Assessing the success of breeding maize inbred lines with contrasting diferulate concentrations

Abstract

Background

The crosslinking of maize cell wall components, particularly mediated by the formation of ferulic acid dimers or diferulates, has been associated with important crop valorization traits such as increased pest resistance, lower forage digestibility, or reduced bioethanol production. However, these relationships were based on studies performed using diverse unrelated inbred lines and/or populations, so genetic background could interfere on these associations.

Results

In the present research, the success of a pedigree selection program aimed to obtain inbred lines from a common antecessor with contrasting diferulate concentration was evaluated. From the 10 inbreds lines developed we could validate the success of the breeding program, obtaining 4 inbred lines with significant contrating values of total diferulate content in the pith tissues (two of each group): high (X̅= 0.69 mg/g of DW) and low (X̅= 0.35 mg/g). Ferulate changes in the same way were also observed: high (X̅= 3.09 mg/g of DW) and low (X̅= 1.62 mg/g). On the other hand, we found strong and positive correlations between DFAT and individual dimers, and moderate negative correlations between total DFAT and a main cell wall component such as cellulose. However, we did not find a significant effect of DFAT on maize valorization traits, except of a negative effect of DFAT on the concentration of sugars released after the enzimatic hydrolysis of the pith tissues. Interestingly, increasing DFAT in the pith does not seem to affect the digestibility of the forage or the saccharification of the stover residue, highlighting that changes in a specific tissue do not encompass correlated changes in other resources.

Conclusions

Overall, we have obtained contrasting inbred lines with diferulates concentration, which could be uselful in further studies focussing in the identification of regions/genes predominantly involved in the hydroxycinnamate biosynthesis pathway and cell wall crosslinking network.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12870-024-05990-w.

Introduction

The primary cell wall of maize is composed of cellulose microfibrils embedded in a matrix of hemicellulose (mainly glucuronoarabinoxylans (GAX), and β-glucans with mixed bonds, hydroxycinnamic acids, pectins (homogalacturonans and rhamnogalacturonans) and structural and signalling proteins, whereas secondary cell wall consists mainly of cellulose, GAX, hydroxycinnamic acids and lignin [1, 2]. The most common hydroxycinnamates in maize cell wall are p-coumaric (pCA) and ferulic acid (FA). Lignins are acylated (primarily syringyl units) at the γ-position by pCA [3]. This acylation having a marked influence on the bonding mode of S lignin units, on the spatial organization of lignins and consequently on their capacity to interact with polysaccharides. On the other hand, FA can undergo dehydrodimerization, and the resulting diferulates would crosslink heteroxylans enhancing cell wall stiffening and strengthening and promoting cell wall growth cessation [4]. Besides, during lignification, the FA and diferulic esters form cross-links through the etherification of the phenolic hydroxyl group to lignin polymers, creating a polysaccharide-lignin matrix that makes cell walls more recalcitrant to degradation (Fig. 1) [5, 6].

Fig. 1.

Linkage between two arabinoxylan chains of the cell wall mediated by the 5–5 diferulic acid and a combination of possible ester linkages between ferulates

The crosslinking of maize cell wall components, especially mediated by dimers of ferulic acid molecules or diferulates, has been associated with important crop characteristics such as stress resistance [7, 8], forage digestibility [9, 10], and bioethanol production [11, 12] (Fig. 2; [13–24]). In this sense, research has emphasized the importance of hydroxycinnamates in plant defences against pests and diseases; resistant genotypes exhibited higher quantities of hydroxycinnamate monomers and various diferulic isomers in their pith tissues compared to susceptible varieties [7]. Additionally, xylan-to-xylan ferulate linking and ferulate-to-lignin cross-links impede the enzymatic breakdown of cell wall polysaccharides. These cross-links increase cell wall recalcitrance to deconstruction, negatively impacting digestibility and limiting energy intake in livestock [10]. Similarly, the structural reinforcement of cell walls through the formation of ether bonds between feruloylated arabinoxylans and lignin G units, creates a polysaccharide-lignin matrix that hinders enzymatic hydrolysis [9, 10]. This resistance to deconstruction underscores the need for targeted strategies to reduce ferulate cross-links, which hold promise for enhancing cell wall degradability. Such approaches are critical not only for improving forage digestibility but also for optimizing cellulosic ethanol production [25]. Improving forage quality often involves reducing cell wall esters to increase polysaccharide accessibility and facilitate hydrolysis, balancing the goal of enhanced digestibility with the structural functions of the cell wall [10].

