Expression and Purification of Integral Membrane Fatty Acid Desaturases

Fatty acid desaturase enzymes perform dehydrogenation reactions leading to the insertion of double bonds in fatty acids, and are divided into soluble and integral membrane classes. Crystal structures of soluble desaturases are available; however, membrane desaturases have defied decades of efforts due largely to the difficulty of generating recombinant desaturase proteins for crystallographic analysis. Mortierella alpina is an oleaginous fungus which possesses eight membrane desaturases involved in the synthesis of saturated, monounsaturated and polyunsaturated fatty acids. Here, we describe the successful expression, purification and enzymatic assay of three M. alpina desaturases (FADS15, FADS12, and FADS9-I). Estimated yields of desaturases with purity >95% are approximately 3.5% (Ca. 4.6 mg/L of culture) for FADS15, 2.3% (Ca. 2.5 mg/L of culture) for FADS12 and 10.7% (Ca. 37.5 mg/L of culture) for FADS9-I. Successful expression of high amounts of recombinant proteins represents a critical step towards the structural elucidation of membrane fatty acid desaturases.


Introduction
Lipids are first synthesized as saturated fatty acids, and double bonds are introduced post-synthetically by oxygen-dependent enzymes known as fatty acid desaturases, in a process that is initiated by abstraction of hydrogen from a methylene group. Fatty acid desaturases are divided into soluble and integral membrane classes, which may have evolved independently [1]. The acyl-ACP desaturases are soluble enzymes found in the plastids of higher plants, whereas the more widespread class of integral membrane acyl-CoA desaturases is found in endomembrane systems in prokaryotes and eukaryotes [2]. Fatty acid desaturases in each class are closely related homologs based on their amino acid sequences, and yet perform highly regio-and stereo-selective reactions on long-chain fatty acids composed of essentially equivalent methylene chains that lack distinguishing landmarks close to the site of desaturation. As pointed out by Nobel Laureate Dr. Konrad Bloch, this region-and stereo-specific removal of hydrogen ''would seem to approach the limits of the discriminatory power of enzymes'' [3].
The membrane class of desaturases consists of enzymes with c5, c6, c9, c12 or v3 regio-selectivity. Structure determination would significantly improve our understanding of the structure-function relationships of this diverse class of proteins; however, there has been little progress, despite decades of efforts, due largely to the difficulty of generating recombinant membrane desaturase proteins for crystallographic analysis. Mammalian cells possess c5, c6 and c9, but lack c12 and v3 desaturases [4,5]. Mortierella alpina belongs to the subphylum of Mucoromycotina [6]. It is an oleaginous fungus that can produce lipids up to 50% of its dry weight. We have recently characterized the M. alpina genome [7] which encodes one c5, two c6, three c9, one c12 and one v3 desaturase ( Figure S1). Therefore, M. alpina has all known regioselective groups of membrane desaturases.
Membrane desaturases have been expressed in various hosts, such as Escherichia coli, Saccharomyces cerevisiae, Aspergillus oryzae, Mortierella alpina and cell-free system, for biochemical characterizations [8,9,10,11,12,13,14,15,16]. However, none of these expression systems could achieve sufficient amount of desaturase proteins for crystal structure analysis. In the present study, we expressed M. alpina c9, c12 and v3 desaturases (FADS9-I, FADS12 and FADS15) in the methylotrophic yeast Pichia pastoris, purified the recombinant proteins and determined their enzymatic activities. High yield of these proteins paves the way for structural characterization of membrane desaturases.

