Maize is by quantity the most important C4 cereal crop; however, future climate changes are expected to increase maize susceptibility to mycotoxigenic fungal pathogens and reduce productivity. While rising atmospheric [CO2] is a driving force behind the warmer temperatures and drought, which aggravate fungal disease and mycotoxin accumulation, our understanding of how elevated [CO2] will effect maize defences against such pathogens is limited. Here we report that elevated [CO2] increases maize susceptibility to Fusarium verticillioides proliferation, while mycotoxin levels are unaltered. Fumonisin production is not proportional to the increase in F. verticillioides biomass, and the amount of fumonisin produced per unit pathogen is reduced at elevated [CO2]. Following F. verticillioides stalk inoculation, the accumulation of sugars, free fatty acids, lipoxygenase (LOX) transcripts, phytohormones and downstream phytoalexins is dampened in maize grown at elevated [CO2]. The attenuation of maize 13-LOXs and jasmonic acid production correlates with reduced terpenoid phytoalexins and increased susceptibility. Furthermore, the attenuated induction of 9-LOXs, which have been suggested to stimulate mycotoxin biosynthesis, is consistent with reduced fumonisin per unit fungal biomass at elevated [CO2]. Our findings suggest that elevated [CO2] will compromise maize LOX-dependent signalling, which will influence the interactions between maize and mycotoxigenic fungi.
Maize (Zea mays) is among the most important food, feed and biofuel crops (Young & Long 2000). In order to meet rising grain demands, the United States has gradually increased the area of land devoted to maize growth (U.S. Department of Agriculture, National Agricultural Statistics Service). However, severe weather conditions, fungal disease and mycotoxin contamination are major limiting factors of maize productivity. In 2012, these factors reduced the yield in bushels per acre to the lowest in this decade (Bienkowski 2012; Schultz 2012; Smith & Mitchell 2012). Rising atmospheric carbon dioxide concentration ([CO2]) is largely responsible for the greenhouse effect leading to climatic conditions that reduce maize yield and increase crop susceptibility to mycotoxigenic fungi. Yet despite future food security and food safety concerns accompanying crop productivity and mycotoxin contamination in the context of global climate change, our understanding of how elevated [CO2] in and of itself will affect fungal disease progression and mycotoxin contamination in maize is limited.
Fusarium verticillioides is the most ubiquitous mycotoxigenic fungus that infects maize (Marasas 2001; Miller 2001; Yates et al. 2003; Murillo-Williams & Munkvold 2008). F. verticillioides can cause rot in all parts of the plant and produces polyketide-derived mycotoxins termed fumonisins. High levels of these carcinogenic compounds render the grain unfit for consumption. Environmental conditions associated with climate change, such as high temperatures and drought, increase F. verticillioides systemic disease development and fumonisin contamination (Shelby et al. 1994; Miller 2001; Battilani et al. 2008).
In contrast to C3 plants, rising atmospheric [CO2] will not directly stimulate photosynthesis and enhance the growth and yield of C4 crops, such as maize. Because of their unique foliar Kranz anatomy and cellular physiology, C4 plants are capable of concentrating [CO2] at the site of the major enzyme involved carbon fixation, ribulose bisphosphate carboxylase-oxygenase, to levels practically near saturation (Allen et al. 2011). Thus, only in combination with drought can the physiological effect of reduced stomatal conductance (which occurs in both C3 and C4 plants at elevated [CO2] ) benefit maize productivity by increasing water-use efficiency (Ghannoum et al. 2000; Long et al. 2004; Leakey 2009).
In the absence of major changes in maize photosynthesis and primary metabolism, the potential effects of elevated [CO2] on maize phytohormone signalling networks and secondary metabolites are of particular importance. Elevated [CO2] has been correlated with increases in Fusarium pseudograminearum crown rot in wheat, a C3 member of the Poaceae grass family (Melloy et al. 2010). However, predictions and inferences from this and other studies are difficult as research investigating the impact of elevated [CO2] on plant–pathogen interactions has been variable and very much dependent on the pathosystem; examples of increased, decreased and unaltered susceptibility have been reported (Chakraborty & Datta 2003; Garrett et al. 2006; Eastburn et al. 2010). More generalities can be made in the context of plant–insect herbivore interactions at elevated [CO2], but most of these again apply to C3 plants, which also exhibit changes in tissue primary metabolism and nutritional quality (Bryant et al. 1983; Lincoln et al. 1986; Herms & Mattson 1992; Schadler et al. 2007). Nevertheless, in pursuit of understanding the inconsistencies in plant secondary metabolism at elevated [CO2], studies have demonstrated that CO2 enrichment can also have a direct effect on plant defence hormones. Based on these studies, elevated [CO2] suppresses the production of jasmonic acid (JA) and lipoxygenase (LOX)-dependent defences while increasing the production of salicylic acid (SA) and corresponding downstream defences (Zavala et al. 2009, 2013; Casteel et al. 2012a; DeLucia et al. 2012). The mechanism by which elevated [CO2] influences the phytohormone defence response is not well characterized. However, the antagonistic effects of SA appear to play an important role in the down-regulation of JA, and it has been hypothesized that changes in the plant redox state at elevated [CO2] activate the regulatory protein non-expressor of pathogenesis-related genes 1 (NPR1), which mediates this antagonism between SA and JA (Spoel et al. 2003; Spoel & Loake 2011).
