Due to the un-sustainable nature and detrimental effects of fossil fuels on climate change, there is increased interest for the development of renewable plant-based alternatives for energy production (Simmons et al., 2008; Furtado et al., 2014). Advanced future biofuels will likely derive from non-edible feedstocks, where structural polysaccharides provide the main substrate for biochemical conversion into fuel. Lignocellulose, or woody biomass, is a potential feedstock for bioenergy production considering its availability, cost of production and scale at which it can be generated (Wang et al., 2012); however, lignocellulose’s natural recalcitrance to deconstruction prevents the economic conversion of biomass into fuel (Studer et al., 2011).

Enzymatic hydrolysis of structural polysaccharides is the most critical and costly aspect of biofuel production due to the structure and chemistry of the plant cell wall (Yu et al., 2011). The major structural components of lignocellulose (cellulose, hemicellulose, lignin) each contribute to biomass recalcitrance (Wang et al., 2016), but in woody biomass lignin has been demonstrated to contribute most negatively to saccharification, as its structure prevents enzymatic access to cellulose and non-specifically binds and immobilizes cellulase enzymes (Leu and Zhu, 2013). Additionally, lignin covalently bonds with hemicellulose, creating lignin-carbohydrate complexes that further inhibit saccharification (Yang and Wyman, 2004). Hemicellulose also affects efficient saccharification of lignocellulose due to its composition of difficult to ferment 5′-carbon sugars and its effect on the porosity of biomass through its cross-linkages with cellulose (Lange, 2007; Nigam and Singh, 2011). Despite being the main target for conversion to biofuel, cellulose also resists efficient hydrolysis due to its crystalline structure resulting in hydrophobic macrofibrils with limited reactive surface area and extensive hydrogen bonding (Yang et al., 2011; Mizrachi et al., 2012; Zhang Y. et al., 2015).

In order for a lignocellulose feedstock to be considered as a sustainable option for biofuel production, it must possess a growth rate suitable for economic harvesting. Eucalypt trees are an ideal candidate as a biofuel crop based on their established silviculture practices, global deployment, rapid growth in marginal soils and wide range of rainfall conditions, and genomic resources dedicated to wood formation and environmental resistances (Shepherd et al., 2011; Grattapaglia et al., 2012; Myburg et al., 2014; Healey et al., 2015). In Queensland, Australia, due to climate and environmental stresses, Corymbia citriodora subspecies variegata (CCV) is the most widely harvested hardwood tree, based on its form, wood quality and tolerances to variable soils, drought, pests, and disease (Lee et al., 2010). Examination of pulpwood traits across multiple trial sites indicate C. citriodora subspecies variegata is also well-suited for pulp and paper production based on predicted Kraft pulp yield (55% pulp per wood volume) and density (756 kg/m³), and moderate trait heritability (0.3 and 0.5, respectively) across multiple trial sites (Brawner et al., 2012). Tree improvement programs have also demonstrated the potential of F1 interspecies Corymbia hybrids (C. torelliana × C. citriodora subspecies variegata), combining desirable forestry traits (form, wood quality, vegetative propagation) within a single genetic background and possessing a superior growth rate [127–287%, diameter breast height (DBH)] as compared to either parental taxa (Lee et al., 2009).

While a high rate of growth is desirable for a variety of forestry applications, it is unclear what effect growth rate plays in altering biomass composition or affecting enzymatic saccharification. Transgenic manipulation of woody biomass has demonstrated that alteration of growth rate influences biomass composition and vice versa. Overexpression of a growth hormone precursor to gibberellin in transgenic poplar trees significantly improved growth rate in seedlings and biomass production within stem tissue. Additionally, transgenic lines also had longer and more numerous xylem fibers, which are commercially desirable for producing wood pulp with higher tensile strength (Eriksson et al., 2000). Disruption of lignin biosynthesis in transgenic poplar hybrids can negatively impact growth form and habit, where transgenic lines with significantly reduced lignin content resulted in brown discolored xylem tissue and dwarfed trees with reduced height, DBH, and growth rate (Leplé et al., 2007). Without lignin reinforcement during stem growth and thickening, xylem fibers are prone to collapse and cavitation being unable to withstand water pressures required for long distance transport (Kawaoka et al., 2006; Coleman et al., 2008; Voelker et al., 2011). Similarly, transgenic manipulation of cellulose biosynthesis can negatively impact growth rate and alter biomass composition. Joshi et al. (2011) demonstrated that the up-regulation of cellulose synthase (CesA8) inadvertently caused sense silencing of the native CesA and transgene, producing trees with little cellulose (10% dry weight) and a proportionate increase in lignin content (35%) and non-cellulosic polysaccharides (55%). As there is a strong interaction between growth and biomass composition, increased biomass production could negatively impact biofuel conversion processes if carbon resources shift toward xylem lignification. Small percentage increases in lignin content greatly affect substrate access for cellulases (Gressel, 2008), resulting in biomass that is inefficiently deconstructed and hydrolyzed. The aim of this study was therefore to examine the effect of growth rate on the main components of lignocellulose composition (glucan, xylan, Klason lignin) and their subsequent effect on enzymatic saccharification within commercial Corymbia interspecies hybrids and parental taxa. Given the significant impact each biomass component contributes toward saccharification, changes in biomass composition that occur in response to a high rate of growth will inform optimal harvest size for industrial use of woody biomass.

