Abstract

Long-chain acyl-CoA synthetase (LACS, EC 6.2.1.3) catalyzes the ATP-dependent activation of free fatty acid to form acyl-CoA, which, in turn, serves as the major acyl donor for various lipid metabolic pathways. Increasing the size of acyl-CoA pool by enhancing LACS activity appears to be a useful approach to improve the production and modify the composition of fatty acid-derived compounds, such as triacylglycerol. In the present study, we aimed to improve the enzyme activity of Arabidopsis thaliana LACS9 (AtLACS9) by introducing random mutations into its cDNA using error-prone PCR. Two AtLACS9 variants containing multiple amino acid residue substitutions were identified with enhanced enzyme activity. To explore the effect of each amino acid residue substitution, single-site mutants were generated and the amino acid substitutions C207F and D238E were found to be primarily responsible for the increased activity of the two variants. Furthermore, evolutionary analysis revealed that the beneficial amino acid site C207 is conserved among LACS9 from plant eudicots, whereas the other beneficial amino acid site D238 might be under positive selection. Together, our results provide valuable information for the production of LACS variants for applications in the metabolic engineering of lipid biosynthesis in oleaginous organisms.

Introduction

Fatty acids are carboxylic acids with highly reduced acyl chains, which act as the major energy reservoir in eukaryotic cells and the building blocks for all cellular lipids, including phospholipids, triacylglycerols (TAGs), isoprenoids, sterols, cutins, suberins and jasmonates. Free fatty acids are activated to their coenzyme A (CoA) derivatives prior to further metabolism in an ATP-dependent process catalyzed by long-chain acyl-CoA synthetase (LACS, EC 6.2.1.3). This reaction involves a two-step ping-pong reaction mechanism, in which firstly an adenylyl from ATP is transferred to a fatty acid forming an enzyme-bound acyl-adenylate intermediate and a pyrophosphate, and secondly, the acyl-adenylate intermediate is attacked by a CoA yielding an acyl-CoA and an AMP [1].

LACS activity was first reported in guinea pig (Cavia porcellus) liver by Kornberg and Pricer [2] and was later found in numerous other organisms [1]. In plants and many other organisms, multiple LACS genes are present, which encode enzymes that appear to function in different aspects of lipid metabolism. For instance, the model plant Arabidopsis thaliana (hereafter Arabidopsis) contains nine LACS (AtLACS) genes with distinct expression patterns, subcellular localizations and functions [3]. AtLACS6 and AtLACS7 are peroxisomal-localized enzymes required for the activation of fatty acids for β-oxidation during seedling development [4,5]. In contrast, AtLACS1 and AtLACS2, which are localized in the endoplasmic reticulum (ER), are involved in surface lipid biosynthesis [69]. In addition, AtLACS4, another ER-bound LACS, functionally overlaps with AtLACS1 in mediating the synthesis of lipids for pollen coat formation [10]. LACS may also play a crucial role in TAG biosynthesis in plants. Indeed, de novo fatty acids synthesized in the plastid are required to be activated to acyl-CoAs by LACS for use in TAG assembly in the ER [11]. The presence of LACS activity in the plastidial envelope was demonstrated [12], and further evidence suggested that LACS was specifically associated with the outer envelope [13,14]. In Arabidopsis, AtLACS9 is the only LACS associated with the plastid outer envelope and thus was regarded as the most probably candidate for activating and exporting plastidially derived fatty acids for TAG assembly [15]. The function of AtLACS9, however, is still debatable as a recent study has shown that AtLACS9 might contribute to lipid trafficking from the ER back to the plastid [16].

Regardless of the multiple roles of LACS enzymes in plants, the applications of LACS genes in engineering oleaginous microorganisms have been widely explored. Overexpressing LACS has been shown to increase the production of fatty acid esters, fatty alcohols, waxes and TAGs in Escherichia coli and yeast (Saccharomyces cerevisiae). For example, metabolic engineering of E. coli to produce fatty acid esters, fatty alcohols and waxes was achieved by overexpression of LACS in combination with other genes [17]. In addition, heterologous overexpression of LACS from diatoms (Phaeodactylum tricornutum and Thalassiosira pseudonana) or higher plants (Arabidopsis and Brassica napus) was applied to facilitate fatty acid uptake and stimulate oil deposition in yeast [1821]. Since LACS provides the substrates for acyl-CoA-dependent acyltransferases to produce various lipids, overexpression of LACS in these microorganisms appears to contribute to directing the carbon flux to the lipid biosynthesis pathways by enhancing the size of the acyl-CoA pool. In addition, improved LACS production may directly promote substrate channeling for lipid biosynthesis considering the possible direct association or cooperation of LACS with the downstream lipid biosynthetic enzymes [22,23]. In these regards, further improvement of the enzyme activity of LACS via protein engineering might represent novel perspectives for engineering lipid production in oleaginous organisms. Indeed, overexpression of cDNAs encoding improved enzyme variants has been demonstrated to be a more efficient strategy to increase lipid production than using wild-type (WT) enzymes in various plant and microorganism species [2429]. In addition, it has been shown that the beneficial mutations identified from one enzyme could be used to improve the performance of an enzyme from another species [24,30], and thus could potentially benefit the in planta improvement of enzyme action using non-transgenic approaches such as CRISPR [31].

A few crystal structures of acyl-CoA synthetases (ACSs) from bacteria and mammals have been solved, including a medium-chain-specific ACS from human (Homo sapiens) [32], a long-chain-specific ACS from Thermus thermophilus [33] and a very long-chain-specific ACS from Mycobacterium tuberculosis [34]. The lack of a detailed three-dimensional structure for AtLACS9, however, does not facilitate a rational design approach to modify the enzyme. Indeed, protein modification by rational design approaches is often constrained in practice since the protein structure and function relationship is intricate and hard to predict [35]. As an alternative, error-prone PCR has been used in generating enzyme variants with improved catalytic properties including increased activity, altered substrate specificity and increased temperature tolerance in the absence of detailed structure information [3639]. This approach is so powerful and robust that it could even identify beneficial mutations that were overlooked in rational design experiments [40,41].

In the current study, our strategy is to engineer enhanced performance in LACS using error-prone PCR and site-directed mutagenesis. Two AtLACS9 variants with multiply amino acid substitutions were generated with increased enzyme activity. The possible function of key amino acid residues affecting enzyme activity was further evaluated through in vitro enzyme assays and evolutionary analysis. The identified amino acid residue substitutions provide valuable information for the modification of LACS from different species.

Experimental

Cloning, random mutagenesis and site-directed mutagenesis of AtLACS9

The coding sequence of AtLACS9 was amplified using a cDNA preparation from Arabidopsis developing seeds and was cloned into the pYES2.1 vector (pYES2.1-V5/HIS vector, Invitrogen, Burlington, ON, Canada) under the control of GAL1 promoter and CYC1 terminator to yield pYES-AtLACS9. The stop codon of AtLACS9 was removed for in-frame fusion with a C-terminal V5 epitope. Random mutagenesis of AtLACS9 was carried out by error-prone PCR using the GeneMorph II Random Mutagenesis kit (Agilent Technologies, Santa Clara, CA, U.S.A.) and pYES-AtLACS9 as a template. The DNA fragment containing AtLACS9-coding region and the regions on the pYES2.1 backbone (250 bp before and 155 bp after the AtLACS9-coding region) was amplified using Mutazyme II DNA polymerase, and 2000 and 200 ng of pYES-AtLACS9 plasmid. PCR was performed for 30 cycles of 95°C for 30 s (denaturation), 55°C for 30 s (annealing) and 72°C for 2 min 40 s (extension). Site-directed mutagenesis within AtLACS9 was conducted using the QuikChange™ Site-Directed Mutagenesis Kit (Stratagene Cloning Systems, La Jolla, CA, U.S.A.) and pYES-AtLACS9 as a template. The primers used for the random mutagenesis and construction of various single-site mutants are listed in Supplementary Table S1.

