Isomaltulose (IM) is a natural isomer of sucrose. It is widely approved as a food with properties including slower digestion, lower glycaemic index and low cariogenicity, which can benefit consumers. Availability is currently limited by the cost of fermentative conversion from sucrose. Transgenic sugarcane plants with developmentally-controlled expression of a silencing-resistant gene encoding a vacuole-targeted IM synthase were tested under field conditions typical of commercial sugarcane cultivation. High yields of IM were obtained, up to 483 mm or 81% of total sugars in whole-cane juice from plants aged 13 months. Using promoters from sugarcane to drive expression preferentially in the sugarcane stem, IM levels were consistent between stalks and stools within a transgenic line and across consecutive vegetative field generations of tested high-isomer lines. Germination and early growth of plants from setts were unaffected by IM accumulation, up to the tested level around 500 mm in flanking stem internodes. These are the highest yields ever achieved of value-added materials through plant metabolic engineering. The sugarcane stem promoters are promising for strategies to achieve even higher IM levels and for other applications in sugarcane molecular improvement. Silencing-resistant transgenes are critical to deliver the potential of these promoters in practical sugarcane improvement. At the IM levels now achieved in field-grown sugarcane, direct production of IM in plants is feasible at a cost approaching that of sucrose, which should make the benefits of IM affordable on a much wider scale.
Saccharum officinarum and Saccharum interspecific hybrids grown as commercial sugarcanes have the unusual ability to accumulate high concentrations of sucrose in stem tissues. Sugarcane is an attractive crop for metabolic engineering aimed at producing high-value materials from sucrose (Birch, 2007). For example, we are interested in using sugarcane for the direct production of isomaltulose (IM) and trehalulose (TH). As foods, these natural sucrose isomers have potential benefits for consumers, including slow digestion and low cariogenicity (Lina et al., 2002; Okuno et al., 2009; Ooshima et al., 1991). They are currently produced by industrial-scale conversion from sucrose, using immobilized bacterial cells with periplasmic sucrose isomerase (SI) enzymes (Schiweck et al., 1991). Cloning and characterization of bacterial genes encoding efficient IM and TH synthases (reviewed in Goulter et al., 2011) supported the potential for application in plants.
Recently, we showed that field-grown sugarcane plants expressing genes for vacuole-targeted SI enzymes accumulated sucrose isomers in whole-cane juice up to 33% of total sugars in the case of IM synthase (ims) UQ68J (Basnayake et al., 2012) or 40% of total sugars in the case of TH synthase (ths) MX-45 (Hamerli and Birch, 2011). Near-complete conversion of sucrose to isomers occurred in basal (oldest) internodes, despite very low SI enzyme levels attributed to rapid proteolysis in the sugar-storage vacuoles. There were no obvious adverse effects on growth in the initial, small-plot, field trials. These observations indicated the potential for higher-level isomer production through a combination of suitable developmental promoters and modifications to the gene sequence, to achieve stable patterns of SI activity and enhanced sugar profiles along stalks, throughout the commercial harvest season.
Practical metabolic engineering requires appropriate developmental regulation of transgene expression. However, few promoters (most notably PUbi from maize) have been shown to drive transgene expression in mature sugarcane plants (Birch, 2013). PUbi appears more resistant than other tested promoters to the full effects of interacting RNAi pathways assumed to develop in concert with the strong post-transcriptional effects observed soon after transgene integration (Birch et al., 2010).
For metabolic engineering applications aimed at conversion of stored sucrose into higher-value products, the optimal expression pattern in sugarcane is likely to be stem specific. It may be necessary to ensure that sucrose levels are not depleted in other sink tissues, including rapidly growing tissues. To this end, we recently isolated and tested a suite of promoters from the mature stem–expressed ScR1MYB1 gene. None of the tested promoters drove reporter transgene activity in mature plants, whereas the activity of the endogenous copies in the transgenic plants remained unchanged (Mudge et al., 2009). We also isolated other sugarcane stem promoters, notably from a mature stem gene ScCIPK described below and from ScLSG that is expressed preferentially in sucrose-loading and mature stem regions (Moyle and Birch, 2013). In the ScLSG promoter suite, we were able to distinguish variants that drive different expression patterns, through a combination of allele-specific qRTPCR and use of a reporter transgene designed to avoid silencing (Moyle and Birch, 2013).
Here, we demonstrate that developmentally controlled expression of a silencing-resistant gene for vacuolar SI activity allows high IM yields across multiple generations in field-grown plants.