The potential of rising [CO2] to influence plant LOXs is of particular interest in the context of mycotoxigenic fungal pathogens because there is growing evidence that oxylipin signalling is involved in both pathogenesis and stimulation of mycotoxin production (Gao et al. 2007; Tsitsigiannis & Keller 2007). Oxylipins are a diverse class of oxygenated fatty acids, derived from the LOX pathway (Feussner & Wasternack 2002; Howe & Schilmiller 2002; Howe 2004; Gao et al. 2007; Gao & Kolomiets 2009). Free fatty acids, typically linolenic and linoleic acids, are oxidized by LOXs at either the 9 or 13 carbon position of the chain. Depending on this regiospecificity they are referred to as 9-LOXs or 13-LOXs. The phytohormone JA is derived from linolenic acid via the activity of 13-LOXs (Howe & Schilmiller 2002). Predominant components of maize biochemical defences are controlled by the JA signalling pathway (Oikawa et al. 2004; Yan et al. 2012). For example, maize zealexin and kauralexin phytoalexins, which are rapidly induced in response to pathogen infection and show biological activity against a variety of pathogens, are partially regulated by JA (Huffaker et al. 2011; Schmelz et al. 2011). The defence-related functions of 9-LOX metabolites are not as well characterized, but there is evidence that plant 9-LOX derivatives play an essential role in host susceptibility to mycotoxin contamination (Burow et al. 1997; Gao et al. 2007; Christensen & Kolomiets 2011). Expression of the maize gene LOX3 is up-regulated by F. verticillioides in lines that accumulate higher levels of mycotoxins (Calvo et al. 1999; Gao et al. 2007). Furthermore, disruption of maize LOX3, which resulted in reduced 9–LOX-deriveved fatty acid hydroperoxides, dramatically reduced fumonisin production on mutant maize kernels.
In this study, we investigated the impact of elevated [CO2] on the interaction between maize and F. verticillioides. The challenges associated with controlled environmental growth that are necessary to conduct regulated pathogen research with mature maize at different [CO2]s were overcome by establishing a correlation between the susceptibility and mycotoxin contamination in maize kernels and stalks. Utilizing the stalk infection assay as a model, we tested our hypothesis that C4 maize grown at elevated [CO2] would be compromised in LOX-dependent hormone signalling, which would influence both maize phytoalexin accumulation and F. verticillioides fumonisin contamination. In addition, we discuss differences in the impact of elevated [CO2] on C3 and C4 plant hormone signalling, and consider the implications of our findings on the future of maize crop productivity in the context of changing climate.
Materials and Methods
Plant material and growth conditions
All experiments were conducted using hybrid Z. mays var. Golden Queen, which is a sweet corn commonly grown for fresh market throughout Florida (Southern States Cooperative, Inc., Richmond, VA, USA). Four seeds were germinated in each plastic pot (10.5 × 10.5 × 12 cm high for stalk experiments or 25.5 cm diameter × 23 cm depth for cob/kernel experiments) filled with MetroMix 200 (Sun Gro Horticulture Distribution, Inc., Bellevue, WA, USA) supplemented with 14-14-14 Osmocote (Scotts Miracle-Gro, Marysville, OH, USA). Plants were watered daily and received bi-weekly nutrient supplement with soluble Peters 20-20-20 (The Scotts Company, Marysville, OH, USA). Mature maize plants utilized for F. verticillioides ear rot experiments were grown in two rooms of a previously described eight-room sunlit greenhouse on the University of Florida campus at Gainesville (Zhang et al. 2014). Temperatures were controlled at 28 °C from 0700 to 1900 h Eastern Standard Time (EST) followed by 1 h of smooth transition to 25 °C overnight. Air relative humidity was controlled between 55 and 60%. Maize for stalk rot experiments were grown in two identical environmental Conviron E15 (Pembina, ND, USA) growth chambers controlled at 28 °C day/25 °C night, 500 μmol m−2 s−1 photosynthetic photo flux density 12 h photoperiod and between 50 and 60% relative humidity (Supporting Information Fig. S1). One chamber/greenhouse room was controlled at current ambient 400 μmol CO2 mol−1 air (1 × [CO2]) and the other at 800 μmol CO2 mol−1 air (2 × [CO2]), which is the [CO2] the global atmosphere is projected to reach between 2080 and 2100 (Solomon et al. 2007).
Maize inoculation with F. verticillioides
F. verticillioides (Northern Regional Research Laboratory stock no. 7415) was grown on V8 agar for 1 week prior to use. Inoculum was prepared by collecting spores with a moist sterile cotton swab which was then stirred into sterile 0.1% Tween 20 (Fisher BioReagents, Fairlawn, NJ, USA). Spores were quantified using a TC10 Automated Cell Counter (Bio-Rad, Hercules, CA, USA).
Maize cobs were inoculated 2 weeks after pollination by pricking 10 kernels (0.5 cm in depth) on each cob with a 25 gauge needle and syringe filled with 1 × 106 spores mL−1F. verticillioides inoculum (or control 0.1% Tween 20 solution). Following inoculation the sheath was pulled back over the cob and covered with a moist paper towel to maintain high humidity. Treated kernels from two cobs were combined for each of four biological replicates. Consistent results were obtained from two independent experiments conducted between March and May and May and July of 2013. [CO2]s were switched between rooms for the repeated experiment.
Maize stalks were inoculated 30 d post sowing at the eight-leaf stage of growth by injecting 0.1 mL of 1 × 106 mL−1F. verticillioides spore solution along a 5 cm vertical incision through the apical meristem. Mock-inoculated (control) stems were identically damaged and treated with sterile 0.1 mL of 0.1% Tween 20 solution. Samples were collected by shaving a 2–3 mm width of tissue from both sides of the inoculation site with a scalpel. Metabolite quantities and the amount of pathogen DNA relative to maize DNA was determined from an average of four biological replicates of which each sample was a pool of eight treated stems. Comparable results were obtained from four independent experiments. Chamber [CO2]s were switched between repeated experiments.
