Metabolic engineering of medium-chain fatty acid biosynthesis in Nicotiana benthamiana plant leaf lipids

Various research groups are investigating the production of oil in non-seed biomass such as leaves. Recently, high levels of oil accumulation have been achieved in plant biomass using a combination of biotechnological approaches which also resulted in significant changes to the fatty acid composition of the leaf oil. In this study, we were interested to determine whether medium-chain fatty acids (MCFA) could be accumulated in leaf oil. MCFA are an ideal feedstock for biodiesel and a range of oleochemical products including lubricants, coatings, and detergents. In this study, we explore the synthesis, accumulation, and glycerolipid head-group distribution of MCFA in leaves of Nicotiana benthamiana after transient transgenic expression of C12:0-, C14:0-, and C16:0-ACP thioesterase genes. We demonstrate that the production of these MCFA in leaf is increased by the co-expression of the WRINKLED1 (WRI1) transcription factor, with the lysophosphatidic acid acyltransferase (LPAAT) from Cocos nucifera being required for the assembly of tri-MCFA TAG species. We also demonstrate that the newly-produced MCFA are incorporated into the triacylglycerol of leaves in which WRI1 + diacylglycerol acyltransferase1 (DGAT1) genes are co-expressed for increased oil accumulation.


Introduction
Vegetable oils, in the form of triacylglycerol (TAG), are an important global commodity with over 150 million tonnes being produced every year (FAO, 2003), which has remained relatively constant over recent years. Vegetable oils are an essential resource for various industries including, human and animal nutrition, the chemical industry and renewable energy. Due to demand for these plant oils, it is predicted that by 2020 worldwide production of vegetable oil is required to increase by 30% (Thoenes, 2011). This increase in production is likely to come in part from increases in arable land at the expense of endangered habitats.
The use of petroleum is widely accepted as being unsustainable. Therefore, it is necessary to develop alternative and sustainable resources to supplement or replace the extensive use of petroleum. In an attempt to expand global vegetative oil production, which currently relies on oilseeds and palm oil, various laboratories are now investigating the production of oil in non-seed biomass such as leaves. A variety of metabolic engineering approaches have resulted in increased biomass oil (Vanhercke et al., 2014b;Wood, 2014) including reducing TAG and fatty acid turnover (Slocombe et al., 2009), ectopic expression of transcription factors that regulate processes of seed development and maturation (Santos Mendoza et al., 2005), down-regulation of metabolic pathways that compete for available carbon (Sanjaya et al., 2011), and the over-expression of acetyl-CoA carboxyalse (ACCase) (Klaus et al., 2004). Furthermore, disruption of SUGAR-DEPENDANT1 (SDP1) in Arabidopsis thaliana, a lipase related pathway, also resulted in increased TAG accumulation in non-seed tissues (Kelly et al., 2013). Recently, much higher levels of oil accumulation in plant biomass was achieved using a combination of biotechnological approaches (Vanhercke et al., 2014a). This accumulation to 15% TAG in Nicotiana tabacum leaves was achieved by the coordinated transgenic expression of the WRINKLED1 transcription factor (WRI1), diacylglycerol acyltransferase (DGAT1) and oleosin genes. This breakthrough in leaf oil accumulation was ascribed to the synergistic increase in both fatty acid synthesis and oil synthesis (via WRI1 and DGAT1, respectively) (Vanhercke et al., 2013) and the formation of stabilized oil bodies (via oleosin). Not only was oil accumulated in leaves, the composition of the leaf oil was dramatically altered, including a reduction in α-linolenic acid in favor of oleic acid.
An important component of global oil consumption is medium-chain fatty acids (MCFA), namely fatty acids with less than 18 carbons, including those with 12, 14, and 16 carbons in length. These MCFA are ideal sources of biodiesel and also for a range of oleochemical feedstocks including, coatings, and detergents. The biochemical pathways for the production of MCFA have largely been elucidated. In plants, de novo fatty acid biosynthesis occurs in the plastid stroma, with the plastidial elongation of acyl chains being terminated by the activity of acyl-acyl carrier protein (acyl-ACP) thioesterases (Tjellstrom et al., 2013). These enzymes are therefore considered to be major determinants of fatty acid chain lengths, and hence the overall fatty acid profile of plants. However, it has been demonstrated that acyl-ACP synthetases can reactivate released fatty acids, hence allowing further elongation (Koo et al., 2005). Acyl-ACP thioesterases are typically classed as being either FatA or FatB, which are responsible for the release of oleic acid (C18:1 9 ), stearic acid (C18:0), palmitic acid (C16:0) and the MCFA, which in this study principally refers to lauric acid (C12:0), myristic acid (C14:0), and palmitic acid (C16:0) (Dörmann et al., 2000). Following export from the plastid, these fatty acids can be esterified to a glycerophosphate backbone via the Kennedy pathway in the endoplasmic reticulum to form TAG. Plant biotechnologists have investigated the accumulation of C12:0 and C14:0 in seed and leaf tissues of transgenic Brassica napus by the transgenic constitutive expression of a C12:0-ACP thioesterase from California Bay Laurel (Umbellularia californica) (Eccleston et al., 1996). This study demonstrated that significant levels of C12:0 could be accumulated in mature B. napus seeds. However, very low levels of C12:0 were observed in leaf tissue, despite high levels of C12:0-ACP thioesterase expression and in vitro activity. Similar results were obtained when the same gene was transformed in A. thaliana (Voelker et al., 1992). This research was extended by the co-expression of the Cocos nucifera lysophosphatidic acid acyltransferase (LPAAT) and thioesterase which resulted in an increased accumulation of total C12:0 as well as an increased fraction of trilaurin in the seeds of B. napus (Knutzon et al., 1999). Therefore, it can be observed that the expression of thioesterases and acyltransferases, enzymes involved in the biosynthesis of fatty acids and TAG, can be used to increase the accumulation of specific fatty acids, and hence tailor the overall fatty acid profile of plants.
Here, we explored whether MFCA production is compatible with high oil production in leafy biomass. We used a combinational expression platform to combine WRI1 and DGAT1 with various thioesterases and oil assembly enzymes to produce MCFA in leaf oil. A variety of thioesterases were found to drive the production of 12-, 14-, and 16-carbon MCFA in the leaves of Nicotiana benthamiana. The co-expression of AtWRI1 in particular increased the levels of leaf oil containing MCFA. Further improvements in MCFA accumulation of WRI1-dependent leaf oils was attained by using the CnLPAAT. High resolution LC/MS techniques were used to trace the accumulation of MCFA from their site of synthesis to their eventual deposition into oil. The biochemical pathways for the accumulation of MCFA in biomass oils and implications of production in stably-transformed crops are discussed.