Expression Vector Construction
M. alpina RNA extraction was performed using Trizol Reagent (Invitrogen, CA) according to the manufacturer's instructions. Total RNA was reverse transcribed with SuperScriptH III First-Strand Synthesis SuperMix (Invitrogen) following the manufacturer's instructions. Using both C-and N-terminal sequences as primers (Table S1), desaturase coding sequences were PCR amplified as follows: denaturation at 95uC for 30 sec, annealing at 55uC for 45 sec and extension at 72uC for 1 min for 35 cycles. The amplified products were cloned into a modified pET19 vector (Novagen) derivative containing a PreScission protease cleavage site (GE Healthcare) between the multiple cloning site and Nterminal His tag ( [17] to construct pET19b-FADS15, pET19b-FADS12 and pET19b-FADS9-I). The desaturase genes, including the His-Tag and PreScission protease cleavage site, were then PCR amplified using primers SF1 and SR1-SR3 (Table S1). The PCR conditions used were the same as the first step for cDNAs. The PCR fragments were then purified and inserted into pPinka-HC to generate the expression vectors pPinka-HC-FADS15, pPinka-HC-FADS12 and pPinka-HC-FADS9-I. The presence of the inserts in the plasmids was confirmed by restriction digestion analysis and sequencing. The strategy used for constructing desaturase expression vectors is shown in Figure S2.

PichiaPink Transformation
Desaturase expression vectors and pPinka-HC (negative control vector) were linearized with restriction enzyme Spe I and transformed into P. pastoris strains (PichiaPink strain 1, 2, 3 and 4) using the MicroPulser Electroporator (Bio-Rad Laboratories, Hercules, CA) according to the User Manual of PichiaPink Expression System (Invitrogen). P. pastoris were incubated with YPDS media (YPD with 1 M sorbitol) in the Gene Pulser Cuvettes at 28uC for 2 hr without shaking, spread onto PAD (Pichia Adenine Dropout) agar selection plates, and then incubated at 28uC for 4 days until distinct colonies were formed. Eight white colonies for each transformation were picked and plasmid integration in the yeast genome was confirmed by PCR.
Isolated clones were individually inoculated into 10 mL of BMGY medium (Buffered Glycerol-complex Medium, 1% yeast extract; 2% peptone; 100 mM potassium phosphate, pH 6.0; 1.34% YNB-Yeast Nitrogen Base; 0.0004% biotin; 1% glycerol) in 50 mL conical tubes. The cells were grown for 48 hr at 28uC with vigorous shaking at 250 rpm. Then, the cultures were centrifuged at 1,500 g for 5 min at room temperature, the cell pellets were resuspended in 2 mL of BMMY medium (Buffered Methanolcomplex Medium, 1% yeast extract; 2% peptone; 100 mM potassium phosphate, pH 6.0; 1.34% YNB; 0.0004% biotin; 0.5% methanol) and cultured at 28uC with shaking at 250 rpm to induce the expression. After continuous cultivation for 72 hr with daily addition of 0.5% methanol, cells were harvested by centrifuging for 10 min at 1500 g. Supernatant was transferred to a separate tube and both the supernatant and cell pellet were stored at 280uC until ready for assay. Supernatants and cell pellets were analyzed for protein expression by SDS-PAGE Coomassie blue staining and Western blot.

Optimized Protein Expression Condition
Individual colonies of P. pastoris-FADS15, FADS12 and FADS9-I were inoculated into 10 mL of BMGY medium in 50 mL conical tubes and cultured for 48 hr at 28uC at shaking speed of 250 rpm. Then, 2.5 mL of culture were inoculated into 50 mL of BMGY medium in 250-mL volume shaker flasks and grown at 28uC for 24 hr at 250 rpm. The cells were collected by centrifugation at 1500 g for 10 min, and resuspended in 10 mL induction medium (BMMY medium with 0.5% methanol) in a 100-mL shaker flask. The induction of protein expression was performed for 96 hr at 28uC with 250 rpm agitation and daily addition of 0.5% methanol. Samples were collected at 0, 6, 24, 48, 72 and 96 hr for measuring cell density at OD 600 , wet cell weight and total protein concentration, and for Western blot analysis of desaturase expression levels.