Selection of positive clones and heterologous expression of AtLACS9 variants in yeast

The product of error-prone PCR was purified and co-transformed with the linearized pYES2.1 vector backbone into S. cerevisiae strain BYfaa1,4Δ (MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0, faa1 Δ::HIS3, faa4 Δ::LYS2) [23] using the S.c. EasyComp Transformation Kit (Invitrogen) for recombination. S. cerevisiae yeast transformants were selected on minimal medium [0.67% (w/v) yeast nitrogen base and 0.2% (w/v) SC-URA] containing 2% (w/v) galactose and 1% (w/v) raffinose, 100 mM oleic acid and 45 mM cerulenin. Tyloxapol 1% (v/v) was also added into plates to disperse the fatty acids. After incubating at 30°C for 2–3 days, the individual colonies grown on the selection plates were then used to inoculate minimal medium containing 2% (w/v) galactose and 1% (w/v) raffinose (refer to as induction medium) in 96-well plates. The yeast cultures were grown at 30°C for 24 and 48 h before subjected to the Nile red assay (using the protocol described below).

For heterologous expression of AtLACS9 variants, the coding sequences of the selected AtLACS9 variants were amplified and re-cloned into the pYES2.1 vector, and the resulting plasmids were sequenced and used to transform S. cerevisiae mutant BYfaa1,4Δ. The recombinant yeast cells were cultured in 2% (w/v) raffinose minimal medium. After overnight growth, the yeast cultures were inoculated into induction medium to an optical density of 0.4 at 600 nm (OD600). Yeast cultures were grown at 30°C with shaking at 220 rpm.

Lipid analysis

Neutral lipid analysis in yeast cells was performed using the Nile red fluorescence assay with a Synergy H4 Hybird reader (Biotek, Winooskit, VT, U.S.A.) as described recently [30]. In brief, 100 µl aliquots of yeast culture were placed in 96-well dark plates and the first fluorescence was measured with excitation at 485 nm and emission at 538 nm. Five microliters of Nile red solution (0.1 mg/ml in methanol) were then added into the yeast culture before the measurement of the second fluorescence under the same conditions. The change in fluorescence from the two measurements (ΔF TAG) is correlated with the amount of neutral lipids in the yeast culture. The Nile red values were calculated based on the amount of neutral lipids (ΔF TAG) as a function of OD600F TAG/OD600).

Protein extraction and Western blotting

Microsomal fractions were recovered from the recombinant yeast cells as described recently [30]. In brief, the recombinant yeast cells were collected at the similar OD600 values (∼7) during the log growth phase if not stated otherwise and then resuspended in lysis buffer containing 20 mM Tris–HCl (pH 7.9), 10 mM MgCl2, 1 mM EDTA, 5% (v/v) glycerol, 300 mM ammonium sulfate and 2 mM dithiothreitol before homogenization using a bead beater (Biospec, Bartlesville, OK, U.S.A.). The crude homogenate was centrifuged for 30 min at 10 000 g, and the supernatant was further centrifuged at 105 000 g for 70 min to separate the microsomal fractions. The microsomal fractions were resuspended in 3 mM imidazole buffer (pH 7.4) containing 125 mM sucrose. All procedures were carried out at 4°C. The protein content was determined by the Bradford assay using BSA as a standard [42].

For Western blotting, 5 µg of microsomal proteins were separated on 10% SDS–PAGE gel and electrotransferred (overnight at 30 mA and 4°C) onto the polyvinylidene difluoride membrane (Amersham, GE Healthcare, Mississauga, ON, Canada). The membrane was first blocked with 2% ECL prime blocking reagent (Amersham) and then was incubated with V5-HRP-conjugated antibody (Invitrogen). HRP-conjugated antibody was detected using the ECL Advance Western Blotting Detection Kit (Amersham) with a FluorChem SP imager (Alpha Innotech Corp., San Leandro, CA, U.S.A.). The band densities of AtLACS9 variants were quantified using ImageJ software [43].

In vitro LACS enzyme assays

The LACS assay was performed as described recently [23] with slight modifications. In brief, the enzyme assay was carried out in a 60-µl reaction mixture containing 100 mM Bis–Tris–propane (pH 7.6), 10 mM MgCl2, 5 mM ATP, 2.5 mM dithiothreitol, 1 mM CoA, 20 µM [1-14C] oleic acid (56.3 mCi/mmol, PerkinElmer, Waltham, MA, U.S.A.) and 2–10 µg of microsomal protein. The reaction was initiated by adding microsomal protein and quenched with 10 µl of 10% (w/v) SDS after incubation at 30°C for 5 min with shaking. The entire reaction mixture was washed four times using 900 µl of 50% (v/v) isopropanol saturated hexane for each wash. An aliquot of the aqueous phase was analyzed for radioactivity by an LS 6500 multi-purpose scintillation counter (Beckman-Coulter, Mississauga, ON, Canada). For substrate specificity assay, 20 µM [1-14C] fatty acids, including palmitic acid (60 mCi/mmol, PerkinElmer), stearic acid (58.9 mCi/mmol, American Radiolabeled Chemicals, St. Louis, MO, U.S.A.), oleic acid and linoleic acid (58.2 mCi/mmol, PerkinElmer) were used in the assay.

Sequence alignment, positive selection and protein three-dimensional structure prediction

Forty-five LACS sequences were collected from different species (Supplementary Table S2). The multiple sequence alignment of LACS proteins was performed using ClustalW in MEGA 7 under the default setting [44]. A neighbor-joining tree with 1000 bootstrap repetitions was built using the same software. A web server PAL2NAL (http://www.bork.embl.de/pal2nal/) was then used to construct a multiple codon alignment based on the corresponding aligned amino acid sequences. The output alignment was imported into the jModelTest 2 program [45] to determine the best-fitting evolutionary model. The general time reversible (GTR) model plus Gamma distribution plus the invariant site model of molecular evolution (GTR + G + I) was determined as the best-fit substitution model based on the lowest value of the Akaike Information Criterion. A maximum-likelihood phylogenetic tree was then constructed with the PhyML webserver (http://www.atgc-montpellier.fr/phyml/; accessed on 06 December 2018) [46,47] according to the best-fit predictive model. The posterior probabilities of sites under positive selection were calculated using CodeML program in the PAML version 4 software [48] based on site-specific Bayes empirical Bayes probabilities [49]. Three sets of models were carried out using the F3X4 codon frequency model, including M0 (one ratio) vs. M3 (discrete); M1 (nearly neutral) vs. M2 (positive selection) and M7 (β) vs. M8 (β + ω). The statistical significance of each pair of nested models was evaluated by the likelihood ratio test (LRT).

The AtLACS9 structure was obtained through homology modeling using the PHYRE2 protein fold recognition server [50]. Several homologous structures were identified as possible templates including carboxylic acid reductases (24–25% identity) and acetyl-CoA synthetases (less than 20% identities). A Nocardia iowensis carboxylic acid reductase [51] exhibiting 24% identity with AtLACS9 was used as a template to generate a model with a 83% sequence coverage and a high confidence level. The following AtLACS9 residues were included in the model: 59–86; 92–320; 328–371; 395–459; 468–525; 530–607; 623–691. A model based on an acetyl-CoA synthetase structure was also obtained to assess the structure. To further verify the quality of the model, I-TASSER was used to predict the 3D structure and the best models from the two softwares were assessed and overlaid [52].

Statistical analysis

Data are means ± standard deviation (SD) for the number of independent experiments as indicated. All statistical analyses were performed using the SPSS statistical package (SPSS 16.0, Chicago, IL, U.S.A.). Significant differences between two groups were determined using a two-tailed Student's t-test. The equality of variances was determined by the Levene's test. When the variances were equal, the unpaired Student's t-test assuming equal variances was performed. When the variances were unequal, the unpaired Student's t-test with Welch corrections assuming unequal variances was used.

Results

Selection and characterization of active AtLACS9 variants

To select active AtLACS9 variants, randomly mutated AtLACS9 cDNA libraries were transformed into S. cerevisiae strain BYfaa1,4Δ (a double mutant with both FAA1 and FAA4 genes knocked out, and it thus contains less than 10% of yeast endogenous LACS activity [18]). This yeast mutant cannot grow on the selection media containing fatty acids and cerulenin (an endogenous fatty acid synthesis inhibitor) due to acyl-CoA deficiency. The growth of yeast cells can be rescued, however, by the introduction of a cDNA encoding an active AtLACS9, which would import the exogenous fatty acids into cells and activate them to acyl-CoAs [23]. The positive colonies grown on the selection plates were cultivated in induction media for 24 and 48 h and subjected to the Nile red assay. The positive colonies were screened based on their abilities to produce neutral lipids as reflected by the Nile red assay. Two colonies were found to produce neutral lipid at levels higher than the yeast expressing WT AtLACS9 (data not shown). The coding sequences of these variant AtLACS9s were sequenced, re-cloned into the pYES2.1 vector and transformed into yeast strain BYfaa1,4Δ for detailed characterization. Yeast transformed with AtLACS9 variant cDNAs showed similar growth rates to the yeast harboring WT AtLACS9 or LacZ control (Supplementary Figure S1). Consistent with the screening results, expression of AtLACS9 variants in yeast resulted in higher or similar levels of neutral lipid accumulation (ΔF TAG/OD600) relative to yeast expressing WT AtLACS9 at the early stationary phase (Figure 1). Moreover, yeast cells producing AtLACS9 and its variants resulted in higher neutral lipid content than the LacZ control (Figure 1).