To validate that pathogen DNA relative to maize DNA was correlated to spread of disease, the lesion area was analysed by inoculating maize stalks as previously described by Gao et al. (2007), with some modifications. Stalks of 4-week-old plants grown at 1 × [CO2] or 2 × [CO2] were wounded 1 cm in depth with a sterile 200 μL pipette tip (Rainin, Columbus, OH, USA) at a 45 angle. The tip was filled with 100 μL of 1 × 106 mL−1F. verticillioides conidial suspension and left in place. The conidial suspension was absorbed within 24 h and the tip fell out on its own as the plant continued to grow. Disease symptoms and severity were assessed 7 d post-inoculation. Stalks were spit longitudinally and lesion area was calculated by ImageJ software (ImageJ 1.46r; Wayne Rasband, NIH, Bethesda, MD, USA). Results represent the combined average of two independent replicate experiments of six plants at each [CO2].
Quantification of F. verticillioides biomass
DNA was extracted from infected maize tissues using a modified CTAB method (Nicolaisen et al. 2009). Briefly, 100–200 mg of homogenized tissue was mixed with 300 μL sterile water and 700 μL CTAB buffer (20 g L−1 CTAB, 1.4 M NaCl, 20 mM Na2EDTA, 0.1 M Tris-HCl) and incubated at 65 °C for 30 min. After centrifugation, the supernatant was transferred to a new tube containing 500 μL of chloroform (Sigma Aldrich, St. Louis, MO, USA), mixed, and separated again by centrifugation. The aqueous phase was then mixed with two volumes CTAB precipitation solution (5 g L−1 CTAB, 0.04 M NaCl, pH 8) and incubated at 4 °C overnight. DNA was then pelleted and dissolved in 350 μL 1.2 M NaCl and 350 μL chloroform. The aqueous phase was again separated and precipitated in isopropanol followed by 70% ethanol. Fungal DNA was extracted from 1-week-old mycelium of F. verticillioides grown on V8 agar using the same technique.
The amount of pathogen DNA relative to plant DNA was estimated with qRT-PCR. To determine the quantity of F. verticillioides DNA the previously described (Visentin et al. 2012) gene-specific primers for beta-tubulin (KC964147) were utilized: Fv_TUB2_F (5′-TGCTCAT TTCCAAGATCCGCG-3′) and Fv_TUB2_R (5′-GTAGTTGAGGTCACCGTAGGAGG-3′). Plant DNA quantification was performed using Ef1α gene-specific (NM_001112117) primers: Zm_Ef1α_F (5′-GCTTCACGTCCCAGGTC-3′) and Zm_Ef1α_R (5′-ATAGGCTTGGTGGGTATCA-3′).
A fivefold dilution series of pure F. verticillioides DNA and pure maize DNA was used to generate standard curves by plotting the Ct values obtained by qRT-PCR against the log( [DNA] ) (Nicolaisen et al. 2009). In order to ensure that the priming efficiency was not altered by mixtures of plant and pathogen DNA, standard curves for both primer pairs were also generated with mixtures containing DNA of both species; however, no significant difference was detected between the curves with pure or mixtures of DNA. The R2 values were 0.993 and 0.990 for the F. verticillioides and maize DNA curves, respectively.
Quantification of fumonisin
The Veratox fumonisin kit (Neogen, Lansing, MI, USA) was used to quantify fumonisin contamination caused by F. verticillioides in maize tissues grown at 1 × [CO2] or 2 × [CO2]. The kit uses a direct enzyme-linked immunosorbent assay to determine the total fumonisins (B1, B2, B3) between the quantization rage of 1 and 6 µg g−1 (Bird et al. 2002). As described earlier, the maize cobs and stalks were inculcated with F. verticillioides, the pathogen was allowed to establish for 7 d, and tissues were collected for analysis. Fumonisin was extracted from 1 g of ground infected tissue with 70% methanol, diluted 10- or 2.5-fold for infected kernel or stalk tissues, respectively, and then quantified following the kit manufacturers protocol. Results were calculated using Neogen's Veratox 3.0 for Windows software. The amount of fumonisin relative to F. verticillioides DNA was estimated by dividing the µg g−1 fumonisin by the pg of pathogen DNA as quantified by qRT-PCR.
Intracellular CO2 content (Ci), photosynthetic rate (Pn) and stomatal conductance (gs) measurements were taken with an open-flow portable leaf photosynthesis system (Li-Cor 6400XT, Li-Cor Inc., Lincoln, NE, USA) equipped with the standard leaf chamber (6 cm2 of leaf area) and [CO2] injection system (model 6400-01, Li-Cor Inc.) adjusted to a constant [CO2] concentration of 400 μmol CO2 mol air−1 or 800 μmol CO2 mol air–1. The environmental conditions in the Li-Cor 6400 chamber were adjusted to match those of the growth chambers. Measurements were performed 48 h after control or pathogen treatment between 1100–1300 h on the fifth leaf from the bottom for six individual plants at both [CO2]s.
Primary metabolite analysis
Soluble carbohydrate and starch concentrations were estimated using the previously described anthrone reagent colormetric method (Hansen & Møller 1975). In short, 200 mg of ground tissue was initially clarified to remove interfering pigments by incubating in 100% acetone. Soluble sugars were then extracted twice in 80% ethanol at 65 °C. The ethanol containing the soluble carbohydrates was then separated from the remaining residue, which contained the insoluble starch. The ethanol was lyophilized and the soluble carbohydrates were analysed following the same set of procedures as for starch analysis. Sugar residues were resuspended in 1.1% HCl and incubated in a 95 °C water bath for 30 min. Samples were then diluted to 10 mL with sterile water. Acidic sugar solution (1 mL) was mixed with 5 mL anthrone reagent (2 g anthrone L−1 of 72% sulphuric acid) and incubated for 11 min at 95 °C. The absorbance at 630 nm was then measured against water. Standard calibration curves were generated for glucose and starch (R2 = 0.995 and R2 = 0.993, respectively). Total soluble carbohydrates were calculated based on the calibration curve of glucose.