In this study, determination of biomass composition of Corymbia F1 interspecies hybrids and parental species was completed using the NREL Laboratory Analytical Procedure-Determination of Structural Carbohydrates and Lignin in Biomass (Sluiter et al., 2011), which corrects for ash content within acid insoluble residue measured gravimetrically after pyrolysis at 575°C. During testing and optimization of this procedure, ash content within Corymbia samples was below an amount that could be reliably measured with an analytical balance, resulting in a protocol modification that shortened the pyrolysis step to clean crucibles before the next use. Ash content is detrimental to liquid fuel conversion processes (as non-biodegradable residue) as well as negatively affecting the calorific value of wood and plant processing costs from thermo-chemical conversion (McKendry, 2002; Jørgensen et al., 2007). Eucalyptus and Corymbia have been characterized as possessing low ash content (<1%), (Guerrero et al., 2005; Magalhães et al., 2011; Çetrinköl et al., 2012), which decreases with tree age (Kumar et al., 2010).

If an advanced biofuel feedstock to be considered sustainable, it must possess a growth rate that warrants continued economic harvesting of that crop (Hinchee et al., 2009). Despite the advantages of lignocellulose for biofuel production, the presence of lignin and the biomass’ natural recalcitrance are substantial barriers to overcome, before this can be realized. During the growth and expansion of the plant cell wall, shifting carbon resources can simultaneously increase lignin content while decreasing polysaccharide content. This has been demonstrated through transgenic manipulation of lignocellulose biosynthesis. While in some instances, lignin reduction results in increased cellulose content, biomass and growth rate (Hu et al., 1999), disruption of lignin biosynthesis typically results in negative pleiotropic effects on growth and form. For example, suppression of the LIM domain transcription factor in Eucalyptus camaldulensis, an positive regulator of several lignin biosynthesis genes, resulted in a transgenic lines with reduced lignin content (17% as compared to 24%-wild-type [WT]) that frequently dropped upper leaves (Kawaoka et al., 2006). Specific targeting of the phenylpropanoid cinnamoyl-CoA reductase (CCR) gene in transgenic poplars, produced trees with less lignin (17% vs. 21%-WT) that that had significantly reduced growth (height, DBH, and growth rate) (Leplé et al., 2007). Additionally, disruption of lignin biosynthetic pathways often accompanies increased deposition of phenolics and extractives within wood tissue resulting in discolouration. Down-regulation of 4-coumarate:coenzymeA ligase (4CL) in poplar also results in reduced total lignin content and discolouration of xylem tissues, with biomass and leaf area reduced by half as compared to WT (Voelker et al., 2011). Adequate lignification of xylem vessels allows long distance transport of water through maintenance of internal water tension. Irregular xylem formation causes vasculature collapse, inadequate water transport, and weakened carbon sink strength (Coleman et al., 2008). Alternatively, overexpression of gibberellin 20-oxidase (a precursor to the gibberellin hormone) in hybrid poplars resulted in trees with faster growth in height and DBH, increased biomass, and more numerous and longer xylem fibers (Eriksson et al., 2000). Considering the increased biomass production of the Corymbia hybrid populations (Lee et al., 2009), it is reasonable to expect increased lignification of xylem tissue correlates with growth, consistent with wood samples taken across each size category (Table 1).

