• Open Access

RNAi suppression of lignin biosynthesis in sugarcane reduces recalcitrance for biofuel production from lignocellulosic biomass

Authors

  • Je Hyeong Jung,

    1. Agronomy Department, University of Florida, IFAS, Gainesville, FL, USA
    Search for more papers by this author
  • Walid M. Fouad,

    1. Agronomy Department, University of Florida, IFAS, Gainesville, FL, USA
    Search for more papers by this author
    • Present address: Department of Biology, School of Sciences and Engineering, The American University of Cairo, PO Box 74, AUC Avenue, SSE Building, Room 2184, New Cairo 11835, Egypt.

  • Wilfred Vermerris,

    1. Agronomy Department, University of Florida, IFAS, Gainesville, FL, USA
    2. Plant Molecular and Cellular Biology Program, University of Florida, IFAS, Gainesville, FL, USA
    3. University of Florida Genetics Institute, Gainesville, FL, USA
    Search for more papers by this author
  • Maria Gallo,

    1. Agronomy Department, University of Florida, IFAS, Gainesville, FL, USA
    2. Plant Molecular and Cellular Biology Program, University of Florida, IFAS, Gainesville, FL, USA
    3. University of Florida Genetics Institute, Gainesville, FL, USA
    Search for more papers by this author
  • Fredy Altpeter

    Corresponding author
    1. Agronomy Department, University of Florida, IFAS, Gainesville, FL, USA
    2. Plant Molecular and Cellular Biology Program, University of Florida, IFAS, Gainesville, FL, USA
    3. University of Florida Genetics Institute, Gainesville, FL, USA
      (fax +352-392-7248; email altpeter@ufl.edu)
    Search for more papers by this author

(fax +352-392-7248; email altpeter@ufl.edu)

Summary

Sugarcane is a prime bioethanol feedstock. Currently, sugarcane ethanol is produced through fermentation of the sucrose, which can easily be extracted from stem internodes. Processes for production of biofuels from the abundant lignocellulosic sugarcane residues will boost the ethanol output from sugarcane per land area. However, unlocking the vast amount of chemical energy stored in plant cell walls remains expensive primarily because of the intrinsic recalcitrance of lignocellulosic biomass. We report here the successful reduction in lignification in sugarcane by RNA interference, despite the complex and highly polyploid genome of this interspecific hybrid. Down-regulation of the sugarcane caffeic acid O-methyltransferase (COMT) gene by 67% to 97% reduced the lignin content by 3.9% to 13.7%, respectively. The syringyl/guaiacyl ratio in the lignin was reduced from 1.47 in the wild type to values ranging between 1.27 and 0.79. The yields of directly fermentable glucose from lignocellulosic biomass increased up to 29% without pretreatment. After dilute acid pretreatment, the fermentable glucose yield increased up to 34%. These observations demonstrate that a moderate reduction in lignin (3.9% to 8.4%) can reduce the recalcitrance of sugarcane biomass without compromising plant performance under controlled environmental conditions.

Introduction

The environmental, political and economic challenges associated with our dependence on fossil fuels have sparked a tremendous interest in developing efficient biomass feedstocks to satisfy the increasing demand for sustainable energy sources (Bouton, 2008; Byrt et al., 2011; Hisano et al., 2009; Tew and Cobill, 2008).

Sugarcane (Saccharum spp. hybrids) is a prime herbaceous biofuel feedstock. Sugarcane’s dry biomass production per unit area (39 Mg/ha, stalk, leaves and tops) is significantly higher than that of maize (17.6 Mg/ha, grain and stover), switchgrass (10.4 Mg/ha; biomass) or Miscanthus (29.6 Mg/ha; biomass) (Heaton et al., 2008; Waclawovsky et al., 2010). Its perennial growth habit and C4 photosynthetic pathway maximize carbon sequestration, while minimizing light requirements, water and nitrogen inputs (Byrt et al., 2011; Somerville et al., 2010). Sucrose accumulates in the stalk internodes of sugarcane and is either utilized for sugar production or readily fermented to the transportation fuel ethanol. The abundant lignocellulosic sugarcane residues are currently underutilized for bioenergy production (Somerville et al., 2010; Tew and Cobill, 2008). Including these residues for second-generation, biofuel production has the potential to boost biofuel yields per unit land area compared to current sucrose-based conversion technologies.

Diminished recalcitrance to enzymatic hydrolysis is a desirable trait for lignocellulosic feedstocks. The presence of lignin in the cell wall exacerbates biomass recalcitrance and limits bioconversion of lignocellulosic biomass into fermentable sugars (Jørgensen et al., 2007; Mansfield et al., 1999; Weng et al., 2008). Energy-intensive thermo-chemical pretreatments are required to degrade the cell wall matrix and to disrupt the crystalline structure of cellulose, thereby increasing binding sites for cellulolytic enzymes (Mosier et al., 2005). Pretreatments are followed by saccharification, during which the cell wall polymers, primarily cellulose and hemicellulose, are enzymatically hydrolysed into monomeric sugars for fermentation (Lu and Mosier, 2008).

