Three month-old oil palm seedlings from the commercial variety Dura × Pisifera (D×P) were obtained from Sime Darby Seeds and Agricultural Services Sdn. Bhd., Selangor, Malaysia. Each seedling was transplanted into a 5 L plastic pot, containing about 3.5 kg of sterilized potting medium (3:2:1 v/v/v topsoil:peat:sand). All seedlings were maintained in a greenhouse (relative humidity of 90 ± 5% and temperature of 33 ± 5°C) for 30 days (approximately 2–3 fully expanded leaves stage) and watered daily.

A pathogenic G. boninense strain (T10) obtained from Applied Agricultural Resources, Sdn. Bhd., Sungai Buloh, Malaysia was used to inoculate the seedlings. A fungal culture grown on potato dextrose agar (PDA) (Difco, Germany) was subjected to PCR verification using our in-house primers which amplify ITS regions specific to G. boninense (Govender et al., 2016). For large-scale inoculum preparation, the primary culture was inoculated on maltose extract agar (MEA) (Merck, Germany) plates to obtain secondary cultures. Rubber (Hevea brasiliensis) wood block (RWB) measuring 6 cm × 6 cm × 6 cm was air dried under the hot sun, soaked in distilled water for 24 h, and then sterilized at 121°C for 15 min. This step was repeated. Thereafter, approximately 100 ml of MEA was swirled evenly over the surfaces of each RWB. The MEA treated-RWBs were sterilized again and cooled to room temperature (28 ± 2°C). Following sub-culturing, the 7-day-old fungal culture was used for subsequent RWB inoculations. Each RWB was inoculated with three pieces of mycelial cubes (1 cm × 1 cm × 1 cm), with each individual cube macerated at three different sides of the RWB. The inoculated RWBs were incubated in a dark chamber for 4 weeks at room temperature (Govender et al., 2016). Only RWBs showing active colonization (all RWB surfaces completely covered by mycelia) were used for artificial infection of oil palm seedlings. For the disease trial, each seedling was carefully excavated from the potting media and the bole was then placed directly on top of a fully colonized RWB inoculum. Next, the roots were dispersed to the sides of the RWB (Nur Ain Izzati and Abdullah, 2008). The seedling together with the attached RWB inoculum, was replanted into the potting media.

An experiment of two treatments (control and infected) with four biological replicates per treatment and seven sampling times at 1, 2, 3, 4, 8, and 12 weeks post-inoculation (wpi) was arranged in a completely randomized design (CRD). Root tissues of the control (non-inoculated) and infected (inoculated) seedlings collected at 1, 2, and 3 wpi were used for gene expression study, while those collected at 4, 8, and 12 wpi were used for lignin content, lignin composition, enzyme activities (PAL, CAD, and POD), growth (height and weight) and disease assessment.

One-way analysis of variance (ANOVA) was performed to assess differences in growth (height and fresh weight), lignin content and enzyme (PAL, CAD, and POD) activities in the infected and control seedlings with the PROC GLM procedure using the SAS (version 9.3; SAS Institute) software. In presence of significant difference, the mean separation was performed using Fisher’s protected least significant difference (LSD) method at p < 0.05.

At the end of the 12-week disease trial, all infected seedlings showed BSR symptoms while the control seedlings were healthy. Only healthy seedlings with at least three green leaves (Figure 1A) were used as biological replicates in this experiment. The DS of the infected seedlings increased gradually over the sampling time points (Figure 1). At 4 wpi (DS at 5%), no external symptoms were observed in the infected seedlings (Figure 1B), however, root surface examination revealed a low level of fungal attachments (data not shown). At 8 wpi, the DS increased to 35%. The infected seedlings showed yellowing on the lower leaves (Figure 1C). At 12 wpi, the DS increased further to 67.5%. The infected seedlings showed yellowing and/or browning on leaves with button-like basidiocarp at the bole (Figure 1D). At >12 wpi, the seedlings showed complete leaf browning (Figure 1E). At the end of the disease trial, the bole cross section and root tissues of the control and infected seedlings were compared. Root tissues of the control seedlings were visually large and healthy-looking in absence of any lesion (Figure 1F), whereas the root tissues of the infected seedlings were relatively small and mycelium attachments were found at random positions (Figure 1G). Cross sections of the control bole showed thick succulent tissues with whitish to yellowish colouration (Figure 1H) whereas the infected bole showed extensive brown colouration (Figure 1I).

