Over-expression of AtPAP2 in Camelina sativa leads to faster plant growth and higher seed yield

Background Lipids extracted from seeds of Camelina sativa have been successfully used as a reliable source of aviation biofuels. This biofuel is environmentally friendly because the drought resistance, frost tolerance and low fertilizer requirement of Camelina sativa allow it to grow on marginal lands. Improving the species growth and seed yield by genetic engineering is therefore a target for the biofuels industry. In Arabidopsis, overexpression of purple acid phosphatase 2 encoded by Arabidopsis (AtPAP2) promotes plant growth by modulating carbon metabolism. Overexpression lines bolt earlier and produce 50% more seeds per plant than wild type. In this study, we explored the effects of overexpressing AtPAP2 in Camelina sativa. Results Under controlled environmental conditions, overexpression of AtPAP2 in Camelina sativa resulted in longer hypocotyls, earlier flowering, faster growth rate, higher photosynthetic rate and stomatal conductance, increased seed yield and seed size in comparison with the wild-type line and null-lines. Similar to transgenic Arabidopsis, activity of sucrose phosphate synthase in leaves of transgenic Camelina was also significantly up-regulated. Sucrose produced in photosynthetic tissues supplies the building blocks for cellulose, starch and lipids for growth and fuel for anabolic metabolism. Changes in carbon flow and sink/source activities in transgenic lines may affect floral, architectural, and reproductive traits of plants. Conclusions Lipids extracted from the seeds of Camelina sativa have been used as a major constituent of aviation biofuels. The improved growth rate and seed yield of transgenic Camelina under controlled environmental conditions have the potential to boost oil yield on an area basis in field conditions and thus make Camelina-based biofuels more environmentally friendly and economically attractive.


Background
Crop plants obtain energy entirely by photosynthesis; this autotrophic mechanism has long been a target of research aiming to maximize crop production [1]. Sugars derived from photosynthesis are energy sources for metabolism and are building blocks of complex carbohydrates [2]. Sugars also regulate expression of photosynthetic genes in mesophyll and interact with other signaling pathways involved in plant growth regulation [3]. Previous studies suggest that photosynthetic performance can be improved by modifying sugar-signaling mechanisms [3]. A large number of protein kinases, protein phosphatases and transcription factors are involved in these pathways [4]. Constitutive expression of Arabidopsis purple acid phosphatase 2 (AtPAP2) by the CaMV35s promoter in Arabidopsis enhances plant growth and seed yield by modulating carbon metabolism [5]. Both sucrose phosphate synthase (SPS) activity and sucrose content are elevated in leaves of transgenic lines. Overexpressed lines grow faster and produce > 50% more seeds per plant than wild type [5]. AtPAP2 is targeted to both chloroplasts and mitochondria by its C-terminal hydrophobic motif; removal of this motif eliminates growthpromoting phenotypes, indicating that its subcellular localization within the two key energy-generating organelles in plants is important for its biological functions [5]. Purple acid phosphatases (PAPs) constitute a family of acidic binuclear metalloenzymes that hydrolyse phosphate esters and anhydrides under acidic conditions [6]. Plant PAPs are generally considered to mediate phosphorus acquisition and redistribution through their ability to hydrolyze phosphorus compounds [7,8]. Some PAPs have biological functions other than those involved in P metabolism. For example, certain PAPs, such as GmPAP3 [9] and AtPAP17 [10], have peroxidase activity. The tobacco purple acid phosphatase NtPAP12 dephosphorylates phosphoryl residues of apoplasticproteins [11]. The enzyme is actively transcribed in the process of cell wall regeneration [12]. AtPAP2 is currently the only PAP known to be involved in carbon metabolism.
Arabidopsis thaliana and Camelina sativa are members of the Angiosperm Family Brassicaceae [13]. Camelina sativa, or false flax, has been cultivated for oil in Europe, Central Asia and North America. Compared with rapeseed, the cultivation of Camelina has become vanishingly small since World War II [14]. Currently, Camelina is being grown to produce aviation biofuels. To date, Japan Airline, Air New Zealand and the US Navy (F/A18) have successfully completed test flights with a 50/50 blend of Camelina-based biofuel and petroleum-based jet fuel. With its low requirements for water and fertilizers, Camelina could be grown on marginal land as a low-input biofuel crop [15]. With an oil content of > 40% polyunsaturated fatty acids (linoleic acid, 18:2n-6 and α-linolenic acid, 18:3n-3), of which, linolenic acid makes up~35%, Camelina seed is also an important source of omega-3 fatty acids [16].
Through overexpression of AtPAP2, transgenic Camelina lines grow faster and produce more seeds than wild-type (WT) plants under controlled environmental conditions. These phenotypes are attributable to elevated photosynthetic rates and SPS activity in leaves. Higher seed yield would make cultivation of Camelina more economical and environmentally-friendly. High seed yield per unit area would enhance biofuel production and reduce consumption of fossil fuels, thus reducing CO 2 , methane (CH 4 ) and nitrous oxide (N 2 O) emissions.

