Article pubs.acs.org/JAFC
Hydroxycinnamate Synthesis and Association with Mediterranean Corn Borer Resistance Rogelio Santiago,*,† Rosa Ana Malvar,‡ Jaime Barros-Rios,§ Luis Fernando Samayoa,‡ and Ana Butrón‡ †
Agrobiologı ́a Ambiental, Calidad de Suelos y Plantas (UVIGO), Unidad Asociada a la Misión Biológica de Galicia (CSIC); Departamento Biologı ́a Vegetal y Ciencias del Suelo, Facultad de Biologı ́a, Universidad de Vigo, Campus As Lagoas Marcosende, 36310 Vigo, Spain ‡ Misión Biológica de Galicia (CSIC), Apartado 28, 36080 Pontevedra, Spain § Department of Biological Sciences, University of North Texas, 1155 Union Circle #305220, Denton, Texas 76203, United States S Supporting Information *
ABSTRACT: Previous results suggest a relationship between maize hydroxycinnamate concentration in the pith tissues and resistance to stem tunneling by Mediterranean corn borer (MCB, Sesamia nonagrioides Lef.) larvae. This study performs a more precise experiment, mapping an F2 derived from the cross between two inbreds with contrasting levels for hydroxycinnamates EP125 × PB130. We aimed to co-localize genomic regions involved in hydroxycinnamate synthesis and resistance to MCB and to highlight the particular route for each hydroxycinnamate component in relation to the better known phenylpropanoid pathway. Seven quantitative trait loci (QTLs) for p-coumarate, two QTLs for ferulate, and seven QTLs for total diferulates explained 81.7, 26.9, and 57.8% of the genotypic variance, respectively. In relation to borer resistance, alleles for increased hydroxycinnamate content (affecting one or more hydroxycinnamate compounds) could be associated with favorable effects on stem resistance to MCB, particularly the putative role of p-coumarate in borer resistance. KEYWORDS: Zea mays, Sesamia nonagrioides, hydroxycinnamic acid, resistance, candidate gene
■
INTRODUCTION The Mediterranean corn borer (MCB, Sesamia nonagrioides Lef.) is the most important maize (Zea mays L.) pest in southern Europe; the species is distributed along the Mediterranean coast and the southern Atlantic coast up to southwestern France.1 In Spain, this Lepidopteran species has two complete generations and a partial third per year. Maize plants are more susceptible to damage by MCB larvae before flowering, but damage by the first generation is scarce because insect population pressure is low.2 However, the increased insect population of the second and subsequent generations could have a great impact on stalk lodging and yield because larvae feed mainly on the pith of maize plants, from tasseling to harvest, reducing stalk strength by tunneling stems and disturbing the movement of assimilates toward the developing kernel.1,3 Conventional breeding for reducing stem tunnel length by MCB has achieved limited gains due to the low heritability of the trait. In addition, yield can be indirectly reduced as a consequence of significant genetic correlation between tunnel length and grain yield in some genetic backgrounds.4,5 The search for structural and biochemical mechanisms used by the plant to resist MCB attack could render basic knowledge about the plant−insect relationship as well as suitable variables for performing indirect selection, if heritability estimates for these variables are higher than for tunnel length and the genetic correlation between resistance and the indirect selection trait is significant. Plants’ antinutritional defenses against insects can occur as both pre-ingestion to limit food supply and post-ingestion to © XXXX American Chemical Society
reduce nutrient value to the attacking insect. Cell wall strengthening via lignin and/or hydroxycinnamate deposition would be considered as a pre-ingestion defense because increased cell wall cross-linkage would act as a physical barrier for impeding insect access to food supplies. A possible role of lignin as an antinutritional defense against corn borer attack has been previously proved in a segregating population for lignin content,6 and studies have been done checking a similar role for hydroxycinnamates.7 The most common hydroxycinnamates found in a wide range of grasses are ferulate (FA) and p-coumarate (pCA). Cell wall hydroxycinnamates are derived from the phenylpropanoid pathway, which originates from phenylalanine and tyrosine (Figure 1). Although structurally related, they seem to have different