Lignin is an aromatic, hydrophobic polymer primarily consisting of p-hydroxyphenyl (H), guaiacyl (G) and syringyl (S) units, which are polymerized by radical coupling of three different monolignols: p-coumaryl, coniferyl and sinapyl alcohol, respectively (Boerjan et al., 2003). During saccharification, cellulases can adsorb irreversibly to lignin, thereby reducing the overall enzymatic activity (Chernoglazov et al., 1988; Yang and Wyman, 2004). The manipulation of lignin biosynthesis in feedstocks is a prime strategy to reduce biomass recalcitrance and improve fermentable sugar yields.

Ten different enzymes catalyse a series of hydroxylation, methylation and side chain reduction reactions of monolignol precursors (Bonawitz and Chapple, 2010). Down-regulation of monolignol biosynthetic genes in transgenic alfalfa reduced total lignin content and altered lignin subunit composition. Fermentable sugar yields, with or without pretreatment, were shown to be negatively correlated with total lignin content (Chen and Dixon, 2007). Down-regulation of enzymes involved in the terminal steps of the monolignol biosynthetic pathway including COMT or CAD generally had little or no negative effects on plant growth (Chen and Dixon, 2007; Fu et al., 2011; Jackson et al., 2008; Pilate et al., 2002; Piquemal et al., 2002). In grasses, biomass conversion is also influenced by lignin subunit composition based on studies in maize and sorghum brown midrib mutants (Dien et al., 2009; Saballos et al., 2008; Vermerris et al., 2007). Recently, transgenic approaches for manipulating lignin biosynthesis have been successfully applied to switchgrass (Fu et al., 2011; Saathoff et al., 2011; Xu et al., 2011).

Caffeic acid O-methyltransferase (COMT) functions late in the monolignol biosynthetic pathway and, despite its name, methylates 5-hydroxyconiferyl aldehyde and 5-hydroxyconiferyl alcohol to form S unit precursors, sinapyl aldehyde and sinapyl alcohol, respectively (Bout and Vermerris, 2003; Guo et al., 2001; Humphreys et al., 1999; Osakabe et al., 1999). In sugarcane, ∼31 different consensus EST sequences are clustered to COMT, which could be potential allelic variants or homo(eo)logous genes (Ramos et al., 2001). The previously identified full-length COMT (GenBank accession no. AJ231133) shows 91% amino acid similarity with maize lignin-specific COMT and is preferentially expressed in stems and roots and in lignifying tissues such as epidermis, xylem and sclerenchyma (Ruelland et al., 2003; Selman-Housein et al., 1999).

Conventional breeding and genetic engineering of sugarcane are challenging because of its highly polyploid genome derived from interspecific hybridization (Dal-Bianco et al., 2011; Lakshmanan et al., 2005). Although most plants examined to date share many of the enzymatic reactions leading to the synthesis of lignin (Xu et al., 2009), limited information is available for sugarcane, and none of the lignin biosynthetic genes have been functionally characterized through forward or reverse genetics (Ramos et al., 2001; Ruelland et al., 2003; Selman-Housein et al., 1999). Biolistic or agrobacterium-mediated gene transfer into sugarcane has typically been achieved by using embryogenic callus as target tissue (Arencibia et al., 1998; Bower and Birch, 1992; Gallo-Meagher and Irvine, 1996). Recent improvements for biolistic gene transfer focused on the reduction in the tissue culture period using direct embryogenesis and by delivering minimal expression cassettes instead of plasmids (Kim et al., 2012; Taparia et al., 2012a,b). A number of transgenes have been introduced into sugarcane to incorporate traits including herbicide resistance, tolerance to abiotic or biotic stress and production of sugar and value-added metabolites (Altpeter and Oraby, 2010). However, lignin modification or the application of RNAi for crop improvement has not been reported in sugarcane.

Here, we describe the generation of transgenic sugarcane with RNAi suppression of COMT and demonstrate that the resulting transgenics are more amenable to biomass conversion for the production of fuels and chemicals.

Results

Generation and molecular characterization of transgenic sugarcane lines

The RNAi-inducing transgene cassette contained a highly conserved region of COMT. This may suppress COMT expression from both targeted COMT and homo(eo)logous COMT’s in the highly polyploidy sugarcane. The targeted region shared 95%–100% nucleotide sequence identity with five tentative consensus (TC) sequences, TC116269, TC120283, TC121675, TC133773 and TC149781, putative homo(eo)logs of sugarcane COMT (DFCI S.officinarum Gene Index; http://compbio.dfci.harvard.edu/cgi-bin/tgi/geneprod_search.pl). The promoter of Oryza sativa C4H gene was used to induce the expression of RNAi-inducing transgene in lignifying tissues.

Biolistic gene transfer was used to produce sugarcane transformants. Following 24 bombardments, regeneration of plants and selection of putative transformants on geneticin and paromomycin, 43 lines tested positive for the NPTII enzyme as determined by ELISA. This resulted in a transformation efficiency of 1.8 transgenic plants produced per bombardment. Thirty-eight of these 43 transgenic lines tested positive for the presence of the COMT RNAi cassette as determined by PCR analysis, resulting in a co-transformation efficiency of 88% for the unlinked nptII and COMT minimal cassettes.