Protein Analyses
The cell pellets and supernatants were collected by centrifuging 100 mL cell culture at 1500 g for 10 min. Cell pellets were resuspended in 100 mL lysis buffer (20 mM Tris.Cl pH7.9, 1 mM EDTA, 5% Glycerol) with an equal volume of 0.5 mm Glass Beads (Biospec products, Inc.), and vortexed for 10 min at 4uC. Cell lysates were mixed with 46SDS sample buffer and heated for 5 min at 95uC. About 5 ml sample was loaded onto Mini-Protein Precast Gels (4-15%, Bio-Rad Laboratories, Cat #456-1086), and ran for 40 min at 150 V. Then, the SDS-PAGE gels were used for Coomassie blue stain, Invision His-Tag in-gel stain (Invitrogen) or Western blot.
The total protein concentration was determined with Pierce BCA protein assay kit (Thermo Scientific). The quantification of target protein on Coomassie blue stained gel was performed using known concentrations of BSA as standard, and analyzed with the AlphaView SA software (Cell Biosciences, Inc.).

Cell Fractionation
All purification procedures were performed at 4uC. Cells harvested from 800 mL of culture were suspended in 800 mL of lysis buffer. After addition of 0.5 mm glass beads to the cell suspension, P. pastoris cells were disrupted by vortexing at 4uC for 10 min. Cell lysis efficiency was usually more than 95% evaluated using a light microscope. Intact cells and cell debris were removed from the membrane suspension by low speed centrifugation (500 g, 10 min at 4uC). Then various centrifugation speeds and time (1,000 g for 10 min; 10,000 g for 10 min; 10,000 g for 20 min; 20,000 g for 10 min; 20,000 g for 20 min) were used to determine the best centrifugation conditions for collecting the membrane fraction.

Protein Affinity Purification
Optimized culture and protein solubilization conditions were used for the subsequent purification process. His Mag Sepharose TM Ni affinity beads (GE Healthcare) were washed with binding buffer (20 mM Tris.Cl, pH 7.9, 500 mM NaCl, 10% glycerol, 0.1 mM EDTA, 0.5% Fos-Choline 16, 5 or 20 mM imidazole) and added to the solubilized fractions after detergent incubation. The bead-protein sample mixtures were incubated for 45 min at 4uC with end-over-end mixing. After washing three times with binding buffer containing 5 mM or 20 mM imidazole, desaturase enzymes were eluted with elution buffer (20 mM Tris.Cl, pH 7.9, 500 mM NaCl, 10% glycerol, 0.1 mM EDTA, 0.5% Fos-Choline 16, 500 mM imidazole). The purified FADS15, FADS12 and FADS9-I proteins were stored at 280uC in aliquots.
The quantity and quality of these purified enzymes were analyzed by SDS-PAGE, mass spectrometry and desaturase activity assay. Protein purity was presented in percentage, dividing the amount of a given desaturase protein quantified on gel by the amount of loaded protein quantified by the BCA protein assay.