Fig. 2.

Interactions of cell wall hydroxycinnamates crosslinking in biotic and abiotic stress and maize valorisation traits. Adapted from Barros-Rios et al. [58]

Direct alterations in cell wall structure can significantly impact the ultimate utilization of maize. Trying to elucidate this hypothesis, a selection program for total diferulates (DFAT) was implemented using a common genetic background. At Misión Biológica de Galicia (CSIC), a divergent selection program for total diferulates content in the pith tissues was initiated from the F₂ population obtained by crossing two maize inbred lines, EP125 (high DFAT content) and PB130 (low DFAT content) [26]. Three selection cycles were obtained for high and low DFAT, as described in Barros-Rios et al. [8, 27]. The selection was successful in varying the concentration of DFAT. Increased diferulate content led to increased cell wall rigidity, which serves as the primary deterrent against corn borer attack and development. The tunnels produced by the larvae of Mediterranean corn borer in the stalk were 29% longer in cycles with the lowest concentrations of diferulates [8]. Moreover, this increase in DFAT concentration was not linked with a significant decrease in yield, a phenomenon commonly observed when borer resistance is the main trait to be improved [28]. However, Jung and coauthors [29] investigated the impact of ferulate crosslinking on dry matter intake, milk production, and in vivo digestibility using the low ferulate sfe mutant, observing that lines harboring the mutation exhibited reduced crosslinking of lignin to arabinoxylans, and cattle fed with this mutant confirmed higher dry matter intake and milk production. In the same way, Barros-Ríos et al. [27] observed negative correlations between diferulate concentration and total cell wall polysaccharides degradability.

Divergent selection made it possible to evaluate the effect of DFAs in the same genetic environment. However, the genetic variability of the cycles makes it difficult to undertake more in-depth studies due to the heterogeneity of the material. Therefore, developing inbred lines from these cycles of selection allow us to study the correlations between traits in a common genetic background and under homozygous conditions. A pedigree selection program was initiated from the third divergent selection cycles to obtain inbreds with contrasting levels for DFAT with a common genetic background ((EP125 x PB130) F₂). Five inbred lines (named EPD1 to EPD5) were obtained after six selfing generations from the third cycle of selection targeting high DFAT content and other five inbreds (named EPD6 to EPD10) from the third cycle targeting low DFAT content (see further description in materials and methods section). All the lines obtained from this program are derived from same original material and have been maximized to differ in DFAT content while in previous studies, different lines derived from distinct original material.

The materials obtained represent a valuable resource for exploring the indirect effects of DFAT selection on cell wall composition and its exploitation for maize valorisation. Accordingly, the objectives of the current research are: (i) to evaluate the uselfulness and succes of the pedigree selection program to generate inbreds with contrasting levels of DFAT; (ii) to study side effects on other structural cell wall components; and (iii) to assess the correlation between DFAT and potential valorization traits such as pest resistance, animal digestibility, or bioethanol production.

Methods

Obtaining inbred lines with high and low concentration of diferulates

The pedigree selection program was initiated in 2010 from the third cycles of selection for high (C3H) and low (C3L) DFAT contents, these cycles previously developed by divergent selection from the (EP125 x PB130) F₂ population [8]. Briefly, the previous divergent selection process began in 2007 in Pontevedra at Misión Biológica de Galicia (Consejo Superior de Investigaciones Cientificas), using the F₂ progeny of the hybrid as the base population (cycle zero, C0). Approximately 150 C0 plants were chain cross-pollinated, where pollen from each plant was used to pollinate the next in sequence. Ears and the second internode below the main ear were harvested 55 days after silking, once the internode development finished and the black layer of the grain (indicator of physiological maturity) was formed. Total ester-linked diferulates (DFAT) in pith tissues of the internodes were quantified [14]. Based on mean DFAT concentrations in pith tissue from both parental plants, ears from the 15 highest and 15 lowest DFAT crosses were selected (10% selection intensity in each direction). Seeds from the selected ears were pooled to create high (C1H) and low (C1L) DFAT populations for cycle one. This methodology was repeated in 2008, with C1H and C1L populations serving as the base material for developing cycle two high (C2H) and low (C2L) populations and in 2009 for developing cycle three high (C3H) and low (C3L) [8].