functional roles within the cell wall.8 Ferulic acid is primarily esterified to arabinosyl residues of arabinoxylan chains, and feruloylated arabinoxylans are later cross-linked to G units of lignins via ether bonds. Maize cell walls can contain up to 5% ferulate monomers plus dimers.9 FA can be coupled by peroxidase- or laccase-mediated oxidative coupling to form a variety of dimers, cross-linking polysaccharide chains.10 Total diferulates (DFAT) usually refers to the sum of 8-O-4, 5-5, and 8-5 diferulates, but other diferulates are also present in maize cell wall, for a more minor part.11 In addition, the ability of ferulic acid to participate in ester linkages and phenol coupling Received: October 6, 2015 Revised: December 16, 2015 Accepted: December 22, 2015
A
DOI: 10.1021/acs.jafc.5b04862 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
Figure 1. Summary of best-known genes involved in the phenylpropanoid pathway. PAL, phenylalanine ammonia-lyase; C4H, cinnamate 4hydroxylase; 4CL, 4-coumarate-CoA ligase; CCR, cinnamoyl-CoA reductase; CAD, cinnamyl alcohol dehydrogenase; HCT, hydroxycinnamoyl-CoA shikimate/quinate hydroxycinnamoyltransferase; C3H, 4-coumarate 3-hydroxylase; COMT, caffeic acid O-methyltransferase; CCoAOMT, caffeoylCoA O-methyltransferase; F5H, ferulate-5-hydroxylase; CALDH*, cinnamaldehyde dehydrogenase (described in Arabidopsis thaliana).
that increased free p-coumaric acid content could be indirectly related to higher resistance by being an intermediate in cell wall strengthening due to increased lignin and/or hydroxycinnamate contents (Figure 1). As a second step, eight maize inbred lines differing in antibiosis against MCB larvae and resistance to stem tunneling by MCB were evaluated for cell wall hydroxycinnamates (ester-linked) concentration 30 days after silking (pCA, FA, and DFAT); the authors found significant relationships between resistance to tunneling by MCB (less tunnel length in cm under field infested conditions) and pCA concentration, as well as between pith antibiosis (reduced weights of larvae reared on the pith under laboratory conditions) and pCA, FA, and DFAT concentrations.22 The average values for hydroxycinnamate concentration in the maize pith of inbreds analyzed were 6495, 2180, and 94.7 μg/g of dry weight for pCA, FA, and DFAT, respectively.22 In addition, recurrent selection for reducing stem tunneling by MCB larvae tended to increase pCA and DFAT concentrations, although the changes were not significant.26 In subsequent studies stem tunneling by corn borers was, once more, negatively correlated with concentrations of pCA, 8-5-diferulate, and DFAT in four inbred lines,23 and three cycles of divergent selection for diferulate content in an F2 derived from a cross between inbreds differing for ester-linked hydroxycinnamate concentration and cell wall content resulted in materials with contrasting resistance to stem tunneling by MCB.24 All of these outcomes suggest a genetic relationship between hydroxycinnamate esters and resistance to stem tunneling by MCB larvae, although it should be confirmed by more precise research. A quantitative trait loci (QTL) mapping experiment on the F2 derived from EP125 × PB130 was designed to colocate genomic regions involved in hydroxycinnamate synthesis and resistance to stem tunneling. The specific pathways for
reactions confers on it the ability to covalently attach polysaccharide with lignin. The resultant product is a ferulate−polysaccharide−lignin complex bonded through ester−ether linkages.12 Thus, FA deposition in the wall may not only lead to cell wall cross-linking during plant growth and development but also regulate the nonrandom pattern of lignin formation within the wall.8,13 Cross-linkages either between lignins and arabinoxylans (G-FA-AX) or between arabinoxylan chains (XA-FA-FA-AX) strongly modify cell wall properties, including tissue stiffness.14 p-Coumaric acid is mainly esterified to the γ-position of the side chains of S lignin units,15 and lignified maize cell walls can contain up to 3% p-coumarate.9 The role of pCA in cell wall matrices is not fully understood. Several studies have shown a positive relationship between pCA incorporation into cell wall matrices and levels of lignin deposition, especially to S unit content in lignin16,17 because pCA may function as a radical transfer agent to aid in