To determine copy number of the COMT RNAi cassette in the transformants, genomic DNA was analysed by Southern blot analysis (Figure 1b). A unique transgene integration pattern was detected for each transgenic line, confirming these were independent transformation events. Among the 12 transgenic lines analysed, two lines (T4 and T18) displayed a relatively simple transgene integration pattern with two copies. Four lines had <8 hybridization signals, and six lines showed a complex integration pattern with more than eight copies.

Figure 1.

COMT RNAi cassette and Southern blot analysis. (a) COMT RNAi cassette consists of inverted repeats of the 346 bp COMT gene fragment separated by 94 bp Paspalum notatum 4CL intron under control of 1994 bp Oryza sativa C4H promoter and CaMV 35S 3′UTR. PF and PR indicate the primer binding site for PCR analysis to confirm transgenic events. The probe indicates the complementary binding site for Southern blot analysis. (b) Southern blot analysis of transgenic sugarcane plants. PC: 9.7 kb linearized COMT RNAi vector; WT, Wild-type sugarcane; T1-T47, Transgenic sugarcane. The displayed image was generated from a single Southern blot. After exposure to x-ray film for different periods of time, lanes with best resolution were merged to minimize overexposure of individual signals.

Small-RNA Northern blot analysis was performed for 17 transgenic plants to examine whether siRNA was produced from the dsRNA precursor generated by the inverted repeats of the COMT fragment. Seven of the 17 transgenic plants (41%) produced COMT siRNA, while siRNA was not detected in the wild-type plant (Figure 2a). The size marker indicated that siRNA from the transgenic plants was ∼21 nt long, whereas another siRNA class of ∼24 nt was not detected in the transgenic plants (Figure 2a).

Figure 2.

 Detection of COMT siRNA and evaluation of COMT expression levels in transgenic sugarcane. (a) Small-RNA Northern blot. The hybridization signal indicates the generation of ∼21-nt-long COMT siRNA by cleavage of the hairpin structure dsRNA. Ethidium bromide-stained 5S rRNA and tRNA were used as a loading control to verify the amount and integrity of the small RNA. WT1 and WT2, Wild-type sugarcane plants; NC, buffer; SM, Size marker corresponding to 21-nt-long RNA; T-, Transgenic sugarcane plants; PC, T17 transgenic plant used as a positive control on a separate membrane. (b) Quantitative RT-PCR analysis of COMT expression level in the transgenic sugarcane. Error bars represent the 95% confidence intervals of the 2−ΔΔCt value of the wild type vs. transgenic plants. Significant differences in COMT expression between the wild type and the transgenic sugarcane are indicated by an asterisk (= 3, < 0.05 in t-test).

Clones of primary transformants were vegetatively propagated and grown in the greenhouse, and quantitative real-time RT-PCR of the transcripts was performed to investigate siRNA-induced suppression of COMT. The expression level was significantly reduced by 67, 97 and 97% in lines T41, T23 and T4, respectively, compared to the wild-type controls (Figure 2b).

Lignin content, composition and enzymatic saccharification

Total lignin content and subunit composition were determined to examine the effect of COMT suppression on lignin biosynthesis in the vegetative progenies of the transgenic plants. Compared to wild type, total lignin was reduced by 3.9, 8.4 and 13.7% in lines T41, T23 and T4, respectively (Table 1). Lignin of transgenic plants was composed of significantly less S unit and similar amount of G unit compared to wild type (Table 1). Transgenic lines had lower S/G ratios ranging from 1.27 to 0.79, while that of wild type was 1.47. There was no difference in total lignin content or lignin monomer composition of the transgenic control plants transformed with nptII compared to wild-type plants (Table 1).

Table 1.  Lignin content, composition and glucose yields after enzymatic saccharification
LinesTotal lignin (mg g−1)G unit (μmol g−1 lignin)S unit (μmol g−1 lignin)S/G molar ratioGlucose yield (mg glucose g−1 biomass)
PretreatmentNo pretreatment
  1. WT, Wild-type sugarcane; TC2, Transgenic control harbouring nptII gene alone; T41, T23 and T4, Transgenic sugarcane.

  2. Values are means ± SE (= 3 for total lignin, = 2 for lignin composition and = 3 for glucose yields).

  3. *Significantly different from the wild-type plants at < 0.05 in t-test.

WT181.4 ± 2.2158.4 ± 2.7233.4 ± 2.01.47190.9 ± 4.996.1 ± 3.0
TC2182.0 ± 2.0154.0 ± 0.9230.0 ± 7.91.49196.9 ± 6.695.6 ± 3.7
T41174.3 ± 4.6*150.4 ± 1.6191.6 ± 7.6*1.27*238.7 ± 3.5*95.2 ± 3.9
T23166.1 ± 1.1*163.6 ± 4.7179.8 ± 5.0*1.10*241.3 ± 1.8*94.4 ± 2.3
T4156.6 ± 1.9*165.8 ± 6.4131.5 ± 2.3*0.79*288.0 ± 1.2*135.6 ± 2.8*

To obtain an estimate of recalcitrance to enzymatic saccharification, enzymatic hydrolysis of lignocellulosic biomass from all samples was performed, with and without dilute acid pretreatments. Transgenic plants with reduced total lignin content and reduced S/G ratios showed improvements in enzymatic digestibility of the biomass compared to the wild-type controls (Table 1). Without pretreatment, the biomass generated from transgenic line T4 yielded 29% more glucose than the wild-type plants, while T41 and T23 showed similar glucose yields as the wild-type plant. With dilute acid pretreatment, significant increases were observed for all of the lines T41, T23 and T4, which yielded 20%, 21% and 34% more glucose, respectively. Pretreatment enhanced the saccharification efficiencies of the transgenics 2.1- to 2.5-fold, compared to 1.9-fold for wild type. No significant differences were detected in glucose yields between the nptII-only and wild-type controls.