Total soluble protein was extracted from 200 mg stem tissue in 1 mL protein extraction buffer (10 mM Na2HPO4, 15 mM NaH2PO4, 100 mM KCl, and 2 mM EDTA). Samples were centrifuged for 15 min at 10 000 g at 4 °C and the quantity of protein in the supernatant was determined using the Pierce BCA Protein Assay Kit (Thermo Scientific, Rockford, IL, USA) following the manufacturer's microplate procedure.
Quantification of maize fatty acids, phytohormones and terpene phytoalexins
The quantification of free fatty acids, phytohormones, kauralexins and zealexins, were performed using a previously developed method (Schmelz et al. 2004, 2011; Huffaker et al. 2011). Metabolites were extracted with 1-propanol and methylene chloride, and collected by vapour-phase extraction following carboxylic acid methylation. Analysis was performed using GC/CI-MS and quantification was based on internal standards. Quantity estimates of total fatty acids, zealexins and kauralexins were based on 13C18-linolenic acid. The concentration of zealexins (1–14 as numbered in Huffaker et al. 2011) and kauralexins (A1–A3 and B1–B3) reported are cumulative total amounts of related metabolites described (Huffaker et al. 2011; Schmelz et al. 2011). Quantity estimates of total JA (trans + cis) and SA were based on corresponding deuterated internal standards.
Zealexin and kauralexin activity assays
Compound isolation and antifungal assays were performed as previously described (Huffaker et al. 2011; Schmelz et al. 2011) with minor modifications for the zealexin mixture. The mixture of oxygenated zealexins was extracted from infected maize stem tissue with ethyl acetate and solid debris were removed by filtration. The organic layer was concentrated under vacuum and then separated by preparative flash chromatography (CombiFlashRf, Teledyne ISCO, Inc.; Lincoln, NE, USA) on a 130 g C18 (RediSepRF Gold, Teledyne ISCO, Inc.) column in a gradient from hexane to ethyl acetate. A fraction consisting almost exclusively of oxygenated zealexins was obtained from this separation. The mixture was composed of approximately, 50% zealexin #14, 40% zealexin #6, 10% zealexin #13 and 5% zealexin #12, following the same compound number system described by Huffaker et al. (2011). Without any further purification this mixture of oxygenated zealexins was dried under vacuum, dissolved in dimethyl sulfoxide (DMSO), and used for the activity assay. Kauralexin B3 was isolated as previously described (Schmelz et al. 2011), dissolved in DMSO as the carrier solvent, and similarly used for analysis. Using 96-well microtiter plates, fungal inoculum (2 × 104 spores mL−1) was incubated with 0.5 μL of either phytoalexin treatment in DMSO or a DMSO control. Fungal growth in broth media was monitored through periodic measurements of changes in 405 nm optical density using a Synergy4 (BioTek Instruments, Inc.; Winooski, VT, USA) reader.
Phytohormone and gene expression time course
The time course experiment was performed by inoculating maize stems as described earlier. Tissue samples were collected immediately after inoculation (time 0 min), 15 min, 30, 60 and 120 min post-inoculation. All samples were frozen in liquid N2 and stored at −80 °C till further analysis. Four biological replicates each derived from two pooled individual plants were collected for each time point except 30 min, for which eight biological replicates were collected also as pool of two plants in each. Phytohormone analysis was conducted as described earlier with all biological samples collected. For gene expression analysis, however, in order to reduce the number of 30 min biological replicates, pairs of two RNA sample extractions were combined so that only four (instead of eight) replicates consisting of a pool of four independent plants were analysed. Samples from other time points were utilized as collected.
Quantitative real-time PCR
Total RNA was isolated from 200 mg ground tissue with NucleoSpin RNA plant kit (Machery-Nagel, Bethlehem, PA, USA) as per manufacturer's protocol. cDNA was synthesized using SuperScript® II reverse transcriptase (Invitrogen, Carlsbad, CA, USA) and random decamers. qRT-PCR was performed in a total 20 μL volume with 10 μL of 2× SsoAdvanced SYBR Green supermix (Bio-Rad, Hercules, CA, USA), 300 nM of each primer and 1 μL of template. Sample reactions were all performed in triplicate. Samples were diluted 10-fold before being used as template in PCR. PCR was performed on a Bio-Rad CFX96 Real Time System using the thermocycling programme, which consisted of an initial denaturation at 98 °C for 2 min, 40 cycles each of 98 °C denaturation for 15 s and 60 °C annealing/elongation for 1 min and a final dissociation curve from 65 to 95 °C.
Threshold cycle (Ct) values of LOX genes, AOS, AOC and NPR1 were normalized to the geometric mean of two endogenous controls, glyceraldehyde-3-phosphate dehydrogenase (X07156) and ribosomal protein L17 (RPL17; AF034948) (Vandesompele et al. 2002). The amount of each gene transcript was calculated relative to its corresponding average for mock-treated controls at time zero. Transcript fold-changes were calculated by the 2−ΔΔCt method (Schmittgen & Livak 2008) using CFX Manager software, version 3.0 (Bio-Rad). Calculations were adjusted according to primer efficiencies. Gene-specific oligonucleotides and their efficiencies are listed in Supporting Information Table S1.
Student's t-tests were performed to compare variable means at 1 × [CO2] with those at 2 × [CO2]. Variables for which comparisons were made included: relative F. verticillioides biomass, lesion area and quantity of fumonisin. For the time course experiment, differences between phytohormone concentrations or transcript levels at 1 × [CO2] in comparison with 2 × [CO2] were determined independently for individual time points with Student's t-tests.