In 2010, 10 plots of 15 plants from each selection cycle (C3H and C3L) were sown in Pontevedra. Seventy-five plants were self-pollinated (Self1) in each cycle (C3H and C3L) and 55 days after flowering, the main ear, and the second internode below the main ear of the corresponding plants were harvested for pith diferulate analysis [26]. The 40 plants/ears with the highest or lowest concentrations within C3H and C3L, respectively, were selected (mean values for high and low DFAT concentrations of 0.36 mg/g DW and 0.23 mg/g DW, respectively). Ten kernels from each selected ear were sent to the winter nursery in Chile (Dr Juan E Gebauer; Winter Nursery Service; Estación Experimental Guindos) to carry out a new generation of self-pollination (Self2). Although all the plants were self-pollinated, only one well-filled ear from each Self1 ear was returned. In 2011, 80 plots of 15 plants/plot corresponding to each of the 80 well-filled ears obtained in Chile were sown in Pontevedra, 40 lines for high and 40 for low diferulatesAll possible self-pollinations (Self3) were carried out within the 15 plant per row plots depending on the number of viable plants each year and, 55 days after flowering, well-seeded main ears and pith samples were collected from 2 plants per plot. DFAT analyses were carried out, and 15 Self2 families with the highest or lowest concentrations were selected according to the mean of the two plants analyzed per family (mean values for high and low DFAT concentrations of 0.51 mg/g DW and 0.18 mg/g DW, respectively). Within each family, the ear of the plant with the highest or lowest concentration in each case was selected. Ten kernels from each selected ear were sent to the winter nursery in Chile to carry out a new self-pollination (Self4). Although plants were self-pollinated, only one well-filled Self4 ear was returned. In 2012, 30 plots of 15 plants/plot corresponding to each of the 30 well-filled ears obtained in Chile were sown in Pontevedra, 15 for high and 15 for low diferulates. All possible self-pollinations (Self5) were carried out within the 15 plant per row plots depending on the number of viable plants each year and, 55 days after flowering, well-seeded main ears and pith samples were collected from 2 plants per plot. Diferulate analyses were carried out, and seven Self4 families with the highest or lowest concentrations were selected according to the mean of the two plants analyzed per family (mean values for high and low DFAT concentrations of 0.64 mg/g DW and 0.28 mg/g DW respectively). In each plot, the ear of the plant with the highest or lowest concentrations in each case was selected. Twenty kernels from each selected ear were sent to the winter nursery in Chile to carry out a new cycle of self-pollination (Self6). As usual, only one well-developed ear from each plot was returned. In 2013, the best 10 well-filled ears, 5 for high and 5 for low diferulates, received from the winter nursery were sown in Pontevedra in two 15-seed rows for seed multiplication purposes (see final pedigree formula in Table 1) (Fig. 3).

Table 1.

Means of the 10 inbred lines obtained for varying total diferulic acid concentration and its corresponding classification

Genotype	Pedrigree	Total Diferulates (mg/g)	CD (5%)	Classification (DFAT content)
EPD1	(EP125 x PB130) F2 DFAT AC3 syn1-23-1-2-1-2-1	0.63 abc	0.2805	INTERMEDIATE
EPD2	(EP125 x PB130) F2 DFAT AC3 syn1-69-1-1-1-1-1	0.47 dc	0.2805	INTERMEDIATE
EPD3	(EP125 x PB130) F2 DFAT AC3 syn1-70-1-1-1-2-2	0.55 abcd	0.3097	INTERMEDIATE
EPD4	(EP125 x PB130) F2 DFAT AC3 syn1-62-1-1-1-1-3	0.66 ab	0.2821	HIGH
EPD5	(EP125 x PB130) F2 DFAT AC3 syn1-59-1-2-1-1-1	0.71 a	0.2805	HIGH
EPD6	(EP125 x PB130) F2 DFAT BC3 syn1-68-1-1-1-2-1	0.39 d	0.2848	LOW
EPD7	(EP125 x PB130) F2 DFAT BC3 syn1-29-1-1-1-2-1	0.50 bcd	0.2821	INTERMEDIATE
EPD8	(EP125 x PB130) F2 DFAT BC3 syn1-14-1-2-1-1-1	0.52 bcd	0.2821	INTERMEDIATE
EPD9	(EP125 x PB130) F2 DFAT BC3 syn1-63-1-2-1-1-1	0.54 abcd	0.2867	INTERMEDIATE
EPD10	(EP125 x PB130) F2 DFAT BC3 syn1-27-1-2-1-2-1	0.32 d	0.3137	LOW

Different letters indicate significant differents at p value < 0.05; The 5% Confidence Difference (CD 5%) was calculated using the t critical value from Student’s t-distribution for a 95% confidence level and 7 degrees of freedom (t = 2.3646), multiplied by the standard error of each estimate

Fig. 3.