Plant phenotype and growth

Clonally propagated progeny of transgenic sugarcane displayed phenotypes similar to wild type under greenhouse conditions, without lodging or excessive tillering (Figure 3a). Microscopic evaluation suggested that vascular bundle tissues and sclerenchyma fibre cells were intact in transgenic sugarcane with reduced total lignin content (Figure 4). In contrast to wild-type or nptII-only control plants, transverse stem sections in the internode region of the transgenic lines revealed a deep brown colour (Figure 3c). The intensity of brown colour appeared to be correlated with the level of lignin reduction. Transgenic line T4, with the highest reduction in total lignin content, displayed the darkest brown colour of all transgenic lines extending all the way from the basal node to the 4th node below the apical meristem. T41 and T23 only showed brown colour in the basal internodes. The midrib of leaves from transgenic plants did not display brown coloration (data not shown).

Figure 3.

 Phenotypic evaluation of transgenic sugarcane in comparison with wild type. WT, Wild-type sugarcane; TC1, nontransgenic plant co-regenerated with transgenic plants from tissue culture; TC2, Transgenic control plant harbouring nptII alone; T41, T23 and T4, Transgenic sugarcane plants. (a) Greenhouse-grown wild type (WT) and transgenic sugarcane plants (T4-T41). (b) Flowering of the transgenic sugarcane plant (T4) under greenhouse conditions with natural photoperiod. (c) Coloration of internodes of transgenic and nontransgenic plants. The basal internode of the transgenic, control and wild-type plant were transversely or longitudinally sectioned, and the picture was taken immediately without fixation and staining.

Figure 4.

 Microscopic and histochemical evaluation of transgenic sugarcane in comparison with wild type. Microscopic observations of transverse sections of the internode, without staining (a and b), stained with Mäule reagent (c and d) or stained with Wiesner reagent (e and f). Note the differences in colour of vascular tissues between wild type (a) and T4 transgenic line (b) in the absence of staining and arrows indicating yellow coloration in the sclerenchyma fibre cells in T4 transgenic line (d) compared with that in the wild type (c). Ph, Phloem; X, Xylem; S, Sclerenchyma; P, Parenchyma. Scale bars indicate 250 μm.

Microscopic evaluation revealed a reddish brown colour of vascular tissues and surrounding sclerenchyma cells of stem tissues from transgenic sugarcane (Figure 4a,b). Following histochemical analysis with Mäule reagent vascular bundles displayed a red colour in both wild type and transgenic sugarcane. However, the sclerenchyma fibre cells surrounding the vascular bundles displayed a yellow colour in transgenic sugarcane, indicating a reduction in S units of lignin (Figure 4c,d). Staining with Wiesner reagent produced similar results in both wild type and transgenic sugarcane, indicating no changes in hydroxycinnamaldehyde end groups of lignin (Figure 4e,f).

The effect of COMT suppression on plant growth was investigated using clonally propagated progenies grown in randomized, complete blocks and analysed by ANOVA. Biomass production and stalk diameters of transgenic lines T41 and T23, with moderate reductions in lignin (3.9% and 8.4%, respectively), were similar to those of the nontransgenic tissue culture control plants (TC1) and nptII-only transgenic control plants (TC2) (Figure 5a,b). Transgenic line T4, with a 13.7% reduction in lignin, produced less biomass and had thinner stalks compared with the controls. The tissue culture control (TC1) and the nptII-only transgenic control (TC2) plants produced 12% and 10% less biomass than the wild-type plants, respectively. Transgenic lines T41, T23 and T4 accumulated 17%, 13% and 35% less biomass than the wild-type plants, respectively. As the controls and transgenic sugarcane exhibited similar or slightly taller stalk lengths compared with the wild-type sugarcane, the reduced biomass production was mostly due to decreased stalk diameters (Figure 5b). Transgenic plant T4 with the most severe lignin reduction flowered in the greenhouse under natural photoperiod and did not display lodging (Figure 3b).

Figure 5.

 Plant growth characteristics under greenhouse conditions. (a) Biomass production. Values represent means of 8 clonally propagated plants, grown in an air-conditioned greenhouse following randomization. Error bars represent standard error. Means with different letter are significantly different at P < 0.05 according to Tukey’s multiple comparisons. Fresh weight was measured from the most mature stalk of 7-month-old plants. (b) Stalk diameter (grey bar) and length (black bar) from the set of plants described under (a). Stalk diameter was measured in the middle of the stalk, and length was measured from the soil surface to the apical meristem of the stalk. Means of stalk diameter (a, b, c) or means of stalk length (x, y) with different letter are significantly different at P < 0.05 according to Tukey’s multiple comparisons.