AtPAP2 overexpression accelerated Camelina growth
Multiple independent overexpression lines of Camelina sativa with AtPAP2 protein expression were obtained through Agrobacterium transformation. After confirmation by genomic PCR and western blotting (Figure 1), three independent T2 AtPAP2 overexpression (OE) lines were selected for characterization. When there was a 3:1 segregation of T 1 transgenic seeds, null-lines were selected for their relatively short hypocotyl lengths on MS plates. Identities of the three null-lines were confirmed by Figure 1 Growth phenotype of transgenic Camelina at day 40. a, All three OE lines produced more branches, grew faster, and flowered earlier than the WT and null-lines. b, Expression of AtPAP2 in OE lines was confirmed by Western blotting using AtPAP2-specific antibodies. OE lines had higher expression of AtPAP2, while there was a homologous protein of AtPAP2 in Camelina.
genomic PCR and western blotting ( Figure 1). A weak western blot signal was observed in protein extracts from the null lines and WT. We attribute this signal to likely cross-reactivity between the AtPAP2 antibody and the endogenous AtPAP2-like protein (91% amino acid sequence identity) that has been detected in transcriptome analysis of Camelina sativa leaves (unpublished data). All plants were grown on MS agar for a week and then transferred to a growth chamber (150 μmol photons/m 2 /s) under a 16 h light 22°C/8 h dark 18°C photoperiod/temperature regimen. The plants were transferred immediately after first flowering to greenhouse conditions at 22°C under Hong Kong's natural day/night photoperiod. Compared to WT and their respective null-lines, all three AtPAP2 overexpressor lines (OE6, OE14 and OE20) grew faster and produced more lateral branches ( Figure 1). Typically, plants of the three OE lines started flowering 7-9 days earlier than WT and null-lines in all three generations. Only the results of the T3 generation are presented in Table 1 for statistical reason. When the WT and three null-lines first flowered at an age of 40-d, AtPAP2 OE lines had already produced more branches (Figure 1).

Longer hypocotyls in OE lines
Homologous seeds of three OE lines, three null-lines and the WT were germinated on MS plates and the lengths of hypocotyls and roots were measured 5 days after germination. The three OE lines had longer hypocotyls than their respective null-lines and WT, but there was no significant variation in lengths of roots among lines ( Figure 2). The longer hypocotyls may have resulted from the larger seed size in overexpressor lines ( Figure 3).

OE lines produced larger seeds and had higher seed yields
Seeds were collected from all lines after full maturity. Seed weights and sizes were measured after drying. The transgenic lines produced larger seeds (Figure 3a-g) and the hundred-seed-weights from OE plants were 501 10% higher than those of WT and null-lines ( Figure  3h), which may be attributable to an elevated supply of photoassimilates in OE plants. Larger seeds may provide more carbon for seed germination and seedling growth, with correspondingly longer hypocotyls ( Figure 2). Generally, total seed yield of OE lines was significantly higher than WT or null-line yields ( Table 2), which we attribute to the higher photosynthetic rate in the overexpressors (Table 3). There was no significant difference in germination rates among WT, null-lines and the OE lines (data not shown).

Higher photosynthetic rates and stomatal conductances in OE lines
Photosynthetic activity and stomatal conductances were measured with a Licor LI-6400XT Portable Photosynthesis System. Third mature leaves below the shoot tips of all plants (35-37-d old) grown in the growth chamber were used for measurements. The OE lines had higher rates of photosynthesis than WT plants and null-lines at photon flux densities over 200 μmol m -2 s -1 (Table 3). In general, photosynthetic activities per unit leaf area were elevated by 10-25% in overexpressors. As the OE lines had both more branches (and thus greater leaf areas) and higher photosynthetic rates per unit leaf area, each of the OE plants was able to fix more CO 2 than plants of other lines. In photosynthetic leaves, stomatal conductance is coordinated with CO 2 requirements of the mesophyll [17]. In our test system, transgenic plants had greater stomatal conductances than WT and null-line plants, in congruence with the elevated photosynthetic rates of overexpressors (Table 4).