Squematic chart of the obtaining program for maize inbred lines with high and low diferulates concentrations

Resistance to Mediterranean Corn Borer (MCB) (Sesamia nonagrioides Lef.) attack

In the northwestern region of Spain, the impact of second-generation stem borer larvae is significant, leading to substantial losses in stover and grain yield [30]. The tunnels created in the pith of the stalk disrupt the movement of nutrients towards the developing ear and contribute to increased lodging rate [31, 32]. In order to study the relationship between pest resistance and DFAT, we infested plants using Mediterranean Corn Borer (MCB) eggs (Sesamia nonagrioides Lef.), the main corn borer species affecting the area of study.

Ten inbred lines for total diferulate concentrations (EPD1 to EPD10) were assessed in the field in Pontevedra, located in North Western Spain (42°25′ N, 8°38′ W, 20 m above mean sea level) (Table 1). The experimental design of trials consisted in a complete random block design with three replications. Each experimental plot comprised three rows, with 15 double-kernel hills per row. The spacing between consecutive hills within a row was 0.18 m, and there was a spacing of 0.8 m between rows, resulting in a final density of approximately 70,000 plants per hectare after thinning to one plant per hill. Local agronomical practices were adhered to throughout the experimentation process.

The evaluation trial took place under two years (2016 and 2017) and two infestation conditions each year (protected with insecticide and artificial infestation with MCB). The trials protected were treated with the insecticide INSECT 5G (Chlorpyrifos 5%) applied manually every 21 days to all plants within the plot. This treatment regimen began upon detection of the first MCB moths and continued until harvest time.

In each plot of the infested trial, plants were artificially infested with approximately 80–120 MCB eggs. These eggs were positioned between the sheath and the stem in one internode below the main ear [33, 34]. At harvest, the stalks of five plants were split lengthwise, and the lengths of the tunnels produced by the larvae damage were measured (in centimeters).

Cell-wall biochemical analyses of the internode pith tissues

The second internode below the main ear was collected from five plants per plot in the protected trials. Samples were collected 55 days after silking, considered as the time when approximately 50% of the plants showed silks. Biochemical analyses were conducted on dry pith tissues. Cell wall biochemical traits such as cellulose, neutral sugars and lignin monomeric composition were examined in isolated cell walls from those internodes, using a cell wall isolation protocol adapted from Mélida et al. [35].

Hydroxicinamates concentration (breeding trait)

The quantification of cell wall-bound hydroxycinnamates was assesed following the protocol described in Santiago et al. [36] Phenolic standards pCA and FA were purchased from Sigma-Aldrich Quimica SL, Madrid, Spain. The identities of FA dimers were confirmed by a comparison with the 5–5 standard or published retention times and UV spectra [37]. DFAT concentration was calculated as the sum of the following three identified and quantified DFAs isomers: DFA 8–O–4, DFA 5–5, and DFA 8–5. The DFA 8–5 concentrations were calculated as the sum of DFA 8 − 5-cyclic (or benzofuran) and 8–5-noncyclic (or open).

Polyssacharides contents

For liberation of sugars from hemicellulose and cellulose, 5 mg of cell walls were weighed and hydrolyzed at 212 °C by adding 2 ml of 2 M trifluoroacetic acid (TFA). Total sugar content was determined using the phenol–sulfuric method with glucose as standard [38]. Uronic acid contents were determined using the m-hydroxybiphenyl method [39], with glucuronic acid as standard.

Cellulose content was determined using the Updegraff method [39]. To assess the concentration of cellulosic sugars, the anthrone method was employed with glucose as standard [40].

Lignin content and monomeric composition

Total lignin content was determined using the Klason Lignin protocol starting from dry matter [41]. Monomer composition of lignin in the cell walls was determined by HPLC after alkaline nitrobenzene oxidation. The quantification of lignin degradation products generated through oxidation with nitrobenzene was conducted at 280 nm. Standards for p-hydroxybenzaldehyde (used as the H-subunit indicator), vanillin (used as the G-subunit indicator), and syringaldehyde (utilized as the S-subunit indicator) were employed for this purpose [13].

Cell wall enzimatic degradability

Pith cell walls were hydrolysed (20 mg/1.5 ml) in a mixture of Cellulase R10 (1%), Macerozyme R-10 (0.5%) and purified Driselase (0.1%) dissolved in sodium acetate 20 mM (pH 4.8). Aliquots were taken at 24 and 96 h, clarified by centrifugation and assayed for total sugars release [27].