Discussion

Genetic improvement of sugarcane through breeding or biotechnology is challenging because of the highly polyploid genome of this interspecific hybrid. To our knowledge, this is the first report of sugarcane improvement via an RNAi approach. Transgenic sugarcane plants with RNAi suppression of COMT had significantly lower lignin contents, ranging from 3.9% to 13.7% reduction relative to wild type. Yields of fermentable glucose were significantly increased up to 29% even without pretreatment and up to 34% with dilute acid pretreatment. These observations are consistent with what was observed for transgenic switchgrass in which COMT had been down-regulated, leading to significantly reduced total lignin content by 11.4%–13.4% and improved saccharification efficiencies by 16.5%–21.5% following mild pretreatment (Fu et al., 2011), as well as maize (Vermerris et al., 2007) and sorghum (Dien et al., 2009; Saballos et al., 2008) brown midrib mutants with reduced COMT activity. The reduction in lignin content combined with the lower S/G ratio in the lignin of these plants results in more efficient enzymatic saccharification, possibly due to altered physico-chemical properties of the lignin, so that a smaller proportion of the cellulases irreversibly adsorb on the lignin. The brown coloration of the stems is also consistent with the phenotype observed in other plants with reduced COMT activity (Fu et al., 2011; Piquemal et al., 2002). The lack of brown midribs in the transgenic sugarcane plants is consistent with the findings in transgenic switchgrass with RNAi suppression of COMT (Fu et al., 2011).

Lignin plays an important role in plant growth and development and serves to protect plants from abiotic and biotic stress (Boerjan et al., 2003; Dixon and Paiva, 1995). Therefore, reducing lignin content could compromise plant performance, stress tolerance or defence mechanisms. Our results indicate that in vitro generated transgenic sugarcane lines, with a moderate reduction in total lignin ranging from 3.9% to 8.4%, showed comparable biomass production to the tissue culture or transgenic control plants. Similarly, partial suppression of COMT through RNAi or antisense strategies in transgenic switchgrass and maize resulted in normal phenotypes under controlled environment conditions (Fu et al., 2011; Piquemal et al., 2002). Suppression of COMT may adversely affect plant growth depending on the impact on lignin content and the genetic background. Transgenic sugarcane line T4, with 97% reduction in COMT transcripts and 13.7% reduction in total lignin content, displayed significantly reduced stalk diameter and biomass production compared with the control and wild-type plants. Reduced COMT activity in brown midrib mutants of maize (bm3) and sorghum (bmr12) resulted in ∼10% reduction in stover and dry matter yields, respectively (Miller et al., 1983; Oliver et al., 2005). A comparison of bmr12 near-isogenic lines (NILs) in different inbred backgrounds showed decreases in dry matter yields ranging from 6% to 22%. However, the bmr12 NIL in the genetic background of Early Hegari-Sart showed the greatest reduction in acid detergent lignin, but was able to produce the same amount of dry matter compared to its wild type-counterpart (Oliver et al., 2005). This suggested that there is an interaction between the genetic background and tolerance to reduced lignin content for biomass production. This was corroborated by analyses of soluble and cell wall-bound aromatics in sorghum bmr6, bmr12 and bmr6-bmr12 mutants in different genetic backgrounds (Palmer et al., 2008). Transgenic sugarcane line T4, with the most severe lignin reduction, was still able to mature and produce flowers. Line T4 displays a simple transgene integration pattern with two copies of the COMT RNAi cassette, and CP 88-1762 is a fertile cultivar. Transferring the reduced lignin trait into energycane, a high-fibre, high-biomass variant of sugarcane (Tew and Cobill, 2008) grown primarily for biomass production is therefore expected to be feasible.

While not affecting lignin content or composition, reduced stalk diameters and biomass production were observed in tissue culture and transgenic control plants when compared with wild-type CP 88-1762. Somaclonal variation resulting in less-than-optimal field performance has been reported for sugarcane (Gilbert et al., 2005; Taparia et al., 2012a,b). A single backcross has the potential to eliminate tissue culture-derived mutations (Bregitzer et al., 2008).

Sugarcane has a high level of genetic redundancy with an average of 12 homo(eo)logous haplotypes (Le Cunff et al., 2008), and most of these homo(eo)logs are considered to be functional (Garsmeur et al., 2011). Because of this functional redundancy, identification of mutant sugarcane plants with substantial changes in lignin content and/or lignin subunit composition is highly unlikely. RNAi-mediated gene silencing allows simultaneous suppression of homo(eo)logs in a high ploidy genome within members of a gene family (Lawrence and Pikaard, 2003; Miki et al., 2005). In this study, the hairpin structure of COMT transgene successfully triggered generation of siRNA and induced suppression of the targeted endogenous COMT gene expression. The reduction in total lignin content and altered S/G ratios suggest that the conserved sequence that was used to suppress COMT may have supported co-suppression of related COMT homo(eo)logous genes.