the formation of sinapyl alcohol and lignin radicals.18 In addition, pCA might induce specific geometrical properties of the lignin polymer due to steric hindrance and oxidative reactivity.8,19 Although incorporation of pCA into grass walls appears to be primarily associated with the lignin fraction, there are also reports of pCA being ester linked to arabinoxylans, although the esterification level of the hemicellulose by pCA is much lower (1:15) than that by FA.20 The putative role of hydroxycinnamates in maize resistance to stalk tunneling by MCB has been the focus of several studies.21−24 The levels of free phenolic acids in the pith of 13 inbreds differing in resistance to stem tunneling were studied, and researchers found that higher concentrations of p-coumaric acid could contribute to general resistance to MCB.21 However, no antibiotic effect of the p-coumaric acid or any other phenolic acid has been observed against MCB larvae.25 It was suggested B
DOI: 10.1021/acs.jafc.5b04862 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
Figure 2. Genetic map of the EP125 × PB130 F2 maize population. Quantitative trait loci (QTL) positions for pith cell wall hydroxycinnamates (pCA, FA, and DFAT refer to p-coumarate, ferulate, and total diferulate, respectively), response to Mediterranean corn borer attack (TL and KR refer to tunnel length and kernel resistance, respectively), and yield. in 1985 from the Instituto de Biologı ́a de Barcelona (IBB) as a version of the Canadian inbred line CO125; however, recent molecular studies have pointed out that the MBG accession and the inbred CO125 of Canadian origin were undoubtedly different (unpublished work)]. The hybrid EP125 × PB130 was self-crossed, and the resulting F2 was planted in 2011. Leaf material for DNA amplifications was collected separately from each individual when plants presented approximately four to five leaves and was frozen at −80 °C until DNA extraction was performed. The F2 plants were self-crossed to obtain 300 F2:3 families that were evaluated at field conditions in 2012 and 2013 by using an augmented design. Three hundred unreplicated F2:3 families were randomly assigned to 10 blocks, whereas 6 checks [EP125 (very high resistance to MCB),
each hydroxycinnamate component in relation to the better known phenylpropanoid pathway, which leads to monolygnol biosynthesis, are highlighted.
■
MATERIALS AND METHODS
Plant Material and Field Trials. The inbreds EP125 and PB130 were selected because they presented contrasting concentrations of hydroxycinnamates;22 PB130, an inbred line derived from the Spanish landrace “Rojo Vinoso de Aragón”, presented lower concentrations for p-coumarate, ferulate, and diferulates in the maize pith than EP125 [EP125 was named CO125 in previous studies because this inbred accessed to the Misión Biológica de Galicia (MBG) germplasm bank C
DOI: 10.1021/acs.jafc.5b04862 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
diferulates identified and quantified by this analytical procedure: 8-O4-diferulate, 5-5-diferulate, and 8-5-diferulate. Population Genotyping and QTL Analysis. DNA of 285 F2 individuals, the parents EP125 and PB130, and the hybrid was extracted according to the method of Liu and Whittier31 with modifications rendering samples for three 96-well plates; simple sequence repeat (SSR) amplifications were performed as described by Butrón et al.32 SSR products were separated after amplification by electrophoresis using 1 TBE on a 6% nondenaturing acrylamide gel (approximately 250 V for 3 h).33 First, the parents and the hybrid were amplified with >500 SSR pairs; 110 resulted clearly polymorphic between both parents, and both bands were amplified in the hybrid. Then, DNA from the 285 F2 individuals plus the parents and the hybrid was amplified with the 110 polymorphic SSR. A linkage map was created using SSR marker data by applying the software package MAPMAKER.34 A LOD (log10 of the likelihood odds ratio) threshold of 2 was used to declare two markers significantly linked, and a maximum distance of 50 (exceptionally 100) cM was used. The software MapChart35 was used to elaborate a graph of the genetic map (Figure 2). Heritabilities in a wide sense (ĥ2) were estimated for each trait on a family mean basis as described previously by Holland et al.36 considering F2:3 families and years as random effects. The