For experiments with multiple explanatory factors (1 × [CO2] or 2 × [CO2] and mock-treated control or F. verticillioides inoculation), a more complex statistical analysis was required. To initially determine which factors contributed to differences among means, a 2 × 2 ([CO2] × F. verticillioides) full factorial analysis of variance (anova) was performed. To improve the power of the test, in the case of no interaction between potential contributing factors, a pair wise Student's t-test was performed on the means of the main effect. (For example, if there is no interaction and F. verticillioides inoculation was the main effect of differences, the values form control treatment at 1 × [CO2] and 2 × [CO2] were combined and then compared with the average of the values from F. verticillioides treatment at 1 × [CO2] and 2 × [CO2]. Thus, only two means were compared and each mean consisted of an average of eight samples instead of four.) If, however, there was a significant interaction between the factors, a Tukey–Kramer honestly significant difference test was performed to determine differences between individual means.
Elevated [CO2] increased maize susceptibility to F. verticillioides, but did not alter fumonisin contamination
The effect of [CO2] on maize susceptibility to F. verticillioides caused ear rot and stalk rot was evaluated by comparing the severity of F. verticillioides infection on maize grown at 1 × [CO2] (400 μmol CO2 mol−1 air) and 2 × [CO2] (800 μmol CO2 mol−1 air) (Fig. 1a,b). Assuming that the relative quantity of F. verticillioides DNA to maize DNA is associated with the amount of F. verticillioides pathogen within the tissue (Nicolaisen et al. 2009); F. verticillioides biomass was approximately 2.5-fold higher in maize kernels and stalks at 2 × [CO2] as determined by qRT-PCR (Fig. 1c,d; P < 0.02). However, fumonisin did not correspondingly increase with fungal biomass, and there was no significant difference in fumonisin contaminants at 2 × [CO2] in either the kernel or the stalk tissues (Fig. 1e,f). Hence, the amount of fumonisin per pg of F. verticillioides DNA was almost twofold lower in kernels and stalks grown at 2 × [CO2] as compared with 1 × [CO2] (Fig. 1g,h; P < 0.01). Although magnitudes varied, the general trend of maize susceptibility to F. verticillioides pathogen proliferation and fumonisin contamination in stalk tissues was analogous to that of kernels on the cob. Given the practical advantages of working with younger vegetative tissues, the stalk infection assay was used to investigate mechanisms behind increased fungal susceptibility.
To initially verify that increased pathogen DNA per unit plant DNA translated into increased disease spread, maize stems grown at different [CO2]s were point inoculated (as described in the Materials and Methods section) and the lesion area was measured 7 d following. Lesion area in maize grown at 2 × [CO2] was approximately threefold greater than at 1 × [CO2] (Fig. 2; P < 0.01), further confirming that maize at 2 × [CO2] was more susceptible to F. verticillioides stalk rot.
The occurrence of temporal events is central to ecological performance. Plant resistance or susceptibility to pathogens is largely dependent upon the speed of defence activation (Kim et al. 2011). Thus early time points at which initial responses could be detected were considered for metabolite analysis. Significant differences in relative F. verticillioides biomass at 1 × [CO2] and 2 × [CO2] could be detected as early as 2 d post-inoculation (Supporting Information Fig. S2; P < 0.01).
Primary metabolites were altered in F. verticillioides-infected tissues under elevated [CO2]
To confirm the occurrence of established patterns of C4 plant responses to elevated [CO2], the intracellular CO2 content (Ci), photosynthetic rate (Pn) and stomatal conductance (gs), were measured for both mock-treated (control) and F. verticillioides inoculated maize. As shown by the anova model, there was no significant interaction between [CO2] and F. verticillioides inoculation. Therefore, since [CO2] was the main effect contributing to differences, a pair wise comparison was performed between the mean Ci or gs at 1 × [CO2] and 2 × [CO2] (Fig. 3). As expected for C4 plants, 2 × [CO2] increased Ci and reduced gs (P < 0.01), but did not affect the Pn (P > 0.05).
Consistent with a lack of a difference in Pn, [CO2] alone was not a contributing factor to changes in soluble carbohydrate, starch or protein concentrations of control maize stem tissues (Fig. 4). However, in combination with F. verticillioides infection, the interaction of both factors resulted in significant differences in soluble carbohydrate and starch contents (P = 0.004 and P = 0.02, respectively). The soluble carbohydrate and starch content increased with F. verticillioides infection at 1 × [CO2], but not at 2 × [CO2]. Total protein content also increased with F. verticillioides infection but [CO2] was a significant contributing factor (anovaP = 0.06).
Since free fatty acids are key precursors in the biosynthesis of both 13-LOX and 9-LOX oxylipins (Calvo et al. 1999; Gao & Kolomiets 2009; Christensen et al. 2013), the concentrations of stearic, oleic, linoleic and linolenic acid were also measured (Fig. 5). The concentration of fatty acids increased with F. verticillioides inoculation at 1 × [CO2], but not at 2 × [CO2]. Hence, the concentration of linolenic acid was twofold less in F. verticillioides inoculated maize at 2 × [CO2] in comparison with corresponding tissues 1 × [CO2]. (Fig. 5c,d; P < 0.01).
Elevated [CO2] dampened the accumulation of JA and SA following F. verticillioides inoculation
To assess differences in defence hormone signalling at 1 × [CO2] and 2 × [CO2] the amount of JA and SA was determined for control and F. verticillioides inoculated tissues at various time points. Control tissues did not exhibit any significant differences in basal phytohormone levels (Fig. 6). Following the same pattern as the fatty acid precursors, at 2 d post-inoculation the concentration of JA increased 1.8-fold (Fig. 6a; P < 0.01). However, at the same time point there was no significant difference between JA levels in control and F. verticillioides inoculated tissues at 2 × [CO2]. At 2 d post-inoculation, [CO2] was not a contributing factor to differences in SA levels. F. verticillioides inoculation was the only effect resulting in the reduction of SA with pathogen infection (Fig. 6b).