Animal digestibility

Digestibility of the organic matter (DOM) was assesed using forage samples, that consisted in whole plants “including” the ears. Plots were harvested 55 days after silking (days from planting until half of the plants in the plot showed visible silks). Dry matter content between 28 and 35%. In each plot, from 2 to 10 plants plants were crushed and a representative sample was taken. Those samples were pre-dried at 35Cº in a forced air camera and finally dried at 60Cº in a stove. Dried samples were ground in a Wiley (Arthur H. Thomas, Philadelphia, PA) mill with a 0.75 mm screen before being analysed. DOM was determined by Near-infrared spectroscopy (NIRs) at Centro de Investigación Agrarias de Mabegondo (CIAM). The spectral information of the dried and grounded (1 mm) samples was obtained using a Foss NIRSystem 6500 monochromator spectrophotometer (Foss NIRSystem, Silver Spring, Washington, USA), located in an isothermal chamber (24 ± 1 ° C), provided with a rotation module that performs reflectance measurements in the spectral region between 400 and 2500 nm, at 2 nm intervals. The collection of the spectral data and the chemometric analysis of the data was carried out using the WinISI II v program. 1.5 (Infrasoft International, Port Matilda, PA, USA). The samples recognized as outliers were analysed by the in vitro digestibility procedure described by Tilley and Terry [42], modified by Alexander & McGowan [43].

Bioethanol potential

Saccharification is known as the hydrolytic degradation of carbohydrates to the constituent sugar monomers in the conversion of lignocellulosic biomass to ethanol. Saccharification efficiency estimates were obtained using stover samples, that include whole plants “without” ears, as main residue in second generation biofuels. Plots were harvested approximately 70 days after silking (days from planting until half of the plants in the plot showed visible silks) for sacharification determination. In each plot, from 2 to 10 plants were crushed and a representative sample was taken. Those samples were pre-dried at 35Cº in a forced air camera and finally dried at 60Cº in a stove. Dried samples were ground in a Wiley (Arthur H. Thomas, Philadelphia, PA) mill with a 0.75 mm screen before being analysed. Saccharification assays were performed as described in Gómez et al. [44]. Samples were pre-treated with 0.5 M NaOH at 90 °C for 30 min, washed four times with 500 µl sodium acetate buffer and subjected to enzymatic digestion (Celluclast CTec2, 7FPU/g) at 50 °C for 8 h. The amount of released sugars (nmol mg-1 material-1 h-1) was assessed against a glucose standard curve using the 3-methyl-2-benzothiazolinone hydrozone method [44].

Statistical analysis

An analysis of variance was made for total diferulate concentration and for all cell wall componens using the generalized linear mixed model procedure (PROC GLIMMIX) of SAS [45]. The comparison of means among inbred lines was done using Fisher’s protected lsmeans. Based on these analyses, inbred lines were categorized into three groups: high, intermediate, and low diferulate concentration. Mean comparisons among high and low groups were conducted for cell wall composition using Fisher’s protected lsmeans.

In addition, genotypic and phenotypic correlations were estimated to study the relationship between DFAT and other cell wall components of the pith tissues following Holland [46]. To explore the relationship between diferulic acid content and valorization traits such as resistance to stem borers, animal digestibility of the forage and suitability for bioethanol production from stover, simple regressions were performed. Only those variables for which the variation between lines had a p < 0.15 in the analysis of variance were considered, since in other cases, any association would be spurious or due to chance. Valorization traits were used as dependent variables, while total diferulate concentration served as the independent variable.

Results and discussion

Based on the biochemical results obtained, we can classify the inbred lines evaluated into three groups based on the total diferulate concentration: high (EPD4 and EPD5), intermediate (EPD1, EPD2, EPD3, EPD7, EPD8, EDP9), and low (EPD6 and EPD10) (Table 1). This classification was based on statistically significant differences in DFAT concentration identified through the analysis of variance, which was the breeding trait. Specifically, the high DFAT group (X̅= 0.69 mg/g of DW) clearly differs from the low DFAT group (X̅= 0.35 mg/g) (Table 1; Fig. 4). The intermediate group was established to include lines that did not significantly differ from either the high or low groups. For instance, the inbred lines classified as intermediate, which were obtained from C3L (EPD7, EPD8 and EPD9) did not differ significantly for DFAT concentration from the line with the lowest concentration (EPD10) nor from EPD4, classified in the high DFAT content group. Similarly EPD1 and EPD3 obtained from C3H did not differ significantly from the line with the highest concentration of DFAT (EPD5). Also, more specifically EPD3 did not differ from the lines with the lowest concentration (EPD6, EPD10). The inbred line EPD2 was selected from breeding cycles for high diferulic content, yet it does not significantly differ from the lines with the lowest diferulic content. In this case, despite the high heritability of DFAT, the selection process, which relied on evaluating only 2 plants for family selection and a single plant within-family, might not have been consistently effective because the limited sampling could introduce a bias due to environmental effects.