Neutral lipid content of yeast producing AtLACS9 variants.

Figure 1.
Neutral lipid content of yeast producing AtLACS9 variants.

Neutral lipid content was analyzed using the Nile red assay and the values are calculated based on the Nile red fluorescence (ΔF TAG) as a function of the optical density (OD600) at 600 nm (ΔF TAG/OD600). Data are means ± SD, n = 3.

Figure 1.
Neutral lipid content of yeast producing AtLACS9 variants.

Neutral lipid content was analyzed using the Nile red assay and the values are calculated based on the Nile red fluorescence (ΔF TAG) as a function of the optical density (OD600) at 600 nm (ΔF TAG/OD600). Data are means ± SD, n = 3.

To analyze the production profiles of AtLACS9 variants in yeast mutant BYfaa1,4Δ, yeast cells producing AtLACS9 variants were collected periodically from the log to the stationary growth phase, and the corresponding microsomal fractions were prepared for the analyses of in vitro LACS activity and protein accumulation by Western blotting. The activity of the recombinant AtLACS9 enzyme and the variants remained at high levels during the log phase, and then decreased after reaching the stationary phase (Figure 2A). AtLACS9 variants displayed the highest activity at the early log phase, whereas the highest activity of WT AtLACS9 occurred at the late log or early stationary phase. Increased LACS activity was observed for variants L12F/C207F/L656F and D238E/P659S. The recombinant AtLACS9 polypeptide accumulation in the microsomal fraction displaying the highest activity from each variant (Figure 2B) was then analyzed by Western blotting. The AtLACS9 variants displayed different polypeptide accumulation levels in yeast. Variant D238E/P659S had higher polypeptide accumulation, while variant L12F/C207F/L656F had lower polypeptide accumulation compared with that of WT AtLACS9 (Figure 2C). After normalizing the enzyme activity to the corresponding protein accumulation [30], both variants displayed 3-fold higher normalized activity relative to the WT enzyme (Figure 2D).

Characterization of AtLACS9 variants.

Figure 2.
Characterization of AtLACS9 variants.

(A) In vitro LACS activities of different AtLACS9 variants. Microsomal fractions from yeasts producing recombinant AtLACS9 variants were harvested at different time points after induction and used for the enzyme assay. The growth curve was monitored by measuring OD600. (B) Relative enzyme activities of AtLACS9 variants. The highest activity of each variant is shown, with the WT AtLACS9 activity set as 1.0. (C) Relative protein abundance. Five micrograms of microsomal protein from the same batch of microsomes used to assess enzyme activity were used for Western blotting analysis. The relative protein accumulation of recombinant WT AtLACS9 was set as 1.0. (D) Normalized enzyme activities of AtLACS9 variants. The normalized relative activity of each enzyme variant was obtained by dividing the enzyme activity value by relative protein abundance, with recombinant WT AtLACS9 activity set as 1.0. For (A), (B), (C) and (D), data are means ± SD; n = 2 for (A), n = 4 for (B), n = 3 for (C) and (D). The asterisks indicate significant differences in activity (B), protein abundance (C) and normalized activity (D) of the microsomes containing recombinant AtLACS9 variants versus recombinant WT AtLACS9 (t-test, **P < 0.01, *P < 0.5). ND, not determined.

Figure 2.
Characterization of AtLACS9 variants.

(A) In vitro LACS activities of different AtLACS9 variants. Microsomal fractions from yeasts producing recombinant AtLACS9 variants were harvested at different time points after induction and used for the enzyme assay. The growth curve was monitored by measuring OD600. (B) Relative enzyme activities of AtLACS9 variants. The highest activity of each variant is shown, with the WT AtLACS9 activity set as 1.0. (C) Relative protein abundance. Five micrograms of microsomal protein from the same batch of microsomes used to assess enzyme activity were used for Western blotting analysis. The relative protein accumulation of recombinant WT AtLACS9 was set as 1.0. (D) Normalized enzyme activities of AtLACS9 variants. The normalized relative activity of each enzyme variant was obtained by dividing the enzyme activity value by relative protein abundance, with recombinant WT AtLACS9 activity set as 1.0. For (A), (B), (C) and (D), data are means ± SD; n = 2 for (A), n = 4 for (B), n = 3 for (C) and (D). The asterisks indicate significant differences in activity (B), protein abundance (C) and normalized activity (D) of the microsomes containing recombinant AtLACS9 variants versus recombinant WT AtLACS9 (t-test, **P < 0.01, *P < 0.5). ND, not determined.

Effect of single-site mutations on enzyme activity

Since the two AtLACS9 variants (L12F/C207F/L656F and D238E/P659S) with increased LACS activity contained more than one amino acid residue substitution, the effect of each amino acid residue substitution on enzyme activity was separately determined. Five single-site mutants (L12F, C207F, L656F, D238E and P659S) were generated and expressed in yeast mutant BYfaa1,4Δ. WT AtLACS9 and LacZ were used as positive and negative controls, respectively. The microsomal fractions containing the recombinant enzymes were used for enzyme assays and Western blotting (Figure 3). Compared with the WT enzyme, increased microsomal enzyme activity and polypeptide accumulation were observed for variants L12F/C207F/L656F and D238E/P659S along with the following variants with single amino acid substitution: C207F, L656F, D238E and P659S (Figure 3A,B). Variant L12F, however, displayed comparable microsomal activity, but decreased protein accumulation to those of WT enzyme. The enzyme activity for each variant was then normalized to the corresponding protein accumulation level (Figure 3C), and single-site mutants C207F and D238E were found to possess the highest normalized activity, which could have mainly contributed to the increased enzyme activity of the original variants with multiple amino acid residue substitutions.

Enzyme activity and corresponding protein abundance of AtLACS9 single-site mutants.

Figure 3.
Enzyme activity and corresponding protein abundance of AtLACS9 single-site mutants.

(A) Relative enzyme activities of single-site mutants. The WT AtLACS9 activity was set as 1.0. (B) Relative protein abundance. Five micrograms of microsomal protein from the same batch of microsomes used to assess enzyme activity were used for Western blotting analysis. The relative abundance of recombinant WT AtLACS9 was set as 1.0. (C) Normalized enzyme activity. The normalized relative activity of each mutant was obtained by dividing the enzyme activity value by relative protein accumulation, with recombinant WT AtLACS9 activity set as 1.0. For (A), (B) and (C), data are means ± SD, n = 3. The asterisks indicate significant differences in activity (A), protein abundance (B) and normalized activity (C) of the microsomes containing recombinant AtDGAT9 variants versus recombinant WT AtLACS9 (t-test, **P < 0.01, *P < 0.05). ND, not determined.

Figure 3.
Enzyme activity and corresponding protein abundance of AtLACS9 single-site mutants.

(A) Relative enzyme activities of single-site mutants. The WT AtLACS9 activity was set as 1.0. (B) Relative protein abundance. Five micrograms of microsomal protein from the same batch of microsomes used to assess enzyme activity were used for Western blotting analysis. The relative abundance of recombinant WT AtLACS9 was set as 1.0. (C) Normalized enzyme activity. The normalized relative activity of each mutant was obtained by dividing the enzyme activity value by relative protein accumulation, with recombinant WT AtLACS9 activity set as 1.0. For (A), (B) and (C), data are means ± SD, n = 3. The asterisks indicate significant differences in activity (A), protein abundance (B) and normalized activity (C) of the microsomes containing recombinant AtDGAT9 variants versus recombinant WT AtLACS9 (t-test, **P < 0.01, *P < 0.05). ND, not determined.

The substrate specificity of AtLACS9 and its single-site variants C207F and D238E was assessed using different radiolabeled fatty acids as substrates (Figure 4). AtLACS9 or its variants was able to utilize all fatty acids tested, with linoleic acid (18:2Δ9cis, 12cis) being the most effective substrate in each case. No significant differences in substrate preference, however, were observed for AtLACS9 and its variants.