RNAi-mediated gene suppression is a useful tool for elucidation of gene function (Miki et al., 2005; Travella et al., 2006). This study confirms that RNAi is an effective method for the suppression of target genes in sugarcane. This finding is consistent with an earlier report on RNAi suppression of the phytoene desaturase in sugarcane (Osabe et al., 2009). The sugarcane COMT gene was previously annotated and characterized based on sequence identity, transcript expression pattern and tissue/cellular localization (Ruelland et al., 2003; Selman-Housein et al., 1999). However, COMT belongs to a large S-adenosyl-L-methionine (SAM)-dependent O-methyltransferase (OMT) family including lignin-specific OMT and other phenylpropanoid-specific OMT genes. Therefore, additional evidence for confirming the correct annotation was needed (Noel et al., 2003; Zhou et al., 2010). The transgenic evidence provided by this study verifies the function of COMT participating in lignin biosynthesis, particularly in S unit formation. Transgenic sugarcane plants displayed a significant reduction in S units without changing quantities of G units, thereby causing a lower S/G ratio compared to wild type. Similarly, wild-type RNAi suppression of COMT in switchgrass also resulted in reduced S/G ratio (Fu et al., 2011).

Reducing the recalcitrance of lignocellulosic sugarcane biomass to enzymatic hydrolysis is expected to enhance the value of this prime biofuel feedstock. Follow-up field evaluations will elucidate the influence of the genetic background of the transgenic sugarcane lines on biofuel yield per land area.

Experimental procedures

Plant growth

Commercially important sugarcane cv. CP88-1762 stalks were collected from the Everglades Research and Education Center, University of Florida, Belle Glade, Florida. Single-node segments from these wild-type plants were transplanted to 15-l pots containing Fafard No. 2 mix (Conrad Fafard, Agawam, MA) and grown to maturity under natural photoperiod in an air-conditioned greenhouse set at 28 °C/22 °C (day/night). Plants were irrigated once a day and fertilized biweekly with Miracle-Gro Plant and Lawn Food (Scotts Miracle-Gro, Marysville, OH).

The primary transgenic lines and wild-type plants were clonally propagated by single-node segments of mature plants and in 15-l pots containing Fafard No. 2 mix. Plants were arranged in a randomized block design, with eight replications, and grown under the conditions as described above. A total of four stalks were maintained in each pot by removing juvenile tillers weekly. Above-ground fresh weight was measured from the most mature tiller of the 7-month-old plants. After removing leaves and leaf sheath, stalk diameter was measured in the middle of the stalk, and length was measured from soil surface to the apical meristem of the stalk. The basal internode of the transgenic, control and wild-type plant was transversely or longitudinally cut and photographed immediately without fixation and staining to display the brown coloration.

Vector construction

The COMT RNAi vector was constructed using the pWFOsC4H::Bg4CLi vector (Fouad et al., 2010). To amplify a 346 bp COMT fragment from the sugarcane cDNA, a pair of primers (forward: 5′-AGAGCTGGTACTACCTCAAGGACG-3′, reverse: 5′-GTTTAAACATGTCCCCGCCGACGTG-3′) was designed to the sugarcane COMT sequence (GenBank accession no. AJ231133; Selman-Housein et al., 1999). This fragment was located on the COMT coding region between nucleotide 410 and 755, spanning part of the highly conserved SAM-binding pocket among plant COMTs (Louie et al., 2010). PCR products were cloned into the pCR2.1-TOPO vector (Invitrogen, Carlsbad, CA) and confirmed by sequencing. XbaI and EcoRV restriction enzyme sites were introduced to the 5′ end of the COMT fragment and cloned into the pCR2.1-TOPO vector. The inverted repeats of Bg4CL fragment in the pWFOsC4H::Bg4CLi (Fouad et al., 2010) were replaced by the COMT fragment with two subsequent cloning steps. The resulting COMT RNAi cassette consisted of the 1994 bp Oryza sativa cinnamate 4-hydroxylase (C4H) gene promoter (GenBank accession no. AC136224), inverted repeats of the 346 bp sugarcane COMT fragment separated by the 94 bp of Paspalum notatum 4-coumarate : CoA ligase (4CL) intron (Fouad et al., 2010), and the CaMV 35S 3′ UTR (Figure 1a). The pJFNPTII vector (Altpeter et al., 2000) provided the selectable marker and contained the neomycin phosphotransferase II (nptII) gene under the transcriptional control of the Zea mays ubiquitin promoter with first intron and CaMV 35S 3′ UTR.

Generation of transgenic sugarcane

Embryogenic sugarcane calli were induced as described by Chengalrayan and Gallo-Meagher (2001). Tissues were sub-cultured on CI-3 media every 2 weeks. Ten weeks after callus induction, particle bombardment was performed using the PDS-1000/He biolistic particle delivery system (Bio-Rad, Hercules, CA) as previously described (Altpeter et al., 1996). For delivery of the minimum linear transgene cassette (MC), the COMT RNAi cassette and nptII expression cassettes were released by restriction enzyme digestion with XmnI and I-SceI, separated by gel electrophoresis and extracted using the QIAquick Gel Extraction kit (Qiagen, Valencia, CA). COMT and nptIIMC DNAs were mixed in a 2 : 1 m ratio, co-precipitated onto 1-μm-diameter gold micro-carriers (Bio-Rad) and accelerated into target tissues as described earlier (Altpeter et al., 2010). Calli were transferred to CI-3 media containing geneticin (G-418, 30 mg l−1) 6 days after bombardment, for 3 biweekly subcultures (Kim et al., 2012). Calli that survived were regenerated into plants on media containing 50 mg l−1thidiazuron (TDZ) for 2 weeks. For further selection, elongation and rooting of shoots, regenerated plantlets were transferred to hormone-free media containing 30 mg l−1 paromomycin for four biweekly subcultures. Regenerated plants with roots were transplanted in 0.65-l pots with Fafard No. 2 mix and placed in a growth room with 80% relative humidity with 16 h photoperiod and 500 μmol m−2 s−1 light intensity.