Fig. 4.

Contrast analysis among DFAT groups [high (EPD 4, 5) and low (EPD 6, 10)] differing in total diferulate concentration. Comparison on the figures were performed using the Games-Howell test for comparisons using the R package ggstatsplot::ggbetweenstats; the bar shows the significance

Overall, the selection has been susscessfull in order to obtain contrasting lines (homozygous material) for DFAT concentration, effectively meeting the primary goal of the study. The genetic gain achieved through this breeding program is evident, with DFAT concentrations in the high group being nearly double those in the low group and approximately three times higher than the average DFAT concentration reported in other genetic inbred materials [8]. Although the breeding program successfully developed inbred lines with significantly contrasting DFAT concentrations, it appears to be a limit to the reduction in DFAT with no response in comparison to previously reported amounts [8], this could be most likely due to the existence of a basal threshold required for cell wall strengthening by diferulates. This threshold may represent a physiological constraint that ensures the structural integrity of the cell wall, preventing further reductions in DFAT concentration below a certain amount.

In agreement with the qualitative classification, high and low groups significantly differed in DFAT, and all individual dimers (Table 2). Differences for all individual lines are showed in Supplementary Table 1. Diferences in DFAT were mainly due to differences in concentration of DFA 8–O–4 (30.49%) and DFA 8–5 dimers (48.57%), which, in this genetic material, are the most abundant dimers, being DFA 5–5 (15.02%) the dimer found in less amount. This in accordance with Barros-Rios et al. [47]who observed that within the pith of diverse maize inbred lines, the most prevalent dimers were DFAT 8–5 (52% of all measured DFAT), followed by DFAT 8–O–4 (35%), and DFAT 5–5 (13%). Besides, we found significant differences among high and low groups and individual inbred lines for FA (high X̅= 3.09 mg/g; low X̅= 1.61 mg/g) (Table 2; Fig. 4, Supplementary Table 1) and high genetic correlation between DFAT and FA in agreenment with previous studies [27, 48]. Dimerization of FA primarily occurs through peroxidase/H2O2 mediated reactions, yielding various types of dehydrodimers that cross-link xylans within the cell wall matrix [3]. As suggested by Barros-Rios et al. [27], the most probable alteration attributable to selection pressure was a potential change in the enzymatic activity of peroxidase or laccase enzymes. Furthhermore, FA is derived from phenylpropanoid metabolism and it plays a central role in feruloylation processes within grass cell walls [49]. Through ester linkages with the C5 hydroxyl of arabinosyl side chains on arabinoxylans, FA becomes incorporated into the cell wall matrix, participating in the cross-linking and polymerization of cell wall components [6, 50]. The emergence of candidate genes responsible for AX feruloylation has offered promising insights [51–54]. Nevertheless, tracking the dimerization within grass cell walls is still challenging. In the current research, due to the same trend of variation for DFAT and FA, we could suggest that variations in DFAT could be assisted with rises in the overall FA quantity, rather than just adjustments in genes particularly involved in the dimerization process. Additionally, we found strong and positive genotypic and phenotypic correlations between DFAT and all individual dimers as expected, but also with FA (Table 3).

Table 2.

Contrast analysis for inbred lines categorised for high and low diferulic acid concentration

	High	Low	p value < 0.05
Pith cell wall components and properties
p -Coumaric acid (mg/g)	10.05	9.92
Ferulic acid (mg/g)	3.09	1.62	*
Diferulic acid 5–5 (mg/g)	0.10	0.06	*
Diferulic acid 8-O-4 (mg/g)	0.21	0.11	*
Diferulic acid 8 − 5 (mg/g)	0.31	0.19	*
Total Diferulates (mg/g)	0.69	0.35	*
Lignin (%)	18.72	20.11
Subunit H (%)	14.19	15.86
Subunit G (%)	44.26	41.26
Subunit S (%)	41.61	43.16
Cellulose (mg/g)	240.14	280.84
Hemicellulose neutral sugars (mg/g)	120.21	144.05
Hemicellulose uronic acids (mg/g)	36.72	33.74
Sugar release after 24 h (mg/g)	232.95	278.66
Sugar release after 96 h (mg/g)	167.21	188.74
Valorization traits
Tunnel length (cm)	21.1	22.6
Saccharification efficiency (nmol mg-1 material-1 h-1)	92.95	85.01
Digestibility of organic matter (%)	64.07	61.94

* indicate significant differences at p value < 0.05

Table 3.

