2.1.1 Regeneration of plants

Sugarcane has a well-established history of in vitro regeneration that began in the late 1960s (Heinz and Mee, 1969). Regeneration of sugarcane can occur via several different pathways (Figure 3) (reviewed by Snyman, 2004; Lakshmanan, 2006; Lakshmanan et al., 2005). The pathways of regeneration by somatic embryogenesis or organogenesis have been well characterized and documented (Ho and Vasil, 1983; Guiderdoni and Demarly, 1988; Taylor et al., 1992). For the regeneration of transgenic plants in vitro, the route of morphogenesis is dependent on the explant targeted for DNA delivery. Criteria for explant choice are: (i) a large number of regenerable cells and (ii) maintenance of regenerative capacity during the selection procedure.

Where embryogenic callus is used as the recipient material for foreign DNA, regeneration is via somatic embryogenesis (Bower et al., 1996; Falco et al., 2000) or organogenesis (Gallo-Meagher and Irvine, 1996). Although transgenic plants have been successfully generated via indirect morphogenesis, limitations include the amount of time taken to regenerate a transgenic plant (36 weeks from DNA delivery to glasshouse transfer) (Bower et al., 1996; Snyman et al., 2000) and the incidence of somaclonal variation as evidenced in agronomic variability when plants are evaluated in the field (Grof and Campbell, 2001; Vickers et al., 2005b).

Alternative tissue targets for transgene delivery have been sought, such as exposed apical meristems from axillary buds followed by shoot morphogenesis (Gambley et al., 1993, 1994; Manickavasagam et al., 2004), leaf roll discs (Snyman et al., 2000, 2001) and pre-emergent inflorescences (Snyman et al., 2006) followed by direct somatic embryogenesis. Although these methods have resulted in transformed plants in a reduced time frame compared to protocols employing indirect embryogenesis, determination of phenotypic fidelity in the field has yet to be published.

2.1.2 Transformation methods

Initial work on sugarcane transformation targeted protoplasts as recipient cells and DNA delivery was via polyethylene glycol (PEG) (Chen et al., 1987) or electroporation (Rathus and Birch, 1992). However, neither technique resulted in the production of transgenic plants, as sugarcane regeneration from protoplasts is notoriously difficult. A significant milestone in sugarcane transformation was achieved when particle bombardment of embryogenic callus was reported and a protocol for recovery of transgenic plants was described (Bower and Birch, 1992). Embryogenic callus proved to be a reliable source of tissue for further refinement of microprojectile bombardment protocols and for the production of transgenic plants in several parts of the world (Bower et al., 1996; Gallo-Meagher and Irvine, 1996; Snyman et al., 1996; Joyce et al., 1998; Ingelbrecht et al., 1999; Falco et al., 2000). In addition, Agrobacterium-mediated DNA delivery was also developed using embryogenic callus as recipient cells (Arencibia et al., 1998; Elliott et al., 1998; Enriquez-Obregon et al., 1998; Liu et al., 2003).

Microprojectile bombardment using a gene gun has been achieved using either the particle inflow gun (PIG) (Finer et al., 1992) or a BioRad device (Heiser, 1993). The principle of this transformation method is to coat either tungsten or gold particles with plasmid DNA containing the gene(s) of interest and to bombard these particles into embryogenic sugarcane callus using pressurized helium combined with a partial vacuum chamber. The first report of successful transformation of sugarcane suspension culture cells using the PIG gun was by Birch and Franks 1991 using the GUS (β-glucuronidase) reporter gene and kanamycin selectable marker gene, although no plants were regenerated. This approach was further refined by Bower and Birch 1992 and is now widely adopted for the production of transgenic sugarcane (Bower et al., 1996; Gallo-Meagher and Irvine, 1993, 1996; Franks and Birch, 1992; Snyman et al., 1996; Joyce et al., 1998; Ingelbrecht et al., 1999; Falco et al., 2000). Transformation efficiencies vary depending on genotype, quality of embryogenic callus, length of time spent in vitro, and selection regime employed, but most laboratories have developed a protocol that suits their requirements.

The desire to minimize transgene copy number and integration complexity, coupled with advances reported in Agrobacterium-mediated transformation in other monocotyledonous crops (reviews by Cheng et al., 2004; Shrawat and Lorz, 2006; Wang and Ge, 2006), made this an appealing system to develop in sugarcane. Although not widely applied yet, Agrobacterium-mediated transformation using embryogenic callus and four different strains of Agrobacterium (Arencibia et al., 1998; Elliott et al., 1998; Enriquez-Obregon et al., 1998; Liu et al., 2003) and axillary buds with two strains (Manickavasagam et al., 2004) has been reported. Although it is difficult to compare transformation efficiencies between the above papers, it appears that pretreatment of the callus prior to Agrobacterium co-cultivation improves transformation efficiency (e.g., a 30min dehydration period). In addition, selection of a particular size of callus (1000μm) (Arencibia et al., 1998), and use of antinecrotic compounds (2mgl^–1 silver nitrate; 15mgl^–1 ascorbic acid (Enriquez-Obregon et al., 1998) were reported to be advantageous. In both papers, Arencibia et al. 1998 and Enriquez-Obregon et al. 1998, the transgene copy number was 1–3. Elliott et al. 1998 applied no selection pressure for up to 6 weeks after co-cultivation. After this period, GFP (green fluorescent protein)-positive calli were visually selected and transferred to bialaphos at 1mgl^–1. Two transgenic plants were regenerated that contained three and seven copies of the GFP gene, respectively.

More recently, transgenic plants were generated using axillary meristems of sugarcane (Manickavasagam et al., 2004). This novel approach resulted in a transformation efficiency of 49.6%. They used two strains of Agrobacterium, LBA4404 and EHA105. The plasmid pGA492 contained the nptII (neomycin phosphotransferase) gene driven by the nos (nopaline synthase) promoter and the bar and gus genes driven by cauliflower mosaic virus (CaMV) 35S promoter. Selection was on 5mgl^–1 bialaphos, applied immediately after the co-cultivation period. Transformation efficiency was greatest in both strains when co-cultivation period was for 3 days in the presence of 50μM acetosyringone. Thousands of plants were produced within 5 months. Although chimeric plants were reported, the incidence was eliminated in secondary shoots after five rounds of selection on Basta^®-containing medium. When these plants were transferred to the greenhouse and sprayed with herbicide, a total of 336 plants (clones of 10 independent transformation events) representing 50% of total plants screened, displayed herbicide resistance. Southern blot analysis on a small subset of these plants further confirmed the presence of 1–2 copies of the transgene in each plant.

2.1.3 Selection of transformed tissues

Studies on transgenic sugarcane for technology development have involved marker and reporter genes. The reporter genes used include GUS (Jefferson et al., 1987; Gnanasambandam and Birch, 2004; Braithwaite et al., 2004), luciferase (luc) (Mudge et al., 1996b; Gnanasambandam and Birch, 2004), maize anthocyanin regulatory elements (R and C1) (ANT) (Ludwig et al., 1990; Bower et al., 1996; Gnanasambandam and Birch, 2004), and GFP (Elliott et al., 1998; Gnanasambandam and Birch, 2004; Gnanasambandam et al., 2007; Petrasovits et al., 2007). Comparative studies by Bower et al. 1996 on the suitability of GUS, luc, and ANT as reporter genes for transient assays in sugarcane transformation indicated that the ANT system is the most suitable reporter system. ANT expression is visible within 8h after bombardment and steadily increases in intensity up to 48h after bombardment. The expressing cells remained visible in the target tissue for 2–3 weeks, before fading or being overgrown. In addition, the results were not confounded by background ANT activity. In contrast, the main disadvantage of using the GUS reporter gene is that the conditions for detection of gene activity are lethal to plant cells and therefore the transformed event is subsequently lost. Detection of both luc and GFP genes require specialized camera and/or detection systems. In addition, GFP detection is confounded by autofluorescence of the callus and chlorophyll in green plant tissues. However, the power of confocal laser scanning microscopy allows precise visualization of fluorescent signals within a narrow plane of focus, and the reconstruction of three-dimensional structures from serial optical sections (Haseloff et al., 1997). This is an advantage of GFP over both GUS and luc (Gnanasambandam and Birch, 2004).

The most widely adopted stable-integration antibiotic marker used for sugarcane transformation is the aphA2 Escherichia coli Tn5-derived nptII. The first successful application of the nptII-based selection system was reported as a step-wise incremental procedure using geneticin (Bower and Birch, 1992). This formed the foundation for subsequent protocols utilizing geneticin (also known as G418) (Bower et al., 1996; Falco et al., 2000; Snyman, 2004) or paromomycin (Joyce et al., 2006a) as the selective agent (Table 1). Less widely used is the hygromycin phosphotransferase (hph) gene with selection on 20mgl^–1 hygromycin (Arencibia et al., 1998; Carmona et al., 2005).

The first paper reporting herbicide resistance in sugarcane also described a selection procedure incorporating a herbicidal agent, bialaphos (Gallo-Meagher and Irvine, 1996). The bar and pat genes from Streptomyces hygroscopicus and Streptomyces viridochromogenes, respectively, encode the phosphinothricin acetyltransferase enzyme that leads to detoxification of phosphinothricin and its derivatives that are ingredients in some commercial herbicides. Selection using herbicides has been used widely, although each laboratory has to determine empirical regimes for selection of different explants and genotypes and formulations of active ingredients differ (Gallo-Meagher and Irvine, 1996; Enriquez-Obregon et al., 1998; Manickavasagam et al., 2004) (Table 1).

The use of the above genes, in addition to selection of transformed cells and plants, has been beneficial in generating information about transgene expression and stability in transgenic sugarcane (Gallo-Meagher and Irvine, 1996; Enriquez-Obregon et al., 1998; Leibbrandt and Snyman, 2003).