Fatty Acid Analysis
Approximately 20 mg of P. pastoris cell pellets were collected and used for each lipid extraction with the method of Bligh and Dyer [18] under acidified conditions with pentadecanoic acid and heneicosanoic acid added as internal standards. The solvent from the extract was removed under a stream of nitrogen. Lipids were saponified in 1 mL of freshly prepared 5% ethanolic potassium hydroxide at 60uC for 1 hr under an argon atmosphere. After cooling, 1 mL of water was added to the samples and nonsaponifiable lipids were extracted into 3 mL of hexane. The aqueous layer was acidified with 220 mL of 6 M hydrochloric acid and the fatty acids extracted into 3 mL of hexane. After removing the hexane in a stream of nitrogen, fatty acids were converted to methyl esters by first treating with 1 mL of 0.5 M methanolic sodium hydroxide at 100uC for 5 min under argon followed by 1 mL of 14% methanolic boron trifluoride at 100uC for 5 min under argon [19]. After cooling, the sample was mixed with 2 mL of hexane followed by 4 mL of saturated aqueous sodium chloride. After separating the phases, aliquots of the hexane layers were diluted 24-fold with hexane and then analyzed by GC/MS. One mL was injected in the splitless mode onto a 30 m6250 mm DB-WAXETR column (Agilent Technologies, Santa Clara, Califor- nia) with 0.25 mm film thickness. The temperature program was as follows: 100uC for 2 min, ramp to 200uC at 16uC per min, hold for one min, ramp to 220uC at 4uC per min, hold one min, ramp to 260uC at 10uC per min, and hold for 11 min. Helium was the carrier gas at a constant flow of 1.5 mL/min. The mass spectrometer was operated in positive-ion electron impact mode with interface temperature 260uC, source temperature 200uC, and filament emission 250 mA. Spectra were acquired from m/z 50 to 450 with a scan time of 0.433 s. Lower-boiling fatty acid methyl esters were quantified using the pentadecanoic acid internal standard, whereas higher-boiling methyl esters were quantified using the heneicosanoic acid internal standard.  In vivo Desaturase Activity Analysis Individual colonies of P. pastoris-FADS15, FADS12 and FADS9-I were cultured as described in the Recombinant protein expression section. Protein expression was induced for 72 hr with 0.5% methanol. Cell pellets were collected by centrifugation and stored at 280uC for fatty acid analysis.
In vitro Desaturase Activity Analysis 20 mL of the purified protein was added to 200 mL of yeast EGY49 cell homogenate, prepared by breaking cells with 0.5 mm glass beads in lysis buffer (20 mM Tris-HCl pH7.9, 1 mM EDTA, 5% Glycerol). The enzyme reactions were performed at 28uC for 3 h with shaking (250 rpm), and the assay mixture (220 mL) were stored at 280uC for fatty acid analysis.

Sequence of M. alpina Fatty Acid Desaturases
Recently, we sequenced the genome and EST of M. alpina ATCC #32222. The cDNA and protein sequences of FADS9-I, FADS12 and FADS15 were compared to published sequences from other strains of M. alpina ( Figure S3-S5). The FADS12 and FADS9-I genes from M. alpina ATCC#32222 are 99.9% and 98.4% identical, respectively, to the corresponding genes from M. alpina 1s-4. The FADS12 and FADS9-I proteins from M. alpina ATCC#32222 are 100% and 99.6% identical, respectively, to these proteins from M. alpina 1s-4. The high similarity of FADS12 and FADS9-I genes between two strains indicates that these genes are highly conserved in M. alpina. Interestingly, the FADS15 gene is much less conserved at both DNA (93.1% identity) and protein (97.9%) levels.

Expression of Fatty Acid Desaturases in the PichiaPink System
After several unsuccessful attempts to express recombinant M. alpina desaturases in bacteria, we tried to express FADS15, FADS12 and FADS9-I in the Pichia pastoris PichiaPink expression system (Invitrogen). Our data showed that PichiaPink strain 2(ade2, pep4) supported the highest level of expression for FADS15, 12 and 9-I. Interestingly, all recombinant desaturase proteins remained on the cell membrane despite the presence of afactor secretion signal, whereas EGFP (enhance green fluorescent protein) was successfully secreted into culture medium ( Figure  S6).
To determine potential toxicity of recombinant proteins, we first examined cell growth density, weight and total protein synthesis of the PichiaPink pPinka-HC-FADS clones. The recombinant PichiaPink pPinka-HC-FADS cells had growth characteristics similar to the control ( Figure 1A). A time course experiment showed that desaturase expression was detectable after 24 hr induction with 0.5% methanol and remained high for at least 72 hr post-induction ( Figure 1B). There were no significant differences in protein expression when cells were induced at different temperatures (16uC, 22uC, 28uC) or with a different concentration of methanol (0.5%, 1%). Therefore, we used an optimized procedure as described in the Materials and Methods for the expression of recombinant desaturase. Under this condition, expression levels of recombinant desaturase proteins reached approximately 130 mg/L of culture for FADS15, 110 mg/L for FADS12 and 350 mg/L for FADS9-I ( Figure 1C and Table 1). FADS15 and FADS12 recombinant proteins overlapped with endogenous background proteins in Coomassie blue staining gels. The InVision TM His-tag In-Gel staining, however, showed clearly the expression of FADS15 and FADS12 over control ( Figure 1D).