In order to further evaluate the observed effects of [CO2] on defence response hormone signalling, a 2 h time course was conducted to track the initial hormone burst following F. verticillioides inoculation. The amplitude and duration of JA accumulation in inoculated tissues was significantly reduced at 2 × [CO2] (Fig. 6c). Within 15 min the concentration of JA in plants at 1 × [CO2] was 45% higher than plants at 2 × [CO2] (P < 0.05). The maximum difference of twofold was observed at 60 min (P < 0.01). The early accumulation of SA was also impacted by 2 × [CO2]. At 30 min post-inoculation, the concentration of SA was 43% less in stem tissues grown at 2 × [CO2] than at 1 × [CO2] (Fig. 6d; P < 0.05).
Elevated [CO2] reduced the induction of both 9-LOX and 13-LOX gene transcription and down-regulated the transcription NPR1 following F. verticillioides inoculation
To test the hypothesis that suppression of JA biosynthesis was regulated at least in part at the level of transcription, qRT-PCR was used to compare transcript levels of LOXs and two key JA biosynthetic genes, allene oxide synthase (AOS) and allene oxide cyclase (AOC), throughout the same time course following pathogen inoculation. At time zero, there was no significant difference between constitutive transcript levels of the JA-related genes analysed in plants at 1 × [CO2] in comparison with those at 2 × [CO2] (Fig. 7). Transcript abundance of LOX, AOS and AOC genes relative to 1 × [CO2] grown tissues at time zero, tended to increase with time following inoculation; however, with only a few exceptions, the up-regulation of LOXs at 1 × [CO2] was significantly higher than at 2 × [CO2] (Fig. 7). Transcript abundance of LOX1, which both has 9-LOX and 13-LOX functionality, was 73% less in 2 × [CO2] grown plants 60 min after inoculation (Fig. 7a). With the exception of LOX2, which was unaffected by atmospheric [CO2] (Fig. 7b), all other 9-LOX genes, LOX3, 4, 5, 6 and 12, had at least one time point at which transcript levels of infected 1 × [CO2] grown plants exceeded those of 2 × [CO2] grown plants. LOX3 transcript levels were approximately 2.5-fold higher 120 min after inoculation at 1 × [CO2] in comparison with corresponding levels at 2 × [CO2] (Fig. 7c). The 13-LOXs (LOX8, 9, 10, 11, 13), which have putative functionality in JA production (Chauvin et al. 2013), similarly showed a dampened transcriptional response at 2 × [CO2]. Consistent with reduced accumulation of JA at 2 × [CO2], the transcript abundance of LOX8, which is directly involved in the production of substrate for JA biosynthesis (Christensen et al. 2013), was 15-fold lower in 30 min inoculated tissues at 2 × [CO2] in comparison with 1 × [CO2] plants at the same time point (Fig. 7g). Transcript abundance of LOX10, which has been implicated in the production of green leaf volatiles (Bacon et al. 2008; Christensen et al. 2013), was the only LOX whose induction was stronger 2 × [CO2] (Fig. 7i). Even though transcript levels followed the same oscillating pattern at both [CO2]s, at 30 and 60 min after inoculation, the transcript level of LOX10 was three- and fivefold higher at 2 × [CO2], respectively. AOS and AOC transcription was not effected by [CO2]. Transcript abundance of both genes increased with time following inoculation, but there was no significant differences between the levels at 1 × [CO2] in comparison with 2 × [CO2] (Fig. 7m,n).
Because the early induction of SA was affected at 2 × [CO2], the transcript abundance of NPR1, a key regulator of SA-mediated systemic acquired resistance, was also evaluated throughout the time course. Constitutive (time zero) NPR1 transcripts were 2.5-fold higher at 2 × [CO2]. However, following F. verticillioides inoculation, NPR1 gene transcription was down-regulated at 2 × [CO2] while it was conversely induced at 1 × [CO2]. This inversion ultimately led to a three- and 3.5-fold higher transcript abundance of NPR1 60 and 120 min, respectively, after inoculation at 1 × [CO2].
Accumulation of phytoalexin defence metabolites was significantly reduced in maize at elevated [CO2]
In order to further evaluate the possible downstream effects of dampened LOX and JA signalling, the concentration of maize phytoalexins in control and F. verticillioides inoculated tissues at 1 × [CO2] and 2 × [CO2] were compared. Treatment with F. verticillioides strongly induced the production of zealexin and kauralexin metabolites at both [CO2]s, but in comparison with inoculated plants at 1 × [CO2], inoculated plants at 2 × [CO2] had 41% less zealexins and 29% less kauralexins (Fig. 8a,b). A mixture of oxygenated zealexins reduced F. verticillioides growth by 14% and purified kauralexin B3 alone reduced F. verticillioides growth by as much as 30% (Fig. 8c).
Our research demonstrates that elevated [CO2] enhances maize susceptibility to F. verticillioides, but does not alter the severity of fumonisin contamination (Fig. 1). The attenuated induction of maize defences at elevated [CO2] is consistent with increased maize susceptibly to F. verticillioides proliferation. In addition, reduced host-derived signals and defence metabolites, which have the potential to aggravate mycotoxin production, may result in the observed decrease in fumonisin per unit of F. verticillioides biomass. Nevertheless, when coupled with increased fungal biomass, even reduced host stimulants were sufficient to cause similar quantities of fumonisin in maize at 2 × [CO2] as compared with infected maize at 1 × [CO2].
Plant–pathogen interaction studies involving other grain crops have likewise shown increased host susceptibility at elevated [CO2] (Chakraborty & Datta 2003; Garrett et al. 2006). For example, growth at elevated [CO2] enhances rice susceptibility to Magnaporthe oryzae (Kobayashi et al. 2006) and wheat susceptibility to F. pseudograminearum (Melloy et al. 2010). However, our study is unique in that it exemplifies a C4 grain, and evaluates the effects of elevated [CO2] on maize defences and mycotoxin accumulation. To our knowledge, this represents the first report of the influence elevated [CO2] has on a C4 crop's defence response in interaction with a mycotoxigenic pathogen.