Enzymatic saccharification of Corymbia biomass demonstrates the strong negative effect lignin content on enzymatic hydrolysis. This is attributed to the structure of lignin physically inhibiting enzymatic access to cellulose microfibrils, forming cross-linkages with hemicellulose, and lignin non-specifically binding and immobilizing cellulases (Yu et al., 2011; Zhao et al., 2012; Leu and Zhu, 2013). Without delignification, up to 70% of cellulases remain immobilized within lignin (Berlin et al., 2005; Jørgensen et al., 2007). This has resulted in development of a variety of pretreatments for eucalypt biomass, to increase efficiency of enzymatic hydrolysis (Yu et al., 2010; Silva et al., 2011; Papa et al., 2012; Santos et al., 2012; Yáñez-S et al., 2013; Zhang C. et al., 2015). Lignin removal also creates pores in the cell wall through which cellulases can gain access to cellulose microfibrils (Yu et al., 2011). The pretreatment method for this dataset was hydrothermal (pressurized hot-water), which does not remove lignin but solubilizes hemicellulose, and disrupts the cellulose-hemicellulose-lignin complex. In this study, nor glucan or xylan content significantly impacted total glucose production from biomass. Cellulose, the primary donor of glucose during saccharification, resists enzymatic hydrolysis through hydrogen bonding and microfibril crystallinity. The highly compact cellulose polymer is hydrophobic, so only the hydrophilic ends of the microfibril are susceptible to enzymatic attack. Without disruption of the microfibril structure (normally achieved through energy intensive ball-milling) which increases porosity and promotes cellulose accessibility, glucan content independently considered may not significantly affect saccharification (Leu and Zhu, 2013) or ethanol production during simultaneous saccharification and fermentation (Vinzant et al., 1997). Additionally, an increase in hemicellulose content has been demonstrated to disrupt cellulose crystallinity in Miscanthus, thereby increasing enzymatic hydrolysis after acidic and alkaline pretreatment (Xu et al., 2012), the opposite effect has been demonstrated in Poplar transgenic experiments, where disruption of glycosyltransferase GAUT12 resulted in transgenic trees with less xylan (17–30% reduction) and increased saccharification yield (4–8% increase in glucose recovery) without a significant reduction in lignin content (Biswal et al., 2015). Future studies of this nature would benefit from the investigation of the transcriptome within the natural trait extremes within populations that cannot yet be transformed in order to discern the genetic mechanisms by which trees compensate for an increased growth rate.

Within Corymbia samples, the highest conversion efficiency was achieved within population HF-51 (M = 74%), with populations HF-153, HF-51, and Corymbia torelliana all containing samples that approached 100% conversion (Figure 6). While higher glucan conversion has been achieved in the literature (Sykes et al., 2015), our intent was to maximize the relative differences among parental species and F1 hybrid populations. C. citriodora subspecies variegata and HF-148 underwent the least efficient conversion of glucan to glucose, unsurprising considering the fixed enzyme dosage, however, HF-148 released the highest mean glucose amount from biomass, a promising result for future investigation with alternative pretreatments designed to increase cellulose accessibility and saccharification.

Given the economic importance of eucalypt taxa for industrial processes such as pulp and paper, their biomass composition has been well researched (Table 3). Klason lignin content among the Corymbia populations is consistent with values found in the literature and while acid soluble lignin was not measured here, it is reasonable to expect a similar range (2–4%) that would also likely have a detrimental effect on saccharification and subsequent fermentation (Ximenes et al., 2010). While the xylan content among Corymbia populations is consistent with those found in the literature for eucalypts, mean glucan content within Corymbia torelliana and hybrid families HF-153, HF-51, and HF-69 are low in comparison to HF-151, C. citriodora subspecies variegata and HF-148.

While mean glucan content in C. citriodora subspecies variegata and HF-148 samples is higher than other Corymbia populations or other literature values, the result is consistent with α-cellulose content (cellulose which remains insoluble) within elite Eucalyptus hybrids (E. urophylla × E. grandis) (Shinya et al., 2016). In their study Shinya et al. (2016) evaluated the wood properties of 918 hybrids and selected two genotypes (AM380 and AM063) for their extreme Klason lignin content (35 and 20%, respectively). Further characterization of AM063 biomass estimated its α-cellulose content was 59% dry weight (native wood). Similarly, the α-cellulose content of genotype AM380 was estimated at 48%, suggesting a negative correlation between lignin and cellulose. While the results glucan content from C. citriodora subspecies variegata and HF-148 were consistent with Shinya et al. (2016), the negative (Pearson) correlation between Klason lignin and glucan content within this dataset (r = -0.1) was not significant (P > 0.05). Despite cellulose being the principle contributor of glucan released from lignocellulose, the primary cell wall of woody species also contains xyloglucan, a matrix polysaccharide which contains glucan and xylan residues (Carpita and Gibeaut, 1993; Harris and DeBolt, 2010). If the high % dry weight of xylan within C. citriodora subspecies variegata and HF-148 can be attributed to xyloglucan, then it would be reasonable to expect that glucan content could also be increased.

