As a biomass and biofuel plant, switchgrass is also considered a potential crop for production of biodegradable plastics as a value-added co-product, which can "reduce petroleum consumption and decrease plastic waste disposal issues" (Somleva et al. 2008). In such an attempt, Metabolix, Inc. introduced bacterial genes into switchgrass to produce such a plastic, polyhydroxybutyrate (PHB) (Somleva et al. 2008). The enzymes encoded by the three transgenes for PHB synthesis were targeted to plastids to enhance PHB yield as previously demonstrated. Transgenic plants containing up to 3.7% dry weight of PHB in leaf tissues and 1.2% dry weight PHB in whole tillers were obtained. The PHB granules were accumulated in chloroplasts of the leaves. Most of the transgenic plants grew normally although affected growth was also observed. Transgenes and PHB production were inherited to offspring plants through both male and female gametes. Although the yield of PHB in these transgenic plants did not meet the 7.5% dry weight threshold estimated by Metabolix to be necessary for profitable commercialization, the authors believe it is the first step towards achieving the goal. It also demonstrated the amenability to introduce multiple genes to alter metabolic pathways in this important biofuel crop.

Current studies suggest that it is feasible to generate low-lignin switchgrass, improve biomass yield, and add value to this biofuel crop. Field tests on these various transgenic plants are needed to support the claims that the low lignin content, normal or increased biomass yield, and other improved traits of the transgenic lines still hold in various field conditions.

Switchgrass can be established with conventional planting techniques— broadcasting or drilling into a well-till seedbed—or with no-till planters that can drill seed into un-worked bare soil, killed sod or crop stubble. Aside from environmental considerations associated with tillage, perhaps the biggest issue with planting method is the ability to achieve proper planting depth. As noted above, seed depth is critical for successful establishment, and poor seed placement has been a common cause of stand failure with both conventional drilling or broadcast and no-till seeding methods.

Broadcast seeding is the least preferred of planting methods and typically only successful when the seedbed is rolled or compacted after the seed are broadcast (Evers and Butler 2000; Monti et al. 2001). Rolling or packing likely pushes some seed to the appropriate depth, allowing for establishment and survival.

Comparisons of switchgrass establishment by drilling into tilled or non-tilled seedbeds are often confounded and inconsistent (Parrish and Fike

2005) with outcomes dependent on year, location and presence of residues (e. g., see King et al. 1989). Success with no-till seeding methods have been possible for some time (e. g., see Wolf et al. 1989) and Rehm (1990) reported no yield differences between the two planting methods. Some studies have reported greater seedling numbers with no-till planting methods (Harper et al. 2004) but seedling numbers may or may not have long-term effects on stand productivity (see section on row spacing below).

Row Spacing

Although less critical than planting depth and timing, appropriate row spacing allows plants to optimize resource capture with decreased seeding rates. Reducing competition can allow more efficient resource use and support increased yields. Results in Texas from row spacing experiments during establishment years indicate that switchgrass produces more robust plants with greater tiller density and mass as row spacing increases (Muir et al. 2001). Increasing row spacing from 20 to 80 cm also increased yields in Alabama without N fertility (Ma et al. 2001). Similar data are reported by Foster et al. (2012) who found that switchgrass planted at row spacings of 17.8 to 88.9 cm had similar DM yield. These data indicate row spacing may not be a major consideration to successful establishment of switchgrass as a biomass energy crop at low latitude locations. Although there is little information to guide such planting decisions at greater latitudes, we would predict similar outcomes. Potential for success with various row widths offers greater flexibility for farmers with different types of drills and who might apply mechanical tools for weed control.

Plant growth-promoting bacterial endophytes can affect growth directly by providing bacterium-synthesized compounds, often plant hormones, and by facilitating the acquisition of compounds from the environment, including atmospheric nitrogen fixation. Endophytes may also act indirectly by decreasing or preventing the colonization or the deleterious effects of pathogenic organisms (Lodewyckx et al. 2002) by producing antibiotics to outcompete plant pathogens (Bibi et al. 2012).

One of the most well-studied bacterial endophyte associations is atmospheric nitrogen fixation by specific endophytes. This symbiosis is well known in leguminous plants (Stacey et al. 2006) where the soil bacteria Rhizobia infect the roots of the host plants, inducing the formation of nodules where they fix atmospheric nitrogen and provide it to the host plant in exchange for carbon compounds (Lodewyckx et al. 2002). Additionally mutualistic associations through the fixation of nitrogen can also be observed in non-leguminous plants, such as rice (Mattos et al.