Substrate specificity of AtLACS9 variants.

Figure 4.
Substrate specificity of AtLACS9 variants.

Enzyme activity data were normalized to activity observed using oleic acid (18:1Δ9cis) as the substrate (i.e. oleic acid-supported activity was set at 100%). The microsomal activities of AtLACS9, C207F and D238E were 4.92 ± 0.09, 6.06 ± 0.25, 8.17 ± 0.14 nmol [14C] oleoyl-CoA/min/mg protein, respectively. Microsomal preparations from the yeast mutant BYfaa1,4Δ producing AtLACS9 variants were used for analysis of enzyme assay. Data represent means ± SD, n = 3. 16:0, palmitic acid; 18:0, stearic acid; 18:1, oleic acid; 18:2, linoleic acid (18:2Δ9cis, 12cis).

Figure 4.
Substrate specificity of AtLACS9 variants.

Enzyme activity data were normalized to activity observed using oleic acid (18:1Δ9cis) as the substrate (i.e. oleic acid-supported activity was set at 100%). The microsomal activities of AtLACS9, C207F and D238E were 4.92 ± 0.09, 6.06 ± 0.25, 8.17 ± 0.14 nmol [14C] oleoyl-CoA/min/mg protein, respectively. Microsomal preparations from the yeast mutant BYfaa1,4Δ producing AtLACS9 variants were used for analysis of enzyme assay. Data represent means ± SD, n = 3. 16:0, palmitic acid; 18:0, stearic acid; 18:1, oleic acid; 18:2, linoleic acid (18:2Δ9cis, 12cis).

Multiple sequence alignment, positive selection prediction for LACS proteins and structure prediction

Given that several amino acid residue substitutions in AtLACS9 were shown to affect AtLACS9 activity (Figures 2 and 3), it is useful to further explore the relationship between the identified beneficial amino acid residue and the putative amino acid residue sites with functional importance in LACS proteins. In this regard, sequence-based approaches, including multiple sequence alignment and positive selection prediction, were performed for various LACS proteins. The multiple sequence alignment is an effective approach to identify conserved functional motifs and subfamily-specific positions. In addition to the conserved sites, some unconserved sites, such as positively selected sites, may also affect protein function. Positive selection or Darwinian selection is considered to drive the sweep and fixation of the advantageous mutations throughout a population [53], and thus may have crucial roles in the evolution of protein function [54]. To test the presence of positive selection in the LACS sequence, the site-specific non-synonymous (dN) to synonymous (dS) substitutions ratio (dN/dS or ω) test was conducted by using three sets of models (M0 versus M3, M1 versus M2, and M7 versus M8) from the PAML version 4 software [48]. The LRT of the comparison between the model pair of M1 (null and neutral) versus M2 (selection) did not give a significant result to reject the null hypothesis of neutral selection (Table 1). However, the comparison between the model pairs, M0 (null and neutral) versus M3 (selection), and M7 (null and neutral) versus M8 (selection), yielded the LRT statistics of 7909.3 and 783.0, respectively, suggesting that certain sites were indeed under selective pressures in LACS proteins (Table 1). No positively selected sites were detected from the model M3 (ω2 = 0.89758), whereas in total 63 amino acid residues were identified as sites of positive selection from the model M8 (ω =1.22349) using Bayes empirical Bayes (BEB) analysis [49].

Table 1
Parameter estimates and likelihood scores of LACS9 for site models

Positive selection by site models was performed using the CODEML program in PAML. The number of positively selected sites is also shown, with the BEB posterior probability in blankets. Abbreviations: df, degrees of freedom; LRT, likelihood ratio test; lnL, log-likelihood scores; 2ΔlnL, twice the log-likelihood difference of the models compared.

Model Estimates of parameters lnL LRT pairs df 2ΔlnL P-value Positively selected sites 
M0: one ratio ω = 0.19727 −63 498.3 M0/M3 7909.3 <0.0001 None 
M3: discrete p0 = 0.82713, p1 = 0.15115, p2 = 0.02172, ω0 = 0.01143, ω1 = 0.18961, ω2 = 0.89758 −59 543.7      
M1: nearly neutral p0 = 0.96980, p1 = 0.03020, ω0 = 0.06927, ω1 = 1.00000 −60 172.6 M1/M2 1.00000 None 
M2: positive selection p0 = 0.96980, p1 = 0.02844, p2 = 0.00176, ω0 = 0.06927, ω1 = 1.00000, ω2 = 1.00000 −60 172.6      
M7: β p = 0.08048, q = 0.83695 −59 838.9 M7/M8 783.0 <0.0001 30 sites (>50%)
15 sites (>95%)
18 sites (>99%) 
M8: β + ω p0 = 0.99029, p = 0.14171, q = 2.31642, p1 = 0.00971, ω =1.22349 −59 447.4     
Model Estimates of parameters lnL LRT pairs df 2ΔlnL P-value Positively selected sites 
M0: one ratio ω = 0.19727 −63 498.3 M0/M3 7909.3 <0.0001 None 
M3: discrete p0 = 0.82713, p1 = 0.15115, p2 = 0.02172, ω0 = 0.01143, ω1 = 0.18961, ω2 = 0.89758 −59 543.7      
M1: nearly neutral p0 = 0.96980, p1 = 0.03020, ω0 = 0.06927, ω1 = 1.00000 −60 172.6 M1/M2 1.00000 None 
M2: positive selection p0 = 0.96980, p1 = 0.02844, p2 = 0.00176, ω0 = 0.06927, ω1 = 1.00000, ω2 = 1.00000 −60 172.6      
M7: β p = 0.08048, q = 0.83695 −59 838.9 M7/M8 783.0 <0.0001 30 sites (>50%)
15 sites (>95%)
18 sites (>99%) 
M8: β + ω p0 = 0.99029, p = 0.14171, q = 2.31642, p1 = 0.00971, ω =1.22349 −59 447.4     

By mapping the detected positively selected sites and putative functional motifs [33,55,56] along the aligned sequences of LACS (Figure 5A), it became apparent that the positively selected sites were mainly located at the N- and C-termini of the enzymes, whereas no sites on the putative functional motifs were observed under positive selection. The beneficial amino acid residue substitutions in the identified AtLACS9 variants were further compared with the positively selected sites and putative functional motifs. Most of the beneficial amino acid residue substitutions were found to reside in the less conserved regions. L12 and D238 were predicted as positive selection sites despite that the posterior probabilities of both sites are only higher than 50%. Furthermore, the phylogenetic analysis revealed that AtLACS9-C207 is highly conserved among the LACS9 from plant eudicots, whereas AtLACS9-D238 is more divergent and the substitution of D238E exists naturally in other LACS sequences (Figure 5B).

Sequence analysis and predicted structure of AtLACS9.

Figure 5.
Sequence analysis and predicted structure of AtLACS9.

(A) Sequence alignment of LACS9 protein from seven typical plant species. Conserved sites are shaded. Positively selected sites with a Bayes Empirical Bayes posterior probability higher than (≥) 50%, higher than (≥) 95% and higher than (≥) 99% are indicated by the amino acid sites in green, blue and red background, respectively. The single mutation sites are indicated by red-filled triangle. The bar above the sequence corresponds to the ATP/AMP signature motifs (I and II) and the fatty acyl-CoA synthetase signature motif (III). The putative active sites are indicated by black-filled star. The alignment was visualized and displayed using Jalview [59]. (B) Amino acid sequence analysis of LACS proteins from different species. Ah, Arachis hypogaea; At, Arabidopsis thaliana; Bd, Brachypodium distachyon; Bn, Brassica napus; Cas, Camelina sativa; Cs, Cucumis sativus; Eg, Elaeis guineensis; Fv, Fragaria vesca subsp. vesca; Gh, Gossypium hirsutum; Gm, Glycine max; Ha, Helianthus annuus; Lu, Linum usitatissimum; Mt, Medicago truncatula; Os, Oryza sativa; Rc, Ricinus communis; Si, Sesamum indicum; Sl, Solanum lycopersicum; Vv, Vitis vinifera; Zm, Zea mays. Phytozome/Genbank accession number for each sequence is shown in brackets. Phylogenetic relationship among protein sequences of LACS was constructed using the neighbor-joining method. Bootstrap values are shown at the tree nodes. The amino acid substitution sites of AtLACS9 variants are marked with red-filled triangle. (C) Homology model of AtLACS9 using PHYRE2 software and a carboxylic acid reductase as a template. The putative binding sites for ATP and fatty acid are shown in blue and green, respectively. The identified beneficial mutation sites C207 and D238 are shown in red. About 83% of the sequence was modeled with high confidence. The first 58 N-terminal residues were not included in the model as the template is a soluble enzyme. Based on TMHMM and SOSUI analyses, this N-terminal segment is predicted to constitute one N-terminal transmembrane domain, which was added to the structure.