Lower sucrose content in leaves of OE lines
Sugars and sugar-derived metabolic signals affect source-sink balances and carbon partitioning in plants [3,18]. Sugars trigger global repression of photosynthetic gene transcription in plants [4]. To determine whether there were changes in sugar level, we used HPLC methodology to analyze sugar content in leaves of 35-day old Camelina. While hexose contents of three AtPAP2 overexpressors were not significantly different from those of the WT and three null-lines, the sucrose contents in leaves of all three transgenic lines were 20~30% lower in the transgenic plants ( Table 5). The low levels of sucrose in leaves may have enhanced photosynthesis, growth rate and biomass accretion [19].

Increased SPS activity in OE lines
Sucrose phosphate synthase (SPS) is a key enzyme in the sucrose synthesis pathway; it catalyzes conversion of frucose-6-phosphate (Fru6P) and UDP-Glucose (UDPG) to sucrose-6-phosphate [20]. Overexpresssion of SPS in Arabidopsis [21] and tobacco [22] affects leaf carbohydrate contents and whole plant development. The phenotypes of Camelina AtPAP2 OE lines may be attributable to enhanced enzyme activity during sucrose synthesis. SPS activity was elevated in leaves of AtPAP2 OE plants at both optimal V max and limiting V limit capacities (Table 6). Higher photosynthetic rates (Table 3) and SPS activity in OE lines may have supplied more sucrose for growth.
Unaltered protein expressions of FBP aldolase SnRK1.1 (and its phosphorylation status), nitrate Reductase, and FBPase in OE lines SnRK1 protein kinase in plants balances cellular energy levels by regulating ATP production and consumption; this enzyme is activated in response to carbon starvation and energy-depleting stress conditions [23]. SnRK1 phosphorylates central enzymes involved in primary metabolism (e.g. SPS, nitrate reductase), leading to their inactivation [3,24]. We analyzed protein expression levels of SnRK1 and its phosphorylation status by western blotting (Additional file 1). There were no differences in SnRK1 protein expression level or its phosphorylation status among the lines, indicating that higher SPS activity in the OE lines had not resulted from the action of SnRK1 on SPS. Western blotting demonstrated that protein expression levels of other metabolic enzymes, including nitrate reductase (NR), FBPase and FBP aldolase, were unaltered in transgenic plants (Additional file 1). Because anti-AtSPS antibodies from Agrisera (Vännäs, Sweden) did not cross-react with Camelina SPS, we were unable to determine the levels of SPS protein.
Life cycle analysis (LCA) of Camelina-based biodiesel (green diesel) produced from seeds of genetically unaltered and transgenic lines SimaPro LCA software is commonly used to construct LCAs of biodiesel produced from different feedstocks. To build a complete life cycle flow tree for the analysis, it is necessary to first build individual life cycle processes in SimaPro. These are then combined into an assembly, and finally into the life cycle as a whole. The functional unit, as well as the point of junction, is one megajoule (MJ) of energy output [25]. Additional file 2 depicts the life cycle flow tree of Camelina-derived fuel from Group 1, namely the group of Camelina seed that were genetically unaltered.
The outputs of genetically modified Camelina seed yields ha -1 were assumed to increase by 10% (Group 2), 20% (Group 3), 30% (Group 4), 50% (Group 5) and 100%    Figure 4 shows GWP, and presents emissions of GHG in terms of CO 2 equivalents for the six Camelina cultivation scenarios. The declining GWP from Group 1 through Group 6 demonstrates that emissions of CO 2 equivalent per MJ energy of oil output decreased with increasing seed yields. Table 7 compares CED across the six scenarios. There is a similar slightly decreasing trend from Group 1 to Group 6, indicating a moderate decrease in energy demand with increasing seed yields. As mentioned, seed yields of OE6 and OE14 lines were improved by about 100% in greenhouse conditions; these two lines would thus fall into Group 6 of the simulation. The LCA results for Group 6 indicate 8.2% and 2.3% reductions in GWP and CED, respectively, in comparison with Group 1. Thus, transgenic lines would be more environmentally friendly if yield improvement is reproducible under field conditions.