Evaluation of NPTII expression

NPTII expression was evaluated by NPTII ELISA using a commercial kit (Agdia, Elkhart, IN). Total protein was extracted from leaves and quantified with the Bradford assay (Bradford, 1976). The ELISA assay was performed using 20 μg of total protein, according to the manufacturer’s instruction. NPTII expression of putative transgenic plants was qualitatively evaluated by colour development in comparison with the supplied NPTII standard and wild-type protein extracts.

PCR analysis

The presence of the COMT RNAi cassette in genomic DNA extracts of transgenic lines was confirmed by PCR. Genomic DNA was extracted from leaves using the CTAB method (Murray and Thompson, 1980), and 75 ng was used per reaction as a template for amplification. Primers PF:5′-CCTGCTAGTCTTCTCTCTCATTGTT-3′ designed to the C4H promoter region and PR:5′-GTGATGATGACCGAGTGGTTCTT-3′ annealing to the sense fragment of COMT (Figure 1a) were used, with an expected amplification product of 550 bp. PCR was performed in the MyiQ cycler (Bio-Rad) with iTaq DNA polymerase (Bio-Rad) under the following conditions: 95 °C for 3 min denaturation, 35 cycles at 95 °C for 30 s, 60 °C for 30 s, 72 °C for 1 min and final extension at 72 °C for 7 min.

Southern blot analysis

High molecular weight genomic DNA was extracted from leaves using the CTAB method (Murray and Thompson, 1980). Twenty microgram of genomic DNA were digested to completion with EcoRI, separated by eletrophoresis on 1.0% agarose gel and transferred onto the Hybond-N+ membrane using the manufacturer’s instructions (GE Healthcare Biosciences, Pittsburgh, PA). Probes were generated by PCR to the C4H promoter region of the COMT RNAi cassette and labelled with 32P-dCTP (Perkin Elmer, Waltham, MA) using the Prime-a-Gene Labeling System (Promega, Mannheim, Germany). Hybridization and washing were performed according to the manufacturer’s instructions, and membranes were exposed to Kodak X-ray film (Fisher Scientific, Atlanta, GA) at −80 °C for 2 days.

Small-RNA Northern blot

The third internode below the apical shoot meristem was collected from primary transgenic or wild-type plants, and total RNA was extracted from 3 g of internode tissue using the modified hot SDS/phenol method (Shirzadegan et al., 1991). Small RNA was separated from the total RNA using the method described in Lu et al. (2007). Thirty microgram of small RNA was separated by electrophoresis in 15% polyacrylamide TBE/urea gel (Bio-Rad) and transferred to the Hybond-N+ membrane (GE Healthcare Biosciences) using a semi-dry transfer cell (Bio-Rad). The 346 bp COMT fragment, used for COMT RNAi vector construction, was labelled with 32P-dCTP (Perkin Elmer) using the Prime-a-Gene Labeling System (Promega) for use as a probe. Hybridization was carried out overnight at 38 °C in Church’s hybridization buffer (Brown et al., 2004). Following hybridization, the membrane was briefly rinsed 2X SSC—0.2% SDS and then washed twice with 2X SSC—0.2% SDS at 50 °C for 20 min each. The membrane was exposed to Kodak X-ray film at –80 °C for 2 days.

Quantitative real-time RT-PCR for quantification of COMT expression

The third internode below the apical shoot meristem was collected from the clonally produced progenies of transgenic and wild-type plants. Total RNA was extracted using Trizol reagent (Invitrogen) according to the manufacturer’s instruction. To prevent genomic DNA contamination, total RNA was treated with RNase-Free RQ1 DNase (Promega). cDNA was synthesized from 500 ng of DNase-treated RNA using iScript cDNA Synthesis kit (Bio-Rad). COMT-specific primers (forward: 5′-TAAATACGCACACCTGCTGCT-3′ and reverse: 5′-ATTCGACAATTTAGAATCCAGAACAT-3′) were designed for the amplification of a specific fragment of the 3′UTR region of the targeted COMT gene. Sugarcane GAPDH primers (forward: 5′- CACGGCCACTGGAAGCA-3′ and reverse: 5′-TCCTCAGGGTTCCTGATGCC-3′) were used to amplify a fragment of the sugarcane GAPDH gene as a reference gene for normalization of transcripts as described by Iskandar et al. (2004). Quantitative real-time PCR of the transcripts was performed in the MyiQ cycler (Bio-Rad) with iQ SYBR Green Supermix (Bio-Rad) under the following conditions: 95 °C for 3 min denaturation, 40 cycles at 95 °C for 10 s and 55 °C for 45 s. Amplicon specificity was verified by melt curve analysis from 55 to 95 °C and by agarose gel electrophoresis. COMT expression levels in transgenic plants relative to wild-type plants were calculated using the 2−ΔΔCt method (Livak and Schmittgen, 2001).