There is limited published work on the use of positive selection systems, such as the E. coli manA phosphomannose isomerase (PMI) gene, in sugarcane. The ubiquitous plant enzyme hexokinase converts mannose to mannose-6-phosphate (Man-6-P). Man-6-P is toxic to plants, but most plants lack PMI and are inhibited by the accumulation of Man-6-P. PMI catalyzes the reversible interconversion of Man-6-P and fructose-6-phosphate, thereby releasing the Man-6-P inhibition and making mannose available as a carbon source for the plant. This system was first demonstrated for transformation of potato, sugar beet, and maize (Bojsen et al., 1998, 1999) and since then it has been used successfully in a variety of other plant species. PMI has no adverse effects in acute mouse oral-toxicity tests, generates no detectable biochemical changes in mannose-associated pathways (Privalle et al., 1999), lacks many attributes known to be associated with allergens (Privalle, 2002), and may thus be considered as an ideal selection protein for plant transformation. For sugarcane transformation, calli were selected on media containing 3gl^–1 mannose, in addition to the 20gl^–1 sucrose present in the media (Jain et al., 2006). Plant regeneration was performed under the same level of selection. An increase of mannose from 1.5 to 3gl^–1 for rooting improved the overall transformation efficiency. The PMI encoding gene, manA, was stably integrated and expressed in almost all of the transgenic lines.

Of the other nonantibiotic systems tested in sugarcane, glutamate-1-semialdehyde aminotransferase was unsuitable, while selection incorporating arabitol showed up to 100-fold lower transformation efficiency. In addition, arabitol is expensive, making it a less desirable selection agent.

2.1.4 Promoters and termination sequences

The two main sources for promoters are from microorganisms (viral or bacterial) or from plants. In addition, there are a few synthetic promoters that have been constructed in the laboratory by combining regulatory and/or enhancer elements from different viral or plant promoters. What is lacking in sugarcane, and in monocots in general, is a promoter that functions as strongly as the CaMV 35S promoter does in dicots. Schledzewski and Mendel 1994 compared transient reporter gene expression in transgenic cells of barley, maize, and tobacco driven by maize polyubiquitin1 (Ubi-1), rice actin1, Emu, or CaMV 35S promoter. CaMV 35S promoter had the highest GUS activity in tobacco (316.71nmolh^–1) compared to maize (9.88nmolh^–1) and barley (1.22nmolh^–1). In contrast, of the four promoters tested, the Ubi-1 promoter showed highest expression in both maize and barley cells, but not in tobacco. However, CaMV 35S in tobacco showed a promoter strength equivalent to 2.5- to 5-fold greater than Ubi-1 in maize (Table 2).

2.1.5 Activity, stability of inheritance, and silencing of transgenes

The high level of ploidy found in all sugarcane may have some important implications for transformation, which are not normally encountered when working with diploid species. For instance, increase in ploidy by genome duplication, which has been suggested to play an important role in angiosperm evolution (Stephens, 1951; Ohno, 1970; Blanc et al., 2000; Initiative, the Arabidopsis Genome, 2000; Paterson et al., 2000, 2003, 2004, 2005; Bowers et al., 2003; Vandepoele et al., 2003; Wang et al., 2005b), is associated with a host of rapid responses, including loss and restructuring of low-copy DNA sequences (Song et al., 1995; Feldman et al., 1997; Ozkan et al., 2001; Shaked et al., 2001; Kashkush et al., 2002), activation of genes and retrotransposons (O'Neill et al., 1998; Kashkush et al., 2003), gene silencing (Chen and Pikaard, 1997a, b; Comai, 2000; Comai et al., 2000; Lee and Chen, 2001), and organ-specific subfunctionalization of gene expression patterns (Adams et al., 2003, 2004). Sugarcane appears to have undergone two such genome duplications in about the past 5 million years and these mechanisms may be especially important in providing raw material for evolutionary change. It is important to understand how stable the sugarcane genome is with respect to: (i) transposon activity; (ii) genome expansion or contraction; (iii) measuring the extent of gene silencing and alterations in gene expression; (iv) assessing in vitro regeneration systems for epigenetic variation in regenerated plantlets; and (v) designing transformation technologies to overcome the above constraints.

Molecular techniques for detection and characterization of transgenes with respect to integration pattern, copy number, and protein concentrations have been conducted and published (Bower et al., 1996; Gallo-Meagher and Irvine, 1996). However, these analyses were carried out on sugarcane plants maintained in glasshouses. Field analysis to determine expression and stability of transgenes is important because of sugarcane's multiple vegetative crop cycles and reports of PTGS.

The transgenesis approach in sugarcane has targeted elite commercial cultivars. It would be advantageous to be able to use transgenic plants as parents in breeding programs especially when the desired trait is not present in the sugarcane gene pool. Transgene inheritance and segregation of the bar herbicide resistance gene and the hut Sorghum mosaic virus (SrMV) coat protein gene was tracked in progeny arising from conventional crosses made between transgenic and nontransgenic parents (Butterfield et al., 2002). The results demonstrated that transgenic plants can be used as parents in a sugarcane breeding program but screening progeny for a characteristic such as virus resistance that relies on PTGS mechanisms may have to be carried out after one vegetative cycle of growth after crossing.

PTGS is known to occur in sugarcane (Ingelbrecht et al., 1999) and is considered to be one of the factors limiting accumulation of recombinant proteins (Wei et al., 2003). It was thought that it could be reversed by retransformation with a gene encoding a viral suppressor of PTGS, such as HcPro from SrMV (Ingelbrecht et al., 2000) and P0 from sugarcane yellow leaf virus (ScYLV) (Wang et al., 2006, 2007). Despite the potential application in transgene silencing control, there are possible negative effects on endogenous gene regulation and virus susceptibility. Initial results indicate that there is no consistent correlation between the RNA expression levels of P0 or HcPro and the expression of a transgene and several miRNA-regulated endogenous genes. Further research is required to characterize the effect of these viral suppressors in transgenic sugarcane (Wang et al., 2006, 2007).

Sugarcane plants transformed via particle bombardment with GM-CSF (human cytokine granulocyte macrophage colony stimulating factor), contain numerous copies of transgenes with complex integration patterns as revealed by Southern blot hybridization (Albert et al., 2003; Wang et al., 2003). Multiple approaches have been evaluated for the introduction of single- or low-copy transgenes, to determine if these methods can reduce transgene silencing (Albert et al., 2004). These methods include the use of insert-only DNA for bombardment (Fu et al., 2000), Cre/lox site-specific recombination to resolve multiple transgene copies (Srivastava and Ow, 2001), and the Ac/Ds transposon system to direct transgene integration by transposition (Koprek et al., 2000, 2001). The use of linear expression cassette-only DNA was reported to produce a high frequency of low-copy transgene insertions in rice, with no evidence of silencing through the R₄ generation (Fu et al., 2000). In contrast, most transgenic sugarcane lines produced by this method contained multiple transgene copies, with only three of 27 selected lines containing three or fewer copies (Wang et al., 2003). Cre/lox lines did contain fewer copies of the transgene as estimated by quantitative PCR (qPCR). However, accumulation of GM-CSF in the low-copy lines was not higher than in multicopy lines (Albert et al., 2004). In addition, a vector system was developed to allow direct plasmid to chromosome transposition using Ac/Ds in monocot cells. This should allow single-copy transposition lines to be produced in a single generation without sexual crosses (Albert et al., 2003). However, no increase in GM-CSF protein level was observed in the Ac/Ds lines.

2.1.6 Adverse effects on growth, yield, and quality

Most evaluation of transgene stability and expression patterns and performance of transgenic sugarcane plants in the field has been limited to a few lines (Gallo-Meagher and Irvine, 1996; Arencibia et al., 1999; Leibbrandt and Snyman, 2003). However, results of recent studies where a larger number of lines were tested (Gilbert et al., 2005; Vickers et al., 2005b) may influence future transgenic approaches.

Initial field trials demonstrated stable expression of a herbicide-resistant transgene over three rounds of vegetative propagation in a single transformant (Gallo-Meagher and Irvine, 1996; Leibbrandt and Snyman, 2003), but no agronomic measurements were taken in the first study. In the latter trial, no differences were found between the transgenic line and wild-type control in phenotypic characters such as stalk height and diameter, agronomic performance indicators such as sucrose yield and fiber content, and disease susceptibility ratings to smut and rust. However, in a field trial where 100 transgenic lines were compared for Sugarcane mosaic virus (SCMV) resistance and yield characteristics, a considerable amount of variability for measured parameters was reported, which the authors attributed largely to the effects of the cultivar used and the tissue culture process (Gilbert et al., 2005; Vickers et al., 2005b). Similarly, preliminary data from metabolomic analysis, comparing leaves from nontransformed sugarcane with leaves from transgenic sugarcane lines producing polyhydroxybutyrate (PHB), found that the vast majority of the variation was a tissue culture effect and was not from the insertion of the PHB metabolic pathway and the selectable marker genes (Purnell et al., 2007).

2.1.7 Subcellular targeting

Most transgene expression constructs used for sugarcane transformation result in production of foreign gene products in the cytosol, but important metabolic activities ranging from photosynthesis to sugar storage are carried out in other compartments. Hence, targeting proteins to subcellular compartments may be necessary for effective resistance to pest and diseases and efficient metabolic engineering in sugarcane. For example, for increased resistance to sugarcane leaf-scald disease, resistance gene products may need to be targeted to the plastids, as the albicidin toxin from the pathogen Xanthomonas albilineans blocks plastid DNA replication and chloroplast development. Similarly, for efficient conversion of sucrose to alternative carbohydrates such as starch and fructans, foreign proteins need to be targeted to the sugarcane vacuoles in mature stem parenchyma cells, where most of the sucrose is stored.

The ability to target recombinant proteins to the correct subcellular location using efficient and appropriate targeting signals is one of the most important requirements for sugarcane metabolic engineering. Various targeting signals have been successfully tested to target heterologous proteins to different subcellular compartments in sugarcane. The availability of visible reporters such as GFP in combination with a transient assay system in sugarcane leaves allowed the testing of the efficiency of various targeting signals with less time and resources (Gnanasambandam et al., 2007). While most of the tested signal sequences are from dicotyledons species, they were effective in the monocotyledon sugarcane. So far, targeting signals for vacuoles, endoplasmic reticulum (ER), plastids, mitochondria, and peroxisomes have been tested in sugarcane (McQualter et al., 2005; Petrasovits et al., 2007; Brumbley et al., 2006b; Gnanasambandam et al., 2007). In future, signals for other compartments and efficiency of signals will be determined.

2.1.7.2 Protein targeting to plastids

The N-terminal plastid transit peptides of RbcS (rubisco small subunit) genes from maize and pea were used to target heterologous enzymes to sugarcane plastids to successfully produce p-hydrobenzoic acid (pHBA; McQualter et al., 2005) and polyhydroxybutyric acid (PHB biopolymer; Petrasovits et al., 2007). The transit peptides from tomato DCL (defective chloroplast and leaves) and tobacco RbcS genes were shown to target GFP to the sugarcane leaf proplastids (Gnanasambandam et al., 2007; Figure 4).