Genotypic and phenotypic correlations between total diferulic acid (DFAT) concentration and other cell wall components

Genotypic correlation

pCA

FA

DFA 5–5

DFA 8-O-4

DFA 8 − 5

KL

H

G

S

CEL

HEMICELNS

HEMICELUA

DFAT

ns

0.98

1

1

1

ns

ns

ns

ns

-0.12

ns

ns

Phenotypic correlation

pCA

FA

DFA 5–5

DFA 8-O-4

DFA 8 − 5

KL

H

G

S

CEL

HEMICELNS

HEMICELUA

DFAT

ns

0.76

0.82

0.92

0.95

ns

ns

ns

ns

-0.35

ns

ns

The correlations showed are significat at p < 0.05

ns: Non significant; pCA: p-Coumaric acid; FA: Ferulic acid; DFA 5–5: Diferulic acid 5–5; DFA 8-O-4: Diferulic acid 8-O-4; DFA8-5: Diferulic acid 8 − 5; DFAT: Total diferulic acids; KL: Klason lignin; H: Subunit H; G: Subunit G; S: Subunit S; CEL: Cellulose; HEMICELNS: Hemicellulose neutral sugars; HEMICELUA: Hemicellulose uronic acids

On the other hand, in the contrast analysis between high and low groups, we did not find differences among DFAT groups for any other main cell wall components (Table 2), proving as well that the modified genetic mechanism is particularly targeted on feruloylation and/or dimerization. Likewise, when studying the differences among all inbred lines, we found significant differences for hydroxycinnamates, but not for the other components of the cell wall (Supplementary Table 1).

Regarding correlations among DFAT and other cell wall components, we only found moderate negative correlations between DFAT and cellulose. Barros Ríos et al. [27] already reported that after two cycles of selection for low DFAT plants exhibited increased concentrations of total cell wall components such as glucose (main sugar of cellulose) [27]. Lower DFAT concentration and thicker cell walls where previsouly observed in the parental inbred line PB130 [14, 43]. Closer spatial arrangements between FA moieties due to thinner cell walls (lower glucose) could contibute in dimerization chances; although no significant diferences between high and low DFAT groups in the contrast analyses where observed for cellulose (high X̅= 240.14 mg/g; low X̅= 280.84 mg/g) (Table 2).

In multiple studies a greater amount of monomers, pCA and FA, and diverse DFAs isomers were found in borer resistant inbred lines compared to susceptible. Despite the fact that one of the objectives of this divergent selection program was to improve pest resistance [27], nor the groups of inbred lines nor the inbreds individually show significant differences for borer resistance parameters such as tunnel length (Table 2, Suplementary Table 1). It must be taken into account that plants have several defense mechanisms against the attack of corn borers, the high concentration of DFAT is one of them, but not the only one [26]. In fact, the inbred lines chosen as parents of the F₂, EP125 and PB130, significantly differ for DFATs but not for resistance to borer. The defense mechanism of EP125 could be the high cross-linking of cell walls, while the resistance of PB130 could be linked to the presence of thicker cell walls [26, 55]. As seen in this research, there is a slight negative correlation between DFAT and cellulose, which would cause that selecting for high DFAT concentrations would indirectly reduce the size of cell walls, blurring the differences for pest resistance by combining two resistance mechanisms in opposite directions.

Moreover, significant increases in stem tunneling resistance may also result from variations in lignin content, structure, and/or the cross-linkage between lignin and hemicelluloses [7]. This includes the formation of undetermined ferulate or diferulate-polysaccharide-lignin complexes bonded via ester-ether linkages. These factors may contribute to the intricate role of just ester DFAT in this specific context, which may be difficult to ascertain. On the other hand, higher concentrations of pCA in both rind and pith tissues have been correlated with increased resistance to corn borer damage [13, 24, 56], as well as increases in pCA exhibited a higher proportion of subunit S lignin, suggesting pCA’s involvement not only in lignin content but also in its composition and structure. S lignin, associated with increased pCA acetylation, has been linked to resistance against various biotic stresses [13]. In this particular material we did not find differences between high and low DFAT groups, but we identified randomly variations for this hydroxycinnamate in individual inbred lines (Suplementary Table 1).