The fresh weight and height of the control and infected oil palm seedlings displayed a similar trend over a period of 12 weeks. The average fresh weight of the control seedlings increased significantly with sampling points. Initially, the average fresh weight of the 4-month-old control seedlings was 7.76 kg and increased by 1.5-fold (11.31 kg) at 4 wpi. At 8 wpi, the average fresh weight of the control seedlings showed a further 3.1-fold increase (35.48 kg), followed by another 2-fold increase (60.76 kg) at 12 wpi (Figure 2A). In contrast, the growth rate of the infected seedlings was generally lower than the control seedlings. At 4 wpi, the average fresh weight of the infected seedlings (9.46 kg) showed no significant difference relative to the control seedlings (11.31 kg). The average fresh weight of the infected seedlings increased by 1.9-fold (20.5 kg) at 8 wpi, followed by another 1.1-fold increase (23.04 kg) at 12 wpi.

The average height of the control seedlings was 29.3 cm, increased significantly to 35.8 cm at 4 wpi, followed by further significant increments with 44 cm at 8 wpi and 50.6 cm at 12 wpi (Figure 2B). At 4 wpi, the average height of the infected seedlings (30.4 cm) showed no significant difference compared to the control seedlings, while at 8 and 12 wpi, the respective height increased significantly to 33 and 34.5 cm. Significant differences in average height as affected by disease development were notable at 8 wpi onwards. The average height increment at 8 and 12 wpi in the infected seedlings were significantly lower than the corresponding control seedlings (Figure 2B). The growth differences (fresh weight and height) observed between the infected and control seedlings at the different sampling points demonstrated BSR disease manifestation.

Since lignification is regulated during growth, the biosynthesis of induced lignin during disease development was taken into consideration only when the lignin content of the infected seedlings was significantly higher than the control seedlings. Lignin content was significantly high (0.15 pM/μg) in the infected seedlings collected at 4 wpi (DS at 5%) and remained elicited (0.14 pM/μg) with no difference at 8 wpi (DS at 34%). Correspondingly, lignin contents of the control seedlings at 4 and 8 wpi were 0.08 and 0.14 pM/μg, respectively. At the advanced stage of BSR progression, 12 wpi (DS at 69%), lignin content of the infected seedlings decreased to 0.11 pM/μg, while the control seedlings had 0.14 pM/μg (Figure 3).

The lignin content of the infected seedlings at 4 wpi (DS at 5%) was significantly higher (0.15 pM/μg) than the control seedlings (0.08 pM/μg) and thus, confirmed the accumulation of disease-associated lignin or induced lignin. The accumulation of lignin at 4 wpi (DS at 5%) also demonstrated the early defense response of oil palm seedlings against G. boninense. At 8 wpi (DS at 34%), lignin content of the infected seedlings (0.14 pM/μg) showed no significant difference as compared to the control seedlings (0.14 pM/μg). Reduced lignin content in the infected seedlings at 12 wpi (DS at 69%) implied the occurrence of active colonization by G. boninense, which had successfully degraded lignin for nutrient acquisition. The lignin content in the control seedlings was significantly low at 4 wpi and then increased at 8 and 12 wpi. The rise in developmentally regulated lignin in the control seedlings at 4 and 8 wpi, were in parallel to the growth profile; the height and fresh weight at 4 wpi showed significantly lower performance than 8 and 12 wpi (Figure 2). Lignin content increase over the sampling time points in the control seedlings corroborated with the growth data (increased height and fresh weight).