Figure 5.
Sequence analysis and predicted structure of AtLACS9.

(A) Sequence alignment of LACS9 protein from seven typical plant species. Conserved sites are shaded. Positively selected sites with a Bayes Empirical Bayes posterior probability higher than (≥) 50%, higher than (≥) 95% and higher than (≥) 99% are indicated by the amino acid sites in green, blue and red background, respectively. The single mutation sites are indicated by red-filled triangle. The bar above the sequence corresponds to the ATP/AMP signature motifs (I and II) and the fatty acyl-CoA synthetase signature motif (III). The putative active sites are indicated by black-filled star. The alignment was visualized and displayed using Jalview [59]. (B) Amino acid sequence analysis of LACS proteins from different species. Ah, Arachis hypogaea; At, Arabidopsis thaliana; Bd, Brachypodium distachyon; Bn, Brassica napus; Cas, Camelina sativa; Cs, Cucumis sativus; Eg, Elaeis guineensis; Fv, Fragaria vesca subsp. vesca; Gh, Gossypium hirsutum; Gm, Glycine max; Ha, Helianthus annuus; Lu, Linum usitatissimum; Mt, Medicago truncatula; Os, Oryza sativa; Rc, Ricinus communis; Si, Sesamum indicum; Sl, Solanum lycopersicum; Vv, Vitis vinifera; Zm, Zea mays. Phytozome/Genbank accession number for each sequence is shown in brackets. Phylogenetic relationship among protein sequences of LACS was constructed using the neighbor-joining method. Bootstrap values are shown at the tree nodes. The amino acid substitution sites of AtLACS9 variants are marked with red-filled triangle. (C) Homology model of AtLACS9 using PHYRE2 software and a carboxylic acid reductase as a template. The putative binding sites for ATP and fatty acid are shown in blue and green, respectively. The identified beneficial mutation sites C207 and D238 are shown in red. About 83% of the sequence was modeled with high confidence. The first 58 N-terminal residues were not included in the model as the template is a soluble enzyme. Based on TMHMM and SOSUI analyses, this N-terminal segment is predicted to constitute one N-terminal transmembrane domain, which was added to the structure.

A three-dimensional structure of AtLACS9 was then obtained to further map the two identified beneficial amino acid residue substitution sites (Figure 5C). The best template was identified as a carboxylic acid reductase [51], which is a soluble enzyme. AtLACS9, on the other hand, has been experimentally demonstrated to reside in the envelope of the chloroplast [6,16]. The first 20 amino acid residues of AtLACS9 are predicted to constitute a membrane-spanning segment [23] despite that the prediction results varied among different programs (Supplementary Figure S2A,B). Indeed, the predicted membrane-associated nature agrees with its microsomal localization in yeast (Supplementary Figure S2C) and chloroplastidial localization in Arabidopsis [6,16]. Since both the experimental results [6,16] and the TargetP1.1 prediction (http://www.cbs.dtu.dk/services/TargetP/; accessed on 06 December 2018) suggest the chloroplastic subcellular localization of AtLACS9, the protein sequence of AtLACS9 was subjected to the signal (chloroplast transit) peptides prediction using ChloroP 1.1 (http://www.cbs.dtu.dk/services/ChloroP/; accessed on 06 December 2018) and SOSUIsignal (http://harrier.nagahama-i-bio.ac.jp/sosui/sosuisignal/sosuisignal_submit.html; accessed on 06 December 2018). Although the N-terminal region (19 amino acid residues) of AtLACS9 is predicted as a signal peptide by SOSUIsignal, no chloroplast transit peptides are predicted from ChloroP 1.1. Therefore, the N-terminal fragment with the putative transmembrane domain was not included in the model as this section did not exhibit any homology to the carboxylic acid reductase and was thus separately added to the structure to show localization in the membrane. The identified beneficial mutation sites are shown in red, whereas the putative ATP and fatty acid-binding motifs in the AtLACS9 model structure are shown in blue and green, respectively. These substrate-binding sites are close to one another, suggesting that these sites may be able to facilitate the transfer of an AMP moiety from ATP to the carboxylate group of fatty acids. The identified beneficial sites, however, are present at distal sites relative to the putative substrate-binding sites, indicating that these mutations do not directly affect substrate binding. Although the AtLACS9 sequence exhibits only 20% sequence identity with an acetyl-CoA synthetase with a reported three-dimensional structure, the homology structure of AtLACS9 using this as a template also gave similar orientations of the putative substrate-binding sites and the beneficial sites for mutagenesis (Supplementary Figure S3). Furthermore, results of I-TASSER modeling confirmed that the two models predicted from different software programs have a similar overall fold as shown by the overlaid structures (Supplementary Figure S4).

Discussion

The current study reports on the generation of performance-enhanced variants of AtLACS9 using protein engineering. To improve the enzyme performance of AtLACS9, we introduced random mutations into the coding sequence by error-prone PCR, transformed the mutagenized AtLACS9 into the yeast mutant BYfaa1,4Δ and screened for the active AtLACS9 variants by identifying positive clones on selection media followed by fluorescence detection of neutral lipid accumulation in yeast cells. After screening, two AtLACS9 variants were identified which slightly increased neutral lipid accumulation in yeast cells (Figure 1) and thus were selected for further characterization by analyzing in vitro LACS activity and polypeptide accumulation using yeast microsomal proteins. These enzyme variants were found to display increased yeast microsomal activity and altered polypeptide accumulation (Figures 2 and 3). To eliminate the influence of differences in protein abundance, the microsomal activity of each variant was then normalized to the corresponding polypeptide accumulation level. The two LACS variants had higher normalized activity compared with that of the WT enzyme (Figure 2D). Since the two activity-improved variants (L12F/C207F/L656F and D238E/P659S) contained more than one amino acid residue substitution, recombinant enzymes with single amino acid residue substitution were generated using site-directed mutagenesis to further elucidate the contribution of each substitution. The single-site mutants C207F and D238E were considered mainly responsible for the increased enzyme activity, although amino acid residue substitutions of L12F and P659S also led to increases in enzyme activity to some extent (Figure 3).

To interpret the effects of amino acid residue substitutions in the two AtLACS9 variants, the substituted amino acid residue sites from each variant and the predicted positively selected sites were mapped onto the multiple-aligned sequences of LACS9 proteins (Figure 5A), since both moderately conserved sites (e.g. subfamily-specific positions) [57] and unconserved sites (particularly positively selected sites) [54] appear to have crucial roles in affecting protein function. The beneficial amino acid residue substitution of C207F is at a moderately conserved site, whereas two additional activity-improved variants (L12F and D238E) were found with amino acid residue substitutions at the predicted positively selected sites (Figure 5A). Indeed, the substitution of D238E is naturally present in LACS from other plant species (Figure 5B). As for the substitution of C207F, the change from the polar C residue to the non-polar (aromatic) F residue appears to be dramatic, especially considering the capability of formation of a disulfide bond between two cysteine residues, the replacement of which might lead to changes in the tertiary structure. It should be noted, however, that although C207 is conserved among LACS9 from eudicots, a non-polar and aromatic Y residue is also found at that position of LACS from many other species. Furthermore, PSIPRED analysis predicts C207 and D238E to be part of β-sheet and loop region, respectively, in agreement with the PHYRE modeling results. C207F mutation may lead to secondary structure changes, whereas D238E may lead to changes in the loop region (Supplementary Figure S5).

