Identification of a Residue Affecting Fatty Alcohol Selectivity in Wax Ester Synthase

Abstract

The terminal enzyme in the bacterial wax ester biosynthetic pathway is the bifunctional wax ester synthase/acyl-coenzyme A:diacylglycerol acyltransferase (WS/DGAT), which utilizes a fatty alcohol and a fatty acyl-coenzyme A (CoA) to synthesize the corresponding wax ester. In this report, we identify a specific residue in WS/DGAT enzymes obtained from Marinobacter aquaeolei VT8 and Acinetobacter baylyi that alters fatty alcohol selectivity and kinetic parameters when modified to alternative residues.

TEXT

Wax esters are high-value neutral lipid compounds that serve a variety of functions in biological systems and have specific industrial uses as cosmetics, high-grade lubricants, and food additives (1–3). While common in many higher organisms such as plants and animals, wax esters are produced in only a small selection of bacteria (4). In bacteria, the enzyme responsible for catalyzing the esterification of a fatty acyl-coenzyme A (CoA) and a fatty alcohol is referred to as the wax ester synthase/acyl-coenzyme A:diacylglycerol acyltransferase (WS/DGAT) (EC 2.3.1.75; Fig. 1) (5). Together with the fatty acyl-CoA reductase and fatty aldehyde reductase enzymes (6–8) which provide the requisite fatty alcohol, the WS/DGAT enzyme is proposed to produce wax esters in bacteria from the fatty acyl-CoA pool (4). The WS/DGAT from Acinetobacter has been extensively studied, while WS/DGAT enzymes from additional species have been characterized to a lesser extent (5, 9–12). One interesting feature of most WS/DGAT enzymes characterized to date is the broad substrate range found for various fatty acyl-CoA and fatty alcohol substrates, though the enzyme does exhibit a certain level of selectivity for specific substrates in preference to others (9, 11, 13). Although WS/DGAT is of interest for potential biotechnology applications (14), further studies to tailor or improve the characteristics of the enzyme activity directed toward specific substrates have been hindered by the lack of structural information for the rational design of experiments.

Fig 1.

Reaction scheme of the wax ester synthase and carnitine o-acetyltransferase reactions. The wax ester synthase/acyl-coenzyme A:diacylglycerol acyltransferase enzyme (EC 2.3.1.75) catalyzes the esterification of a fatty acyl-CoA and a fatty alcohol, yielding free coenzyme A and the corresponding wax ester. Carnitine O-acetyltransferase (EC 2.3.1.7) catalyzes the esterification of acetyl-CoA and carnitine, yielding free coenzyme A and O-acetylcarnitine.

WS/DGAT belongs to a broader family of enzymes classified as acyltransferases (EC 2.3.1.X) and includes enzymes such as the polyketide-associated protein involved in the production of phthiocerol and phthiodiolone dimycocerosate esters (15, 16), carnitine acyltransferases (Fig. 1) that serve important roles in fatty acid transport and beta-oxidation (17–19), and xenobiotic acetyltransferases such as chloramphenicol acetyltransferase that are involved in antibiotic resistance (20). Although these enzymes share little similarity in protein primary sequences, common folds have been found in the structures of many of these enzymes, with specific residues composing the catalytic active site found in similar orientations (15, 18, 20).

Structural analysis.

We previously compared and contrasted the rates of activity and selectivity of five different WS/DGAT enzymes and presented the findings from those studies such that differences in activity along with a primary sequence alignment of the various enzymes could be utilized as a potential blueprint for identifying key residues affecting the activity and selectivity of these diverse enzymes (9). One key factor hindering potential applied studies of WS/DGAT enzymes is the lack of a published WS/DGAT structure (12), and homologies between WS/DGAT enzymes and current structures for enzymes sharing the active site motif are low (13). Our approach here used several enzymes that catalyze similar reactions to the WS/DGAT enzyme (Fig. 1) and share a very minimally conserved protein primary sequence (Fig. 2). Protein structures were visualized and manipulated using Discovery Studio Visualizer (Accelrys, Inc., San Diego, CA). For this study, an approach was taken using the polyketide-associated protein (PapA5) crystal structure (Protein Data Bank [pdb] file 1Q9J [15]) and three-dimensional alignments of carnitine acetyltransferases with bound substrates (pdb files 2RCU, 1NDI, 2H3W, and 2H3P [17–19]) to visualize where various substrate binding sites might exist within the PapA5 structure in a manner similar to what has been reported previously (15). As several of the structures utilized in alignments have bound substrates or analogs, this approach further provides some idea of the general location where substrates might bind in the WS/DGAT enzyme. One primary assumption made in this approach is that carnitine and fatty alcohols would bind to similar locations within the fold if the mechanism of the enzyme catalysis and general orientation of substrates were conserved. In this manner, the location of the carnitine bound in specific structures serves as a point of reference for selecting residues to target within the protein. Further, an alignment based on the minimal similarity between PapA5 and the WS/DGAT from M. aquaeolei VT8 (Ma1) was performed using MultAlin (21) and utilized to select residues lining putative binding pockets of the enzyme structure in an attempt to block access to specific substrates as illustrated in Fig. 2. While the approach is based on a number of broad assumptions, it is demonstrated here to be successful in identifying a potential residue lying more than 200 amino acids downstream of the active site motif (HHXXXDG) that has been shown to be essential to enzyme activity (12).