Solubilizaton and Purification of Recombinant Desaturases
In order to solubilize and purify the recombinant desaturases from cell membrane for in vitro enzymatic activity, we first tested conditions to enrich the cell membrane containing recombinant FADS15, FADS12 and FADS9-I. Different centrifugation speeds and times were examined for the separation of the membrane fractions containing target proteins. Efficient recovery of each recombinant desaturase produced in P. pastoris was achieved by centrifuging the cell homogenates at 500 g for 10 min to remove cell debris, then at 10, 000 g for 10 min to collect membranes ( Figure 2).
Solubilization of membrane proteins requires the presence of detergents. Therefore, we tested the conditions for solubilization of the recombinant FADS15, FADS12 and FADS9-I from enriched cell membrane fractions using a panel of detergents: Tween-20, Tween-80, NP-40, n-Dodecyl-b-D-maltoside (DDM), Fos-Choline 12 or Fos-Choline 16. After treatment with 1% (w/v) of Fos-Choline 12 or Fos-Choline 16, FADS9-I and FADS12 were totally solubilized, and approximately 50% and 80% of FADS15 was solubilized with Fos-Choline 12 and Fos-Choline 16, respectively ( Figure 3A). Tween-20, Tween-80, NP-40 and DDM had little effect on extracting these desaturase enzymes from the membrane. In addition, we noticed that FADS9-I protein degradation occurred during protein solubilization. This phenomenon was visible for proteins solubilized by both Fos-Choline 12 and 16. Thus, we investigated detergent incubation time during solubilization to optimize for the least protein degradation. Our results showed that the solubilization of FADS9-I protein reached its maximum level after incubation with detergent for 1.5 hr. Degradation of desaturase protein increased after more than 3 hr of incubation ( Figure 3B). To maximize the ratio of intact vs. degraded proteins, we used 1.5 hr as our standard detergent incubation time for protein solubilization. We also compared the effect of detergent concentrations on protein solubilization efficiency and found that 0.5%, 1% or 2% of Fos-Choline 16 had similar effects. Taken together, our results indicate that all three recombinant desaturase enzymes can be solubilized efficiently from the cell membrane with 0.5% Fos-Choline 16 for 1.5 hr at 4uC. Solubilized FADS15, FADS12 and FADS9-I were affinitypurified on His Mag Sepharose Ni beads (GE healthcare) with aims of high purity or high yield. High purity (.95%) was achieved after one step purification using the His Mag Sepharose Ni beads with high stringency wash before elution ( Figure 3C, S7). High yield (2-fold higher than that in the high purity process) was achieved with low stringency wash. Yield and quantity of each desaturase enzyme are summarized in Table 1. Our estimated yields of desaturases with purity .95% are approximately 3.5% (Ca. 4.6 mg/L of culture) for FADS15, 2.3% (Ca. 2.5 mg/L of culture) for FADS12 and 10.7% (Ca. 37.5 mg/L of culture) for FADS9-I.
Saccharomyces cerevisiae is an excellent experimental system to study fatty acid desaturation, as it provides a eukaryotic endoplasmic reticulum, cytochrome b5 and NADH and lacks polyunsaturated fatty acid [20]. We used yeast EGY49 cell homogenate for our in vitro assay of recombinant desaturase activity. Our results showed that purified recombinant FADS12 converted C18:1 D9 to C18:2 D9,12 in vitro, and C18:2 D9,12 level was increased 116% compared to the control ( Table 2). Activities of purified FADS9-I and FADS15 were relatively low in vitro.
We noticed that the size of FADS9-I was smaller on SDS-PAGE than its predicted molecular weight. To determine whether a cleavage had occurred on the N-terminus which has a His-tag and Precision protease (PP) cleave site, the purified recombinant proteins were digested with PP (GE Healthcare) and analyzed by Coomassie staining of SDS-PAGE gel. The reduction in molecular weight supported that all three desaturases had the His-tag and Precision protease cleave site which were removed by PP digestion. Therefore, FADS9-I might have been cleaved on its C-terminus. Molecular weight of the FADS9-I protein was determined by mass spectrometry. Result suggests that the cytochrome b5 domain was cleaved off from FADS9-I during protein solubilization and purification (Figure 4).