The lack of an increase in primary metabolites following F. verticillioides inoculation correlates with the dampened defence response in maize at elevated [CO2]
Carbohydrate signalling plays a critical role in the mobilization of plant defences and can even determine the outcome of plant–pathogen interactions (Bolouri Moghaddam & Van den Ende 2012). During a pathogen attack, an increase in endogenous sugar levels can enhance plant resistance (Roitsch et al. 2003; Morkunas et al. 2005; Gomez-Ariza et al. 2007), and many phytohormone-inducible defence-related genes are co-regulated by carbohydrates (Reinbothe et al. 1994; Rolland et al. 2002; Thibaud et al. 2004; Rolland & Sheen 2005). A considerable amount of energy is required for the cascade of defence reactions to be activated, thus a localized increase of carbohydrate supply is important in providing for this additional metabolic demand. The lack of an increase in soluble carbohydrates at 2 × [CO2] with F. verticillioides inoculation (Fig. 4a) could therefore compromise the defence response and intensify maize susceptibility to fungal proliferation.
The lack of an increase in free fatty acids following pathogen infection at 2 × [CO2] (Fig. 5) may contribute to impaired JA and 9-LOX oxylipin biosynthesis at the level of substrate supply. Previous reports have demonstrated that free fatty acid levels increase after wounding, herbivory or elicitor application (Conconi et al. 1996; Ryu & Wang 1998), and this increase of substrate (linolenic acid) is essential for the burst of JA which initiates the defence signalling process (Kallenbach et al. 2010). Interestingly, NPR1 is required for the activation mechanism that controls the influx supply of linolenic acid (Kallenbach et al. 2010); therefore, the down-regulation of NPR1 following F. verticillioides inoculation at elevated [CO2] (Fig. 7o) may influence this influx of substrate.
The attenuation of induced 9-LOX and 13-LOX gene transcription is consistent with increased F. verticillioides growth and reduced fumonisin per unit fungal biomass
Although functional elucidation of individual maize LOXs is still in progress, it is evident that they play an important role in resistance to fungal pathogens and mycotoxin accumulation (Gao & Kolomiets 2009; Christensen et al. 2013; De La Fuente et al. 2013). Elevated [CO2] reduces the induction of most LOX genes including LOX8 (Fig. 7g), which directly mediates JA production (Acosta et al. 2009). The burst of JA following pathogen perception is part of the phytohormone signalling process that regulates the accumulation of zealexin and kauralexin phytoalexins, which have the potential to reduce fungal growth (Fig. 8c) (Huffaker et al. 2011; Schmelz et al. 2011). Other LOXs not directly involved in providing substrate for JA biosynthesis have also been implicated in regulating maize resistance to pathogens. For example, disruption of LOX5 or LOX12 activity has been shown to increase maize susceptibility to F. verticillioides (Park et al. 2011; De La Fuente et al. 2013). Thus the depression of induced LOX gene transcription at 2 × [CO2] is consistent with enhance maize susceptibly to F. verticillioides growth. Furthermore, in support of the hypothesis that F. verticillioides fumonisin production can be stimulated by host 9-LOX oxylipins (Gao et al. 2007; Tsitsigiannis & Keller 2007; Christensen & Kolomiets 2011), the dampened induction of 9-LOX gene transcription following infection at 2 × [CO2] (Fig. 7c,d,e,f) coincides with less fumonisin production per unit F. verticillioides biomass (Fig. 1h). On lox3 mutant kernels, F. verticillioides vegetative growth was not impaired, but fumonisin production was reduced (Gao et al. 2007, 2009). The abated induction of LOX3 at 2 × [CO2] may similarly reduce fumonisin production with respect to fungal biomass.
Differences in transcriptional regulation are common in defence-related gene families and reflect their non-redundant/diverse functionality in plants. Maize LOXs are affected by and potentially function in response to wounding [LOX5 (Park et al. 2010); LOX8 (Christensen et al. 2013); LOX10 (Nemchenko et al. 2006)], fungal infection [LOX3 (Gao et al. 2008); LOX4 and 5 (Park et al. 2010); LOX6 (Gao et al. 2007); LOX10 (Nemchenko et al. 2006) ], herbivory [LOX5 (Park et al. 2010)], temperature stress [LOX10 (Nemchenko et al. 2006) ], drought stress [LOX11 (De La Fuente et al. 2013) ], circadian rhythm [LOX10 (Nemchenko et al. 2006)], and elevated [CO2] (based on the present study). Casteel et al. (2008) observed analogous regulatory differences in the soybean LOX gene family. While the expression of most soybean LOXs were affected by [CO2] and herbivory, others were not (Casteel et al. 2008). In maize the induction of LOX2 (presently unknown function) does not appear to be influenced by elevated [CO2] (Fig. 7b).