2008) , maize (Montanez et al. 2009), sugarcane (Oliveira et al. 2009), wheat (Webster et al. 1997), strawberries (de Melo Pereira et al. 2012), and grasses (Reinhold-Hurek et al. 1993; Kirchhof et al. 2001).

Nitrogen-fixing bacteria have been studied extensively in the bioenergy crop sugarcane, and include Gluconacetobacter spp., Azospirillum spp., Herbaspirillum spp. and Burkholderia spp. (James and Olivares 1998; James et al. 2001; Suman et al. 2005; de Carvalho et al. 2011). In fact, in Brazil, the cultivation of sugarcane uses only a small amount of fertilizer (de Carvalho et al. 2011) without showing nitrogen deficiency symptoms (Rosenblueth and Martinez-Romero 2006), and there is evidence that a significant amount of nitrogen is obtained from plants associated with bacterial endophytes (de Carvalho et al. 2011). To date, there are no reports on nitrogen-fixing bacterial endophytes in switchgrass, and screening for diazotrophic bacteria that inhabit switchgrass is under way in our laboratory.

There have been numerous publications on plant growth promotion by bacterial endophytes (see review in Berg 2009; Mei and Flinn 2010). In switchgrass, young seedlings of the cultivar Alamo inoculated with Burkholderia phytofirmans strain PsJN, isolated from onion roots (Frommel et al. 1991), showed significant growth promotion with an increase of root and shoot length of 35.6% and 32.8%, respectively, as well as an increase of fresh weight of 83.6% compared with control plants (non-inoculated) after one month under in vitro conditions (Kim et al. 2012). The same pattern was observed under growth chamber and greenhouse conditions, where plants inoculated with the B. phytofirmans strain PsJN showed persistent growth vigor with significant increases in fresh and dry weights, and an increase in the number of early tillers (Kim et al. 2012). Also, results showed that B. phytofirmans strain PsJN has potential in the development of a low input and sustainable switchgrass feedstock production system on marginal lands as higher biomass yields were observed under sub-optimal growth conditions with PsJN inoculated plants over control (Kim et al. 2012). However, PsJN growth promotion is genotype specific in switchgrass as the upland cultivar Cave-in-Rock did not respond to inoculation. We are currently isolating bacterial endophytes from switchgrass tissues and have made progress in screening and selecting beneficial bacterial endophytes which have a broad spectrum of growth promotion in various switchgrass cultivars.

SND1 is a higher order activator expressed in xylem cells that activates the biosynthesis of cellulose, matrix polysaccharides, and lignin (Zhong et al. 2006). Repression of SND1 leads to abnormal Arabidopsis plants lacking vascular and interfascicular fibers; whereas, overexpression lines display ectopic expression of genes involved in secondary wall biosynthesis. Zhao et al. (2010) also found that SND1 directly regulates the expression of F5H, one of the key enzymes in lignin biosynthesis. The genes encoding MYB46, MYB83, MYB103, MYB32, SND3 and KNAT7 possess the Secondary wall NAC Binding Element (SNBE) cis-element and appear to also be direct targets of SND1 (Zhong et al. 2007; Zhong et al. 2011; Zhong et al. 2012).

Other NAC transcription factors, including NST1/2, VND6, and VND7, also play a key role in regulation of secondary cell wall synthesis. They all positively regulate similar downstream targets compared with SND1, including MYB46, MYB83, MYB103, MYB58, and SND3 (Zhong et al. 2010). NST1 and NST2 are involved in regulating secondary wall thickening in anther walls as well as stems (Mitsuda et al. 2007). NST2 especially is strongly expressed in anther tissue. VND6 and VND7 act as a key regulator of xylem differentiation. Overexpression of the VNDs prompts the differentiation of non-vascular tissues into treachery elements (Kubo et al. 2005). VND6 physically binds to the Trachery Element Regulating cis-Element (TERE), which is possessed by a number of genes involved in tissue-specific trachery cell wall biosynthesis and programmed cell death. VND7 is negatively regulated by VNI2 (VND-INTERACTING2 NAC PROTEIN2), which is another recently characterized NAC domain transcription factor (Yamaguchi et al. 2010). The secondary wall regulatory network that functions in xylem differentiation also includes ASL19 (ASSYMETRIC LEAVES2-LIKE19) and ASL20. Expression of these proteins is activated by VND6 and VND7 and also forms a positive feedback loop in turn up-regulating expression of the VND genes (Soyano et al. 2008).