N. benthamiana Transient Assay
Transient expression in N. benthamiana leaves was performed as previously described (Wood et al., 2009), with some minor modifications. A. tumefaciens strain AGL1 harboring each binary vector (p19 viral suppressor protein and the chimeric gene of interest) was grown at 28 • C in LB broth supplemented with the appropriate antibiotics. Cultures were centrifuged and gently resuspended in two volumes of infiltration buffer (5 mM 4morpholineethanesulfonic acid (MES), 5 mM MgSO4, pH 5.7, 100 µM acetosyringone) and grown for a further 3 h. The optical densities of each culture were measured and adjusted to a final OD 600 equal to 0.1 prior to infiltrations. The final mixture of AGL1 cells was infiltrated via syringe into the underside of leaves of 5 week old N. benthamiana plants (Voinnet et al., 2003). Infiltrated areas of leaves, commonly ∼3 cm in diameter, were circled by a permanent marker. Infiltration experiments were conducted in quadruplicate using different leaves of similar age. The samples being compared were randomly located, with a p19 control infiltrated for each plant. After infiltrations, the N. benthamiana plants were grown for a further 5 days before leaf discs were cut from the infiltrated areas (included the injection zone), freeze-dried, weighed and stored at −80 • C. Each analyzed sample consisted of discs from four independent infiltrations.