Elevated [CO2] influences phytohormone signalling pathways: a comparison between C3 soybean and C4 maize
Elevated [CO2] similarly increases the susceptibility of C3 soybean plants to herbivory by compromising LOX gene transcription, JA biosynthesis and JA-dependent antiherbivore defences; however, this is typically accompanied by an increase in SA production (Casteel et al. 2008; Zavala et al. 2008, 2009; Casteel 2010; Casteel et al. 2012a), which does not occur in maize (Fig. 6; Casteel 2010). Opposing the hypothesis of a potential antagonistic effect of SA on JA signalling, our findings suggest that elevated [CO2] also compromises the induction of SA (Fig. 6d). Furthermore, following F. verticillioides inoculation, the transcription of NPR1, which plays a central role in the activations of SA-dependent defence genes, is down-regulated instead of induced (Fig. 7o). Surprisingly, constitutive NPR1 transcripts are significantly higher in maize grown at 2 × [CO2], but it is possible that NPR1 is kept in an inactivated form under non-induced conditions (Mou et al. 2003). NPR1 transcription is not differentially regulated by [CO2] in soybean (Casteel et al. 2008), but because NPR1 mediates the antagonistic relationship between JA and SA signalling pathways it is hypothesized that NPR1 is post-transcriptionally activated through changes in the cellular redox state at elevated [CO2] leading to the down-regulation of JA-dependent defence gene expression (Casteel et al. 2012b). There does not appear to be a down-regulation of constitutive transcripts and/or phytohormone levels in maize that would suggest such a reconfiguration of phytohormone based defences as shown in soybean. The phytohormones are merely inadequately induced following F. verticillioides inoculation suggesting more of a desensitized or compromised state of maize grown at elevated [CO2] rather than a reconfiguration of defence signalling processes.
Potential implications of attenuated maize defences at elevated [CO2]
It is of increasing concern that predicted global environmental changes, including drought, warmer temperatures and increased insect herbivory, coincide with the major factors affecting maize productivity, predisposition to F. verticillioides infection, and fumonisin contamination (Dowd 2003; Chakraborty & Newton 2011; Wu et al. 2011; de Sassi & Tylianakis 2012). Our results support these concerns and indicate that increasing atmospheric [CO2] will enhance maize susceptibility to F. verticillioides by compromising the phyto-immune response. Fortunately, increased fungal biomass at 2 × [CO2] does not correspond with an increase in fumonisin thus resulting in an overall reduction in fumonisin per unit pathogen (Fig. 1). However, despite this reduction, our results do not suggest that elevated [CO2] will ameliorate the issues of fumonisin contamination in the future. With the augmentation of fungal biomass, the amount of fumonisin in infected tissues at 2 × [CO2] is not different from quantities in tissues at 1 × [CO2] (Fig. 1e,f) indicating that similar food safety issues will exist in the future at elevated [CO2].
The lack of an increase in mycotoxin contaminants and the reduced fumonisin per unit fungal biomass should not lessen the concerns of increased maize susceptibility to F. verticillioides disease at elevated [CO2]. Particularly in the context of a C4 crop, which will not benefit from a substantial stimulation in photoassimilation at elevated [CO2], increased productivity to meet growing crop demands will rely heavily on disease control. At the current [CO2], Fusarium rots are responsible for pre-harvest losses of approximately 8.6% of the global maize production (Nayaka et al. 2010). Crop losses could substantially escalate if the plants defence response is weakened at elevated [CO2] emphasizing the issues of food security. Furthermore, an increase in pathogen biomass per unit plant tissue could further increase transfer of inoculum between successive growing seasons, which would in turn aggravate disease development in future crops (Melloy et al. 2010).
While the main focus of our research has been centred on the maize-F. verticillioides interaction, a weakened defence response will likely also increase maize susceptibility to other biotic and abiotic stressors. LOXs function in the response to a variety of biotic and abiotic factors as described earlier. Similarly, JA is critical in the mediation of wound and anti-herbivore defences (Kim et al. 2003; Oikawa et al. 2004; Schmelz et al. 2009, 2011; Dafoe et al. 2011; Yan et al. 2012; Huffaker et al. 2013), but also plays a role in responses to drought, salinity and thermotolerance (Xin et al. 1997; Clarke et al. 2009; Yoon et al. 2009). Furthermore, maize zealexin and kauralexin phytoalexins are involved in resistance against multiple pathogens, including Fusarium graminearum, Aspergillus flavus and Rhizopus microsporus (Huffaker et al. 2011; Schmelz et al. 2011). Whether SA-dependent resistance against pathogens, such as Ustilago maydis, will be similarly effected in maize at elevated [CO2] requires confirmation, but our results suggest that they will be influenced.
The combined effects of elevated [CO2] and other abiotic and biotic stress factors on maize susceptibility to infection and mycotoxin contamination warrant further experimentation. If drought, heat and insect damage enhance fumonisin production during increased susceptibility to F. verticillioides proliferation at elevated [CO2], the current predictions, which already foresee increased fumonisin contaminants in the future maize crop (Miraglia et al. 2009; Chakraborty & Newton 2011; Wu et al. 2011; Magan et al. 2012), could be underestimated. However, it is also possible that the combination of elevated [CO2] and drought will negate the impact of elevated [CO2] alone, leaving maize susceptibility unchanged as demonstrated in soybean (Casteel et al. 2012a). Given the heightened climate change concerns and the paucity of knowledge regarding the issue, we hope that this study will instigate additional research regarding maize productivity, disease development, and mycotoxin contamination. Ideally in a more organic setting, which combines projected climatic scenarios with multiple abiotic and biotic stress factors in a field-based setting, such as those provided by free air gas concentration enrichment (FACE) experiments. Only through a better understanding of crop responses to individual and combined climatic conditions can effective agricultural management decisions and agronomic practices to be made for adaption to climate changes with minimal agro-economic loss and associated consequences.
We would like to thank Bevin Forguson for greenhouse management, Hoang Tang for technical support and Margaret Martin for logistical assistance. We gratefully acknowledge Zaid Abdo for advice on the statistical analysis. We would also like to especially acknowledge Steve Willms for keeping all the equipment operational, without his assistance none of this research would have been possible. We are exceptionally grateful to Dr. Paul Shirk for his selfless mentorship. The use of trade name, commercial product or corporation in this publication is for the information and convenience of the reader and does not imply an official recommendation, endorsement or approval by the U.S. Department of Agriculture or the Agricultural Research Service for any product or service to the exclusion of others that may be suitable.