To evaluate the monomeric composition of lignin, root tissues from each treatment (control and infected seedlings at different sampling points) were subjected to alkaline nitrobenzene oxidation analysis. Vanillin and syringaldehyde are non-condensed fractions of lignin obtained from this procedure (Chiang and Funaoka, 1988). Both vanillin and syringaldehyde are derived from the guaiacyl and syringyl monomers present in lignin, respectively. The relative abundance of syringyl and guaiacyl monomers is denoted as the S/G ratio. The nitrobenzene oxidation products from root tissues of oil palm seedlings yielded guiacyl and syringyl monomers. The S/G ratio of the control and infected seedlings ranged from 0.42–1.97 to 0.35–0.42, respectively (Table 2).

The S/G ratio of the infected seedlings at 4 wpi (DS at 5%) was 1.31, about three times higher than the control seedlings (0.42). The results demonstrated an apparent accumulation of syringyl-enriched lignin in the infected seedlings. As the disease advanced at 8 wpi (DS at 34%), the S/G ratio of the infected seedlings was 0.64, also a two-times increase than the control seedlings (0.36). At 12 wpi (DS at 69%), the S/G ratio of the infected and control seedlings were 0.5 and 0.36, respectively.

Generally, the PAL enzyme showed the highest activity among the enzymes studied at 4, 8, and 12 wpi; concentration of PAL was four times higher than CAD and POD (Figure 4). In the phenylpropanoid pathway, PAL catalyzes the first step of phenylpropanoids biosynthesis, which are then further modified into an array of phenolic compounds. The activity of PAL confirms its role as the precursor for a broad range of phenolics such as flavonoids, tannin, and pectin.

At 4 wpi, the activities of PAL and CAD in the infected seedlings (DS at 5%) were significantly high in comparison to their respective control seedlings, while POD showed no difference between the treatments. The concentrations of PAL and CAD in the control seedlings were 0.75 and 0.17 μM, respectively, whereas their respective concentrations in the infected seedlings were 0.79 and 0.22 μM. At 8 wpi (DS at 34%), the activity of PAL was significantly higher in the control seedlings (0.76 μM) in comparison to the infected seedlings (0.68 μM). The activity of CAD showed no difference between the treatments while the activity of POD in the infected seedlings decreased significantly (0.14 μM) as compared to the control seedlings (0.16 μM). At 12 wpi, both PAL and CAD displayed a similar trend; the enzyme activities in the infected seedlings (DS at 69%) were significantly lower than the control seedlings. The PAL and CAD concentrations in the infected seedlings decreased to 0.60 and 0.14 μM, respectively, whereas the concentrations of the control seedlings were 0.86 and 0.23 μM, respectively. POD activity showed no significant difference between the infected and control seedlings (Figure 4).

Elicitation of PAL and CAD activities in the infected seedlings at 4 wpi (DS at 5%) were parallel to the accumulation of induced lignin, suggesting both enzymes may have exerted a collective role in the biosynthesis of induced lignin. CAD, responsible for monolignol production, could be closely linked to the concomitant biosynthesis of S-enriched lignin. The POD activity in the infected seedlings showed no apparent elicitation as compared to the control seedlings at 4, 8, and 12 wpi. Therefore, polymerization of monolignols by POD could not be related to the biosynthesis of induced lignin in this study.

The expression patterns of PAL, CAD, and POD genes in the root tissues of oil palm seedlings were profiled at 1, 2, and 3 wpi. Transcript levels in the infected seedlings were measured as fold-change (FC) relative to the control seedlings according to Livak and Schmittgen (2001). At 1 wpi, the transcript levels of PAL, CAD, and POD in the infected seedlings were up-regulated, at 1.8, 1.7, and 1.3 FC, respectively. Similarly, at 2 wpi, the transcript levels of PAL, CAD, and POD increased further to 2, 2.1, and 1.4 FC, respectively, and at 3 wpi, the transcript levels PAL, CAD, and POD genes were 2.5, 3.2, and 1.6 FC, respectively. In brief, throughout the sampling points (1–3 wpi), each gene showed gradual increases in transcript levels. The CAD gene had the highest transcript abundance, followed by PAL and POD genes at 3 wpi (Figure 5).