Total Lipid Extraction and Fatty Acid Profile Analysis
Total lipids were extracted from freeze-dried N. benthamiana leaves. Freeze dried leaf tissue was ground to powder in a microcentrifuge tube containing a metallic ball using Reicht tissue lyser (Qiagen) for 3 min. at 20 frequency/s. Chloroform:methanol (2:1, v:v) was added and mixed for a further 3 min. on the tissue lyser before the addition of 1:3 (v:v) of 0.1 M KCl. The sample was then mixed for a further 3 min. before centrifugation (5 min. at 14,000 g), after which the lower lipid phase was collected. The remaining phase was washed once with chloroform, and the lower phase extracted and pooled with the earlier extract. Lipid phase solvent was then evaporated completely using N 2 gas flow and the lipids resuspended in 5 µL chloroform per mg of original dry leaf weight.
Fatty acid methyl esters (FAMEs) of total lipids were produced by incubating extracted lipid in 1 N methanolic-HCl (Supelco, Bellefonte, PA) at 80 • C for 3 h. FAMEs were analyzed by an Agilent 7890A gas chromatograph coupled with flame ionization detector (GC-FID, Agilent Technologies, Palo Alto, CA), on a BPX70 column (30 m, 0.25 mm inner diameter, 0.25 µm film thickness, SGE) essentially as described previously (Zhou et al., 2011), except the column temperature program. The column temperature was programmed as an initial temperature at 100 • C holding for 3 min, ramping to 240 • C at rate of 7 • C/min and holding for 1 min. NuChek GLC-426 was used as the external reference standard. Peaks were integrated with Agilent Technologies ChemStation software [Rev B.04.03 (16)].

Results
A Survey of Thioesterase Activity in the Production of MCFA in Leaves Seven thioesterases from organisms known to produce MCFAs were infiltrated along with the silencing suppressor gene (p19) to determine whether MCFA could be produced in N. benthamiana leaf tissue. Leaf spots were harvested and freeze-dried after 5 days, after which the total fatty acid profiles were determined by GC (Table 1). Control leaf was found to contain trace or non-detectable levels of C12:0 and C14:0, whereas C16:0 was present at 14.9 ± 0.6% in total leaf lipids. C12:0 accumulation was observed at low levels with expression of Cocnu-TE3 (1.2 ± 0.1%) and Umbca-TE (1.6 ± 0.1%). Expression of all the thioesterases tested resulted in the accumulation of C14:0 in N. benthamiana leaf with Cinca-TE giving the greatest response of 11.3 ± 1.0%. Similarly, all of the thioesterases resulted in increased C16:0 accumulation with the exception of Umbca-TE. The greatest C16:0 accumulation was observed with Cocnu-TE1 expression (35.4 ± 4.7%).

The Co-Expression of the Coconut LPAAT Increases MCFA Production in Leaf
The C. nucifera LPAAT (CnLPAAT) had previously been shown to increase MCFA incorporation on the sn-2 position of TAG (Knutzon et al., 1999). This gene was co-infiltrated with the thioesterases described above. The co-infiltration of CnLPAAT and Umbca-TE increased the accumulation of C12:0 to 3.3 ± 0.5% and C14:0 was found to accumulate to 14.9 ± 1.6% in the Cinca-TE + CnLPAAT sample ( Table 1). The highest C16:0 levels were observed after co-expression of Cocnu-TE1 and CnL-PAAT (40.2 ± 2.8%). Some leaf chlorosis was observed following Cocnu-TE1 infiltration, so Cocnu-TE2 was used for further experiments since it also resulted in high C16:0 accumulation (38.3 ± 1.1%).

Lipid Species Analysis by LC-MS
LC-MS was used to gain a better understanding of MCFA accumulation and whether the newly-produced MCFA were accumulating in the neutral, polar or plastidial lipids.