2.2.1 Disease resistance

Reports of pathogen-derived resistance to virus diseases include resistance to SCMV in otherwise susceptible sugarcane clones (Joyce et al., 1998; Ingelbrecht et al., 1999). Both groups used the coat protein gene of the virus driven by the Ubi-1 promoter with the nos terminator. Plants transformed with the coat protein gene of the SrMV strain SCH displayed a range of phenotypes, including immune, resistant, recovery, and susceptible plants, when challenged with virus. These observations enabled insights to the RNA-mediated PTGS resistance mechanism (Ingelbrecht et al., 1999). Virus induced gene silencing (VIGS) has been suggested as the mechanism of resistance against viral infections. Interestingly, transgenic plants having the same transgene integration pattern as determined by Southern blot analysis (i.e., clones) displayed a different response to mechanical SCMV infection. The reason for this peculiar result is still not clear (Ingelbrecht et al., 1999).

Sugarcane yellow leaf disease, which is characterized by yellowing of the leaf midribs followed by tissue necrosis, is caused by ScYLV (Borth et al., 1994; Schenck et al., 1997; Vega et al., 1997; Comstock et al., 1998). Economic losses from sugarcane yellow leaf disease of up to 50% have been reported (Vega et al., 1997). The ScYLV coat protein gene in the sense orientation driven by Ubi-1 was used to generate virus-resistant transgenic sugarcane. Resistance levels were evaluated with inoculation of viruliferous aphids. Virus titers were determined using tissue blotting and quantitative reverse-transcriptase polymerase chain reaction (qRT-PCR). Two resistant transgenic sugarcane lines were identified based on the tissue blot analyses. However, virus RNA can still be detected using qRT-PCR in these lines. Greenhouse tests are currently being conducted to compare the yield difference between nontransformed and transgenic lines.

The transgenic approach for providing resistance to the more devastating viral disease in sugarcane, Fiji leaf gall has been tested (McQualter et al., 2004). Resistance to Fiji leaf gall was produced by microprojectile-mediated transformation with a transgene encoding a translatable version of Fiji disease virus (FDV) segment 9 ORF 1 under the control of the maize Ubi-1 promoter. The molecular phenotypes of the transgenic plants at both the DNA and RNA levels were not entirely consistent with a resistance mechanism based on PTGS. Transgenic plants showed very low steady state messenger-RNA (mRNA) levels under normal conditions, but many of these plants failed to show resistance upon challenge with the FDV virus, suggesting that the virus possessed a mechanism for overriding the post-transcriptional silencing mechanism. Further research required to achieve complete immunity to FDV includes: (i) additional characterization of the FDV genome, since only a superficial knowledge of the virus life cycle and replication strategy is currently known; (ii) a shotgun approach where all 12 ORFs contained within the FDV genome are used as transgenes, either singly or in various combinations (RNAi (RNA interference) silencing constructs should be employed in this case); and (iii) determining if FDV possesses a dsRNA (double-stranded RNA) binding protein with the ability to suppress gene silencing. If so, then a more effective approach to achieve pathogen derived resistance to FDV in sugarcane might require deactivation of this protein.

Leaf scald is a serious disease of sugarcane caused by the bacterium Xanthomonas albilineans. The bacterium produces a toxin (albicidin) that blocks plastid DNA replication of the sugarcane plant. Zhang et al. 1999 identified a bacterium that could survive in the presence of albicidin because it carried an albicidin detoxification gene (albD). This gene was cloned from the bacterium and introduced into sugarcane. Some of the resulting transgenic plants were resistant to leaf scald disease (Hansom et al., 1999). It is interesting to note that another bacterial pathogen of sugarcane Leifsonia xyli subsp. xyli (Lxx) like Xanthomonas albilineans colonizes the xylem vessels, was recently shown to have a gene homologous to the one in Xanthomonas albilineans encoding an enzyme to pump albicidin out of its cells (Monteiro-Vitorello et al., 2004).

2.2.2 Insect resistance

The potential for using a modified Bt gene encoding the δ-endotoxin CryIA(c) from Bacillus thuringiensis strain kurstaki to combat damage by the lesser corn-stalk borer (LCB) was demonstrated by Fitch et al. 1996. The Bt gene and the selectable marker gene nptII, were under the control of the CaMV 35S promoter. Bombarded sugarcane calli were selected stepwise on 50–200mgl^–1 of G418, an aminoglycoside antibiotic similar in structure to gentamicin B1. Insect bioassays indicated that LCB larvae fed on some lines of calli or leaves from the regenerated plants weighed less and showed higher mortality than those fed on nontransgenic tissues. Mortality was higher when LCB were fed transgenic calli as opposed to transgenic leaves, possibly due to different levels of the CryIA(c) protein in the two different tissues, although the protein levels were not reported. Leaves from other putatively transformed lines had no detectable effect on larvae survival. Southern hybridization indicated that there were 1–5 copies of the Bt genes in the two most resistant lines.

Resistance to the other lepidopteran sugarcane stalk borers has also been reported using the δ-endotoxin gene from B. thuringiensis. Larval mortality and reduced levels of damage from Diatraea saccharalis in the field and Proceras venosatus in the glasshouse have been reported (Arencibia et al., 1997; Weng et al., 2006). In the short term, it is possible that intellectual property restrictions may limit widespread use of this technology in sugarcane. Consequently, the use of other antimetabolic compounds for improving plant resistance, such as proteinase inhibitors (Allsopp et al., 1996; Falco and Silva-Filho, 2003) and lectins (Sétamou et al., 2002a, 2002b, 2002c), has been tested with promising results from laboratory-based insect bioassays. These products impact on a wider range of insect/pests and studies on the effects on the Australian coleopteran whitegrub have been underway for several years (Allsopp et al., 1996).

Canegrubs are a major pest in the Australian sugar industry causing yield losses up to AU $80 million. The PinII and the snowdrop lectin (Gna) genes have both been used to generate transgenic sugarcane in attempts to control these pests (Allsopp et al., 1996; Nutt et al., 1999). Plants containing the pinII gene grew more slowly than the nontransformed control plants. This may have been due to metabolic disruption within the plants, as no pinII was found in the cell vacuole. Earlier work with artificial feeding trials had shown that avidin could reduce larval growth and increase larval mortality of Antitrogus parvulus (Childers canegrub) (Allsopp and McGhie, 1996). More recently, the gene for avidin, a biotin binding protein from chicken egg white, has been introduced into sugarcane for the control of canegrubs (Nutt et al., 2006). Transgenic sugarcane plants have been regenerated, which contain avidin concentrations of up to 0.06% of total protein.

2.2.3 Sucrose metabolism

A suite of physiological processes and enzymes involved in sucrose accumulation have been identified and characterized over the last 50 years. These processes include: leaf reactions, such as photosynthetic reactions, sucrose synthesis, metabolism and carbon partitioning across various membranes into different pools; phloem reactions, such as phloem loading in leaf, translocation to and unloading in various sink tissues (including primary storage in parenchyma cells of the stalk); stalk reactions, such as membrane transport, sucrose metabolism, carbon partitioning, and remobilization of stored sucrose; genetic and developmental controls, such as timing of maturation; and environmental perception and signal transduction pathways to coordinate plant development (Moore, 2005). Some of the genes encoding these enzymes have been cloned and used to transform sugarcane with the goal of altering sucrose accumulation (reviewed by Grof and Campbell, 2001). However, this reductionist approach has fallen short of expectation in almost all of the cases because of the complexity among the multitude of simultaneous processes and parallel pathways.

As tissue culture methods became established, they were exploited to elucidate the physiology and biochemistry of sugarcane carbohydrate accumulation, with particular emphasis on sugars (Komor et al., 1981; Thom and Komor, 1984). Transgenic sugarcane cell lines were also used to study the effect of invertase expression in different cellular compartments on sucrose accumulation (Ma et al., 2000). Overexpression of a yeast invertase gene (SUC2) in the apoplast led to rapid hydrolysis of sucrose and accumulation of hexoses, both in the medium and the cells, suggesting that hexose uptake, not hexose availability, was the limiting factor for sucrose accumulation. Cells transformed for overexpression of invertase in the cytoplasm did not show a significant change in the sugar composition in the medium, but did significantly reduce the sucrose content in the cells. Partial inhibition of the soluble invertase activity was achieved by transforming with a sugarcane soluble acid invertase complementary DNA (cDNA) (SCINVm) in the antisense orientation to result in increased sucrose accumulation. Intra- and extra-cellular sugar composition was very sensitive to changes in invertase activities in this tissue culture system (Ma et al., 2000).

Pyrophosphate-dependent phosphofructokinase (PFP) activity in sugarcane is inversely correlated to sucrose concentration in maturing internodal tissues, but no clear physiological role in sucrose metabolism has emerged (Whittaker and Botha, 1999). If endogenous PFP activity were to be down-regulated by antisense or co-suppression technologies, then sucrose concentration could be increased. In sugarcane transformed with the catalytic subunit, PFP-β, endogenous PFP gene expression was reduced by up to 40% and 80% in leaf roll and internodal tissue, respectively (Groenewald and Botha, 2001, 2008). Sucrose concentrations in these lines were significantly increased in immature internodes. This finding could make a valuable contribution to the productivity of sugarcane cultivars and elucidate the role of PFP in sucrose accumulation.

Vickers et al. 2005a attempted to modify the endogenous polyphenol oxidase (PPO) activity in sugarcane by introducing sense and antisense constructs of the endogenous sugarcane PPO gene driven by the Ubi-1 promoter. The rationale behind this approach was that the inhibition of PPO activity in juice by chemical inhibitors, elevated pH or heat significantly reduced the color of the cane juice and the subsequent color intensity in sugar crystals. All transgenic lines, irrespective of the orientation of the PPO gene, showed higher PPO activity and more color units than the nontransformed commercial clones. Although higher levels of PPO activity were correlated to juice with a darker color, the converse was not true in this study. It has been hypothesized that a lowering of PPO activity will lead to a reduction in the color of juice and consequently raw sugar (Vickers et al., 2005a). As pale sugar has a market premium over dark sugar, this has potential economic benefits for the sugar industry.

While most transgenic attempts to improve sugar accumulation in sugarcane have met with limited success, a notable exception was the introduction of a bacterial gene encoding a sucrose isomerase (Wu and Birch, 2007) to produce a high-value sugar isomaltulose.