Results and discussion
Isolation of mature-stem promoters from sugarcane
Sugarcane EST MCSA222H07 (Casu et al., 2004) comprises 469 nt encoding the C-terminal region of a CBL-interacting protein kinase (CIPK), corresponding to ScCIPK-21 (Papini-Terzi et al., 2009). Northern blot analysis showed that the corresponding transcript is strongly up-regulated in mature versus immature stem (Figure 1a) and also detectable in mature roots (Figure 1b), but not in leaf tissue. Northern blot analysis using RNA from vascular, parenchyma and rind tissue dissected from internode eight of field-grown Q117 revealed the transcript in all three tissue types (Figure 1c). This is similar to the expression pattern of the ScR1MYB1 gene described (Mudge et al., 2009).
Promoter–reporter transgene activity in mature sugarcane plants
A 2.1-kb promoter region from ScCIPK-21 and a 5.7-kb promoter region from ScR1MYB1, referred to as PH07A and PA157, respectively, were chosen for functional analysis in transgenic sugarcane in comparison with constitutive PUbi and sucrose-induced PRol. Camera assays on fresh tissue samples from mature (12–18 month old) glasshouse-grown plants revealed LUC reporter activity in approximately 75%–80% of independent transgenic lines, independent of the promoter used. Taking only the expressing lines, the average LUC activity in mature internodes differed in the ratio PUbi 50 : PRol 3 : PA157 2 : PH07A 1 (Table 1), with the PUbi activity being significantly higher than the other promoters (P < 0.001; one-way ANOVA with Bonferroni post-test). All of the promoter cassettes include their native 5′ UTR, which in the case of PUbi includes an intron that is likely to serve as an enhancer of coupled transgene expression (Lu et al., 2008).
Table 1. Luciferase (LUC) activity in stem tissue from mature transgenic plants expressing luc* under the control of PUbi, PA157, PH07 and PRol
Only lines with detectable LUC activity were included in this analysis.
As measured in a 500-s exposure. The strongest PUbi-driven lines were assayed using a 60-s exposure (to avoid pixel saturation), and the result was multiplied by 8.33 to obtain a 500-s equivalent.
7.9 × 104 ± 2.8 × 104
2.7 × 103 ± 2.0 × 103
1.6 × 103 ± 0.7 × 103
3.2 × 103 ± 2.0 × 103
PUbi drove LUC activity in leaves, young stems, mature stems and roots (Figure S1), with no significant difference in activity between young and mature internodes (Figure 2a). In contrast, PA157 and PH07A drove LUC activity preferentially in mature stems (Figure 2a,b) and to a lesser extent in roots, with no expression detected in leaf tissue (Figure S1). For both PA157 and PH07, activity in mature stem (internodes 10 and older) was significantly higher than that in young stem (P < 0.05; paired ratio t-test), and the patterns of LUC activity were consistent with the transcript levels for the corresponding genes (Casu et al., 2004; Mudge et al., 2009; Papini-Terzi et al., 2009).
PRol was intermediate in expression pattern between constitutive PUbi and the sugarcane-derived, mature-stem promoters. It drove significantly higher LUC activity in mature stem than young stem (P < 0.01; paired ratio t-test; Figure 2b), but showed substantial activity also in leaves and roots (Figure S1). Whereas activity of PA157 and PH07 increased significantly from internode 10 to basal internodes, that of PRol more closely reflected the pattern of sucrose accumulation in the stem, with activity nearing maximum levels by internode 10, which has near-complete sucrose loading (Rae et al., 2005; Wu and Birch, 2007). Sucrose-responsiveness of PRol has been demonstrated in dicot species (Nilsson et al., 1996; Yokoyama et al., 1994), which most likely explains the expression pattern observed in sugarcane.
Using the same silencing-resistant luc* reporter, we showed previously that most ScLSG promoters drove peak activity in internodes of the sucrose-loading zone and substantial activity in mature-stem internodes (Moyle and Birch, 2013). ScLSG promoter alleles varied in strength in the sucrose-loading internodes (up to 20% of PUbi) and in stem : root expression ratio (Moyle and Birch, 2013). Activity in stem storage parenchyma, coupled with low expression in off-target tissues (expanding internodes, leaves and roots), makes the PA157, PH07 and several ScLSG promoters interesting for transgenic traits including the conversion of stored sucrose into higher-value products. The higher off-target activity of PRol makes it less suitable for applications requiring expression only in sucrose-accumulating stem tissues.
Other sugarcane-derived promoters that drive transgene expression predominantly in stems of tested immature plants have recently been described (Damaj et al., 2010), but these promoters drive expression predominantly in vascular tissues and are therefore not likely to be useful for metabolic engineering applications aimed at conversion of stored sucrose into higher-value products.