FIGURE 2 | Full dataset of triacylglycerol (TAG) species for infiltrated
Nicotiana benthamiana leaf samples (n = 4). The label "No-TE" represents the N. benthamiana control leaf that was infiltrated with p19 only. With the expression of each thioesterase, in combination with CnLPAAT, AtWRI1 and AtDGAT1, significant increases in total TAG levels were observed. The data has been normalized using the internal standard tri-C17:0 TAG (51:0) (y-axis), being presented in units of 1 µg/mg leaf dry weight.
suggesting that lipid handling pathways in N. benthamiana leaf have lower specificity for chains shorter than C16:0. Coexpression of the AtDGAT1 consistently decreased DAG species with MCFA indicating that further increases in MCFA accumulation may require a DGAT with preference for MCFA substrates.
Each row represents DAG species containing at least one of the corresponding acyl chains (labeled on right). The data is represented according to LC-MS response (y-axis). A. thaliana DGAT1 (n = 4). The label "No-TE" represents the N. benthamiana control leaf that was infiltrated with p19 only. The percentage of each PC species for each sample is shown on the y-axis. samples were 28:0 and 30:0, indicating that a significant proportion of the C14:0 in MGDG was either di-C14:0 or co-located with C16:0. The C12:0-containing and C14:0-containing MGDG species were not detected in the p19 control sample. In contrast, C16:0-containing MGDG species tended to be reduced in the Cocnu-TE2 samples. The major MGDG species in the wildtype samples (MGDG 34:6 C16:3/C18:3, and MGDG 36:6 C18:3/C18:3) all tended to be reduced by the expression of the transgenes. This reduction was greatest in the presence of the WRI+DGAT combination. Only trace levels of C12:0-containing DGDG species were observed in the Umbca-TE samples. The major C14:0-containing DGDG species observed in the Cinca-TE samples were 28:0 and 30:0, both of which were absent in the control. These species were also observed at elevated levels in the Cocnu-TE2 samples but only at trace levels in the Umbca-TE samples. The major DGDG species in the wildtype samples (DGDG 34:3 C16:0/C18:3 and DGDG 36:6 C18:3/C18:3) all tended to be reduced by the expression of the transgenes. This reduction was greatest in the presence of AtWRI.
We were also interested in the correlation between the proportions of plastidial galactolipids and TAG (Figure 9). The control samples (p19 only, red circles in Figure 9) show that galactolipids are typically much more abundant than TAG (p19 only total TAG, 8.0 ± 2.4 µg/mg dry weight; total galactolipid, 36.6 ± 3.7 µg/mg dry weight) (Supplementary Table 2). In the case of the U. californica lauroyl-ACP thioesterase, this is not greatly changed by the co-expression of CnLPAAT + Umbca-TE (red vs. blue squares) (8.4 ± 1.9 µg/mg dry weight) or the further addition of AtDGAT1 (green square) (13.0 ± 3.7 µg/mg dry weight). The relative proportion of TAG is, however, seen to increase with the further addition of AtWRI1 (purple square, orange square) (with CnLPAAT+AtWRI1, 40.8 ± 8.0 µg/mg dry weight; with CnLPAAT + AtWRI1 + AtDGAT1, 56.3 ± 3.2 µg/mg dry weight). This is in contrast with both the results observed for the C. camphora C14:0-ACP and C. nucifera C16:0-ACP thioesterases in which the co-expression of CnLPAAT + Cinca-TE (blue diamond) or Cocnu-TE2 (blue triangle) increased the relative proportion of TAG to galactolipids even without the addition of the AtWRI1 or AtDGAT1. TAG was found to be most abundant relative to galactolipids after the co-expression of AtWRI1 + CnLPAAT + Cinca-TE (111.9 ± 23.4 µg/mg dry weight) or Cocnu-TE2 (96.4 ± 10.4 µg/mg dry weight). Complete data for the correlation between TAG and galactolipids is found in Supplementary Table 2.