2.2.4 Alternative products

High-value sugars. Isomaltulose is a sucrose isomer with α-1,6 linkages instead of α-1,2. Although it is only 42% as sweet as sucrose (Li et al., 2004), it is a high-value sugar because it is digested 4–5 times slower than sucrose and so has major health benefits to consumers (Lina et al., 2002). Not only does isomaltulose reduce the highs and lows in blood sugar (Lina et al., 2002), it also has dental health benefits (Ooshima et al., 1983). Wu and Birch 2007 demonstrated that sugarcane can be engineered to target the production of isomaltulose in the storage vacuoles in sugarcane stalks (see Section 2.1.7). In transgenic lines, when sucrose isomerase was targeted to the vacuole, isomaltulose accumulated without a reduction in sucrose accumulation. Up to a twofold increase in total sugar accumulation was observed in some lines (Wu and Birch, 2007). The additional sugar was not produced at the expense of structural carbohydrates, as levels of fiber were the same between wild type and transgenic lines. Additional benefits were that the levels of isomaltulose reduced leaf senescence resulting in more leaf biomass on the plants and a reduction in the percent water in stalk tissue from 70% for the nontransformed control mature stalk internode to 60% in the best isomaltulose producing transgenic line (Wu and Birch, 2007). Field trials of these high-sugar lines are currently underway.

Sorbitol. The primary photosynthate of the members of the Rosaceae family including apples (Malus domestica), pears (Pyrus sp.), peaches and nectarines (Prunus persica), plums (Prunus subgenus Prunus), cherries (Prunus subgenus Cerasus) is sorbitol. Sorbitol has intrinsic value as a noncaloric sweetener and is also used to manufacture ascorbic acid and personal care products (Kirschner, 2004). Sorbitol synthesis in sugarcane is achieved by a single enzyme conversion step sorbitol-6-phosphate dehydrogenase (S6PDH), which catalyzes the reduction of D-glucose-6-phosphate (G6P) to sorbitol-6-phosphate (S6P). The cytosolic expression of the M. domestica sorbitol-6-phosphate dehydrogenase gene (mds6pdh) resulted in high accumulation of sorbitol (up 61% of the soluble sugars or 12% of the leaf dry weight) in sugarcane leaves but with 10-fold less sorbitol in the culm (Chong et al., 2007; Brumbley et al., 2004). Sugarcane leaves developed necrosis in a pattern characteristic of early senescence and the severity was related to the relative quantity of sorbitol accumulated. Vacuolar targeting of the same enzyme, while not attempted, may lead to normal plants and higher accumulation of sorbitol in the culm as observed for isomaltulose (Wu and Birch, 2007). However, protein stability within the vacuole will continue to be a major challenge for efficient metabolic engineering of the sugar storage compartment in sugarcane. A possible solution would be to modify the enzymes, possibly by using directed evolution (Stemmer, 1994a, 1994b; Chica et al., 2005; Kaur and Sharma, 2006) to function optimally at the pH and to be more resistant to the protease activity in the environment of the storage vacuole.

In addition to the above mentioned sugars, trehalose, a nonreducing disaccharide of glucose, has also been produced in sugarcane (Brumbley et al., 2006b; O'Neill et al., 2006; Zhang et al., 2006). Trehalose may play a role in carbohydrate metabolism impacting on glucose, fructose, and sucrose levels (Rontein et al., 2002).

Biopolymers. To test the ability of sugarcane to be a biofactory, the products from the Ralstonia eutropha PHB biosynthetic pathway were targeted to several subcellular compartments of sugarcane (Brumbley et al., 2002, 2004, 2007; Petrasovits et al., 2007; Purnell et al., 2007). PHB is the best-studied member of the polyhydroxyalkanoate (PHA) family, of which approximately 130 naturally occurring members have been identified (see Section 3). PHB is produced from acetyl-coenzyme A (acetyl-CoA) by the successive action of three enzymes [ketothiolase (PHAA), acetoacetyl-reductase (PHAB), and PHB synthase (PHAC)], which are encoded by the genes phaA, phaB, and phaC, respectively (Peoples and Sinskey, 1989a, 1989b). Each gene was on a separate transformation vector and gene expression was controlled by Ubi-1 promoter and nos terminator sequences. The three vectors with the PHB biosynthetic pathway were biolistically transformed into sugarcane callus simultaneously along with a construct containing the selectable marker gene nptII (Petrasovits, 2005; Petrasovits et al., 2007).

Previous attempts to produce PHB at high levels in plants resulted in either low levels of PHB accumulation or severe negative phenotypic effects. In sugarcane, the polymer accumulated in the leaves of chloroplast-targeted lines at levels up to 2.5% of dry weight and 0.01% in stems (Figure 5) (Petrasovits et al., 2007). Purnell et al. 2007 conducted a replicated glasshouse trial using a random block design with six independent PHB producing transgenic sugarcane lines and found that stalk height and weight and sugar levels were not affected by PHB accumulation.

Sugarcane has also been evaluated as a production platform for p-hydroxybenzoic acid (pHBA) using two different bacterial proteins (a chloroplast-targeted version of E. coli chorismate pyruvate-lyase and a 4-hydroxycinnamoyl-CoA hydratase/lyase from Pseudomonas fluorescens) (McQualter et al., 2005; Brumbley et al., 2004). Both lines provide a one-enzyme pathway from different naturally occurring plant intermediates. The substrates for these enzymes are chorismate (a shikimate-pathway intermediate that is synthesized in plastids) and 4-hydroxycinnamoyl-CoA (a cytosolic phenylpropanoid intermediate). Although both proteins have previously been shown to elevate pHBA levels in plants (Siebert et al., 1996; Mayer et al., 2001), they had never been evaluated concurrently nor had they been used simultaneously in the same plant. The pHBA was quantitatively converted to glucose conjugates by endogenous uridine diphosphate (UDP)-glucosyltransferases and was stored in the vacuole. The largest amounts detected in leaf and stem tissue were 7.3% and 1.5% of dry weight (DW), respectively, and there were no observable phenotypic abnormalities. However, as a result of diverting carbon away from the phenylpropanoid pathway, there was a reduction in leaf chlorogenic acid, subtle changes in lignin composition, and an apparent compensatory up-regulation of phenylalanine ammonia-lyase (PAL; McQualter et al., 2005).

2.2.5 High-value protein production

GM-CSF, which is used in clinical applications for treatment of neutropenia and aplastic anemia, was used to test the feasibility of using sugarcane as a biofactory for high value protein production (Wang et al., 2005a). Two promoters, Ubi-1 and sugarcane ubi9 were tested in transgenic sugarcane lines resulting in production of up to 0.02% total soluble protein as GM-CSF. No significant difference was observed between the Ubi-1 or ubi9 promoter line. To achieve higher accumulation of GM-CSF in sugarcane, a C terminal HDEL tag for ER targeting had to be added to the gene construct. The sugarcane-produced GM-CSF showed identical biological activities when compared to commercially available purified protein. In a 14-month old field trial, no abnormal phenotype was observed and the accumulation levels remained relatively stable. This is the first report of GM-CSF production in field-grown plants. No flowering of the trial plants occurred and no pollen or seed was produced during the trial period. Drying, burning, and burial of the test plants effectively blocked possible routes for the transgenic sugarcane to enter the food supply or environment. If GM-CSF accumulation in sugarcane had been as high as Sétamou et al. 2002c reported for the snowdrop lectin (1–1.25% total soluble protein), commercial production would probably have been economically viable because of the high commercial value of this protein.

Sugarcane has also been used for the production of bulk protein, collagen. In vertebrates, the collagen family is made up of approximately 27 proteins (Myllyharju et al., 2000; Kielty and Grant, 2002; Gelse et al., 2003; Myllyharju and Kivirikko, 2004; Pakkanen et al., 2006). Collagens make up 25% of the total proteins in humans and function to act as connecting structures and to give mechanical stability to the entire body. They form triple-helical structures, giving them commercial importance for the food industry and for a range of medical applications, such as tissue repair and cosmetic surgery (Bulleid et al., 2000). Gelatin is denatured collagen. The yeast Pichia pastoris has been engineered successfully for the high level production of triple-helical collagen (Pakkanen et al., 2006). Production has also been attempted in sugarcane and a fragment of human collagen was produced at very low levels (0.025% total soluble protein) (E. Mirkov, personal communication).

2.2.6 Abiotic stress

One strategy for generating abiotic stress-resistant plants has been to express genes for the production of trehalose. In planta trehalose production confers stress tolerance in plants (Garg et al., 2002; reviewed in Penna, 2003; Rolland et al., 2006). Abiotic stress normally causes a reduction in photosynthesis. However, trehalose protects the photosynthetic machinery in the plant cells and allows them to remain functional for longer periods of time under a range of abiotic stress conditions (Garg et al., 2002; Rolland et al., 2006). By using two copies of the CaMV 35S promoter in tandem to drive expression of the Grifola frondosa trehalose synthase gene in S. officinarum, Zhang et al. 2006 achieved trehalose accumulation in sugarcane cells at 0.9–1.2% of fresh weight. This low-level production in transgenic plants has been seen in other plant species and has been attributed to native trehalase activities (Garg et al., 2002; Penna, 2003). The transgenic lines showed no negative effects on growth, were drought tolerant, and produced higher yields under drought conditions.

2.2.7 Flowering

In sugarcane cultivars, the time and intensity of flowering are important as they influence the yield and quality of cane (Premachandran, 2006). Sugarcane yield is reduced by cessation of vegetative growth in flowering canes and profusely flowering clones are undesirable for commercial cultivation, especially when used for late-season crushing.

The identification of genes with major effects on flower induction could lead to new ways to control flowering in sugarcane cultivars (Ulian, 2006). Flower formation depends on the transition of a vegetative meristem to an inflorescence meristem and, subsequently, to a floral meristem. This transition requires the action of a number of genes, including LEAFY (LFY), a gene that is expressed early and considered to be important in the flowering process (Weigel et al., 1992). LFY interacts with another floral control gene, APETALA1 (AP1), to promote the transition from inflorescence to floral meristem. The ectopic expression of LFY induces the ectopic expression of AP1 in leaf and axillary flower primordia (Parcy et al., 1998). LFY and AP1 are pivotal for the switch to the reproductive phase, where, instead of leaves, the shoot apical meristem produces flowers. The AP1 promoter is a direct target of LFY (Wagner et al., 1999). LFY also directly targets the AGAMOUS (AG) promoter that contains a LFY-responsive enhancer necessary for its activity (Busch et al., 1999). AG encodes a transcription factor that regulates genes determining stamen and carpel development in wild-type flowers (Yanofsky et al., 1990). LFY, therefore, plays pivotal roles in the specification of flowers and in the patterning of floral organs (Hempel et al., 2000).