The cell walls of plants contain cellulose, hemicellulose, and lignin. The content and composition of these polymers vary among the different plant organs and within a specific tissue in the plants (Zaprometov et al., 1993; Campbell and Sederoff, 1996; Schuetz et al., 2013; Lourenço et al., 2016). In addition, lignin variation is affected by environmental conditions and plant growth stages (Ralph and Hatfield, 1991; Buranov and Mazza, 2008). In our study, the BSR development was characterized into three distinct stages. In the first stage, the asymptomatic plants showed low level of fungal attachment on the root surfaces, later at the second stage, the yellowing leaves indicated exhaustion of starch reservoirs resulting from biotrophic interaction of the pathogen at the expense of host and in the final stage (necrosis), the whole plant became affected, with leaves and stems turning brittle prior to plant death. The observation was comparable with previously reported findings (Flood et al., 2003; Rees et al., 2009). Oil palm early defense response at 4 wpi (DS at 5%) showed accumulation of lignin at two-fold higher than the control seedlings. In addition, the accumulated induced lignin showed a higher syringyl (S) to guaiacyl (G) ratio (S/G) in the infected seedlings compared to the control seedlings, demonstrating that S-lignin was selectively synthesized by the host plant during the G. boninense infection. Although such findings have been commonly observed in many plant–pathogen interactions (Menden et al., 2007; Moura et al., 2010; Barros-Rios et al., 2011; Bi et al., 2011; Eynck et al., 2012; Cheng et al., 2013; Ali and Mc Near, 2014; Bagniewska-Zadworna et al., 2014; Bivi et al., 2014), our study is the first to report concomitant alteration of lignin content and composition (S/G ratio) in G. boninense infected oil palm seedlings. Lignin accumulation at 4 wpi (DS at 5%) may have limited the host cell wall digestibility and restricted G. boninense infection. Despite a significant accumulation of lignin at 4 wpi (DS at 5%), the host plant was succumbed to attack at 8 wpi (DS at 34%) and 12 wpi (DS at 69%) and subsequently died.

At 4 wpi (DS at 5%) and 8 wpi (DS at 34%), the biosynthesis of S monomers was higher than G monomers, as indicated by the increased S/G ratio in the infected seedlings. The accumulation of either S or G monomers has been reported to be species-specific (Novo-Uzal et al., 2012). In carnation, the G-S lignin in control plants changed to mainly G-rich lignin during defense response against Fusarium oxysporum (Steijl et al., 1999). Contrarily, induced lignin with significant accumulation of S monomers has been reported during disease development in wheat (Tiburzy and Reisener, 1990; Mitchell et al., 1999).

Lignin polymers are chemically flexible, that the monomeric variation is affected by cell and environmental conditions (Faix et al., 1985; Boudet, 2000). A decrease in G monomers may imply a preferential degradation of G monomers compared to S monomers in oil palm root tissues colonized by G. boninense. Several studies have closely linked degradation of lignin to the types of monomer present in the wood tissues (Highley, 1982; Blanchette, 1991; Syafii and Yoshimoto, 1991; Skyba et al., 2013). Both the S and G monomers differ in the rate of methoxylation. The former is methoxylated at the 3′ and 5′ position of the aromatic ring while the latter carries only one methoxyl at the 3′ position, leaving the 5′ position free for participation in branching reactions (Sarkanen and Ludwig, 1971). A linear structured lignin is more recalcitrant to oxidizing agents compared to a branched structural pattern (Stewart et al., 2009). Biosynthesis of lignin is controlled by a complex phenylpropanoid biochemical grid, nevertheless, the regulatory roles posed by each member in the pathway cumulatively fix the composition and content of lignin synthesized at any given time (Rogers and Campbell, 2004; Yang et al., 2007). An increase in the S-rich lignin also implies the need to understand other specific enzymes such as ferulic acid 5-hydroxylase (F5H) and 5-hydroxyferulic acid o-methyltransferase (COMT), which catalyze the formation of sinapaldehyde, which in turn results in the biosynthesis of S monolignols. Further studies on enzymes specific to S monomer lignin need to be taken up to comprehend the inter-play between the members of the complex phenylpropanoid pathway.