Microscopic and histochemical analysis

The transverse stem sections were hand-cut with a razor blade from the sixth internode below the shoot apical meristem in both wild type and transgenic line T4. Mäule and Wiesner staining were performed according to Vermerris and Nicholson (2006). Sections were observed using an Olympus BH-2 light microscope (Olympus, Tokyo, Japan). Images were taken and recorded in the Infinity 1 camera and Infinity analyzer (Lumenera Corporation, Ottawa, ON, Canada).

Sample preparation and determination of lignin content and composition

Stalks with 12 internodes were collected from clonally propagated transgenic or wild-type plants, and internodes 1–3 below the shoot apical meristem, all leaves and leaf sheaths were removed. The remaining mature portion of the stalks was dried at 45 °C and ground using a Wiley mill (Thomas Scientific, Swedesboro, NJ) with a 1.0-mm sieve. The ground samples were passed through 0.42-mm sieve to remove irregular particles. Following three successive extractions with 50% ethanol (v/v), under sonication at 45 °C for 30 min, samples were dried at 45 °C until constant weight.

The modified acetyl bromide method was used to determine lignin content (Foster et al., 2010; Hatfield et al., 1999). Two milligram of extract-free dried sample was placed in a 2-mL polypropylene tube, and 1 mL of freshly prepared 25% (w/w) acetyl bromide/glacial acetic acid solution was added. The tubes were incubated in a water bath at 50 °C for 4 h, and during the last hour, samples were thoroughly mixed at 15-min intervals and placed on ice for 30 min. One hundred microlitre of each reaction mixture was transferred into a 2-mL polypropylene tube containing 200 μl of 2 m NaOH and 1.7 mL glacial acetic acid. The absorbance of the solution was determined at 280 nm using an Evolution 300 spectrophotometer (Thermo Fisher Scientific Inc., Waltham, MA). The lignin content was calculated by employing the molar extinction coefficient of 21.5 L g−1 cm−1 for milled sugarcane vascular bundle lignin (He and Terashima, 1990).

Lignin monomer composition was determined by thioacidolysis as described by Robinson and Mansfield (2009) using 10 mg of extractive-free sample. Thioacidolysis products were identified and quantified using GC/MS on a Varian 3800 (Varian, Walnut Creek, CA) GC coupled to a Varian 1200 MS. The sample was injected onto a Factor-4 VF-5ht column (35 m, 0.25 mm i.d.) with helium (1.2 mL/min) as carrier gas. The injector temperature was 250 °C. The oven temperature was held for 3 min at 130 °C and increased to 250 °C at 3 °C/min, then held constant for 5 min. The mass spectrometer was operated in electron impact (EI) mode at 70 eV. The detector was operated at 1.2 kV. The mass range included m/z 50-550 and was scanned every 0.2 s. Data were analysed using MS workstation software. Lignin monomers were quantified employing response factors of each monomer against internal standard as described in Yue et al. (2012)

Dilute acid pretreatment and enzymatic hydrolysis

Dilute acid pretreatment and enzymatic hydrolysis was performed according to published protocols (Chen and Dixon, 2007; Selig et al., 2008). Extractive-free samples (0.15 g) were soaked in 1.35 mL of dilute sulphuric acid [final concentration 1.3% (w/w)] and autoclaved at 121 °C for 40 min. After autoclaving, the pretreated samples were washed twice with 25 mL distilled water. The pretreated or native biomass samples were placed in 50-mL polypropylene tubes and suspended in 5 mL of 0.1 m sodium citrate buffer (pH 4.8). Autoclaved distilled water containing 100 μL of 2% (w/w) sodium azide, 6 FPU of Kerry Biocellulase W (Kerry Bioscience, Cork, Ireland) and 6.4 pNPGU of Novozyme 188 β-glucosidase (Sigma, Saint Louis, MO) was added to bring the total volume to 10 mL. The hydrolysis was performed for 120 h in a shaking incubator at 50 °C and 250 rpm. After enzymatic hydrolysis, 1 mL of the each sample was collected and filtered through a 0.2-μm syringe filter, and glucose yields were analysed using a YSI glucose analyzer (YSI Life Science, Yellow Springs, OH).

Statistical analysis

ANOVA was performed using Proc GLM in SASTM Version 9.3 (SAS Institute Inc., Cary, NC). Statistical significance among means for biomass, stalk diameter and length was determined using Tukey’s multiple comparisons at < 0.05. T-tests were performed to determine whether the means of total lignin content, S/G ratios and glucose yields were statistically significant between the transgenic plants and the wild-type, tissue culture or nptII-only transgenic controls (< 0.05).

Acknowledgements

We would like to thank USDA-NIFA (grant # 2009-10001-05117) and Syngenta Biotechnology Inc. for their financial support of this work, Dr. Robert Gilbert EREC, Belle Glade FL for providing donor plants of sugarcane cultivar CP88-1762, Dr. James Preston, Department of Microbiology and Cell Science University of Florida for providing Kerry Biocellulase W and Novozyme 188 β-glucosidase, Jeff Seib for training Je Heyong Jung in safe handling of radio isotopes, Janice Zale for her help with editing this manuscript and the Conrad Fafard Inc. Apopka FL for donation of plant growing media. The University of Florida is gratefully acknowledged for providing funding for the GC–MS equipment.

Ancillary