MYB46 and MYB83, which are controlled by NACs, act as positive regulators of secondary cell wall synthesis (Zhong et al. 2012). Among the transcription factors downstream of these MYB proteinss, MYB52, MYB54, MYB58, and MYB63, are important for secondary cell wall synthesis. Promoter deletion coupled with transactivation analysis revealed the Secondary wall MYB Responsive Element (SMRE), a cis-element that is enriched in the promoters of known targets of MYB46 and MYB83. In a further regulatory layer, MYB58 and MYB63, controlled by both NACs and MYB46/83, are implicated in regulating lignin biosynthesis. These proteins target AC-rich elements, which are enriched in the promoters of at least some lignin biosynthesis genes (Lois et al. 1989; Zhong et al. 2012).

More recently a handful of transcriptional regulators in protein families other than NAC and MYB R2R3 have also been determined to have central roles in regulating secondary wall biosynthesis. For example, Wang et al. found that WRKY12 appears to function as a high level negative regulator of secondary cell wall biosynthesis in stems. Medicago and Arabidopsis wrky12 mutants have thickened cell walls in stem pith cells and increased biomass with abnormal deposition of lignin, xylan, and cellulose (Wang et al. 2010). Another example is KNAT7 which is a KNOX-type homeodomain transcription factor that also negatively regulates cell wall thickening and lignin biosynthesis (Li et al. 2011). Loss-of-function knat7 mutants exhibit increased cell wall synthesis gene expression. Besides SND1, MYB75 physically interacts with KNAT7 to restrain cell wall biosynthesis (Bhargava et al. 2010).

A number of secondary wall regulators in grasses have been characterized via heterologous expression in Arabidopsis, but very few have been examined in situ. Heterologous overexpression of ZmMYB31 and ZmMYB42 in Arabidopsis leads to reduced lignin content (Fornale et al.

2010) . Arabidopsis lines that overexpress ZmMYB42 exhibit reduced plant stature, leaf size, tertiary venation, and S-unit lignin content. ZmMYB31 directly interacts with an element similar to the AC-element present in the ZmCOMT promoter (Sonbol et al. 2009). The orthologs of the Arabidopsis SWNs and MYB46 from rice and maize are able to activate secondary wall biosynthesis in Arabidopsis (Zhong et al. 2011). Consistent with conservation of regulatory mechanisms, the promoters of OsMYB46 and ZmMYB46 contain SNBE cis-elements and the rice and maize SWNs directly bind these elements to activate gene expression (Zhong et al. 2011). Another example is expression of an AP2 transcription factor from Arabidopsis, AtSHN2, which in rice significantly enhances cellulose content while reducing lignin content and resulting in improved saccharification yields (Ambavaram et al. 2011). Promoter analysis and binding assays suggest that AtSHN2 may repress expression of the rice orthologs of SND1, NST1/2 and VND6, and activate expression of MYB20 and MYB43.

Dixon and colleagues recently reported that the switchgrass protein, PvMYB4A, an ortholog to the Arabidopsis MYB4 and the maize MYB31 proteins, acts as a repressor of lignin biosynthesis in switchgrass (Shen et al. 2012). Overexpression of PvMYB4 in switchgrass reduced the total lignin content and the amount of cell wall ester-linked p-coumarate. The efficiency of sugar release from transgenic biomass was increased by almost 3-fold. An element similar to the AC-element found in dicots is also the probable binding site of PvMYB4. This discovery demonstrates that manipulating transcription factors that control enzymes that function in cell wall biosynthesis is a good alternative way to reduce the recalcitrance of switchgrass and improve lignocellulosic biomass. Still, the differences in cell wall content between grasses and dicots might be consistent with some divergence in the factors that regulate cell walls. Certainly, given the high ploidy level of switchgrass (4n, 6n, or 8n) and its outcrossing nature, heterozygosity that could have functional consequences has evolved. For example, five distinct, but closely related PvMYB4 sequences were identified from a single switchgrass genotype (Shen et al. 2012; Shen personal communication).

Independent of the possibility of grass-diverged mechanisms of regulation, the continual flow of new publications about cell wall regulators suggests that all factors, and certainly the interactions among them, have yet to be uncovered. Due to space constraints we are not able to elaborate on posttranscriptional and posttranslational regulatory mechanisms in cell wall synthesis, the study of which is still in its infancy (Humphrey et al. 2007; Wolf et al. 2012).

