three is minor in groundsel, more prominent in chickpea and still more dominant in dandelion. The prevailing modes of action by which macrocidins induce chlorosis differ between dandelions, groundsel and chickpea. Both the multiplicity of modes of action of macrocidins and the fact that different modes of action are more important in different species and at different points in time suggest that resistance development would be difficult. Author contributions M. Hubbard, R.K. Hynes and K.L. Bailey designed all experiments except the metal binding of macrocidins, which were designed by W.G. Taylor. M. Hubbard performed experiments, with technical assistance from Jo-Anne Derby, Dan Hupka, Annika Kolar, Parker

Laurence and Jon Geissler, with the exception of the metal binding of macrocidins. The latter experiments were performed by W.G. Taylor with technical assistance from Stephen Walter, Dan Sutherland and Lawrence Hogge. Data on metal binding of macrocidins was analyzed by W.G. Taylor. All other data analysis was carried out by M. Hubbard. M. Hubbard wrote the paper, incorporating written materials and methods and discussion provided by W.G. Taylor on the metal binding of macrocidins. R. K. Hynes, K.L. Bailey and W.G. Taylor read and offered critical feedback on the paper. Acknowledgements This work was supported from funding from Agriculture and Agri-Food Canada and The Scotts Company, Marysville, Ohio. The authors thank Dr. Shiguo Chen (Nanjing Agricultural University, Nanjing 210095, China) for reviewing the manuscript and providing advice on the OJIP data and Jo-Anne Derby, Dan Hupka, Annika Kolar, Parker Laurence, Jon Geissler, Stephen Walter, Dan Sutherland and Lawrence Hogge (Agriculture and Agri-Food Canada) for technical assistance. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.envexpbot. 2016.08.009. References Abadia, J., Morales, F., Abadia, A., 1999. Photosystem II efficiency in low chlorophyll, iron-deficient leaves. Plant and Soil 215, 183–192. Bailey, K.L., Derby, J., 2001. Fungal isolates and biological control compositions for the control of weeds, US Patent Application, USA. Bailey, K.L., Pitt, W.M., Falk, S., Derby, J., 2011a. The effects of Phoma macrostoma on nontarget plant and target weed species. Biol. Control 58, 379–386. Bailey, K.L., Pitt, W.M., Leggett, F., Sheedy, C., Derby, J., 2011b. Determining the infection process of Phoma macrostoma that leads to bioherbicidal activity on broadleaved weeds. Biol. Control 59, 268–276. Bailey, K.L., Falk, S., Derby, J.-A., Melzer, M., Boland, G.J., 2013. The effect of fertilizers on the efficacy of the bioherbicide, Phoma macrostoma, to control dandelions in turfgrass. Biol. Control 65, 147–151. Bertamini, M., Nedunchezhian, N., Borghi, B., 2001. Effect of iron deficiency induced changes on photosynthetic pigments ribulose-1,5-bisphosphate carboxylase, and photosystem activities in field grown grapevine (Vitis vinifera L. cv. Pinot noir) leaves. Photosynthetica 39, 59–65. Capon, R.J., Skene, C., Lacey, E., Gill, J.H., Wadsworth, D., Friedel, T., 1999. Geodin A magnesium salt: a novel nematocide from a southern Australian marine sponge, Geodia. J. Nat. Prod. 62, 1256–1259. Chen, S., Dai, X., Qiang, S., Tang, Y., 2005. Effect of a nonhost-selective toxin from Alternaria alternata on chloroplast-electron transfer activity in Eupatorium adenophorum. Plant Pathol. 54, 671–677. Chen, S., Xu, X., Dai, X., Yang, C., Qiang, S., 2007. Identification of tenuazonic acid as a novel type of natural photosystem II inhibitor binding in Q(B)-site of Chlamydomonas reinhardtii. Biochim. Biophys. Acta 1767, 306–318. Chen, S., Yin, C., Dai, X., Qiang, S., Xu, X., 2008. Action of tenuazonic acid, a natural phytotoxin, on photosystem II of spinach. Environ. Exp. Bot. 62, 279–289. Chen, S., Strasser, R.J., Qiang, S., 2014. In vivo assessment of effect of phytotoxin tenuazonic acid on PSII reaction centers. Plant Physiol. Biochem. 