Discussion
Earlier work described the constitutive expression in B. napus of the U. californica lauroyl-ACP thioesterase (Eccleston et al., 1996), the Umbca-TE gene used in this study. This resulted in the accumulation of significant levels of C12:0 in seed oil but only trace levels in leaf despite high transcript abundance and lauroyl-ACP thioesterase activity. In contrast, C12:0 synthesis was observed during in vitro feeding of [1-14 C] acetate to isolated leaf plastids from these transgenic events. Our study showed similar results in that accumulation of C12:0 was found to be very low (1.6% of total leaf fatty acids) after expression of Umbca-TE alone ( Table 1). The co-expression of the CnLPAAT, reported to have a preference for MCFA , with the individual thioesterases increased the accumulation of C12:0, C14:0, and C16:0, all of which are found in C. nucifera oil (Laureles et al., 2002). This result indicated that the native N. benthamiana LPAAT was either not highly expressed in leaf or did not have high specificity for C12:0, C14:0, and C16:0 substrates.
The co-expression of the AtWRI1 with the CnLPAAT and thioesterases had an additional positive effect on C12:0 and C14:0 accumulation in leaf tissue. The further addition of AtDGAT1 to the thioesterase + CnLPAAT + AtWRI1 combinations did not further increase MCFA accumulation. In fact, in most cases the addition of AtDGAT1 decreased MCFA accumulation, suggesting that this enzyme may not have good activity on MCFA substrates (Aymé et al., 2014). However, it is important to note that tri-C12:0, -C14:0, and -C16:0 TAG species were detected, FIGURE 9 | Correlation between normalized levels of TAG (x-axis) and galactolipids (y-axis), from Nicotiana benthamiana leaf infiltration samples. The blue line represents the linear model, with 95% confidence. Each data point represents the mean value (n = 4) for each treatment. The shape of each data point is representative of the thioesterase used. The color of each data point represents the corresponding gene combinations.
indicating that the CnLPAAT and native N. benthamiana lipid assembly enzymes were able to utilize MCFA substrates to some extent. In contrast with the C12:0-and C14:0-ACP thioesterases, it was consistently noted that the C16:0-ACP thioesterase samples were not as strongly affected by the co-expression of At WRI1. The strongest WRI1-effect was observed with the C12:0-ACP thioesterase, with the C14:0-ACP thioesterase having an intermediate response to the addition of AtWRI1. We speculate that this could be due to relatively rapid KASI-mediated condensation of C4:0 to C14:0 in the plastid with C12:0 being relatively inaccessible to a C12:0-ACP thioesterase for cleavage from ACP.
Lipid species analysis by LC-MS was also informative although it is important to understand that the transient nature of the assay means that results are not necessarily representative of stably transformed plants. The apparent reductions in levels of plastidial lipids MGDG and DGDG could have been due to increased export of fatty acids from the plastid before they could be incorporated into plastidial glycerolipids by the native N. benthamiana plastidial GPAT. The presence of MCFA in MGDG indicated that the plastidial GPAT was capable of accessing MCFA-ACP that remained in the plastid as substrate. Alternatively, the increased extra-plastidial MCFA could have been imported back into the plastid from the ER following thioesterase export. It is worthwhile noting the substantial accumulation of MCFA in PC which could be occurring via the fatty acids first being incorporated into de novo DAG, which is then converted to PC via the action of choline phosphotransferase (CPT) or phosphatidylcholine diacylglycerol phosphotransferase (PDCT) (Bates et al., 2013). It is also possible that MCFA could be incorporated in PC via direct acylation of sn-1 LPC via the action of lysophosphatidylcholine acyltransferase (LPCAT) (Lager et al., 2013). The presence of di-C12:0 PC and di-C14:0 PC does suggest that the above two routes of MCFA incorporation into PC might be operating under the transient leaf assay settings.
It remains to be seen how stable transgenic plants incorporating MCFA in PC fare, as the presence of MCFA could alter membrane fluidity significantly. Some chlorosis was observed in infiltrated leaves during the transient assay, similar to the phenotype observed following monoacylglycerol acyltransferase (MGAT1) transient expression (Divi et al., 2014), possibly as a result of MCFA accumulation in PC. The reduction in total galactolipids, MGDG (Figure 7) and DGDG (Figure 8) was also interesting since these lipid pools are essential to the efficiency of photosynthetic light reactions (Dörmann, 2013). Further investigations into sequestering MCFA into TAG are required. It is also possible that in a stable setting MCFA might be actively edited out of PC by the very active unusual fatty acyl editing mechanisms present in plant cells and sequestered in TAG (Millar et al., 2000;Lager et al., 2013). Eccleston et al. (1996) hypothesized that the lack of MCFA accumulation observed in B. napus leaf was due to the breakdown of the unusual fatty acids by β-oxidation, a concept supported by an observed increase in isocitrate lyase activity. If WRI1 expression does not adequately down-regulate such breakdown pathways in a leaf context it is possible that the co-expression of WRI1 simply increased fatty acid synthesis with a corresponding increase in plastidial export by the transgenic thioesterases. It was interesting to observe the inverse relationship between TAG and galactolipids which occurred after co-expression of the A. thaliana acyltransferases with the U. californica lauroyl-ACP thioesterase and CnLPAAT, or even in their absence in the case of the C. camphora C14:0-ACP and C. nucifera C16:0-ACP thioesterases.
Much of the C12:0, C14:0, and C16:0 were found to accumulate in TAG. This was important since the production of industrially-relevant levels of leaf TAG have already been demonstrated in tobacco (Vanhercke et al., 2014a). Our study offers a potential extension of this result by demonstrating that it is possible to tailor leaf oil profiles to produce industrially-relevant fatty acids in leaves with elevated energy content.

Conclusion
The experiments performed used the rapid N. benthamiana assay to demonstrate the activity of a series of thioesterases in leaf tissue. The study demonstrated that the over-expression of AtWRI1and CnLPAAT along with the thioesterases significantly increased the accumulation of C12:0 and C14:0 in leaf. Importantly, newly-produced fatty acids were effectively incorporated into leaf TAG. Collectively, these results indicate that high oil-containing plant leaves could be a good host for MCFA production.