genotypic (rg) and phenotypic (rp) correlations between traits were computed following Holland.37 The genotypic variance for each trait was computed with the Proc VARCOMP of SAS, method REML.27 The combined analysis of variance of the inbred checks (A509, EP42, EP125, and PB130) across years was performed; inbreds were considered as fixed effects and years and replications as random effects. The QTL analysis of individual traits was performed with 245 F2:3 families that had data for all traits in both years. The composite interval mapping (CIM) approach was conducted for QTL detection and estimation of the QTL effect, position, and variability explained using the software package PlabMQTL.38 All markers preselected by PlabMQTL as cofactors were used in QTL analysis using the statement cov SEL, and additive and dominant effects were included in the model. A permutation test with 1000 random reshuffles resulted in a LOD threshold of 2.7 for declaring significant a putative QTL assuming an experiment-wise error of 0.30. QTLs detected in the preliminary fit are shown. Following Utz et al.,39 a 5-fold crossvalidation (CV/G) approach using 1000 cross-validation runs was employed for determining QTL frequency. 40 The program JointQTL38 was used for co-localizing QTLs with additive effects for the three hydroxycinnamates (pCA + FA + DFAT) and for resistance traits and relevant hydroxycinnamates (tunnel length + pCA + DFAT and kernel resistance + pCA + DFAT) using a LOD threshold of 5. The experiment-wise error expected is 90% damaged, 2 = 81−90% damaged, 3 = 71−80% damaged, 4 = 61−70% damaged, 5 = 41−60% damaged, 6 = 31−40% damaged, 7 = 21−30% damaged, 8 = 1−20% damaged, and 9 = 0%;29 tunnel length as the average length in centimeters of stem tunnels made by borers in the infested plants; grain moisture at harvest measured as grams of water in 100 g of kernels; grain yield under insecticide protection [Y(I)] estimated on the subplot treated with insecticide; and grain yield under infestation with MCB [Y(S)] estimated on the infested subplot as Mg ha−1 at 140 g H2O kg−1. Five noninfested plants per family were collected for hydroxycinnamate analyses 55 days after silking, the time when all genotypes have reached the physiological maturity stage and a relationship between hydroxycinnamates and borer resistance has been observed in previous studies.21−24 The second elongated internode below the main ear was collected and frozen (−20 °C) for analysis. Each internode was dissected into pith and rind fractions, and subsequently pith samples were lyophilized and ground through a 0.75 mm screen in a rotor mill. Hydroxycinnamate extraction followed the procedure described by Santiago et al.22 with some minor modifications. Freeze-dried samples (500 mg) were extracted in 30 mL of 80% methanol and homogenized for 60 s with a Polytron mixer (Brinkman Instruments, Westbury, NY, USA). Samples were extracted for 1 h in the dark and then centrifuged for 10 min at 1000g. The pellet was hydrolyzed with 20 mL of 2 N NaOH and shaken for 4 h in the dark under a nitrogen atmosphere. The hydrolysis reaction was stopped by adding 6 N HCl, and the pH was adjusted to 2.0. Samples were centrifuged as above, and the supernatant was collected. The pellet was washed twice with 10 mL of distilled water and centrifuged, and both supernatants were combined. Then, samples were extracted twice with ethyl acetate (50 and 35 mL each). The organic phase was dried, dissolved in 1.5 mL of HPLC grade methanol, and stored at −20 °C. All standards and samples were filtered through a 20 μm pore poly(tetrafluoroethylene) filter (Chromatographic Specialties, Brockville, Canada) before analysis. HPLC analyses were performed on the basis of a previously described protocol,22 using a 2690 Waters separations module (Waters, Milford, MA, USA) equipped with a Waters 996 photodiode array detector and a Waters YMC ODS narrow-bore column (100 by 2 mm i.d., 3 μm particle size). Retention times were compared with freshly prepared standard solutions of p-coumaric and ferulic acids (Sigma, St. Louis, MO, USA) and of 5-5-diferulate synthesized by Dr. John Ralph’s group (Department of Biochemistry, University of Wisconsin, Madison, WI, USA). Identities of ferulic acid dimers were confirmed by comparison with the authentic 5-5 standard or retention time and UV spectra previously published.30 Total diferulate concentration was calculated as the sum of three regioisomers of