High isomer yields in VG1 field trials with silencing-resistant transgenes
We tested the PUbi, PA157, PH07, PLSG1 and PRol promoters for engineered IM production, using an NTPP sequence encoding a vacuole-targeting leader peptide fused to native ims coding sequence or the silencing-resistant ims* sequence encoding the same SI protein. In the first vegetative generation, plants with high levels of IM under field or greenhouse conditions were obtained from expression of vacuole-targeted SI from each of the tested promoters (Tables 2 and 3).
Table 2. Isomaltulose (IM) production under field conditions, in whole-cane juice from transgenic sugarcane expressing NTPP-ims* controlled by different mature stem–expressed promoters
IM yield expressed as a molar concentration in juice, or as a proportion of total sugars in whole-cane juice, has been observed to increase in PUbi-NTPP-ims transgenic lines, for several months after total sugar levels peak in field-grown sugarcane (Basnayake et al., 2012). Therefore, caution is required in comparison of isomer yields between trials with different seasonal and age effects. The relative performance of different developmental promoters may also change at different plant ages within trials, but there are practical constraints on timing for commercial harvest of mature sugarcane. With these cautions in mind, it is interesting that PUbi-NTPP-ims* and PA157-NTPP-ims* performed similarly across total populations of 92–99 independent transgenic lines per construct tested in VG1 field trials harvested at 10–11 months of age in 2009 and 2010. The highest IM yields were in the range 253–339 mm, or 55%–60% of total sugars in whole-cane juice (Table 2). IM levels were significantly lower in populations with NTPP-ims* expressed from PH07 (P < 0.0001) or PRol (P < 0.05). Lines with NTPP-ims* expressed from PLSG1 were handled separately (in a containment greenhouse) during VG1 as described below.
The IM levels measured using PUbi-NTPP-ims* in VG1 field trials were substantially above those recorded previously using PUbi-NTPP-ims (up to 217 mm or 33% of total sugars, at 16 months of age in September 2007; Basnayake et al., 2012). Among 52 tested PUbi-NTPP-ims lines in the current study, the highest IM yield was 52 mm, which is within the range of 38–217 mm reported previously from 504 lines in six cultivars at 13–16 months after planting (Basnayake et al., 2012). Total sugar concentration was unaffected by IM level (Figure S2), as observed previously up to a lower IM yield from use of NTPP-ims under field conditions (Basnayake et al., 2012).
The value of constructs designed to eliminate triggers of silencing is even clearer with promoters that are more susceptible than PUbi to progression of the silencing cascade. There was no detectable IM in mature stems of any of 230 transgenic sugarcane lines with NTPP-ims driven by PA157 or PH07 (S.R. Mudge, unpublished data), in stark contrast to the results using these promoters with NTPP-ims* (Table 2).
Stability of the high-isomer trait over vegetative generations
Because of constraints on timing for field planting, PLSG1-NTPP-ims* lines were tested in VG1 under containment greenhouse conditions, and promising lines with IM comprising 63%–90% of sugars in middle internodes were selected for testing along with PUbi and PA157 lines in a replicated VG2 field trial. All selections retained high-level IM production in the replicated trial, and for each tested promoter, the highest isomer yield was in the range 70%–72% of total sugars (390–427 mm IM) in whole-cane juice at 11 months of age in September 2011 (Table 3).
To further analyse transgene expression stability in selected high-IM lines from the VG2 trial, we analysed IM content in five individual stalks from each of two stools, 11 months after planting. Isomer levels in individual stalks of a transgenic line were highly consistent, with a coefficient of variation of 5.2% for the PA157 promoter lines and 7.9% PLSG1 promoter lines (Figure 3). Greater variability in Ubi promoter lines (coefficient of variation 12%–37%) probably reflects the environmental responsiveness of this promoter (Hansom et al., 1999).
Germination and vigour of high-IM sugarcane lines
The high-IM lines described above appeared vigorous in VG1 and VG2 field trials, but reliable measurements of cane yield will require replicated, multirow, field trials after several cycles of field propagation (Basnayake et al., 2012). Germination and shoot height 4 weeks after planting of setts were unaffected by IM accumulation up to the maximum levels around 500 mm in juice from the internodes flanking these planted setts (Figure 4). This is understandable because sugarcane uses very little of the sugar in stalks for shoot germination or growth. High sugar levels in cultivars are a result of human selection for sweetness, not natural selection of stored reserves for vegetative propagation. Wild relatives such as Saccharum spontaneum germinate efficiently, with relatively little sugar accumulation in their stalks (Bull and Glasziou, 1963).