To study flowering in sugarcane, Ulian 2006 analyzed the SUCEST database and found a DNA EST with significant similarity to the Arabidopsis LFY sequence with two regions highly conserved in the LFY family of transcription factors. The putative LFY gene from sugarcane was expressed in the antisense orientation in transgenic sugarcane clone SP87-432 under the control of the constitutive maize Ubi-1 promoter. Ulian 2006 observed that silencing of the LFY gene suppressed flowering in transgenic sugarcane, supporting the notion that the gene plays an important role in sugarcane flower development.

3.1.1 Improved nuclear transformation efficiency

Currently, the biolistic method is the most efficient method for the production of transgenic sugarcane plants (Bower and Birch, 1992). However, transformed plants obtained through this method have several shortcomings, including the presence of numerous copies of transgenes with complex patterns of integration (Albert et al., 2003; Wang et al., 2003). This creates some difficulties in obtaining regulatory approval for transgenic crop plants. Multiple transgene integrations may also inhibit transgene expression, cause gene silencing, and/or promote transgene rearrangements. Additionally, biolistic transformation entails incorporation of the entire plasmid into the plant nuclear genome, which is a cause for public concern. Thus, genes essential for plasmid replication (ori) in bacteria as well as antibiotic resistance (e.g., ampicillin) required for selection of the transformed bacterial colony are co-introduced along with the gene(s) of interest. To minimize these problems, two technologies are currently being tested—Agrobacterium-mediated transformation (see Section 2) and the use of linearized plasmid DNA (LDNA) containing only the promoter-ORF-terminator cassette.

Agrobacterium-mediated transformation generally produces transgenic plants with fewer copies of the transgenes. Agrobacterium-mediated transformation of sugarcane has been reported in a few laboratories (Section 2). Most reports in sugarcane have shown the number of transgene copies to be generally between one and three (see Section 2). This may lead to higher and more stable expression of the transgenes as has been reported in rice (Dai et al., 2001) and maize (Shou et al., 2004) when compared to biolistic transformants. However, the stability of transgene expression using Agrobacterium, compared to biolistic transformation, has yet to be verified in sugarcane. Also, the efficiency of transformation is comparatively lower than that for microprojectile bombardment. Future research must aim to increase the efficiency of this transformation method in sugarcane.

In Agrobacterium-mediated transformation, transfer of genes of interest (present in the T-region of the Ti plasmid of Agrobacterium) into the plant nucleus usually occurs between the right and left T-DNA (transfer DNA) borders of the plasmid (Hellens et al., 2000). Generally, this region only contains the regulatory sequences along with the genes of interest. Thus, the transfer is believed to be precise, with none of the plasmid sequences flanking the border regions being transferred. Recently, however, Shou et al. 2004 analyzed transgenic maize to determine whether vector backbone sequences of the Ti plasmid were also transferred. They showed that in 75% of the R₁ progeny of primary transformants there was some portion of the backbone present. This event may have occurred due to the specific transformation conditions used in the experiment. This needs to be determined when transforming sugarcane with Agrobacterium. Field trials of sugarcane plants transformed with Agrobacterium are proposed to commence in Australia in 2007 to address many of the concerns raised above (P. Joyce, personal communication).

Concerns raised by activists regarding the release of microbial genes into the environment coincidently with the biolistic transformation procedure has led scientists to use LDNA containing only the promoter-ORF-terminator cassette region for plant transformation. This LDNA method has been tested for gene expression and copy number analysis in rice (Fu et al., 2000; Loc et al., 2002). There was a significant reduction in the number of copies integrated into the rice genome. In addition, a low frequency of transgene rearrangements without any deleterious effect on expression pattern was observed (Fu et al., 2000). Moreover, when two minimal linear genes were co-bombarded, the co-transformation efficiency was similar to that using intact circularized plasmids. Similar experiments in sugarcane, however, showed that it behaved differently to rice (Albert et al., 2004; Joyce et al., 2006b).

3.1.2 Plastid transformation

The current nuclear transformation technology has several limitations, including semi-random and unpredictable incorporation of foreign genes into the genome, and the operation of genetic mechanisms that can switch off foreign genes, resulting in widely varying expression of introduced genes in sugarcane (Ingelbrecht et al., 1999; Gilbert et al., 2005; Vickers et al., 2005a, 2005b). A powerful new strategy, plastid transformation, has been developed in some plants, including Arabidopsis, tomato, potato, cotton, carrot, and rice (Bock and Khan, 2004; Koya et al., 2005), with the potential to circumvent these problems by introducing genes into plastid DNA rather than the plant nuclear genome (Maliga, 2002; Daniell et al., 2005b). The plastid transformation strategy utilizes two targeting sequences that flank the foreign genes and that inserts them through homologous recombination at a precise, predetermined location in the chloroplast genome (Maliga, 2004).

Despite its tremendous biotechnological potential, plastid transformation has only been used routinely in tobacco (Maliga, 2004) and attempts to develop plastid transformation in sugarcane have been unsuccessful to date. However, a recent patent issued on plastid transformation in monocots (Daniell, 2006) indicates the potential exists to develop the technology in sugarcane. The complete nucleotide sequence of the chloroplast genome of sugarcane has been determined recently (Asano et al., 2004). The plastome of sugarcane is a circular double-stranded DNA molecule, 141182bp in size, and is composed of a large single copy of 83048bp, a small single copy of 12544bp, and a pair of inverted repeat regions of 22795bp each. The sugarcane chloroplast genome is similar to maize, but not to rice or wheat. The availability of sequences for sugarcane chloroplast genome (Asano et al., 2004) could facilitate the development of plastid transformation technology in sugarcane.

3.1.3 Avoiding antibiotic resistance gene in transgenic plants

3.1.4 Suppressing transgene silencing

There are convincing reports indicating that sugarcane displays a high level of PTGS (Liu et al., 2003; Wei et al., 2003). Strategies to suppress this include the use of viral suppressor proteins (Mangwende et al., 2005), addition of putative matrix attachment regions (MARS) (Wei et al., 2003), and presence of introns. The use of viral suppressors, including P0 from ScYLV and HcPro from SrMV, to advance our knowledge of RNA silencing in sugarcane are being explored (Wang et al., 2006, 2007). More research is needed to understand and suppress transgene silencing.

3.1.5 Tissue-specific promoters

Despite several years of research, an effective stem-specific promoter is not yet available in sugarcane. Though a stem-specific promoter has been reported in sugarcane (Hansom et al., 1999), its usefulness remains to be established. Similarly, other tissue-specific promoters that are operative in sugarcane (for example, root-specific promoters) are not available. Attempts to find new promoters for strong expression especially in the mature parenchyma cells of sugarcane stem are ongoing. Recently, Mudge et al. 2006 reported the isolation of an EST clone (SMS04) that was selected based on its expression pattern. Further analysis revealed that at least eight variants of this gene existed in sugarcane and that no two had the same sequence. The promoters associated with these genes were isolated and activity tested using the GUS reporter gene. Most were functional soon after introduction into the explant, but none maintained activity in mature regenerated plants.

To avoid gene silencing, Potier et al. 2006 isolated tissue-specific promoters from the upper stem or roots of maize and sorghum rather than sugarcane. These promoters (with and without the first intron) were used to test the expression of GUS gene in sugarcane callus. In general, the presence of an intron showed higher expression in all the promoters studied. Further research should identify effective tissue-specific promoters that can maintain activity in mature transgenic plants. With the available sorghum genome sequences, the search and solution to finding new tissue-specific promoters is more tangible.

3.1.6 Transcriptional regulators

Transcription factors are proteins that interact with the promoter regions of target genes in a sequence-specific manner. They influence the manner in which RNA Polymerase II initiates mRNA synthesis by enhancing or repressing gene expression and, hence, are important regulators of gene expression. Transcription factors tend to control multiple steps in metabolic pathways, meaning that a single protein can affect the expression pattern of a large number of downstream genes (Kinney, 2006). This is an important consideration for plant transformation strategies aimed at significantly affecting the level of end-product accumulation. Although this can be a successful strategy, a lack of knowledge about a particular pathway may mean that additional and hitherto unknown rate-limiting steps may be present that may frustrate the attempt to significantly modify end-product accumulation. Overexpression of transcription factors holds the promise of significantly altering the overall flux through a metabolic pathway by controlling the transcription of multiple genes. The potential of using transcription factors to enhance current and future work into the genetic enhancement of sugarcane is highlighted in Sections 2.2.6, 3.2.6, 3.2.7, and 3.2.9.

3.1.7 Modifying more complex traits

Very few traits that are simply inherited have been described in sugarcane (Hogarth, 1987; Daugrois et al., 1996; Mudge et al., 1996a, b; Raboin et al., 2006), suggesting that most phenotypic characteristics, such as cane yield and sugar content, are controlled by polygenes. In other organisms, genes that contribute to complex traits (QTL) pose special challenges that make gene discovery and subsequent modification of such traits more difficult. Such problems include locus heterogeneity, epistasis, low penetrance, variable expressivity, and pleiotropy (Glazier et al., 2002). Approaches are even more difficult in sugarcane, where there is the possibility of segregation for three or more alleles at a locus, a lack of chromosome preferential pairing and high levels of heterozygosity (Ming et al., 2001a, b; Aitken et al., 2006), which generally results in numerous alleles of small effect, some of which cannot be positioned on a genetic map. These factors work together to make claims of linkage discovery notoriously difficult to verify and candidate gene identification challenging. However, prospects for success can be improved through genome sequencing and comparative genomics. Such resources will allow more rapid dissection of complex traits through candidate gene identification and subsequent testing using functional genomics approaches.