■
RESULTS AND DISCUSSION In the analysis of variance of the inbred checks, there were significant differences between the parental inbreds (EP125 and PB130) of the F2 population for all hydroxycinnamates, resistance to stem tunneling by MCB, days to silking, grain moisture, and yield under infestation with MCB (Table 1); the year × inbred interaction was not significant except for yield under insecticide treatment (data not shown). The level of resistance of EP125 to stalk tunneling by MCB was comparable D
DOI: 10.1021/acs.jafc.5b04862 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
LSMean comparisons of the parental inbreds (EP125 and PB130) are also shown. bpCA, FA, and DFAT refer to p-coumarate, ferulate, and total diferulate concentrations per gram of dried matter (DM). Kernel resistance was scored on a subjective visual scale from 1 to 9, in which 1 indicates completely damaged and 9 indicates no damage by the larvae. d(I), under insecticide treatment; (S), under infestation with Sesamia nonagrioides. eĥ2, heritability in a wide sense calculated following Holland.36 fThe asterisk (*) indicates that estimates for genotypic variance significantly differed from zero at 0.05 probability level using the Proc VARCOMP, method REML.27
to that of A509 and significantly higher than those of PB130 and EP42 (means ± standard deviations of EP42, PB130, A509, and EP125 for tunnel length were 54 ± 5, 27 ± 2, 22 ± 2, 18 ± 2, respectively). Low genotypic variability was detected among F2:3 families for traits related to resistance to MCB attack (Table 1), and this could be a consequence of the high severity of damage by MCB in 2012 (the average tunnel length was 36.4 cm in 2012 and 16.9 cm in 2013) that would mask the moderate differences for tunnel length detected under worse insect performance. Difference between EP125 and PB130 for tunnel length across years was significant, but no difference was detected between these inbreds in 2012. Even though artificial infestation guarantees a homogeneous distribution of the pest across genotypes evaluated in the same year, variances in the severity of damage among years are inevitable if environmental conditions differ substantially.42 However, as we are interested in finding out genetic effects stable across environments, QTL analyses were performed with average phenotypic values across years, and the stability of QTL detected was tested using the program PlabMQTL (data not shown). Heritabilities in a wide sense for traits related to resistance to MCB attack were low; meanwhile, heritabilities for agronomic and biochemical traits were moderate (Table 1). The pCA concentration could be partially responsible for differences in resistance to MCB because this hydroxycinnamate presented high genotypic correlation coefficients with kernel resistance (rg = 0.78) (Table 2). Previous studies have found a negative correlation between stem tunneling by MCB and pCA concentration in the pith stem,22,23 but no studies have been reported until now on the possible relationship between resistance to kernel damage by MCB and pCA concentration. Generally, MCB larvae enters into the maize ear through the shank after feeding on the pith for a period; therefore, any mechanism that interferes with pith consumption could secondarily reduce kernel damage.43 In addition, increased pCA content in the pith could contribute to reduce the risk of mycotoxin contamination in the kernels because kernel MCB damage influences fungal infection by Fusarium species.44 FA and DFAT concentrations were moderately correlated, but they did not show any significant correlation coefficient with traits related to damage by MCB (Table 2). On the other hand, Garcı ́a-Lara et al.45 found a low but significant genotypic correlation between FA and DFAT concentrations in the kernel pericarp and kernel damage by the maize weevil, Sitophilus zeamais (Motsch.). There was a significant but low relationship between DFAT content and days to silking, suggesting some linkage between genes for precocity and diferulate biosynthesis. These results are in agreement with results from other studies that evaluate divergent populations for DFAT concentration, with high DFAT populations showing a significant reduction in the flowering date (