It is possible that the increased enzyme activity was caused by a more favorable conformation in support of catalysis, such as improved affinity to the substrates. However, mapping of the beneficial amino acid residue substitution sites onto the predicted three-dimensional structure of AtLACS9 further revealed that these beneficial sites are apparently remote from the putative substrate-binding sites (Figure 5C). Indeed, beneficial substitutions at residues far away from the putative functional sites are not unusual in variants generated via random mutagenesis, although these sites are easily overlooked in rational design [40,41]. For example, directed evolution of diacylglycerol acyltransferase 1 from American hazelnut (Corylus americana) [24] and canola-type B. napus [30,25] also identified several amino acid residue substitutions affecting enzyme activity which are far away from the enzyme's functional sites. Mutations in these distal sites may influence enzyme conformation or regulatory sites resulting in more active conformations. Regulatory sites can be found distal from active site motifs. For example, it was recently shown that several amino acid residues substitutions within the hydrophilic N-terminal domain of B. napus diacylglycerol acyltransferase 1 could increase enzyme activity [25]. This N-terminal domain represents a regulatory domain that is located away from the putative active sites near the C-terminal region [58]. Though far from the putative catalytic sites, amino acid residues substitution within the acyltransferase's regulatory domain resulted in significant changes in enzyme activity [25]. It should also be noted that the beneficial substitutions may affect enzyme stability and thus lead to increased enzyme activity. Nonetheless, it would be worthwhile to further explore the effects of the beneficial substitutions on the enzyme's conformation and stability.

The observed fatty acid specificity of microsomal AtLACS9 (Figure 4) is generally consistent with previously reported substrate specificity of AtLACS9 [3]. The only difference is that AtLACS9 was observed to prefer linoleic acid slightly more than oleic acid in this study, whereas Shockey et al. [3] observed that AtLACS9 showed a slightly higher preference towards oleic acid over linoleic acid. It should be noted that different sources of recombinant AtLACS9 were used in the assays in these two studies. Shockey et al. used the E. coli mutant strain K27 to produce recombinant AtLACS9, which provided a clean background for LACS assay due to the complete loss of endogenous LACS activity in E. coli K27. In our study, AtLACS9 and its variants were expressed in yeast stain BYfaa1,4Δ, which contains less than 10% LACS enzymes in yeast background [18], and as a result, the control microsomes (LacZ) displayed a low level of LACS activity (see Figures 2A and 3A). The low level of LACS activity from the yeast background would therefore only have a limited influence on substrate specificity data depicted in Figure 4. Recently, AtLACS9 was shown to be involved in establishing a specific linoleoyl-CoA pool, which is connected to lipid trafficking from the PC to the plastid [16]. Indeed, in the current study, recombinant AtLACS9 produced in yeast exhibited an enhanced specificity for linoleic acid relative to palmitic, stearic and oleic acid (Figure 4).

As the only LACS associated with the outer envelope of the plastid, AtLACS9 appears to function in the activation of free fatty acids and transport of fatty acyl moieties between the plastid and extra-plastidial compartment [15,16]. Although the direction of AtLACS9-mediated transport of fatty acyl moieties is still debatable, increasing AtLACS9 activity through metabolic engineering could potentially shed light on the specific contribution of AtLACS9. In addition, AtLACS9 variants with enhanced enzyme activity may also provide potential candidates for engineering of oleginious organisms to produce fatty acid-derived compounds, including TAG. Indeed, introduction of various LACS from diatoms or higher plants into yeast has been shown to increase oil deposition [1821]. The current study also suggested that the introduction of activity-improved AtLACS9 variants led to a slight increase in the yeast neutral lipid accumulation (Figure 1). Combined heterologous expression of an AtLACS9 variant with cDNA encoding one or more of the acyl-CoA-dependent acyltransferases of the Kennedy pathway involved in seed oil biosynthesis (see Chapman and Ohlrogge [11]) may even result in greater accumulation of TAG in yeast. A similar metabolic engineering strategy might also be explored as means of increasing the TAG content of seeds or vegetative tissue. Furthermore, the beneficial amino acid substitutions of AtLACS9 provide valuable information for systematically engineering LACS from other species since these amino acid substitutions are at sites conserved among different species (Figure 5B).

In conclusion, two activity-enhanced variants of AtLACS9 were developed via random mutagenesis combined with site-directed mutagenesis. The recombinant AtLACS9 variants produced in yeast were characterized by analysis of in vitro enzyme activity and polypeptide accumulation in microsomes. The beneficial amino acid substitution sites were further analyzed by sequence and structural analyses. Amino acid residue substitution at the moderately conserved site C207 and the predicted positively selected site D238 were responsible for the increases in enzyme activity of corresponding enzyme variants. These findings provide valuable information for improving LACS enzymes from plants which may be useful in the in vivo alteration of acyl-CoA pools.

Abbreviations

     
  • ACS

    acyl-CoA synthetase

  •  
  • AtLACS

    Arabidopsis thaliana LACS

  •  
  • CoA

    coenzyme A

  •  
  • BEB

    Bayes empirical Bayes

  •  
  • ER

    endoplasmic reticulum

  •  
  • GTR

    general time reversible

  •  
  • LACS

    long-chain acyl-CoA synthetase

  •  
  • LRT

    likelihood ratio test

  •  
  • SD

    standard deviation

  •  
  • TAG

    triacylglycerol

  •  
  • WT

    wild-type

Author Contribution

R.J.W., Y.X. and G.C. designed the research; G.C. and R.J.W. supervised the experiments; Y.X. performed most of the experiments, analyzed the data and drafted the manuscript. K.M.P.C. performed the analysis of the predicted 3D structure of AtLACS9. K.M.P.C., R.H., E.M., J.O., G.C. and S.M.R. contributed valuable discussion during the present study. All co-authors contributed in further revising the manuscript.

Funding

The research was supported by Genome Canada, Genome Prairie, Dow AgroScience (Corteva Agriscience, Agriculture Division of DowDupont), Alberta Innovates Bio Solutions (R.J.W.), the Natural Science and Engineering Research Council of Canada (NSERC) Discovery Grants (Discovery grant no. 163306 to J.O.; RGPIN-2016-05926 to G.C. and RGPIN-2014-04585 to R.J.W.) and the Canada Research Chairs Program (R.J.W. and G.C.). We also acknowledge the support of the VEGA agency grant (No. 2/0180/12) and Slovak Research and Development Agency under the contract No. APVV-0785-11.

Competing Interests

The Authors declare that there are no competing interests associated with the manuscript.