3.1.8 Functional and comparative genomics

3.2.1 Improved sugar

Sugarcane is unusual amongst plants in storing up to 62% dry weight (and 16–25% fresh weight) of sucrose as the primary source of carbon and energy reserve in the storage parenchyma of the mature culm or stalk (Bull and Glasziou, 1963). These are approximate levels obtained in S. officinarum, the major source of commercial hybrid germplasm. In contrast, some of the wild relatives of sugarcane store less than 2% of the fresh weight as sucrose. As the photosynthetic rates on an area basis of S. spontaneum have been reported as nearly twice those of S. officinarum and 30% greater than that of hybrid cultivars (Irvine, 1975), these striking differences in sink accumulation and storage activity cannot be explained by differences in photosynthetic rates in the leaves (Moore, 2005). Despite many years of research, the process of sugar accumulation is not fully understood.

As more than 70% of the sugar harvested for human consumption is derived from sugarcane (FAO, 2006), improving sugar content in sugarcane will remain as an important focus for future biotechnological innovations. Traditional breeding has improved the yield of sucrose mostly through increased cane yield. Increasing sucrose concentration has not occurred recently through traditional breeding in countries such as Australia, where the levels are already relatively high (Jackson, 2005). Attempts to increase sucrose content in sugarcane cultivars through the modification of plant genes involved in sugar metabolism have been mostly unsuccessful (reviewed by Lakshmanan et al., 2005). For example, a 70% reduction in activity of acid invertase, a key enzyme involved in sucrose breakdown, resulted in no significant change in sucrose yield or purity in the immature internodes of transgenic sugarcane (Botha et al., 2001). A better understanding of sugar-transport pathways will be required for strategies to increase the flux of sugars and, ultimately the amount of sucrose in the storage tissue (Grof and Campbell, 2001; Rae et al., 2005).

The doubled sugar content observed in transgenic sugarcane while attempting to produce isomaltulose (Wu and Birch, 2007) indicates that there is potential to improve sugar content in sugarcane through metabolic engineering of the vacuolar compartment. However, most vacuolar-targeting studies in many crops involving stable transformants have been limited by difficulties in detecting vacuolar-targeted enzyme activities (Ebskamp et al., 1994; van der Meer et al., 1994; Caimi et al., 1996; Ma et al., 2000; Gnanasambandam and Birch, 2004), indicating that the stability of targeted proteins is an important factor for effective metabolic engineering. For isomaltulose production in sugarcane, a highly efficient sucrose isomerase had to be used (Wu and Birch, 2007).

Reports of experimental efforts to protect foreign gene products from vacuolar degradation are scarce. To understand how to genetically engineer a protein for stability in the vacuole, the “enemies” within the vacuoles—the proteolytic enzymes and their sites of action—need to be characterized (http://www.proweb.org/other.html). A method to isolate protoplasts from sugarcane mature stem parenchyma cells has been reported (Gnanasambandam and Birch, 2006), but further optimization is needed for successful vacuole isolation for characterization of proteolytic enzymes. Alternatively, if vacuole isolation proves to be difficult, fluorescent substrates for proteases (from Molecular Probes, Inc.) can be used with isolated protoplasts. Such characterization on proteolytic enzymes may allow for informed modification on foreign proteins to remove sites for attack by sugarcane proteases. Alternatively, directed evolution can be applied to rapidly evolve enzymes for optimal activity under vacuolar conditions (Stemmer, 1994a, 1994b; Chica et al., 2005; Kaur and Sharma, 2006). There may be particular sequences within a protein that are sensitive to proteolytic enzymes, and altering these sensitive sequences may increase protein stability. For example, mutation of protease-sensitive regions of the firefly luc yielded larger and more stable signals compared with the wild-type protein in E. coli (Thompson et al., 1997).

3.2.2 Alternative or novel sugars

There have already been examples of alternative sugars produced in genetically modified (GM) sugarcane (Section 2). In addition to isomaltulose (Wu and Birch, 2007) and sorbitol (Chong et al., 2007), there is potential to engineer synthesis of a range of other sugar complexes and sugars into sugarcane, for example alternative carbohydrates such as starch and fructans (already demonstrated in sugarbeet).

Tagatose (Kim, 2004) has properties closer to sucrose than any of the other sugar substitutes. D-Tagatose is an isomer of D-galactose and it can function as a low-Joule sweetener. Although it has relatively the same sweetness as sucrose, it is detected on the tongue sooner.

Xylitol is a sugar alcohol of considerable commercial interest (Granström et al., 2007a, 2007b). Its metabolism in humans is independent of insulin and it is anticariogenic. It also has uses as a pharmaceutical. Xylose is a major component of sugarcane hemicellulose, making up 25% of the w/w dry matter and 91% of the nonfermentable sugars (Pessoa Jr. et al., 1997). Research is currently underway to optimize the conversion xylose to xylitol. Candida guilliermondii has been used as a test species on sugarcane bagasse hydrolysate, because it naturally metabolizes xylose (da Silva and Felipe, 2006). However, Candida cannot be used in food production, because many Candida species are pathogens. An alternative is to use bacteria. When a xylose reductase and a xylose transporter were engineered into the bacterium Lactococcus lactis, levels of xylitol production were equivalent to those in the best-producing Candida line (Nyyssölö et al., 2005). An alternative would be to engineer a xylose reductase into sugarcane. One strategy might be to have the enzyme incorporated into the sugarcane cell wall structure, so that it is only functional when it is released during sugarcane processing at the mill. Another might be to regulate it with a wound-inducible promoter, so that it is produced in the sugarcane cells after harvesting. In both of these scenarios, the enzyme(s) will have to be engineered to function at the temperatures and pH of a sugar mill and in particular under mill conditions designed to break down hemicellulose (Pessoa Jr. et al., 1997).

Though there has been much research toward the development of low-Joule nonsucrose sweeteners (Modi and Borges, 2005), two such sweeteners (saccharin and cyclamates) were found to be carcinogenic under some circumstances, while others (such as aspartame) exhibit temperature and low pH instability that limit their utility (Knight, 1994). Hence, it may be useful to produce a low-Joule nonsucrose product from sugarcane. Alternatively, there are a range of sweet proteins that may be worth considering. Monellin, Thaumatin, and Brazzein are 100000, 3000, and 500 times sweeter than sucrose, respectively. These would be a natural replacement for synthetic sugars because they are low-Joule sweeteners and would not trigger insulin production (Kant, 2005).

3.2.3 The biorefinery—biofuels and biochemicals

Sugarcane is expected to become a platform for biorefinery production of fuels and chemicals. Recent demand for renewable energy sources has increased the interest within the research community on transgenic breeding for biomass production. Sugarcane is the crop having the most favorable input/output balance for bioenergy, and with the growing body of research on its genome, may allow the development of a low-input energy crop with key outputs as biofuel (ethanol) and bioenergy (electricity). The production of an alternative energy crop, with high biomass potential, will be a significant step toward the implementation of a biofuels, bioenergy, and, in the not too distant future, bioproducts industries. If biorefineries, producing dozens of products (Figure 6) are integrated into sugar mills, one can envision a sugarcane industry functioning more like petrochemical plants do today. Unlike petrochemical plants, the resource driving this new bioeconomic revolution will be renewable.

In nature, the most prevalent forms of carbon are the biopolymers cellulose and hemicellulose (Houghton et al., 2006). The third most abundant biopolymer is lignin (van Dam et al., 2005). These are considered to be the raw materials for future bio-based industries creating opportunities for sugarcane industries to diversify. Aside from water, sugarcane plants are primarily comprised of cellulose, hemicellulose, and lignin. Traditionally, this resource has been combusted, either in the field to remove the leaf material or at the mill to utilize the waste product bagasse to generate the necessary energy to power mill operations. Since the mid-1980s, it has been proposed that “energy cane” should be introduced as a multiple product alternative to sugarcane grown largely for sucrose (Alexander, 1984, 1985). The time may have come to implement this vision with the inclusion of advances in the fields of biotechnology. Changing the sugarcane industry from primarily a one-product to a multiple-product industry is an enormous challenge, but there are multiple indicators that should make both sugarcane farmers and millers optimistic.

Of all the agricultural crops, sugarcane may be the most ideally suited to benefit from a shift to a bio-based economy because of its large biomass and, therefore, large quantities of cellulosic and hemicellulosic sugars and lignin (Alexander, 1984, 1985; Brumbley et al., 2004, 2006b, 2006c, 2007). There is good reason for simultaneously developing sugarcane both as a biofactory and as a feedstock for biorefinery processes. As technologies are developed to produce a range of new and diverse products in the sugarcane plant, other technologies are being developed to dismantle the cellulose and hemicellulose to release the large quantities of fermentable sugars from biomass. The breakdown of the tough structural elements that make up sugarcane cell walls will facilitate the extraction of new bioproducts such as the bioplastics described in Section 2.

The biofuel ethanol is becoming increasingly more important as the world attempts to wean itself off its addiction to petroleum-based fuels. The second largest producer of ethanol in the world is Brazil, and sugarcane is their feedstock (Pessoa Jr. et al., 2005; Lorentzen, 2006). In the near future, another bio-based fuel (butanol) may enter the market. DuPont and British Petroleum recently formed a joint venture to produce this biofuel commercially. Butanol resolves many of the problems of ethanol. Not only does it have a higher energy content, it is not as corrosive, uses an air/fuel ratio that is closer to that of gasoline, can be shipped through existing fuel pipelines, is safer to handle, and can replace gasoline at any percentage up to 100%.

Governments and private industry are investing in technologies to convert biomass into biofuels and biochemicals. Cellulose, hemicellulose and lignin interact to form plant cell walls and are responsible for giving plants their structural integrity. Compared to starch, getting to the sugars that make up these polymers is far more difficult and more expensive (Bayer et al., 2004; Houghton et al., 2006). A diverse range of bacteria have been discovered that have multienzyme complexes that bind to and degrade cellulose and hemicellulose. This complex has been termed the cellulosome and the research on these enzyme complexes was reviewed recently (Bayer et al., 2004). Researchers are looking for new organisms with cellulose degrading capabilities. Once discovered, activity and performance of these enzymes and complexes can be improved using directed evolution (Joyce, 2004; Neylon, 2004; Hibbert and Dalby, 2005).

The pool of sugars currently locked up in cellulosic biomass could be used to feed biorefineries both for the production of biofuels and bioproducts. White biotechnology (Paster et al., 2003) is the use of biological processes for the production of chemicals such as plastics, adhesives, and paints. Fermentation technologies could be used to produce building block chemicals that will be the basis for the range of secondary and intermediate chemicals (Figure 6). Currently, the United States and European Union annually produce €1000 billion worth of fine and bulk chemicals from a petroleum feedstock. They plan to shift over 90% of this production from a petroleum to a biorenewable base over the next 50 years. To accomplish this, new enzymes will be required to convert the sugars within the cellulosic biomass to the renewable building blocks required by tomorrow's manufacturing sector.