F1 generation hybrids are typically intermediate with regard to parental trait values, but occasionally hybrids will either resemble one parent or exceed both (Rosenthal et al., 2002; Barbour et al., 2003). Despite lignin content of hybrid populations being intermediate to parental species, structural polysaccharide content in Corymbia hybrids HF-153, HF-51, HF-69 resembled the Corymbia torelliana parental species, whereas HF-151 was typically intermediate, and HF-148 was transgressive for glucan content (Figure 2). Other investigation of stem and leaf attributes among controlled-cross F1 Corymbia torelliana × section Maculatae crosses (to which C. citriodora subspecies variegata belongs), hybrids typically resembled the Corymbia torelliana maternal parent or were transgressive (Abasolo et al., 2012). Based on maternal parentage and the moderate to high heritability of cellulose content (h² = 0.42–0.86) (Kube et al., 2001; Apiolaza et al., 2005; Stackpole et al., 2010) it would be expected that Corymbia hybrids with shared genetics (HF-148 and HF-153) should be similar in structural polysaccharide content. This, however, is not the case. Additionally, for the hybrid families that have at least one parent in common (HF-148 and HF-69; HF-51 and HF-151) parentage of the common male pollen parent among the hybrids failed to elucidate genetic patterns underlying polysaccharide content. Population HF-148 is clearly distinguished from the other hybrids by having transgressive glucan content, while historical investigation of data collected from the Amamoor plantation site also confirms heterosis for growth rate (Figure 7). Between ages 1.9 and 8.6, each hybrid family at Amamoor, in terms of DBH, outperformed parental species with HF148 distinguished having the fastest growth beyond age 2.9.

Specific consideration of hybrid family HF-69 found each of its traits relating to biomass composition responded significantly different from other Corymbia populations. With the inclusion of the XL DBH tree data, Klason lignin, glucan, and xylan content as predicted by DBH of HF69 was less impacted by increasing tree size. Interestingly, evaluation of historical growth records from the Amamoor plantation also indicate HF-69 displayed an odd growth habit, displaying nearly double the variation in DBH of any other Corymbia population at age 8.6 (Figure 7). While HF-69 was not a top performer regarding total glucose production (M = 149 mg/g biomass) or glucan to glucose conversion (M = 63%), its biomass stability and large growth variation for selection and improved breeding may be desirable for other forest industries, such as timber production, where product consistency is highly regarded.

As advanced future biofuel feedstocks require a high rate of growth to justify harvesting at an industrial scale and efficient deconstruction and conversion is dependent on biomass composition, a key consideration for a renewable feedstock is the impact of growth on lignocellulose formation. This investigation identified that in response to growth in Corymbia populations, major structural components of biomass were significantly impacted by tree size, shifting toward increased recalcitrance through increased Klason lignin content and deceased polysaccharide content. This research suggests that fast growing trees harvested under fast rotations would be best suited for lignocellulosic biofuel production. Given current forestry management practices involve thinning trees planted at a high stocking rate to promote growth in high value trees, traditional forestry and bioenergy applications could be combined if thinned trees are removed for biofuel use before lignification is complete.

AH: Responsible for overall experimental design, sample collection, analysis and manuscript writing. DL: Responsible for overall experimental design, sample collection, interpretation of results and manuscript editing. JL: Responsible for overall experimental design, processing samples for analysis, interpretation of results and manuscript writing. GP: Responsible for two-step acid hydrolysis design, collecting and interpreting composition data, and manuscript writing. JG: Responsible for high-throughput saccharification design, collecting and interpreting glucose data, and manuscript editing. LC: Responsible for high-throughput design and collection of structural polysaccharide data and manuscript editing. FA: Responsible for hydrolysis and saccharification design, interpretation of results and manuscript editing. SS: Responsible for two-step acid hydrolysis design, interpretation of results and manuscript editing. BS: Responsible for hydrolysis and saccharification design, interpretation of results and manuscript editing. RH: Responsible for overall experimental design, interpretation of results and manuscript writing. All authors have agreed on the final version of this manuscript and are accountable for the research therein.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.