References

References
1
Watkins
,
P.A.
(
1997
)
Fatty acid activation
.
Prog. Lipid Res.
36
,
55
83
2
Kornberg
,
A.
and
Pricer
,
W.E.
(
1953
)
Enzymatic synthesis of the coenzyme A derivatives of long chain fatty acids
.
J. Biol. Chem.
204
,
329
343
PMID:
[PubMed]
3
Shockey
,
J.M.
,
Fulda
,
M.S.
and
Browse
,
J.A.
(
2002
)
Arabidopsis contains nine long-chain acyl-coenzyme A synthetase genes that participate in fatty acid and glycerolipid metabolism
.
Plant Physiol.
129
,
1710
1722
4
Fulda
,
M.
,
Schnurr
,
J.
,
Abbadi
,
A.
,
Heinz
,
E.
and
Browse
,
J.
(
2004
)
Peroxisomal acyl-CoA synthetase activity is essential for seedling development in Arabidopsis thaliana
.
Plant Cell
16
,
394
405
5
Fulda
,
M.
,
Shockey
,
J.
,
Werber
,
M.
,
Wolter
,
F.P.
and
Heinz
,
E.
(
2002
)
Two long-chain acyl-CoA synthetases from Arabidopsis thaliana involved in peroxisomal fatty acid beta-oxidation
.
Plant J.
32
,
93
103
6
Schnurr
,
J.
,
Shockey
,
J.
and
Browse
,
J.
(
2004
)
The acyl-CoA synthetase encoded by LACS2 is essential for normal cuticle development in Arabidopsis
.
Plant Cell
16
,
629
642
7
Tang
,
D.
,
Simonich
,
M.T.
and
Innes
,
R.W.
(
2007
)
Mutations in LACS2, a long-chain acyl-coenzyme A synthetase, enhance susceptibility to avirulent Aseudomonas syringae but confer resistance to Botrytis cinerea in Arabidopsis
.
Plant Physiol.
144
,
1093
1103
8
,
S.
,
Song
,
T.
,
Kosma
,
D.K.
,
Parsons
,
E.P.
,
Rowland
,
O.
and
Jenks
,
M.A.
(
2009
)
Arabidopsis CER8 encodes LONG-CHAIN ACYL-COA SYNTHETASE 1 (LACS1) that has overlapping functions with LACS2 in plant wax and cutin synthesis
.
Plant J.
59
,
553
564
9
Weng
,
H.
,
Molina
,
I.
,
Shockey
,
J.
and
Browse
,
J.
(
2010
)
Organ fusion and defective cuticle function in a lacs1 lacs2 double mutant of Arabidopsis
.
Planta
231
,
1089
1100
10
Jessen
,
D.
,
Olbrich
,
A.
,
Knüfer
,
J.
,
Krüger
,
A.
,
Hoppert
,
M.
,
Polle
,
A.
et al.  (
2011
)
Combined activity of LACS1 and LACS4 is required for proper pollen coat formation in Arabidopsis
.
Plant J.
68
,
715
726
11
Chapman
,
K.D.
and
Ohlrogge
,
J.B.
(
2012
)
Compartmentation of triacylglycerol accumulation in plants
.
J. Biol. Chem.
287
,
2288
2294
12
Roughan
,
P.G.
and
Slack
,
C.R.
(
1977
)
Long-chain acyl-coenzyme A synthetase activity of spinach chloroplasts is concentrated in the envelope
.
Biochem. J.
162
,
457
459
13
Andrews
,
J.
and
Keegstra
,
K.
(
1983
)
Acyl-CoA synthetase is located in the outer membrane and acyl-CoA thioesterase in the inner membrane of pea chloroplast envelopes
.
Plant Physiol.
72
,
735
740
14
Block
,
M.A.
,
Dorne,
A.-J.
,
Joyard
,
J.
and
Deuce
,
R.
(
1983
)
The acyl-CoA synthetase and acyl-CoA thioesterase are located on the outer and inner membrane of the chloroplast envelope, respectively
.
FEBS Lett.
153
,
377
381
15
Zhao
,
L.
,
Katavic
,
V.
,
Li
,
F.
,
Haughn
,
G.W.
and
Kunst
,
L.
(
2010
)
Insertional mutant analysis reveals that long-chain acyl-CoA synthetase 1 (LACS1), but not LACS8, functionally overlaps with LACS9 in Arabidopsis seed oil biosynthesis
.
Plant J.
64
,
1048
1058
16
Jessen
,
D.
,
Roth
,
C.
,
Wiermer
,
M.
and
Fulda
,
M.
(
2015
)
Two activities of long-chain acyl-CoA synthetase are involved in lipid trafficking between the endoplasmic reticulum and the plastid in Arabidopsis
.
Plant Physiol.
167
,
351
366
17
Steen
,
E.J.
,
Kang
,
Y.
,
Bokinsky
,
G.
,
Hu
,
Z.
,
Schirmer
,
A.
,
McClure
,
A.
et al.  (
2010
)
Microbial production of fatty-acid-derived fuels and chemicals from plant biomass
.
Nature
463
,
559
562
18
Guo
,
X.
,
Jiang
,
M.
,
Wan
,
X.
,
Hu
,
C.
and
Gong
,
Y.
(
2014
)
Identification and biochemical characterization of five long-chain acyl-coenzyme A synthetases from the diatom Phaeodactylum tricornutum
.
Plant Physiol. Biochem.
74
,
33
41
19
Tan
,
X.
,
Zheng
,
X.
,
Zhang
,
Z.
,
Wang
,
Z.
,
Xia
,
H.
,
Lu
,
C.
et al.  (
2014
)
Long chain acyl-coenzyme A synthetase 4 (BnLACS4) gene from Brassica napus enhances the yeast lipid contents
.
J. Integr. Agric.
13
,
54
62
20
Tonon
,
T.
,
Qing
,
R.
,
Harvey
,
D.
,
Li
,
Y.
,
Larson
,
T.R.
and
Graham
,
I.A.
(
2005
)
Identification of a long-chain polyunsaturated fatty acid acyl-coenzyme A synthetase from the diatom Thalassiosira pseudonana
.
Plant Physiol.
138
,
402
408
21
Pulsifer
,
I.P.
,
Kluge
,
S.
and
Rowland
,
O
. (
2012
)
Arabidopsis long-chain acyl-CoA synthetase 1 (LACS1), LACS2, and LACS3 facilitate fatty acid uptake in yeast
.
Plant Physiol. Biochem.
51
,
31
39
22
Xu
,
N.
,
Zhang
,
S.O.
,
Cole
,
R.A.
,
McKinney
,
S.A.
,
Guo
,
F.
,
Haas
,
J.T.
et al.  (
2012
)
The FATP1-DGAT2 complex facilitates lipid droplet expansion at the ER-lipid droplet interface
.
J. Cell Biol.
198
,
895
911
23
Xu
,
Y.
,
Holic
,
R.
,
Li
,
D.
,
Pan
,
X.
,
Mietkiewska
,
E.
,
Chen
,
G.
et al.  (
2018
)
Substrate preferences of long-chain acyl-CoA synthetase and diacylglycerol acyltransferase contribute to enrichment of flax seed oil with α-linolenic acid
.
Biochem. J.
475
,
1473
1489
24
Roesler
,
K.
,
Shen
,
B.
,
Bermudez
,
E.
,
Li
,
C.
,
Hunt
,
J.
,
Damude
,
H.G.
et al.  (
2016
)
An improved variant of soybean type 1 diacylglycerol acyltransferase increases the oil content and decreases the soluble carbohydrate content of soybeans
.
Plant Physiol.
171
,
878
893
25
Chen
,
G.
,
Xu
,
Y.
,
Siloto
,
R.M.P.
,
Caldo
,
K.M.P.
,
Vanhercke
,
T.
,
El Tahchy
,
A.
et al.  (
2017
)
High performance variants of plant diacylglycerol acyltransferase 1 generated by directed evolution provide insights into structure-function
.
Plant J.
92
,
167
177
26
Kamisaka
,
Y.
,
Kimura
,
K.
,
Uemura
,
H.
and
Shibakami
,
M.
(
2010
)
Activation of diacylglycerol acyltransferase expressed in Saccharomyces cerevisiae: overexpression of Dga1p lacking the N-terminal region in the Deltasnf2 disruptant produces a significant increase in its enzyme activity
.
Appl. Microbiol. Biotechnol.
88
,
105
115
27
Xu
,
J.
,
Francis
,
T.
,
Mietkiewska
,
E.
,
Giblin
,
E.M.
,
Barton
,
D.L.
,
Zhang
,
Y.
et al.  (
2008
)
Cloning and characterization of an acyl-CoA-dependent diacylglycerol acyltransferase 1 (DGAT1) gene from Tropaeolum majus, and a study of the functional motifs of the DGAT protein using site-directed mutagenesis to modify enzyme activity and oil content
.
Plant Biotechnol. J.
6
,
799
818
28
Greer
,
M.S.
,
Truksa
,
M.
,
Deng
,
W.
,
Lung
,
S.-C.
,
Chen
,
G.
and
Weselake
,
R.J.
(
2015
)
Engineering increased triacylglycerol accumulation in Saccharomyces cerevisiae using a modified type 1 plant diacylglycerol acyltransferase
.
Appl. Microbiol. Biotechnol.
99
,
2243
2253
29
O'Quin
,
J.B.
,
Mullen
,
R.T.
and
Dyer
,
J.M.
(
2009
)