The vast majority of microbes have never been characterized. This enormous and diverse population is thought to be a rich source of new genes, metabolic pathways, and enzymes. However, most of these organisms are not culturable. Metagenomics is the study of the collective genomes of organisms, independent of growth in culture. The tools for doing metagenomics are advancing rapidly (Tringe and Rubin, 2005; Green and Keller, 2006). For example, Venter et al. 2004 used metagenomics to identify over 1.2 million new genes from at least 1800 species, including 148 previously unknown bacterial groups, just in the Sargasso Sea. As our understanding of protein domains advances (Portugaly et al., 2007), it may be possible to design “lab-on-a-chip” devices (Hong et al., 2004) that can be used to rapidly screen different environmental samples for genes encoding specific enzyme domains. Using systems biotechnology, newly discovered enzyme's can be characterized rapidly and their utilities for industrial bioprocessing determined (Lee et al., 2005). Directed evolution tools, such as domain shuffling (Hibbert and Dalby, 2005), can be utilized to combine the new enzyme's functional domains with those of other related enzymes to create an enzyme that will function in the precise environment it is needed, whether that is a fermentation tank inside the cytoplasm, storage vacuole, mitochondria, peroxisome, or plastid of a sugarcane cell.

3.2.4 Biopolymers

A potential major industrial use for sugarcane is the production of biodegradable polymers. These polymers can be produced by fermentation technologies and, as shown in Section 2, can be produced in sugarcane. For example, PHAs are polyesters synthesized naturally by many species of bacteria as carbon sources and energy reserves. PHAs are accumulated to levels as high as 90% of the cell dry weight in bacteria. PHAs have the properties of thermoplastics and elastomers and they are biodegradable (Madison and Huisman, 1999). These classes of biopolymers offer a renewable and environmentally friendly alternative to the 150t per year of plastic currently produced from petrochemical resources.

In nature, over 130 PHA variations have been identified that produce plastics with properties varying from hard and brittle, too elastomeric, gluelike, and even rubbery (Steinbüchel and Valentin, 1995). PHAs are classified based both on the length of the side chains on the polymer (i.e., whether the side chains are short (C4-6) or medium (C6-16) in length), and on whether the polymer is a homopolymer, or a co-polymer of PHAs with different side chains. In the case of co-polymers, the ratio of medium (mcl) to short side (scl) chains also affects the properties. Co-polymer PHAs of commercial interest range from 2% mcl (hard, brittle) to 20% mcl (soft, elastic). PHB, the most studied of the PHAs, is a short-side-chain homopolymer and displays properties similar to polypropylene (Hocking and Marchessault, 1994).

In Brazil, an important precedent has been established on how to diversify the product base of a sugar industry, first with ethanol, and then with bioplastics (Lorentzen, 2006; Velho and Velho, 2006). In 1992, initial projects were funded to develop a novel platform for the production of PHB from sugarcane through fermentation (Velho and Velho, 2006). The Brazilian company Copersucar (http://www.copersucar.com.br) first engineered the bacterium Ralstonia eutropha to use sucrose as a carbon source. A pilot plant was then built, integrated into a sugarcane mill (http://www.biocycle.com.br), and ultimately produced 50t per year of high purity PHB. This refinery receives its feedstock (sugar syrup) directly from the sugar mill, which also produces all of the energy necessary to run the fermentation, extraction, and purification facilities. The solvents for PHB extraction are also sugar mill products sourced from the ethanol production facility (Velho and Velho, 2006). This model is an excellent example of how an existing resource industry can, with government and corporate support, build on its knowledge base to create new business opportunities and a help establish bio-based economy in a developing country (Lorentzen, 2006).

Biopol^TM (Choi and Lee, 1999) is the only PHA co-polymer that has been produced commercially. However, Metabolix, Inc. has formed a strategic alliance with Archer Daniels Midland (ADM) to commercialize PHA natural polymers. Metabolix and ADM claim that they can make these natural plastics by fermentation at costs that are competitive, or near competitive, with petrochemical-based plastics (http://www.metabolix.com). Metabolix argues that producing these plastics in the cells of living plants will reduce this cost even further. Hence, agricultural crops are regarded as a promising low-cost alternative for the production of PHAs on a large scale (Snell and Peoples, 2002). High-level production (approximately 40% of dry weight) of PHB has been achieved in plants by introducing three R. eutropha genes encoding a β-ketothiolase (phaA), an acetoacetyl-CoA reductase (phaB), and a PHA synthase (phaC) (Bohmert et al., 2000). The success in the production of PHB in sugarcane (Petrasovits et al., 2007; Purnell et al., 2007) indicates that other biopolymers can be produced in sugarcane.

3.2.5 Natural products/pharmaceuticals/proteins

There are opportunities to engineer the metabolism of sugarcane to increase production of natural products (Dixon, 2005). For instance, the most functionally and structurally diverse group of plant metabolites is the isoprenoids (also called terpenoids). These diverse compounds have a commercial value as flavors, pigments, polymers, or drugs (reviewed in Rodríguez-Concepción, 2006; Withers and Keasling, 2007). There are other good reasons to learn how to manipulate and/or regulate isoprenoid production in sugarcane, as they play primary roles in respiration, photosynthesis, and regulation of growth and development. As secondary metabolites, they function in protecting plants against herbivores and pathogens, attract pollinators and seed dispersing animals, and influence competition among plant species (Rodríguez-Concepción, 2006).

As described in Section 2, some work has already been done to engineer sugarcane to produce pharmaceuticals and high-value and bulk proteins. Monoclonal antibodies (mAbs) are one class of pharmaceutical proteins with considerable potential for production in plants. The estimated market size for mAbs is expected to exceed US $20 billion per year (Reichert et al., 2005). Both IgG- and IgA-type antibodies have been successfully produced in plants (Giritch et al., 2006). Recently a new transient expression technology, “magnifection”, for production of gram quantities (0.5gkg^–1 fresh leaf biomass) of mAbs in plant cells was reported (Giritch et al., 2006). The time from gene delivery to production of fully assembled mAb is 14 days and to production of gram quantities is 14–20 days, making it the fastest production platform available (Hiatt and Pauly, 2006). This system uses vectors from two different noncompeting plant viruses, one for producing the heavy chain and one for the light chain, and transfers them to the plant via Agrobacterium-mediated transformation.

One area that may hold promise is the production of industrial enzymes. Future biorefineries are going to require a host of enzymes for converting base material into the raw materials required by the manufacturing sector (Figure 6). Some of the enzymes can be produced directly in the sugarcane cells, and potentially at high concentrations once plastid transformation technologies are developed for sugarcane.

3.2.6 Drought tolerance

In most crops, past breeding efforts for drought tolerance have been hindered by the quantitative genetic basis of the trait and the poor understanding of the physiology (reviewed in Tuberosa and Salvi, 2006). For breeders, it is important to select genotypes that are able to optimize water uptake and water use efficiencies and minimize damage incurred by drought stress. This ultimately will lead to maximized yields. In an effort to select for improved drought tolerance, breeders have identified a myriad of morpho-physiological QTLs acting from the cellular to the whole crop level (reviewed in Tuberosa and Salvi, 2006). However, given the genetic complexity of sugarcane, a marker-assisted selection approach for a myriad of improved drought tolerance QTLs would appear a very difficult proposition. An alternative way forward would be to make use of the knowledge gained in simpler systems and identify candidate genes to test in sugarcane. Candidate genes for drought tolerance have been divided into three broad areas including: (i) functional proteins such as enzymes involved in osmotically active compounds, transporters, chaperones, and reactive oxygen species (ROS) scavengers; (ii) transcription factors involved in the plant response to drought stress; and (iii) signaling factors upstream of transcription factors, including protein kinases and proteins involved in phospholipid metabolism, calcium sensing, and protein degradation (reviewed in Umezawa et al., 2006; Valliyodan and Nguyen, 2006). Most of these validation studies have been undertaken with model plants in laboratory or glasshouse conditions. An important next step is to trial these approaches in crop species under field conditions.

In other plants, a variety of transgenes have been used to improve drought tolerance. Examples include, the barley HVA1 gene (Fu et al., 2006), manganese superoxide dismutase (Wang et al., 2005c), regulator of G-protein signalling protein (RGS) (Chen et al., 2006), and aldehyde dehydrogenase (Rodrigues et al., 2006). Perhaps the most widely and successfully used genes have been the CBF/DREB group of transcription factors (Agarwal et al., 2006). CBF/DREB is a class of transcription factors that binds to drought responsive cis-acting elements and is important in regulating gene expression in response to drought, high salt, and cold stress (Yamaguchi-Shinozaki and Shinozaki, 1994). The ability of CBF to enhance tolerance to cold stress (Jaglo-Ottosen et al., 1998) and drought stress (Haake et al., 2002) was first demonstrated in Arabidopsis and has subsequently been shown to enhance tolerance to abiotic stress in commercially important crops such as tomato (Hsieh et al., 2002; Lee et al., 2003b) and rice (Ito et al., 2006).

Sugarcane lines have been produced containing a gene encoding drought-induced Arabidopsis transcription factor, AtCBF4 (McQualter and Dookun, 2007). Under glasshouse conditions, plants containing AtCBF4 driven by the constitutive maize polyubiquitin promoter showed growth retardation. Research is now focused on using drought-inducible promoters to reduce this negative phenotype.

As mentioned in Section 2, trehalose has already been used to engineer drought tolerance in sugarcane (Zhang et al., 2006). There are also a number of other potential osmoprotectants that can be trialed in sugarcane. These include betaines, the amino acids proline and ectoine, and polyols mannitol and sorbitol (Rontein et al., 2002). As mentioned above, transgenic sugarcane plants have been generated that produce sorbitol (Chong et al., 2007) and are being progressed to field trials.