Dayong Li,[13]‘[14] Man Zhou,2 Zhigang Li2 and Hong Luo2’*

Introduction

With the rapid development of genomics and bioinformatics, recent studies have suggested that the number of protein-coding genes is similar in many model eukaryotes whose whole genome sequences have been obtained and analyzed in detail (Matera et al. 2007; Ponting et al. 2009). Genome-wide transcriptional analyses have identified large numbers of non-coding RNAs (ncRNAs) in humans, animals and plants (Hirsch et al. 2006; Ravasi et al. 2006; The ENCODE Project Consortium 2007; Guttman et al. 2009; Amor et al. 2009; Jouannet et al. 2011). Based on their length, ncRNAs can be arbitrarily divided into small ncRNAs, intermediate-size ncRNAs and long ncRNAs (Amor et al. 2009; Jouannet et al. 2011; Liu et al. 2013). To date, the best characterized of all the ncRNAs has been small RNAs (sRNAs). Endogenous small RNAs are about 19-30 nucleotides (nt) RNA molecules that modulate

gene expression at the transcriptional and/or posttranscriptional levels and play key roles in many developmental and physiological processes in eukaryotic organisms (Zamore and Haley 2005; Bonnet et al. 2006; Zhang et al. 2006; Ramachandran and Chen 2008; Poethig 2009).

In plants, sRNAs can mainly be classified into small interfering RNAs (siRNAs) and microRNAs (miRNAs) based on their precursor structures and biogenesis processes (Vazquez 2006; Vaucheret 2006; Sunkar and Zhu 2007; Ramachandran and Chen 2008; Jin and Zhu 2010; Vazquez et al. 2010). The siRNAs are derived from double stranded RNA precursors and can be divided into heterochromatic siRNAs (hc-siRNAs), trans-acting siRNAs (ta-siRNAs), long siRNAs (lsiRNAs), natural antisense transcripts-derived siRNAs (nat-siRNAs), and others (Bonnet et al. 2006; Zhang et al. 2012). The miRNAs are distinguished from the siRNAs since they are derived from the processing of longer primary miRNA transcripts, which fold into hairpin-like stem-loop structures (Bartel 2009; Chen 2009; Chuck et al. 2009; Poethig 2009; Voinnet 2009; Zhu et al. 2009).

Switchgrass (Panicum virgatum L.) is a warm-season perennial grass and has been recognized as a dedicated cellulosic biofuel crop because of its broad adaptation to marginal lands and high biomass production (Vogel 2004; McLaughlin and Kszos 2005; Bouton 2007; Li and Qu 2011; Mann et al. 2012). Although switchgrass has attracted great attention, little is known about its many aspects on basic biology, including ncRNAs. In this chapter, we will provide an overview of the miRNAs in this biofuel plant species and discuss their potential applications in switchgrass genetic improvement.

Mowing. Mowing for weed control in forages is generally not very effective (Miller and Strizke 1995) because it is non-selective and may occur too late to reduce competition between weeds and the seedlings. Mitchell et al. (2010a) recommended mowing just above the switchgrass canopy (typically 20 to 30 cm) near the 4th of July to reduce the leaf area of both grassy and broadleaf weeds in newly-seeded switchgrass stands. Mowing can reduce competition for light, and can prevent weeds from going to seed and contributing to the soil seed bank. It is sometimes the only option to suppress grassy weeds, especially when trying to establish switchgrass where herbicides are not effective.

Mob grazing. Mob-grazing is stocking a high density of animals in an area for a short duration (up to 1 wk). It reduces selective grazing by livestock, and thus, can be effective in the control of grass weeds and allowing sunlight to the new seedlings (Miller and Strizke 1995). However, grazing must be delayed until seedling roots are well established or the seedlings can be uprooted. Often, as with mowing, the efficacy of mob-grazing is only moderate, because it is applied too late to have maximum benefit in reducing weed competition for moisture, sunlight, and nutrients, and damage to the soil from foot traffic may be significant. Further, unpalatable weeds might not be grazed and the seedling forage may be preferred to weed species.

Cultivation (Tillage). The main use of cultivation in switchgrass has been for seedbed preparation to remove all vegetation prior to planting to help ensure good seed to soil contact. It is important to consider that tillage can bring dormant weed seeds to the surface, so it is often best plant into a stale seedbed after tillage. Under certain unique circumstances, tillage could be used post-planting to non-selectively control weeds between rows of switchgrass planted on wide row spacings.