84, 10–21. Chen, S., Kang, Y., Zhang, M., Wang, X., Strasser, R.J., Zhou, B., Qiang, S., 2015. Differential sensitivity to the potential bioherbicide tenuazonic acid probed by the JIP-test based on fast chlorophyll fluorescence kinetics. Environ. Exp. Bot. 112, 1–15. Contin, D.R., Soriani, H.H., Hernández, I., Furriel, R.P.M., Munné-Bosch, S., Martinez, C.A., 2014. Antioxidant and photoprotective defenses in response to gradual water stress under low and high irradiance in two Malvaceae tree species used for tropical forest restoration. Trees 28, 1705–1722. Cunningham, F.X.J., Gantt, E., 1998. Genes and enzymes of carotenoid biosynthesis in plants Annu. Rev. Plant Physio. 49, 557–583. Domonkos, I., Kis, M., Gombos, Z., Ughy, B., 2013. Carotenoids, versatile components of oxygenic photosynthesis. Prog. Lipid Res. 52, 539–561. Farquhar, G.D., Wong, S.C., Evans, J.R., Hubick, K.T., 1989. Photosynthesis and gas exchange. In: Jones HG, F.T.a.J.M. (Ed.), Plants Under Stress. Cambridge University Press, Cambridge, pp. 47–69. Garcia-Casal, M.N., Leets, I., 2014. Carotenoids, but not vitamin A, improve iron uptake and ferritin synthesis by Caco-2 cells from ferrous fumarate and NaFeEDTA. J. Food Sci. 79, H706–712.

M. Hubbard et al. / Environmental and Experimental Botany 132 (2016) 80–91 Graupner, P.R., Carr, A., Clancy, E., Gilbert, J., Bailey, K.L., Derby, J.A., Gerwick, B.C., 2003. The macrocidins: novel cyclic tetramic acids with herbicidal activity produced by Phoma macrostoma. J. Nat. Prod. 66, 1558–1561. Hirakil, M., van Rensen, J.J.S., Vredenberg, W.J., Wakabayashi, K., 2003. Characterization of the alterations of the chlorophyll a fluorescence induction curve after addition of Photosystem II inhibiting herbicides. Photosynth. Res. 78, 35–46. Howes, C.D., 1974. Nicotine inhibition of carotenoid cyclization in Cucurbita ficifolia cotyledons. Phytochemistry 13, 1469–1471. Hubbard, M., Hynes, R.K., Erlandson, M., Bailey, K.L., 2014. The biochemistry behind biopesticide efficacy. Sustainable Chem. Process. 2 (18) doi:http://dx.doi.org/ 10.1186/s40508-014-0018-x. Hubbard, M., Hynes, R.K., Bailey, K.L., 2015. Impact of macrocidins, produced by Phoma macrostoma, on carotenoid profiles of plants. Biol. Control 89, 11–22. Jahns, P., Holzwarth, A.R., 2012. The role of the xanthophyll cycle and of lutein in photoprotection of photosystem II. Biochim. Biophys. Acta 1817, 182–193. Jiang, C.-D., Gao, H.-Y., Zou, Q., Shi, L., 2007. Effects of iron deficiency on photosynthesis and photosystem II function in soybean leaf. J. Plant Physiol. Mol. Biol. 33, 53–60. Kaufmann, G.F., Sartorio, R., Lee, S.H., Rogers, C.J., Meijler, M.M., Moss, J.A., Clapham, B., Brogan, A.P., Dickerson, T.J., Janda, K.D., 2005. Revisiting quorum sensing: discovery of additional chemical and biological functions for 3-oxo-Nacylhomoserine lactones. P. Natl. Acad. Sci. USA 102, 309–314. Kitajima, M., Butler, W.L., 1975. Quenching of chlorophyll fluorescence and primary photochemistry in chloroplasts by dibromothymoquinone. Biochim. Biophys. Acta 376, 105–115. Krohling, C.A., Eutrópio, F.J., Bertolazi, A.A., Dobbss, L.B., Campostrini, E., Dias, T., Ramos, A.C., 2016. Ecophysiology of iron homeostasis in plants. Soil Sci. Plant Nutr. 62, 39–47. Larbi, A., Abadia, A., Morales, F., Abadia, J., 2004. Fe resupply to Fe-deficient sugar beet plants leads to rapid changes in the violaxanthin cycle and other photosynthetic characteristics without significant de novo chlorophyll synthesis. Photosynth. Res. 79, 59–69. Liu, Z., Yan, H., Wang, K., Kuang, T., Zhang, J., Gui, L., An, X., Chang, W., 2004. Crystal structure of spinach major lightharvesting complex at 2.72 Å resolution. Nature 428, 287–292. Morales, F., Abadia, A., Abadia, J., 1990. Characterization of the xanthophyll cycle and other photosynthetic pigment changes induced by iron deficiency in sugar beet (Beta vulgaris L.). Plant Physiol. 