Fig 2.

Structure of the active site from polyketide synthase. A model used to select potential residues for modification was assembled by first aligning pdb file 2H3P to pdb file 1Q9J using the active site motif residue HHXXXDG (EHXXXEG in pdb 2H3P) and conserved fold elements as the basis for alignment. This figure shows only a cutaway of the protein backbone and active site residues H123, H124, D128, and G129 along with the targeted F331 residue from pdb file 1Q9J. Acetyl-CoA and carnitine (from pdb 2H3P) based on the alignments are also shown. Sequence alignments of the polyketide synthase (pdb 1Q9J) and M. aquaeolei VT8 wax ester synthase (Ma1) primary sequences were determined using multalin. Shown below the structure are a protein primary sequence alignment of the active site region and the region targeted for the modification to A360 in M. aquaeolei VT8 (corresponding to F331 in 1Q9J). The three-dimensional image was generated using POV-Ray (Persistence of Vision Pty. Ltd., Williamstown, Victoria, Australia).

Plasmid construction, protein purification, and activity assays.

The plasmids utilized were described previously (9). Site-specific mutagenesis was performed using the Stratagene method (Agilent Technologies, Santa Clara, CA) and the expression plasmid containing the Marinobacter aquaeolei VT8 gene (YP_957462) as a template. Primers to modify the residue at position 360 were based on the sequence 5′ GACCCT GGCGCCGGCC XXX TTCCACCTGC TGAC 3′ and a complementary primer, where XXX represents ATC (A360I), TTC (A360F), or GTG (A360V). An additional primer (5′ GTATA TGGCCCTGCA ATA CTCAACATAA TTTCTG 3′) and complementary primer were used to modify the plasmid containing the Acinetobacter baylyi gene (YP_045555) to isoleucine at position 355 (G355I). Prior to expression, all plasmids were confirmed by sequencing the entire inserted region. Proteins were expressed in Escherichia coli TB1 (New England BioLabs, Ipswich, MA) and purified using the rapid two-step affinity purification scheme described previously (9). Specific activity measurements and selectivity assays followed by gas chromatography analysis were all performed as described previously using wild-type enzymes as controls (9). All results presented are based on a minimum of three samples.

Enzyme selectivity assays.

The rapid gas chromatography (GC) assay described previously (9) was utilized as an initial screen of modified WS/DGAT enzymes for substrate selectivity. One specific change of residue 360 of Ma1 from alanine to isoleucine (Ma1-A360I) resulted in a shift in the selectivity toward the chain length of the fatty alcohol substrates (Fig. 3A). In the wild-type enzyme (Ma1), the results of the selectivity assay indicate that undecanol and dodecanol were the most favored substrates, whereas in Ma1-A360I, decanol and undecanol were now favored, and significant increases in selectivity were found for several smaller fatty alcohols. Further substitutions to larger and smaller residues (phenylalanine and valine) also resulted in similar but distinct shifts in the selectivity profile. As a further test of this region of WS/DGAT, the corresponding residue 355 from the A. baylyi enzyme (Ac1) was also changed from glycine to isoleucine (Ac1-G355I), resulting in a similar shift in the substrate profile (Fig. 3B). While the difference in selectivity was subtle, the change was consistent and reproducible.

Fig 3.

Wax ester product distribution indicating fatty alcohol selectivity of WS/DGAT enzymes in the presence of palmitoyl-CoA. Shown are the percentages of fatty alcohol-derived wax esters formed when 50 nmol of each of 10 different fatty alcohols was combined with 200 nmol of palmitoyl-CoA (C16-CoA). Assays were run for 15 min in the presence of each enzyme, and then the lipids were extracted from the assay and analyzed by gas chromatography. The area of each peak was converted to a percentage of the total of all wax esters obtained. The x axis is labeled as to the alcohol component of the wax ester (i.e., nonyl hexadecanoate derived from nonanol and palmitoyl-CoA [hexadecanoyl-CoA]). The upper graph (A) shows the various modifications to the alanine at position 360 in Ma1 from M. aquaeolei VT8 versus the wild-type enzyme. The lower graph (B) shows the wild-type enzyme Ac1 for A. baylyi along with the modification to glycine at position 355 in this enzyme.