Companion Crops. Companion crops are planted along with switchgrass to provide protection from wind and water erosion. Hintz et al. (1998) also reported companion crops could reduce weed competition during switchgrass establishment. For example, they reported that corn planted in perpendicular orientation at reduced seeding rates of 24,700 to 49,400 seeds ha1 on either 76 or 114 cm spacing did not reduce switchgrass establishment. They concluded that atrazine reduced weed emergence early season while the corn shaded the weeds during late season. It is important to illustrate that in this study, they planted following soybeans, and although grass weeds (foxtail) were present, they were not detrimental to the switchgrass in the control plots. Companion crops generally are not recommended except in extreme environments and conditions; cover crops may be best used when terminated prior to shading of the switchgrass seedlings. In Oklahoma, cowpea (Vigna unguiculata L. Walp) and forage sorghum (Sorghum bicolor (L.) Moench.) planted in perpendicular or alternating drill row orientations were too competitive with switchgrass seedlings and complete stand failure occurred when these crops were harvested at the end of season (T. J. Butler, unpublished data). Cowpea could be a viable cover crop if planted with an alternating row pattern, since it can be removed with 2,4-D amine once it begins to shade the switchgrass seedlings (Fig. 3). This alternating row pattern can be accomplished with drill containing two seed boxes, one for each species, and plugging every other hole of each drill box. However, care should be taken to ensure each planter unit is calibrated at the appropriate depth for each species.

Both bacterial and fungal endophyte-plant interactions involve modifications of plant gene expression and overall plant physiology/biochemistry to beneficially impact growth and stress tolerance. While monitoring specific gene expression during beneficial endophyte-sugarcane interactions, Arencibia et al. (2006) identified 47 differentially expressed sequence tags (EST) using cDNA-AFLP analysis. The transcripts showed significant genetic homologies to major signaling pathways such as the ethylene signaling pathway. For example, PYK10 encodes for a root — and hypocotyl-specific P-glucosidase/myrosinase and is important during the endophyte P. indica and Arabidopsis beneficial bio-control against herbivores and pathogens (Sherameti et al. 2008). NoxA was found to be crucial in regulating hyphal morphogenesis and growth in the mutualistic symbiotic interaction between the fungal endophyte Epicho festucae and perennial ryegrass (Tanaka et al. 2008). Functional genomics research will help scientists understand and elucidate mechanisms under which beneficial microorganisms promote host plant growth and enhance stress tolerance. Currently we are carrying out studies of mechanisms of plant growth promotion by bacterial endophytes using the responsive switchgrass cultivar Alamo and non-responsive cultivar Cave-in-Rock to Burkholderia phytofirmans strain PsJN (Kim et al.

2012) . Comparative global gene expression profiling is being conducted using both cultivars following B. phytofirmans strain PsJN inoculation with DOE-funded switchgrass EST microarray chips by Genomics Core Facility in the Noble Foundation. Approximately 35,200 switchgrass ID probes were identified to show significant differences between switchgrass cultivars Alamo and Cave-In-Rock after B. phytofirmans strain PsJN inoculation. Using the rice genome as a model for the analysis of the data along with the MapMan (Usadel et al. 2005) and the PageMap (Usadel et al. 2006) software, we are currently analyzing this large data set. Results showed that in Alamo almost 2000 genes were unique up-regulated at 0.5 day. On the other hand, in Cave-in-Rock, the number of unique up-regulated genes for 0.5 day was only 901. The significant changes are found in transcription factor genes, plant hormone and cell wall metabolism (unpublished data).

Bacterial and Fungal endophytes exhibit a diverse range of growth promoting mechanisms. In many cases, endophytes, primarily bacteria, possess multiple mechanisms of action and differentially express these traits at different stages of plant growth and development. Under stress conditions, endophytes help the host plant survive and flourish, as in the case of ACC deaminase activity and bio-control compound production. Under normal conditions, endophytes help fix atmospheric di-nitrogen and produce plant hormones to help the plant grow to its maximum potential.

Together, under both stress and normal conditions, endophytes ensure its host plant thrives, and its nutrient rich environment is maintained.

To optimize improvements to deconstruction due to overexpression of lignocellulolytic enzymes in planta, researchers have sought to increase the percent of plant total soluble protein (TSP) that the enzymes represent and to decrease detrimental plant phenotypes caused by enzyme accumulation. In many studies, hydrolase levels remain relatively low (1-5% TSP) as do the % increases in saccharification. Figure 7 summarizes strategies being used to improve in planta expression of GHs. These include organelle targeting, codon optimization, promoter enhancement, and transient expression. Some of these methods as well as other approaches to regulate hydrolase activity, such as use of inteins, are also being tested for their ability to mitigate plant dwarfism or other physical defects that can accompany in planta GH expression.