94, 607–613. Mozzo, M., Dall'Osto, L., Hienerwadel, R., Bassi, R., Croce, R., 2008. Photoprotection in the antenna complexes of photosystem II: role of individual xanthophylls in chlorophyll triplet quenching. J. Biol. Chem. 283, 6184–6192. Msilini, N., Zaghdoudi, M., Govindachary, S., Lachaal, M., Ouerghi, Z., Carpentier, R., 2011. Inhibition of photosynthetic oxygen evolution and electron transfer from

91

the quinone acceptor QA- to QB by iron deficiency. Photosynth. Res. 107, 247– 256. Murkin, A.S., Manning, K.A., Kholodar, S.A., 2014. Mechanism and inhibition of 1deoxy-D-xylulose-5-phosphate reductoisomerase. Bioorg. Chem. 57, 171–185. Percival, G.C., 2005. The use of chlorophyll fluorescence to identify chemical and environmental stress in leaf tissue of three oak (Quercus) species. J. Arboric 31, 215–227. Prakash, J.S.S., Srivastava, A., Strasser, R.J., Mohanty, P., 2003. Senescence-induced alterations in the photosystem II functions of Cucumis sativus cotyledons: probing of senescence driven alterations of photosystem II by chlorophyll a fluorescence induction O-J-I-P transients. Indian J. Biochem. Bio. 40, 160–168. Qiang, S., Wang, L., Wei, R., Zhou, B., Chen, S., Zhu, Y., Dong, Y., An, C., 2010. Bioassay of the herbicidal activity of AAC-toxin produced by Alternaria alternata isolated from Ageratina adenophora. Weed Technol. 24, 197–201. Rohacek, K., 2002. Chlorophyll fluorescnce parameters: the definitions, photosynthetic meaning, and mutual relationships. Photosynthetica 40, 13–29. Ross, A.R.S., Ikonomou, M.G., Thompson, J.A.J., Orians, K.J., 1998. Determination of dissolved metal species by electrospray ionization mass spectrometry. Anal. Chem. 70, 2225–2235. Schobert, R., Schlenk, A., 2008. Tetramic and tetronic acids: an update on new derivatives and biological aspects. Bioorg. Med. Chem. 16, 4203–4221. Schofield, A., Paliyath, G., 2005. Modulation of carotenoid biosynthesis during tomato fruit ripening through phytochrome regulation of phytoene synthase activity. Plant Physiol. Biochem. 43, 1052–1060. Stirbet, A., Riznichenko, G.Y., Rubin, A.B., Govindjee, 2014. Modeling chlorophyll a fluorescence transient: relation to photosynthesis. Biochemistry (Mosc). 79, 291–323. Strasser, R.J., Srivastava, A., Tsimilli-Michael, M., 2000. The fluorescence transient as a tool to characterize and screen photosynthetic samples. In: Yunus, M., Pathre, U., Mohanty, P. (Eds.), Probing Photosynthesis: Mechanism, Regulation and Adaptation. Taylor and Francis, London, UK, pp. 443–480. Umena, Y., Kawakami, K., Shen, J.R., Kamiya, N., 2011. Crystal structure of oxygenevolving photosystem II at a resolution of 1.9 A. Nature 473, 55–60. Verbruggen, N., Hermans, C., 2013. Physiological and molecular responses to magnesium nutritional imbalance in plants. Plant Soil 368, 87–99. Wightman, P., Hayes, C., 1985. The mode of action and basis of selectivity of diflufenican in wheat, barley and selected weed species. British Crop Protection Conference – Weeds, Volume 1. British Crop Protection Conference, Croydon, England, pp. 171–178. Yeh, D.M., Lin, L., Wright, C.J., 2000. Effects of mineral nutrient deficiencies on leaf development, visual symptoms and shoot  root ratio of Spathiphyllum. Sci. Hortic.-Amsterdam.