3.2.7 Disease and insect resistance

Estimates of the cost of control and loss due to major sugarcane pests and diseases in Australia are in the order of 10% of the total value of the sugarcane crop (McLeod et al., 1999). Of these diseases and pests, soil-borne pathogens caused an estimated 75% of the losses, followed by canegrubs (larvae of Scarabaeidae) (10%), ratoon stunting disease (caused by Lxx.) (6%), and brown rust (caused by Puccinia melanocephala H&P Sydow) (3%). Occasionally, disease outbreaks due to the breakdown of disease resistance occur, which can result in much larger losses (40% reduction in cane yield in some sugarcane growing regions in Australia as a result of orange rust (Magarey et al., 2001)). Diseases such as ratoon stunting disease, pineapple disease, and red rot can be largely controlled by farm hygiene processes such as sterilizing cutting and harvesting equipment and by the use of fungicides (Magarey, 2005). However, for most sugarcane diseases, genetic resistance is the most effective method of disease management and sugarcane breeders have done an excellent job in controlling diseases through the use of resistant cultivars.

In the future, it is possible that resistance may break down, and aside from introgressing wild germplasm with novel resistance (if this exists in the gene pool), alternative strategies will be needed. Such approaches should provide durable resistance, preferably against a range of pathogens. Genetic engineering has the potential to achieve this through inserting carefully selected and possibly multiple genes as transgenes that (i) confer durable broad-spectrum resistance; (ii) are safe for all other organisms; and (iii) cause no yield penalty in the plant (Gurr and Rushton, 2005a). Such approaches have been successful with chewing insect pests, where expression of an insecticidal protein from B. thuringiensis has lead to increased yields and reductions in insecticide applications in cotton, soybean and maize (reviewed Gurr and Rushton, 2005b). Recent work in transgenic sugarcane has also indicated that Bt proteins are active against stem borers (Weng et al., 2006). Strategies for resistance to sugarcane viruses have also met with some success through RNAi-mediated approaches against the single-stranded RNA SCMV and the double-stranded RNA FDV (Joyce et al., 1997; Ingelbrecht et al., 1999; McQualter et al., 2004). However, similar successes for durable resistance to bacteria and fungi have not been forthcoming (reviewed Gurr and Rushton, 2005a).

Strategies have attempted to boost disease resistance through constitutive overexpression of defense components or single-antimicrobial proteins and have led to a variety of problems including poor quality plants, yield reduction, poor efficacy, or short durability (reviewed by Gurr and Rushton, 2005b). New strategies are being explored to overcome these problems. For example, a two-component system has been suggested, where the manipulation of “master switch” genes (such as kinases and transcription factors), which regulate entire signaling pathways are coupled with pathogen-inducible promoters to precisely regulate spatial and temporal gene expression (Yamamizo et al., 2006). Other strategies have come from the genomics field, for instance the complete genome of Lxx, has been fully sequenced (Monteiro-Vitorello et al., 2004) and opportunities now exist to try and control or even eradicate this costly pathogen from sugarcane (Brumbley et al., 2006a). Global transcriptional analysis of plant pathogen interactions has also revealed pivotal steps in defense responses that may be targeted in future strategies (Zabala et al., 2006). Further disease control strategies may involve pyramiding two or more antimicrobial proteins to improve efficacy and durability of resistance.

Work on transcription factors in other plants may offer some future promise for sugarcane. WRKY proteins are a large family of transcription factors that mainly participate in plant biotic-stress responses (Liu et al., 2007). They share a DNA-binding domain consisting of about 60 amino acids, as well as additional conserved features that distinguish individual subgroups (Eulgem et al., 2000). WRKY transcription factors can bind to domains in the promoters of pathogen-response genes, termed W boxes (TTGACC) (Maeo et al., 2001; Turck et al., 2004). While overexpression of some WRKY transcription factors has resulted in enhanced resistance to various bacterial or fungal pathogens, others can suppress the expression of defense-related genes.

Chen and Chen 2002 transformed Arabidopsis plants with AtWRKY18 under control of the CaMV 35S promoter. Transgenic AtWRKY18 plants showed marked increase in the expression of pathogenesis-related genes and resistance to the bacterial pathogen Pseudomonas syringae. Thus, AtWRKY18 was shown to positively modulate defense-related gene expression and disease resistance. Conversely, AtWRKY7 transcription factor plays a negative role in defense responses to P. syringae. Kim et al. 2006 studied the biological function of the Arabidopsis WRKY7 gene in both loss-of-function T-DNA insertion and RNAi mutants and gain-of-function transgenic overexpression plants. The T-DNA insertion and RNAi mutant plants displayed enhanced resistance to a virulent strain of the bacterial pathogen P. syringae. Transgenic Arabidopsis plants that constitutively overexpress WRKY7 showed reduced expression of defense-related genes, including PR1 and developed more severe disease symptoms than wild-type plants.

A large number of WRKY genes have also been identified in rice. Ryu et al. 2006 isolated WRKY transcription factors whose expressions were altered upon attack of the fungal pathogen Magnaporthe grisea, the causal agent of rice blast disease. Among 45 tested genes, the expression of 15 was increased in interactions between rice and M. grisea. Twelve of this subset of genes were also differentially regulated in rice plants infected with the bacterial pathogen X. oryzae pv. oryzae (X.o.o) 13751. Ryu et al. 2006 suggested that a large number of WRKY DNA-binding proteins are involved in the transcriptional activation of defense-related genes in response to rice pathogens. Several pathogenesis-related genes were induced in OsWRKY03-overexpressing transgenic rice plants (Liu et al., 2005), while overexpression of the OsWRKY71 gene in rice resulted in enhanced resistance to virulent bacterial pathogen X.o.o. 13751 (Liu et al., 2007).

As mentioned above, there are extensive databases at Harvard University covering plants, animals, fungi, and protists. The TIGR has the Comprehensive Microbial Resource (CMR) (http://cmr.tigr.org/). Within this free database are the complete genome sequences of 354 bacteria, including a number of plant pathogens. There is currently an effort to reduce the cost of sequencing the genome of a bacterium to under US $1000, and this is likely to be accomplished in the next decade. When that happens, databases such as the one at TIGR and Harvard University will contain the genomes of the vast majority of plant pathogenic bacterium, fungi, viruses, and mycoplasmas as well as the insect pests that plague agriculture around the world. Studying the plant–microbe or plant–insect interaction in the future will be done at the whole genome, transcriptome, proteome, and metabolome level of the host, and the pests and pathogens.

3.2.8 Modified plant architecture and development

Shoot architecture describes the way in which the aerial portions of plants develop. It is influenced predominantly by shoot branching and plant height. Agronomic interest in shoot architecture has stemmed mainly from potential yield increases associated with some plant ideotypes. For instance, height reduction in cereals has led to yield improvement through decreased lodging, whilst reduction of tillering in some cereals may lead to yield improvements (Duggan et al., 2005).

In sugarcane, complex interrelationships exist among shoot architecture, stalk characteristics, cane yield, and sucrose content (Milligan et al., 1990). For instance, two elements of sugarcane shoot architecture, the number and weight of stalks, determine cane yield. However, stalk number and stalk weight are negatively correlated and, thus, selecting for either trait alone may not result in increased cane yield (Milligan et al., 1990; Bell et al., 2004; Bell and Garside, 2005). Stalk architecture can also affect cane sucrose content. Taller canes are more likely to lodge under wet or windy conditions, causing a reduction in sucrose content (Singh et al., 2000, 2002), whilst clones with a propensity for producing the low-sugar-content suckers reduce the sucrose content of the harvested crop (Bonnett et al., 2001, 2004a).

Recently, several genes controlling shoot architecture traits, such as stalk height and tillering, have been described (McSteen and Leyser, 2005; Wang and Li, 2006). There is now an opportunity to test how modifying stalk characteristics with these genes will affect cane yield and sucrose content in the same sugarcane cultivar without altering other genes in the plant. Three areas of research have been identified which may lead to increased yields, accelerated biomass production, controlling lodging, and reducing suckering. The genes that are currently being used to address these research areas are the TB1 gene, which has been shown to affect tiller number and stem width in maize and rice (Doebly et al., 1997; Takeda et al., 2003), the MAX3 gene that affects axillary meristem outgrowth in rice (Zou et al., 2006) and Arabidopsis (Booker et al., 2004), and several gibberrellin oxidases that regulate stem elongation (Hedden and Phillips, 2000). Transgenic sugarcane plants have been produced and are currently undergoing extensive glasshouse and field trials to test what effects these genes produce (Pribil et al., 2007).

3.2.9 Other products

3.3.1 Human health and environment

There is no reason to assume that genetic engineering has a greater potential to harm human health or the environment than any techniques used in classical breeding (Miller, 2007). If there is a lesson we have learned from the enormous experience of commercially growing GM crops on over 1 billion acres (400 million hectares) around the world for over two decades is that if the trait(s) encoded by the transgenes are safe, then there are more than sufficient safeguards in traditional breeding programs to ensure that the GM-derived crop is safe (Bradford et al., 2005).

However, there is still considerable resistance to food derived from GM plants, even though there is no scientific basis for this concern (Miller, 2007). Because of this resistance, the large buyers of raw sugar have been requiring guarantees from the suppliers that the sugar supply is not derived from GM plants. This has created a wait-and-see approach, where sugarcane biotech research and development is ongoing but no one wants to be the first into the market. However, it is likely that sugar from sugarbeets engineered with herbicide and pest resistance traits will enter the market first and that will hopefully open the door for sugarcane. If GM sugarcane lines with doubled sugar production (Wu and Birch, 2007) perform well in field trials, pilot-scale production will most likely be trialed.

Refined white sugar (>99.98% sucrose) is the most chemically pure food produced from agriculture, thus, reducing any foundation for concern about foreign genes or gene products in this food (Birch and Maretzki, 1993). Taylor et al. 1999 analyzed the various products during the crystallization process of sugar from sugarcane juice for the presence of transgene DNA. Their results showed that no PCR amplifiable DNA was present in sugar crystallized from SCMV transgenic plants.

However, when sugarcane is engineered as a biofactory to produce fine chemicals or pharmaceuticals, it is critical that clear safeguards are in place to ensure that there is no conceivable harm to public health or the natural environment. Safety evaluations will have to be done on a case-by-case basis. For instance, sugarcane producing bioplastic such as the pHBA and PHAs discussed in this chapter could still be used for production of sugars for human consumption. Animals already make pHBA and have the capacity to digest it (McQualter et al., 2005), and PHAs have a 40-year history of safety in a variety of medical applications, especially because of biodegradability and biocompatibility. PHAs have been used for surgical sutures and wound dressings. There is ongoing research for the commercialization of PHA-based products for tissue engineering and for controlled release systems, and for bone repair (Zhijiang, 2006). Because these biopolymers in nature are used as a carbon sinks, they are fully biodegradable. So producing them in sugarcane should not be dangerous to human health or the environment.