Organelle targeting is a key factor in optimal accumulation of cell wall degrading enzymes in plants. Most previous studies have examined accumulation of enzymes in the apoplast, cytosol, and vacuole (Sticklen 2006; Taylor et al. 2008; Sainz 2009). Targeting the transgene to incorporate into the chloroplast genome has also been shown to induce accumulation of large amounts of foreign proteins (Oey et al. 2009). For example, a collection of GHs and related proteins from T. reesei expressed in the chloroplast had higher activity and pH and temperature stability compared with the same proteins expressed in E. coli (Verma et al. 2010). A crude cocktail derived from lines expressing each enzyme released up to ~3500% more glucose from biomass compared with a commercial enzyme cocktail, though it is not clear that the study was conducted with the same amount of protein in both samples (Verma et al. 2010). Another study showed that with chloroplast targeting, tobacco could accumulate four cell wall degrading enzymes at levels up to 40% TSP (Petersen et al. 2011). However, if selected for homoplasty, meaning all chloroplasts were transformed, this resulted

Figure 7. Strategies for improving glycosyl hydrolase expression in plants to enhance the quality of biomass for biochemical production of biofuels.

in pigment-deficient mutants unable to grow autotrophically. Even if heteroplastic, transformant had to be grown on a sucrose-rich medium. This is in contrast to the 70% TSP achieved without any severe phenotype when the tobacco expressed a protein antibiotic (Oey et al. 2009). The authors suggest that the chloroplastic GHs might sequester intermediates needed in plant metabolism. Based on these results, chloroplast transformation harbors the following promises: a significantly better biological enzyme factory than bacteria; an environment suitable for high accumulation of enzymes versus the apoplast, vacuole, cytosol, etc.; and a setting amendable for producing plants that could express a multitude of cell wall degrading enzymes. However, further strategies are needed to combat deleterious growth effects.

Codon optimization is a simple way to enhance heterologous expression levels. All organisms and DNA-containing organelles contain particular preferences for codon usage. Foreign genes may not adhere to these preferences and likely due to depletion of rare tRNAs these codons may diminish expression. Arabidopsis expressing codon optimized

D. thermophilum XynA and XynB resulted in a TSP of 14% and 3%, respectively, and a 14% improvement in xylose release compared with the wild-type plant (Borkhardt et al. 2010). The codon optimized expression of XynA and XynB in the apoplast allowed for significantly higher accumulation of these enzymes than what was seen in earlier un-optimized examinations with similar xylanases from S. olivaceoviridis (Yang et al. 2007).

Increasing promoter activity is another key target to improving TSP levels of heterologous GHs. The ideal promoters are those with strong location specific and/or inducible activity (Taylor et al. 2008). For instance, accumulation of 5.8% TSP was obtained by driving the gene encoding a thermostable fi-glucosidase, BglB from Thermotoga mamma, under the control of the rbcS-1A promoter and with targeting to the chloroplast (Jung et al. 2010). This is a light-regulated promoter that controls the transcription of Ribulose-1,5-bisphosphate carboxylase oxygenase small subunit. In another study, a TSP of 6.1% for the E1 cellulase under the control of the synthetic Mac promoter was achieved for apoplast targeting in rice (Chou et al. 2011). This was higher than the 4.9% TSP obtained in another study using the Cauliflower Mosaic Virus 35S promoter (Oraby et al. 2007).

Besides enhancing expression of heterologous hydrolases, other challenges associated with GH expression are developmental defects including stunted growth, severe pigment deficiency, enhanced disease susceptibility, poor seeds, and poor growth (Dai et al. 1999; Skjot et al. 2002; Harholt et al. 2010; Gray et al. 2011). Common strategies to avoid these deleterious effects are conceptually similar to those used to enhance expression and include selectively targeting hydrolase expression to storage organelles, use of inducible or developmentally regulated promoters, as well as employing thermophilic enzymes with low activity at ambient temperatures, as mentioned previously (Taylor et al. 2008; Jung et al. 2012). Inteins are another effective means to regulate in planta GH activity. An intein is a protein that can catalyze its own removal from and the subsequent rejoining of two flanking protein segments, i. e., the exteins (Sharma et al.

2006) . Inteins have been engineered to be activated by a variety of different external stimuli, including pH, temperature, and small molecules, providing good potential for use as a means to control activity of celluolytic enzymes in plants (Skretas et al. 2005; Sharma et al. 2006). Recently, a thermo-regulated intein was used to control the activity of a thermostable xylanase, XynB from Dictyoglomus thermophilum, expressed in maize (Shen et al. 2012). Without the intein the xylanase greatly reduced the plant seed mass and fertility, but these deleterious phenotypes were largely restored by the insertion of the intein. Xylanase activity was retained after the temperature was elevated to induce intein excision. Biomass from the xylanase-intien expressing plants released approximately 45% more sugar than that from the unmodified plants (Shen et al. 2012). Future directions might be to express a cocktail of hydrolases without any significant effects on phenotype.

Transgenic approach manipulating gene expression for trait modification is one of the effective strategies for switchgrass breeding and genetic improvement (Sticklen 2006; Gressel 2008; Li and Qu 2011; Mann et al.