In the present study, an increase in total lignin content with enriched S monomer at 4 wpi (DS at 5%) was preceded by up-regulation of PAL, CAD, and POD genes at 1, 2, and 3 wpi. In addition, the activities of enzymes in the phenylpropanoid pathway were in parallel to the gene expression patterns. Knock-out studies conducted in Pinus taeda (Mackay et al., 1997) and Medicago sativa (Baucher et al., 1999) have demonstrated the role of CAD in modulating lignin composition and content. A reduced expression of CAD in a mutant P. taeda significantly decreased the total lignin content while an increased expression correspondingly increased the total lignin content. Likewise, a reduction in the CAD activity observed in M. sativa yielded a lower syringyl to guaiacyl (S/G) ratio in comparison to the control (Baucher et al., 1999). Extensive studies performed on transgenic species with lost CAD activity had demonstrated intense changes in lignin composition rather than lignin content (Yahiahui et al., 1998; Eudes et al., 2006). The present results demonstrated the role of CAD in regulating defense associated-lignin, composed of predominantly, the S monomer. The results were comparable to other similar studies in forage grass (Giordano et al., 2014), tobacco (Chabannes et al., 2001), wheat (Mitchell et al., 1999), and Arabidopsis (Tronchet et al., 2010).

Phenylalanine is the precursor in the complex phenylpropanoid pathway (Mitchell and Walters, 1995; Kato et al., 2000; Zang et al., 2015). In the present study, PAL, together with CAD and POD genes, were collectively up-regulated at the early stage of Ganoderma infection (1, 2, and 3 wpi). Among the three genes, the transcript level of PAL was highest at 1 wpi, however, at 2 and 3 wpi, CAD consistently showed the highest expression. The transcript level of POD gene was the lowest at 1, 2, and 3 wpi. The specificity of lignin polymerization has been reported to exhibit a spatiotemporal expression pattern (Francoz et al., 2014). Peroxidase and laccase genes polymerise individual monolignols to form the complex lignin (Fagerstedt et al., 2010; Berthet et al., 2012). While laccase had been reported to display G-specific polymerization (Berthet et al., 2011), contrarily, the peroxidase showed a broader selectivity; polymerization of p-hydroxyphenyl (H), syringyl (S), and guaiacyl (G). The peroxidase comprises a group of multiple isoenzymes with adverse substrate specification for polymerization. In this study, the isoform of the peroxidase gene studied may not play a critical role in the biosynthesis of S-enriched lignin in oil palm seedlings during defense response against Ganoderma. The monolignol production evaluated by transcript abundance of CAD was ultimately higher than the degree of polymerization (transcript abundance of POD). The results were in agreement with Barrière et al., 2015, whom demonstrated that the multigene families responsible for monolignol polymerizations in maize could not be considered as a significant determinant factor affecting lignin variation. Further studies evaluating each member of the oil palm peroxidase multigene family need to be carried out to better comprehend the role of peroxidases in the production of S-lignin during G. boninense root colonization.

In the present research, we demonstrated rapid lignification with increased S/G ratio in oil palm seedlings (root tissues) as a possible defense response against Ganoderma infection. We also identified PAL and CAD as potential phenylpropanoid-biosynthetic enzymes and genes involved in regulating the biosynthesis of S-enriched lignin during the early stage of Ganoderma infection. In the quest for efficient BSR management strategy, further studies on host lignin compositional difference and its influence on Ganoderma infection need to be undertaken at a large scale. Phenomic studies which integrate agronomic and lignin data should be conducted to understand the physicochemical relationships of these parameters. Based on these results, PAL and CAD, in particular, may potentially be targets for up-regulation, ideally to confer S-enriched defense lignin.