Relevance of ‘stem-specific’ promoters
Several transgenic lines with high IM levels at 11 months were chosen for analysis of developmental profiles of IM accumulation down the stalk at 13 months after planting. The selected lines were UQ08-4164, UQ09-4525 and UQ10-6337 (in which NTPP-ims* is driven by PUbi, PA157 and PLSG1, respectively) from Table 3. In each case, the isomer content increased with plant age (27%–38% increase between 11 and 13 months). The highest observed IM level in whole-cane juice was 483 mm (81% of total sugars) at 13 months. Although these promoters drive different patterns of (cytosolic) LUC reporter activity, all showed a similar pattern of increasing isomer content with internode age (Figure 5), consistent with previous observations from pUbi-driven expression of vacuole-targeted SI (Basnayake et al., 2012; Hamerli and Birch, 2011). We interpret this to indicate that the isomer accumulation pattern is strongly influenced by developmental changes in (i) the stability of the sucrose isomerase enzyme in the sugarcane vacuole, (ii) sucrose substrate concentration, and/or (iii) the kinetics of sucrose and IM hydrolysis. Some of these factors may be amenable to manipulation, to further increase IM accumulation in harvested stalks, while avoiding the possibility of a growth penalty from sucrose depletion in other developmental stages.
Despite the discovery that endogenous IM is a metabolizable sugar for plants (Wu and Birch, 2011), we remain cautious that high SI activity outside of the mature sugar-storage compartment could have adverse developmental effects. High-level cytosolic SI activity is very disruptive (Wu and Birch, 2007), and we consider that high-level vacuolar SI activity in tissues such as meristems, expanding internodes, leaves, roots and sucrose-transport tissues is also likely to be disruptive. The absence of obvious disruption to growth in lines expressing vacuole-targeted SI from the constitutive Ubi promoter is most likely explained by rapid inactivation of SI enzyme in the vacuoles (Wu and Birch, 2007), coupled with higher SAI activity in tissues that have not matured for sugar storage (Hatch and Glasziou, 1963; Vorster and Botha, 1999; Zhu et al., 2000). This is consistent with the low IM levels measured in these tissues, even in lines with high IM levels in mature internodes. Strategies to further increase IM levels through reduced proteolysis of SI in sugar-storage vacuoles are likely to require tighter developmental control of transgene expression.
Further agronomic testing of the high-IM lines is important. In this context, it is very encouraging that vigorous sugarcane lines with very high levels of IM accumulation can be obtained by expression of a silencing-resistant construct encoding a vacuole-targeted SI enzyme, from promoters that are preferentially expressed in the sucrose-loading and sucrose-storage tissues of sugarcane stems.
Production of IM at up to 483 mm (81% of total sugars) in vigorous sugarcane plants, grown under typical commercial field conditions, is perhaps the clearest demonstration to date of the potential through metabolic engineering to enhance crop plants for the production of materials valuable to human consumers. In the case of IM, the product is a natural nutritional sugar with advantages of lower glycaemic index due to slower digestion, and lower cariogenicity than sucrose.
At the current levels, direct production of IM in plants appears to be feasible at a cost approaching that of sucrose. This should make the benefits of IM affordable on a much wider scale, compared with the current expensive production by conversion from sucrose in fermentation facilities. Taking into account the differences in properties, such as solubility, of other sucrose isomers including TH that can be obtained using different SI enzymes (Hamerli and Birch, 2011), it seems inevitable that sucrose isomers produced directly in plants like sugarcane or sugar beet will substantially replace sucrose for many uses where the individual isomers and their blends have preferred properties.
Sampled aerial organs of sugarcane were numbered from +1 for the top visible dewlap, attached leaf, node and internode below, with consecutively higher numbers for older (lower) phytomers (Van Dillewijn, 1952).
RNA extraction and Northern blot analysis
RNA was extracted from field-grown sugarcane cultivar Q117 and used for northern blot analysis with a probe corresponding to EST MCSA222H07 (Casu et al., 2004), which encodes part of a CIPK protein. As detailed in the results, northern blot analysis showed that transcripts of this ScCIPK gene were much more abundant in mature internodes than in young internodes, and they were not detectable in leaves.