The biological conversion of plant cell wall carbohydrates into fermentable sugars would depend on polysaccharide contents and is hindered by the embedding and/or cross-linkages of carbohydrates with lignins or p-hydroxycinnamic acids that impede the hydrolysis of carbohydrates [53]. However, in the present study, we did not find any significant relationship among the total diferulate content and dry matter digestibility or saccharification efficiency, nor in the contrast analysis where high and low groups did not differ, neither in the variance analysis and the simple regression. Mean values for DOM or SACC were lower than the obtained in other studies using maize hybrids, populations or great diversity panels (64% vs. 70% and 88 nmol/ mg material hour vs. 140 nmol/mg material hour, respectively) [57], although we have to consider that we used related and homozygous inbred lines. Originally, the selection program was carried out to study the influence of DFAT of the maize pith on pest resistance, since the pith is the tissue the larvae feed on [8]. It is important to highlight that digestibility and saccharification estimations were carried out using the forage (entire plant) or the stover, respectively, and the quantification and selection for DFAT concentration was performed in pith tissues. Maize stem constitutes as much as 46% of the plant’s dry weight and comprises separate tissues distributed between the rind and pith sections. Usually, the pith accounts for only a minor portion of the maize plant’s overall dry weight, typically lower than 5% [22]. The concentration of cell wall components in the maize pith could not be link to the final composition of the whole plant or stover, tissue-specific genetic control may be operating. Interestingly, when we focus in the pith tissues, 49% of variation for the amount of sugars released after 96 h hydrolysis was significantly explained by DFAT variations (Table 4). In this case, DFAT had a negative effect on sugar liberation in agreement with previous studies that reported that the crosslinking mediated by dimers could render the cell wall more resistant to enzymatic deconstruction [10, 12, 21, 27, 47].

Table 4.

Simple regression model and equation for significant traits in the anova (p < 0.15) according to total diferulates concentrations (DFAT)

Dependent variable	Model (Intercept +/- estimate)	p value	R2
Digestibility of organic matter (%)	68.31–6.54*DFAT	0.0758	0.18
Sugar enzymatic release after 96 h (mg/g)	162.20-138.50*DFAT	0.0023*	0.49

* significative at p < 0.05

Conclusion

As concluding remarks, we have obtained maize inbred lines contrasting for ferulate and diferulates concentrations in the stem pith, achieving a difference of up to threefold between lines without significantly affecting other cell wall components. Aditionally, these changes do not have interference (increasing or decrasing) in pest resistance, animal digestibility or ethanol production. This genetic material represents a valuable resource for future genetic studies aimed at identifying regions or genes involved in the biosynthetic pathways of feruloylation or dimerization.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

pCA

p-Coumaric acid

FA

Ferulic acid

DFA 5–5

Diferulic acid 5–5

DFA 8-O-4

Diferulic acid 8-O-4

DFA 8 − 5

Diferulic acid 8 − 5

DFAT

Total diferulates

KL

Klason lignin

H

Subunit H

G

Subunit G

S

Subunit S

CEL

Cellulose

HEMICELNS

Hemicellulose neutal sugars

HEMICELUA

Hemicellulose uronic acids

TL

Tunnel lenght

DOM

Digestibility of organic matter

SACC

Saccharification efficiency

Author contributions

R.A.M., R.S. conceived the study. R.A.M., R.S., and C.S. participated in its design and carried out the field trial and harvested the samples; R.S. and A.L.M. participated in sample collection and biochemical analysis; A.L.M., R.A.M., R.S. and A.B. performed data analysis; A.L.M. wrote the original draft; R.A.M., R.S., C.S. and A.B. got the financial support; all authors read and approved the final version of the manuscript.

Funding

This research was funded by the “Plan Estatal de Ciencia y Tecnología de España” (projects PID2021-122196OB-C21, and PID2021-122196OB-C22 cofinanced with European Union funds under the FEDER program). A. Lopez-Malvar’s scholarship for the PhD fulfilment has been granted by University of Vigo and by a contract charged to the project RTI2018–096776-B-C22. The funding body played no role in the study design, data analysis and manuscript preparation.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable in this study.

Consent for publication

Not applicable in this study.

Competing interests

The authors declare no competing interests.