Addition of an N-terminal epitope tag significantly increases the activity of plant fatty acid desaturases expressed in yeast cells
.
Appl. Microbiol. Biotechnol.
83
,
117
125
30
Xu
,
Y.
,
Chen
,
G.
,
Greer
,
M.S.
,
Caldo
,
K.M.P.
,
Ramakrishnan
,
G.
,
Shah
,
S.
et al.  (
2017
)
Multiple mechanisms contribute to increased neutral lipid accumulation in yeast producing recombinant variants of plant diacylglycerol acyltransferase 1
.
J. Biol. Chem.
292
,
17819
17831
31
Belhaj
,
K.
,
Chaparro-Garcia
,
A.
,
Kamoun
,
S.
and
Nekrasov
,
V.
(
2013
)
Plant genome editing made easy: targeted mutagenesis in model and crop plants using the CRISPR/Cas system
.
Plant Methods
9
,
39
32
Kochan
,
G.
,
Pilka
,
E.S.
,
von Delft
,
F.
,
Oppermann
,
U.
and
Yue
,
W.W.
(
2009
)
Structural snapshots for the conformation-dependent catalysis by human medium-chain acyl-coenzyme A synthetase ACSM2A
.
J. Mol. Biol.
388
,
997
1008
33
Hisanaga
,
Y.
,
Ago
,
H.
,
Nakagawa
,
N.
,
Hamada
,
K.
,
Ida
,
K.
,
Yamamoto
,
M.
, et al.  (
2004
)
Structural basis of the substrate-specific two-step catalysis of long chain fatty acyl-CoA synthetase dimer
.
J. Biol. Chem.
279
,
31717
31726
34
Andersson
,
C.S.
,
Lundgren
,
C.A.K.
,
Magnúsdóttir
,
A.
,
Ge
,
C.
,
Wieslander
,
Å.
,
Molina
,
D.M.
et al.  (
2012
)
The Mycobacterium tuberculosis very-long-chain fatty acyl-CoA synthetase: structural basis for housing lipid substrates longer than the enzyme
.
Structure
20
,
1062
1070
35
Siloto
,
R.M.P.
,
Truksa
,
M.
,
Brownfield
,
D.
,
Good
,
A.G.
and
Weselake
,
R.J.
(
2009
)
Directed evolution of acyl-CoA:diacylglycerol acyltransferase: development and characterization of Brassica napus DGAT1 mutagenized libraries
.
Plant Physiol. Biochem.
47
,
456
461
36
Baek
,
S.C.
,
Ho
,
T.-H.
,
Lee
,
H.W.
,
Jung
,
W.K.
,
Gang
,
H.-S.
,
Kang
,
L.-W.
et al.  (
2017
)
Improvement of enzyme activity of β-1,3-1,4-glucanase from Paenibacillus sp. X4 by error-prone PCR and structural insights of mutated residues
.
Appl. Microbiol. Biotechnol.
101
,
4073
4083
37
Porter
,
J.L.
,
Boon
,
P.L.S.
,
Murray
,
T.P.
,
Huber
,
T.
,
Collyer
,
C.A.
and
Ollis
,
D.L.
(
2015
)
Directed evolution of new and improved enzyme functions using an evolutionary intermediate and multi-directional search
.
ACS Chem. Biol.
10
,
611
621
38
Ford
,
T.J.
and
Way
,
J.C.
(
2015
)
Enhancement of E. coli acyl-CoA synthetase FadD activity on medium chain fatty acids
.
PeerJ
3
,
e1040
39
McCullum
,
E.O.
,
Williams
,
B.A.R.
,
Zhang
,
J.
and
Chaput
,
J.C.
(
2010
) Random mutagenesis by error-prone PCR. In
In Vitro Mutagenesis Protocols. Methods in Molecular Biology (Methods and Protocols)
(
Braman
,
J.
ed.), Vol. 634.
Humana Press
,
Totowa, NJ
40
Meyer
,
A.
,
Schmid
,
A.
,
Held
,
M.
,
Westphal
,
A.H.
,
Röthlisberger
,
M.
,
Kohler
,
H.-P.E.
et al.  (
2002
)
Changing the substrate reactivity of 2-hydroxybiphenyl 3-monooxygenase from Pseudomonas azelaica HBP1 by directed evolution
.
J. Biol. Chem.
277
,
5575
5582
41
Fortin
,
P.D.
,
MacPherson
,
I.
,
Neau
,
D.B.
,
Bolin
,
J.T.
and
Eltis
,
L.D.
(
2005
)
Directed evolution of a ring-cleaving dioxygenase for polychlorinated biphenyl degradation
.
J. Biol. Chem.
280
,
42307
42314
42
Bradford
,
M.M.
(
1976
)
A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding
.
Anal. Biochem.
72
,
248
254
43
Schneider
,
C.A.
,
Rasband
,
W.S.
and
Eliceiri
,
K.W.
(
2012
)
NIH image to ImageJ: 25 years of image analysis
.
Nat. Methods
9
,
671
675
44
Kumar
,
S.
,
Stecher
,
G.
and
Tamura
,
K.
(
2016
)
MEGA7: Molecular Evolutionary Genetics Analysis version 7.0 for bigger datasets
.
Mol. Biol. Evol.
33
,
msw054
45
Santorum
,
J.M.
,
Darriba
,
D.
,
Taboada
,
G.L.
and
Posada
,
D.
(
2014
)
Jmodeltest.org: selection of nucleotide substitution models on the cloud
.
Bioinformatics
30
,
1310
1311
46
Guindon
,
S.
,
Lethiec
,
F.
,
Duroux
,
P.
and
Gascuel
,
O.
(
2005
)
PHYML Online — a web server for fast maximum likelihood-based phylogenetic inference
.
Nucleic Acids Res.
33
,
W557
W559
47
Guindon
,
S.
,
Dufayard
,
J.-F.
,
Lefort
,
V.
,
Anisimova
,
M.
,
Hordijk
,
W.
and
Gascuel
,
O.
(
2010
)
New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of PhyML 2.0
.
Syst. Biol.
59
,
307
321
48
Yang
,
Z.
(
2007
)
PAML 4: phylogenetic analysis by maximum likelihood
.
Mol. Biol. Evol.
24
,
1586
1591
49
Yang
,
Z.
,
Wong
,
W.S.W.
and
Nielsen
,
R.
(
2005
)
Bayes empirical Bayes inference of amino acid sites under positive selection
.
Mol. Biol. Evol.
22
,
1107
1118
50
Kelley
,
L.A.
,
Mezulis
,
S.
,
Yates
,
C.M.
,
Wass
,
M.N.
and
Sternberg
,
M.J.
(
2015
)
The Phyre2 web portal for protein modeling, prediction and analysis
.
Nat. Protoc.
10
,
845
858
51
Gahloth
,
D.
,
Dunstan
,
M.S.
,
Quaglia
,
D.
,
Klumbys
,
E.
,
Lockhart-Cairns
,
M.P.
,
Hill
,
A.M.
et al.  (
2017
)
Structures of carboxylic acid reductase reveal domain dynamics underlying catalysis
.
Nat. Chem. Biol.
13
,
975
981
52
Yang
,
J.
,
Yan
,
R.
,
Roy
,
A.
,
Xu
,
D.
,
Poisson
,
J.
and
Zhang
,
Y.
(
2015
)
The I-TASSER Suite: protein structure and function prediction
.
Nat. Methods
12
,
7
8
53
Biswas
,
S.
and
Akey
,
J.M.
(
2006
)
Genomic insights into positive selection
.
Trends Genet.
22
,
437
446
54
Yuan
,
H.
,
Wu
,
J.
,
Wang
,
X.
,
Chen
,
J.
,
Zhong
,
Y.
,
Huang
,
Q.
et al.  (
2017
)
Computational identification of amino-acid mutations that further improve the activity of a chalcone-flavonone osomerase from Glycine max
.
Front. Plant Sci.
8
,
1
8
55
Black
,
P.N.
,
Zhang
,
Q.
,
Weimar
,
J.D.
and
DiRusso
,
C.C.
(
1997
)
Mutational analysis of a fatty acyl-coenzyme A synthetase signature motif identifies seven amino acid residues that modulate fatty acid substrate specificity
.
J. Biol. Chem.
272
,
4896
4903
56
Weimar
,
J.D.
,
DiRusso
,
C.C.
,
Delio
,
R.
and
Black
,
P.N.
(
2002
)
Functional role of fatty acyl-coenzyme A synthetase in the transmembrane movement and activation of exogenous long-chain fatty acids
.
J. Biol. Chem.
277
,
29369
29376
57
Suplatov
,
D.A.
,
Besenmatter
,
W.
,
Švedas
,
V.K.
and
Svendsen
,
A.
(
2012
)
Bioinformatic analysis of alpha /beta-hydrolase fold enzymes reveals subfamily-specific positions responsible for discrimination of amidase and lipase
.
Protein Eng. Des. Sel.
25
,
689
697
58
Caldo
,
K.M.P.
,
Acedo
,
J.Z.
,
Panigrahi
,
R.
,
Vederas
,
J.C.
,
Weselake
,
R.J.
and
Lemieux
,
M.J.
(
2017
)
Diacylglycerol acyltransferase 1 is regulated by its N-terminal domain in response to allosteric effectors
.
Plant Physiol.
175
,
667
680
59
Waterhouse
,
A.M.
,
Procter
,
J.B.
,
Martin
,
D.M.A.
,
Clamp
,
M.
and
Barton
,
G.J.
(
2009
)
Jalview Version 2—a multiple sequence alignment editor and analysis workbench
.
Bioinformatics
25
,
1189
1191