Isolation of promoter regions from sugarcane
To amplify Q117 genomic sequences upstream from the EST MCSA222H07 sequence, we used three rounds of GenomeWalker PCR (Clontech, Mountain View, CA) on libraries prepared using restriction enzymes DraI, EcoRV, PvuII, ScaI, StuI, BsrBI, Ecl136II and SspI. Several promoter haplotypes were amplified, the largest of which spanned 2.1 kb. This promoter (GenBank JX514704) is referred to as PH07A. In addition, we isolated from a sugarcane Q200 BAC clone a longer (5.7 kb) version of the ScR1MYB1 A1 promoter (Mudge et al., 2009). This sequence (GenBank JX514703) is referred to as PA157. Isolation and analysis of ScLSG promoters, which drive expression preferentially in sucrose-loading and mature stem regions, are described (Moyle and Birch, 2013).
Expression vectors and sugarcane transformation
Isolated PH07A and PA157 promoter regions were inserted into a promoter-test vector, to express the firefly luciferase reporter protein (LUC) from a coding sequence luc* (GenBank KC147725) that avoids motifs which trigger efficient silencing in sugarcane (Moyle and Birch, 2013). Control promoter regions were constitutive PUbi from maize ubi1 (Christensen and Quail, 1996) and sucrose-inducible PRol from Agrobacterium rolC (Yokoyama et al., 1994).
For vacuole-targeted SI activity, we used the sporamin NTPP signal peptide fused to the SI enzyme as described (Wu and Birch, 2007). We refer to the native bacterial coding sequence for the mature UQ68J SI enzyme (Wu and Birch, 2005) as ims and use ims* for a sequence (GenBank KC147726) encoding the same protein while avoiding motifs that trigger efficient silencing in sugarcane. PUbi, PA157, PH07, PLSG1 and PRol promoters were used to drive expression of NTPP-ims*. The expression cassette in all cases included three contiguous plant transcriptional terminator regions intended to block read-through transcription in either direction.
Constructs were transferred into sugarcane recipient genotypes Q117, Q138 and KQ228, with conditions for selection of transgenic lines, plant regeneration, and growth under glasshouse and field conditions as described (Basnayake et al., 2012).
In vivo luciferase assays
LUC activity was measured in transverse internode sections, leaf pieces and detached roots by immersing samples in 0.4 mm luciferin and imaging as described (Mudge et al., 1996), except that light emission was captured using a PIXIS 102A camera and WinView32 software (Princeton Instruments, Trenton, NJ).
The average pixel intensity within a box covering the brightest region of each internode (approximately 500 pixels) was calculated, and the average background level (from a nearby area without any tissue sample) was subtracted. To compare developmental expression patterns across lines with diverse peak expression levels, the raw data from internode samples were transformed into relative expression levels, by expressing average pixel intensity in each sample as a percentage of that in the basal internode (typically internode 25). For statistical analysis of the data relating to tissue specificity, we used a paired ratio t-test (Graphpad Prism, La Jolla, CA). This allows comparison between tissues across transgenic lines with very different absolute expression levels that can arise through effects of transgene copy number and integration position.
Field trials and sugar analysis
All field trials described in this report were undertaken at Sucrogen Kalamia Estate in North Queensland and conducted in accordance with licences for limited field release of specified GM sugarcane lines, issued by the Australian Office of the Gene Technology Regulator. Depending on the timing of regeneration in relation to suitable conditions for field planting during 2008–2009, transgenic plantlets were either grown for a first vegetative generation (VG1) in containment greenhouse conditions, or taken to Kalamia in tissue culture and established in pots prior to field planting as a single stool per independent transgenic line. A replicated, randomized trial of selected high-IM lines was planted in 2010 using setts derived from greenhouse- or field-grown plants. Trials were sampled at times detailed in the results. Unless specified otherwise, juice was extracted from sampled stalks and analysed for sugar composition as described (Basnayake et al., 2012).
We speculated that high isomer accumulation at the expense of sucrose might be detrimental for bud germination and early shoot growth, in the routine vegetative propagation of sugarcane from nodal cuttings (setts) taken from mature stalks. To test this concern, 19 independent transgenic lines with high IM levels (including six lines each with PUbi, PA157, and PLSG1 constructs, and a single PH07 line), four transgenic lines with low IM levels and three parental cultivars with no IM in juice were sampled from the field trials (three replicate stalks). Juice was collected from each internode, using a Carver press, and analysed for sugars by HPLC (Wu and Birch, 2007). Corresponding nodal regions were planted as single-eye setts in peat pots in a randomized block layout in a greenhouse and assessed at intervals for germination and early growth of shoots.
The authors acknowledge the excellent technical assistance of Asha Kakkanat, Lilian Chou and members of the Sucrogen Technical Field Department. This research was supported through a collaboration between CSR Sugar Limited (Sucrogen) and The University of Queensland under the Australian Research Council's Linkage scheme.