Discussion
Sugar metabolism plays important roles in vegetative growth and reproductive development of plants. Sugars affect expression of genes that encode carbon metabolic enzymes [3,26]. Developmental processes, such as flowering [27], fruiting [28] and the sink to source transition [29], may be improved by overexpression of SPS. In the present study, overexpression of Arabidopsis AtPAP2 in transgenic Camelina boosted activity of SPS in leaves, similar to the effects in transgenic Arabidopsis [5]. Although SPS activity in leaves was elevated, there was a surprising decrease in leaf sugar contents in the OE lines, contrary to results for transgenic Arabidopsis [5]. The sucrose levels in leaves of 35-d old Camelina were reduced by 20%, suggesting an elevated sucrose export rate in the transgenic lines. The rate of sucrose synthesis under the control of SPS is correlated with the rate of export from leaves [30] and also with the rate of photosynthesis [31]. Higher photosynthetic rates lead to increases in leaf sucrose, while the rate of sucrose export from leaves determines the abundance of sucrose at the source [32]; low leaf sucrose content in overexpressed Camelina lines may have resulted from highly efficient transport of sucrose. High sucrose content in leaves suppresses photosynthesis. Reduced sucrose content in source leaves may alleviate feedback inhibition of photosynthesis and SPS activity caused by high sugar concentrations. Overall, this would enhance carbon fixation and SPS activity. As the major transport carbohydrate in plants, sucrose links carbon assimilation in source tissues to carbon utilization in sink tissues [18,33]. Elevated photosynthesis produces more carbon skeletons, while enhanced efficiencies of sucrose synthesis and transport increase sugar supply to sinks, resulting in earlier flowering, faster growth and improved seed yield in transgenic lines.
LCA analysis indicated that improved seed yield of transgenic Camelina sativa lines may prove useful in the production of resource-economical green biodiesel with reduced Cumulative Energy Demand and Global Warming Potential. Compared to petroleum jet fuel, the GHG emission of Camelina-based biofuels was reduced by 68-75% (> 67%) at the current yield level from genetically unaltered plants. This could be further improved to 75-80% by increasing seed yield per unit area. While the transgenic Camelina sativa produced a higher seed yield under  controlled environmental conditions, whether the transgenic lines could perform well under field conditions needs further experiments. Growth density, availability of water, competition of nutrition, pest damages, may adversely affect the growth and yield of the transgenic lines. Nonetheless, with expected future gains in seed yield, Camelina production in USA has the potential to provide three billion liters per year of high quality, environmentally-friendly jet fuel [25]. Commercial use of Camelina-based jet fuel is expected in 2012.

Conclusion
Overexpression of AtPAP2 promoted Camelina growth through effects on carbon assimilation and distribution. More importantly, improved seed yield could enhance the value of Camelina as a low-input crop for biofuel production and for various industrial and agricultural applications.

Plant materials, growth conditions, and vector
Camelina seeds were thrice surface-sterilized with 20% (v/v) bleach for 15 min; the seeds were then washed with sterile distilled water thrice before plating onto MS [34] medium supplemented with 2% (w/v) sucrose for 5 d.

Agrobacterium-mediated transformation
Six-week old WT Camelina plants that first started flowering in the growth chamber were used for transformation following methodology of Lu et al. [14]. After first transformation, new flowers in inflorescences were used for second and third transformations to improve the transformation ratio. After the third transformation, new inflorescences were cut to reduce background.

Generation of Camelina overexpression lines and nulllines
Homologous AtPAP2 transgenic lines of Camelina were selected on Basta plates (5 g/l Basta, Riedel-deHaen, Germany, 500 g/l carboncillin, 100 seeds/plate); WT Camelina was used as the negative control. Resistant plants were transferred to soil for growth to maturity. Their transgenic status was confirmed by genomic PCR and western blotting analyses using anti-AtPAP2 antibodies. Homozygous T 3 seeds of the transgenic plants were used for further analyses. When there was 3:1 segregation of T 1 transgenic seeds on Basta plates, nulllines were selected by their shorter hypocotyl lengths on MS plates. Identities of the three null-lines were confirmed by genomic PCR.

Seed viability, root and hypocotyl length measurements
Camelina seeds were surface sterilized with 20% (v/v) bleach for 15-20 min, washed with sterile distilled water and then plated on MS medium (50 seeds/plate). The  experiment was carried out in a growth room (~100 μmol photons/m 2 /s) under a 16 h light (22°C)/8 h dark (18°C) temperature/photoperiod regimen. The lengths of roots and hypocotyls were measured 5 d after germination. Viability tests on Camelina seeds were performed in 10-cm diameter Petri plates each containing 20 ml of MS medium supplemented with 2% (w/v) sucrose; results were recorded after 5-6 d.

Western blot analysis
Camelina leaves were finely ground in 1.