2012) . As one of the important regulatory factors in plants, miRNA genes and their targets are potential candidates for this purpose. Although little is known about the functions of miRNAs in switchgrass, several miRNA families are evolutionarily conserved in plant species, so is the miRNA — mediated regulatory mechanism (Jones-Rhoades et al. 2006; Chen 2009; Voinnet 2009; Cuperus et al. 2011). Therefore, it is possible to make use of the miRNAs whose functions have been identified in other species to modify this bioenergy crop.

Switchgrass, along with big bluestem (Andropogon gerardii), little bluestem (Schizachyrium scoparium), and indiangrass (Sorghastrum nutans), were the grasses that accounted for nearly all of the aboveground primary production of the tallgrass prairie. Tallgrass prairie once covered 56 million ha of the central USA, but today less than 4% remains in native vegetation (Rahmig et al. 2008). Switchgrass has a long history of grazing as a component of the tallgrass prairie and grazing switchgrass monocultures has occurred for more than 40 years (Kreuger and Curtis 1979). The first switchgrass cultivar, ‘Nebraska 28’, was released jointly by the USDA and the University of Nebraska in 1949. It was developed from native seed collections in Holt County, Nebraska for livestock forage production and conservation purposes. Nebraska 28 produced steer average daily gains (ADG) of 0.93 kg hd-1 d-1 and body weight gains of 147 kg ha-1 (Kreuger and Curtis 1979). Switchgrass ADG was less than that for indiangrass, but greater than that for big bluestem. Grazing switchgrass with monogastrics is not recommended due to potentially lethal concentrations of saponins, especially for horses (Lee et al. 2009). Commercially available switchgrass cultivars bred specifically for livestock forage production includes ‘Trailblazer’ (Vogel et al. 1991) and ‘Shawnee’ (Vogel et al. 1996). Trailblazer had a 23% increase in body weight gain per hectare when compared to the cultivar ‘Pathfinder’ (Vogel et al. 1991) and Shawnee had greater in vitro dry matter digestibility (IVDMD) than ‘Cave-in-Rock’ and greater forage yield than Trailblazer (Vogel et al. 1996).

Switchgrass biofuel development began in 1984 when the U. S. Department of Energy (DOE) funded field evaluations for about 34 herbaceous species at 31 sites in seven states for their suitability for biomass production (Wright 2007; Vogel et al. 2011; Parrish et al. 2012). Switchgrass was one of the top biomass producers at most of the sites and was selected as a model biofuel feedstock by DOE in 1991 (Wright 2007). Switchgrass is a broadly-adapted native with cultivars available for most US regions, it can be grown from seed, there is an existing seed industry, and it can be grown and harvested with available hay equipment (Vogel et al. 2011). The DOE funded switchgrass production and breeding research through the Biofuels Feedstock Development Program from 1992 to 2002 (McLaughlin and Kszos 2005; Wright 2007; Vogel et al. 2011). In 2002, DOE discontinued the Feedstock Development Program (McLaughlin and Kszos 2005; Sanderson et al. 2006) and focused on crop residues like maize stover for bioenergy because of the assumed availability and low cost (Vogel et al. 2011). USDA-ARS expanded funding for bioenergy in 2002 to include switchgrass genetics, breeding, and management, focusing on its potential use on marginal croplands similar to land that is currently held out of production in CRP (Vogel et al. 2011). Vogel et al. (2011) reported new research programs were initiated on perennial energy crops like switchgrass in 2006 in the USA and estimated that over $1 billion has been invested in the USA on biomass energy research since 2006 by both government and commercial companies.

The increased scientific importance of switchgrass is demonstrated by the number of publications focusing on switchgrass. Parrish et al. (2012) reported that the number of switchgrass publications has increased exponentially since 1990, but the volume of switchgrass research is very small compared to other crops. Reports on switchgrass averaged 8 per year from 1990 through 1994, but increased to 16 articles per year in the second half of the decade (Parrish et al. 2012). In 2010, 165 publications were reported for switchgrass, whereas 5,610 publications were reported for maize (Parrish et al. 2012). Research efforts are not limited to the USA, but have been reported in more than 20 countries including Australia, Canada, China, Colombia, Germany, Greece, Ireland, Italy, the Netherlands, Spain, and the United Kingdom, as well as other areas throughout Europe and Asia (Parrish et al. 2012).