Specific activity assays.

In addition to the GC assay, specific activity measurements were also performed with several fatty alcohols to determine if the modifications had altered kinetic parameters. The activity found for dodecanol was similar to what has been reported previously (9) and was similar for each modified protein versus that of the wild-type enzyme (Table 1). Further, the activity found for hexadecanol, a natural substrate found in most wax ester-accumulating bacteria tested previously (9), remained similar as well. Since activity measurements fluctuate slightly in different aliquots of thawed protein, each set of assays were performed with the same enzyme stock, so that direct comparisons between different alcohols could be made for each protein. The ratio of activity of hexadecanol versus dodecanol, which is the best measure of substrate selectivity, is further included in Table 1, and indicates that the ratio of activity for hexadecanol versus dodecanol remained relatively constant in the modified proteins. However, when nonanol was utilized in the specific activity measurements, the rate of reaction and ratio of activity for nonanol versus dodecanol were found to increase for all three modifications made to residue 360 in Ma1 and paralleled the results from the selectivity assays for this alcohol, further supporting the idea of the importance of this residue in substrate selectivity. A similar shift in the ratio of selectivity for nonanol versus dodecanol was found for the A. baylyi WS/DGAT and the corresponding modification at position 355.

Table 1.

WS/DGAT specific activity with various fatty alcohols

Protein	Sp act with indicated fatty alcohol substrate (nmol of product min−1 [mg of protein]−1)a			C9OH/C12OHb (%)	C16OH/C12OHc (%)
	Dodecanol	Nonanol	Hexadecanol
Ma1—wild type	46,500 ± 1,900	3,600 ± 290	2,700 ± 270	7.7	5.8
Ma1—A360V	35,900 ± 1,800	16,600 ± 860	1,640 ± 90	46.3	4.6
Ma1—A360I	34,300 ± 600	19,600 ± 500	1,600 ± 40	56.9	4.7
Ma1—A360F	34,600 ± 2,600	7,800 ± 230	1,850 ± 40	22.5	5.4
Ac1—wild typed	960 ± 45	140 ± 10	70 ± 5	14.9	7.0
Ac1—G355Id	1,700 ± 90	360 ± 15	80 ± 5	21.4	4.7

^aSpecific activities are based on the results determined with at least three samples. All assays utilized palmitoyl-CoA as the second substrate added and were run as described in the text.

^bCalculated by dividing the average specific activity for nonanol by the average specific activity for dodecanol and converting to a percentage.

^cCalculated by dividing the average specific activity for hexadecanol by the average specific activity for dodecanol and converting to a percentage.

^dIn the Ac1 enzymes analyzed for nonanol, two rates were observed. The initial rate began lower but more than doubled as the substrate was consumed. The rate shown is the initial rate. This phenomenon was found only for this enzyme and substrate.

Study implications.

The modification initially selected for position 360 (Ma1-A360I) was intended to potentially block the putative substrate pocket where it is presumed a fatty alcohol would be positioned. Contrary to our expectation, this modification did not block accessibility of larger fatty alcohols such as hexadecanol but did improve the rate of reaction found for smaller fatty alcohols (Table 1) such as nonanol (increasing the specific activity by about 8-fold), which resulted in an increased production of nonyl hexadecanoate during the initial screen of wax ester products (Fig. 3). Further mutations to other residues at this same position resulted in slight alterations to the selectivity based on the obtained wax ester profile, but none of these modifications blocked hexadecanol or dramatically decreased rates of activity directed toward this substrate (Table 1). The results support a proposal where the modifications made simply improve binding of smaller substrates to the active site and further support the potential for future efforts to improve activity with a number of other smaller or larger substrates. Such alterations could be beneficial for specific biotechnological applications (14).

Specifically for the region of the protein primary sequence in Ma1 where the alanine 360 residue is located, the conservation level is relatively poor among those WS/DGAT enzymes previously studied (9). However, it is expected that the fold of WS/DGAT is likely conserved and thus that residues in other enzymes might be in the same general location in the other WS/DGAT enzymes. The extension of the result from Ma1 to Ac1 further supports the likelihood that the two residues lie in similar locations within the folds of these enzymes and demonstrates that approaches based on related acyltransferase enzymes and minimal alignments could be beneficial in identifying potential residues for studies in the absence of a published protein structure for a WS/DGAT enzyme.

Others have reported differences in activity when probing additional residues in WS/DGAT enzymes related to the active site motif (12). The modifications made here are important, as they lie a significant distance downstream of the conserved active site motif (starting at 135 and ending at 141 in Ma1) and did not result in a significant drop in activity, indicating that modifications could be made to improve substrate selectivity without negatively affecting enzyme stability and catalytic efficiency.