Measurement of photosynthetic rates and stomatal conductances
Photosynthetic rates and stomatal conductances were measured using a portable photosynthesis system (LI-COR, LI-6400, Nebraska, USA) in the morning (08:30-12:30) under a fixed blue-red light-emitting diode (LED) light source. Ten measurements were made on each fully expanded leaf collected from the shoot tips of 35-37-d old plants; at least 3 plants from each line were used for   measurements. The leaf area of the standard broadleaf chamber was 6 cm 2 (2 × 3 cm) and a fixed light source was provided by an array of red and blue LEDs. P vs. I curves were plotted using the instrument's auto-program function. Measurements were taken at light intensities of 200, 400, 600, 800 and 1000 μmol photons/m 2 /s. Photosynthetic rate (A) was calculated as A = F(Cr-Cs)/ 100S-CsE, where A = net assimilation rate (μmol CO 2 / m 2 /s 1 ), F = molar flow rate of air entering the leaf chamber (μmol/s), Cr = mole fraction of CO 2 in the reference IRGA (infra-red gas analyzer) (μmol CO 2 /mol air), Cs = mol fraction of CO 2 in the sample IRGA (μmol CO 2 /mol air), S = leaf area (cm 2 ), and E = transpiration rate (mol H 2 O/m 2 /s).
Stomatal conductance is the rate at which water evaporates from stomata; it is directly related to the relative size of the stomatal aperture. Conductance is calculated from gsw = 1/(1/gtw-Kf/gbw), where gsw = stomatal conductance to water vapor (mol H 2 O/m 2 /s), gtw = total conductance to water vapor (mol H 2 O/m 2 /s), Kf = stomatal ratio, and gbw = boundary layer conductance to water vapor (μmol H 2 O/m 2 /s).

HPLC analysis of sugar content
For measurement of sucrose and hexose of plant tissues, 0.1 g aliquots of freeze-dried tissue powder were resuspended in 1 ml of 70% (v/v) ethanol, incubated at 70°C for 90 min and centrifuged at 13,000 × g for 10 min. After passing through a 0.22 mm filter, a 10 μl sample was injected into a CarboPac PA 1 column (4 × 250 mm) connected to a Dionex (Sunnyvale, CA, USA) LC 20 Chromatography system and sugar contents were analyzed by high performance anion exchange chromatography with pulsed amperometric detection [36]. Standard curves were prepared with 0-0.1 mg/ml sucrose, fructose and glucose solutions in 70% (v/v) ethanol.
Implementation of SimaPro software for life cycle analysis of green diesel from unaltered Camelina and its genetically modified counterparts SimaPro was used to construct a LCA (life cycle analysis) of bio-diesel from wild type and genetically modified plants with different seed yields. The general processes for initiating a LCA study in SimaPro include: i. Determining the goal and scope of the specific study, which define the LCA system boundary. ii. Building of the life cycle flow tree, which specifies each stage involved in the life cycle. All environmental inflow and outflow inventory data are collected at this stage to build the life cycle processes; relevant life cycle assembly is then undertaken, followed by final construction of life cycles. Waste scenarios may be added into the life cycle flow tree, if available. iii. Choosing appropriate life cycle impact assessment methodology to calculate prospective results. iv. Interpreting results.

Fuel pathways
Life cycle analysis was conducted for biofuel derived from unmodified Camelina seeds (Group 1) and for biofuels derived from genetically modified Camelina with enhanced seed yields. We assumed that all input parameters were fixed and that the outputs of genetically modified Camelina seed yield increased by 10% (Group 2), 20% (Group 3), 30% (Group 4), 50% (Group 5) and 100% (Group 6) over output from the unmodified Camelina farming process. The whole life cycle was divided into several stages, including Camelina farming, seed oil production, refined oil production and end use oil output, which also set the system boundary defined for the study. The life cycle stages were used to organize data for software calculations.

Functional unit and life cycle inventory
The function unit in this study was one MJ of energy content in the fuel output. The choice was appropriate since energy content is the fundamental measurement for all environmental and energy flows in the life cycle analysis. The life cycle inventory for every stage came from the public Camelina database for Camelina in Montana, USA and the database in SimaPro. The input data for Camelina farming, seed oil extraction, refined oil production and transportation are shown in Additional file 3, 4 and 5.
As shown in Additional file 3 Group 1 represented the original Camelina seed yield, and Groups 2, 3, 4, 5 and 6 represented the yield increases in genetically modified plants, with seed yields increasing by 10%, 20%, 30%, 50% and 100% per unit area, respectively. The cumulative inputs of potassium (as K 2 O), thomas meal fertilizer (as P 2 O 5 ) and urea-N were fixed for the one-hectare farming input. The quantity of fossil fuel diesel used remained unchanged on a unit area basis but changed with changing Camelina seed output. As for GHG emissions, methane and dinitrogen monoxide emissions per kilogram of seed output remained unchanged. Carbon dioxide emissiona from diesel changed linearly with changing fossil diesel consumption, but at the same time remained fixed per unit area farmland output.