As indicated previously, switchgrass is productive on sites that are poorly suited to annual crop production. The perennial nature of switchgrass will make stands productive for at least 10 years with good management (Mitchell et al. 2012a). However, the economic feasibility of switchgrass for bioenergy hinges on establishing stands with a harvestable yield in the planting year (Perrin et al. 2008). In the planting year, it is feasible to produce and harvest 50% of the yield potential of the cultivar after a killing frost and produce and harvest 75% to 100% of the yield potential of the cultivar in the first full growing season after planting (Vogel et al. 2011; Mitchell et al. 2012a, b). Adequate weed control and uniform stands in the planting year allows for full biomass potential one to two years after planting (Schmer et al. 2006). With upland cultivars, 4 to 5 Mg ha1 after a killing frost is typical during the planting year if precipitation is near the long-term average (Mitchell et al. 2012a, b). In the first year after seeding, fields can be near full production, producing 8 to 13 Mg ha1 after a killing frost in the central Great Plains (Mitchell et al. 2010). A switchgrass stand is considered mature and at full production in the second full production year (third growing season). Lowland cultivars like ‘Kanlow’ and ‘Alamo’ originated in southern latitudes and are typically adapted to areas south of 40° N. latitude and have not been evaluated at the field scale in the Great Plains and Midwest (Vogel et al. 2011). Alamo average biomass production fertilized at 168 kg N ha-1 yr-1 was 14.5 and 10.7 Mg ha-1 yr-1 at two Texas locations (Muir et al. 2001). Without applied N, biomass production declined over the years. Small plot trials of lowland ecotypes in Nebraska produced an average of 10.1 Mg ha1 in the year after seeding, with Kanlow producing 11.7 Mg ha1 in the year after seeding (Mitchell et al. 2010). Wullschleger et al. (2010) compiled a switchgrass biomass production database of 39 field sites in 17 states which supported the single harvest for bioenergy. Mean biomass yield across all locations was 8.7 ± 4.2 Mg ha1 for upland cultivars and 12.9 ± 5.9 Mg ha1 for lowland cultivars and the yield difference between ecotypes was significant. Yield trials in Nebraska indicate new material developed specifically for biomass provides a 2.2 Mg ha1 yield increase. Upland x lowland hybrids are promising for biomass energy, with hybrids increasing yield by 32 to 54% compared with parental lines (Vogel and Mitchell 2008). Deploying hybrid switchgrass to the field scale will result in potential harvestable yields of greater than 20 Mg ha1 in the Great Plains and Midwest (USDOE 2011).

Harvesting biomass removes large quantities of nutrients from the system (Mitchell et al. 2008, 2012b). Since nitrogen (N) is the most limiting nutrient for switchgrass production and is the most expensive annual production input, reducing N removal from the switchgrass production system has a positive effect on the economic and environmental sustainability of the system (Mitchell and Schmer 2012). Nitrogen removal is a function of N concentration and biomass yield, with biomass N concentration increasing as N fertilizer rates increase (Vogel et al. 2002). Biomass has been optimized when switchgrass is harvested at the boot to post-anthesis stage and fertilized with 120 kg N ha-1, with N removed similar to the N applied (Vogel et al. 2002). The interaction of N rate and harvest date must be considered to replace only the N needed to prevent over-fertilization. Harvesting 10 Mg ha-1 of switchgrass DM with whole — plant N of 1% removes 100 kg of N ha-1, but delaying harvest until after frost reduces whole-plant N to 0.6%, resulting in the removal of only 60 kg of N ha-1 (Mitchell et al. 2012b). Few studies have quantified the nutrient removal associated with growing switchgrass for bioenergy. In the Pacific Northwest under irrigation, Collins et al. (2008) reported that each kg of N produced 83 kg of biomass and the macronutrient export averaged 214 kg N ha-1, 40 kg P ha-1, 350 kg K ha-1, 15 kg S ha-1, 60 kg Ca ha-1, 38 kg Mg ha-1, and 6 kg Fe ha-1. Averaged across cultivars, switchgrass removed less than 1 kg ha-1 of B, Mn, Cu, and Zn. In southern Oklahoma, biomass yields of switchgrass averaged 17.8 Mg ha-1 and removed 40 to 75 kg N ha-1, 5 to 12 kg P ha-1, and 44 to 110 kg K ha-1, an indication of its utility as a low-input bioenergy crop (Kering et al. 2012). In the Northeast USA, delaying harvest until spring reduced ash content and leached nutrients from the vegetation (Adler et al. 2006). An evaluation of switchgrass harvest and storage management was published recently and covers this topic in more detail (Mitchell and Schmer 2012). Additional research is needed to match harvest date, nutrient removal, nutrient composition, and conversion platform to optimize nutrient management and limit over-fertilization and unnecessary nutrient contaminants in the feedstock production stream. Chapter 2 contains a more detailed discussion on switchgrass agronomics.