Discussion
No structural information is available to explain the regio-and substrate-selectivity of the membrane class of fatty acid desa-turases. Our initial goal was to express high amounts of recombinant desaturase as secreted proteins in the methylotrophic yeast Pichia pastoris. The pre-pro a-factor from Saccharomyces cerevisiae, the most commonly used signal sequence for targeting protein secretion, was used in our recombinant desaturase expression. However, the expressed FADS proteins remained membrane-bound, whereas recombinant EGFP (enhance green fluorescent protein) proteins were efficiently secreted in our expression system. This membrane association of FADS proteins necessitates an effective solubilization procedure for protein purification.
There is no reference that we can find in the literature about detergents used for solubilization of integral membrane desaturases. Detergents, such as Tween-20, Tween-80, NP-40, and DDM have been commonly used for solubilization of membrane proteins, and their efficiency could be explained by their polar head group structure ( Figure S8). Our results showed that none of them had any effect on solubilization of our recombinant FADS. Instead, we found that FADS9-I and FADS12, and FADS15 were efficiently solubilized in detergent Fos-Choline 12 and 16. It is possible that the structural similarity of Fos-Choline 12 and 16 to fatty acids may contribute to their ability to solubilize membrane fatty acid desaturase. The Fos-Choline detergents have also been successfully used in other membrane protein studies [21].
In vivo analysis showed high desaturase activity of FADS9-I. In contrast, the purified FADS9-I had relatively low desaturase activity. Some fatty acid desaturases possess a cytochrome b5 domain ( Figure S1) used for electron transfer, whereas others do not have such domain and rely on external source of cytochrome b5. The molecular weight of purified FADS9-I was smaller than the predicted, and mass spectrometry data indicated a spontaneous removal of the cytochrome b5 domain (Figure 4). Therefore, the loss of internal cytochrome b5 may explain the low FADS9-I activity in vitro. FADS12, on the other hand, uses external source of cytochrome b5, and thus, retains high activity.
Cloning, expression, purification and functional characterization of FADS9-I, FADS12 and FADS15 from M. alpina ATCC#32222 represent a critical step towards the structural elucidation of membrane fatty acid desaturases. Mortierella alpina is an oleaginous fungus which can produce lipids accounting for up to 50% of its dry weight in the form of triacylglycerols. It is used commercially for the production of arachidonic acid. Understanding the regio-and substrate-selectivity of membrane class fatty acid desaturases may also be useful for the genetic engineering of strains producing higher levels and different constituents of dietary fat.  Figure S2 Diagram of the cloning strategy for desaturase expression vectors. FADS coding sequences were PCR amplified using primers listed in Table S1. PCR fragment were digested with indicated restriction enzymes, column purified and inserted into the pET-19b(PP) vector linearized with corresponding restriction enzymes. The FADS coding sequence plus His tag and Precision protease recognition sequence were PCR amplified and inserted into the pPinkalpha-HC vector. TRP2: TRP2 gene, AmpR: ampicillin resistance gene, pUC ori: oriental promoter of pUC, PAOX1:59AOX1 promoter region, a-factor: a-mating factor secretion signal, CYC1 TT: CCY1 transcription termination region, PADE2 HC: high-copy ADE2 promoter region,