First evidence for a multienzyme complex of lipid biosynthesis pathway enzymes in Cunninghamella bainieri

Malic enzyme (ME) plays a vital role in determining the extent of lipid accumulation in oleaginous fungi being the major provider of NADPH for the activity of fatty acid synthase (FAS). We report here the first direct evidence of the existence of a lipogenic multienzyme complex (the lipid metabolon) involving ME, FAS, ATP: citrate lyase (ACL), acetyl-CoA carboxylase (ACC), pyruvate carboxylase (PC) and malate dehydrogenase (MDH) in Cunninghamella bainieri 2A1. Cell-free extracts prepared from cells taken in both growth and lipid accumulation phases were prepared by protoplasting and subjected to Blue Native (BN)-PAGE coupled with liquid chromatography–tandem mass spectrometry (LC-MS/MS). A high molecular mass complex (approx. 3.2 MDa) consisting of the above enzymes was detected during lipid accumulation phase indicating positive evidence of multienzyme complex formation. The complex was not detected in cells during the balanced phase of growth or when lipid accumulation ceased, suggesting that it was transiently formed only during lipogenesis.

of NADPH, ME has been suggested to play a key role as the sole NADPH provider for fatty acid biosynthesis catalysed by FAS in three oleaginous fungi; Mucor circinelloides, Mortierella alpina 9 and Cunninghamella bainieri 2A1 10 . (The biochemistry of no other oleaginous fungus has been studied in similar detail). In each of these fungi, the amount of lipid that is accumulated reaches a finite level. In Mcr circinelloides, this is about 25%, in Mta alpina it is about 50% 9 and in C. bainieri it is about 30% 10 . Cessation of lipid accumulation in all three fungi directly correlates with the loss of ME activity even though other NADPH-generating enzymes continue to be active 9,10 .
In Mcr circinelloides 11 and Rhodotorula glutinis 12 , substantial increases in lipid content occur when the ME gene coding for ME (EC 1.1.1.40) was cloned and over-expressed. Furthermore, the lipid content of Mcr circinelloides decreased significantly from 24 to 2% (w/w) when sesamol (specific inhibitor of ME activity) was added to the cultures 7 . These results have been interpreted as strongly implicating a direct role of ME in lipid biosynthesis in oleaginous fungi.
It has been proposed that the NADPH requirement for FAS reactions in Mcr circinelloides, Mta alpina and C. bainieri 2A1 is specifically supplied by ME and not from the general "pool of NADPH" 4,10 . A possible explanation for this is the occurrence of a multienzyme complex involving ME and FAS that would allow direct channelling of NADPH during lipid biosynthesis only from ME but not from other NADPH generating enzymes 5,10 . Enzymes that are involved in sequential reactions in a pathway are often physically associated with each other, forming temporary or stable larger multienzymes complexes 13,14 . This physical association of enzymes involved in the same pathway allows direct channeling of intermediates through the sequence of reactions in the pathway as products generated by each of the enzymes are immediately transferred to the active site of the subsequent enzymes without equilibrating with the global aqueous system where the complexes are located in cells. This allows the working concentration of intermediates of the enzymes to be efficiently achieved. Thus, physical association of the enzymes could increase the efficiency, specificity and speed of metabolic pathways in cells. To date many multienzyme complexes have been elucidated such as the TCA cycle, triacylglycerol (TAG) biosynthesis and respiratory complexes [15][16][17][18] .
The existence of a lipogenic multienzyme complex consisting of ME, FAS, ACL, acetyl-CoA carboxylase (ACC), pyruvate carboxylase (PC) and malate dehydrogenase (MDH) has been hypothesized 5 (as shown in Fig. 1). The hypothesis involves a "transhydrogenase cycle" consisting of PC, MDH and ME that operates independently in the generation of NADPH and the malate-citrate shuttle that generates acetyl-CoA. Both NADPH and acetyl-CoA are subsequently channelled to FAS 8 .
Several reports regarding physical association of enzymes implicated in lipid biosynthesis in non-microbial systems have been reported. The physical association of ACC, ACL and FAS in the microsomal fractions of rat liver, suggesting the microsome as being the major site for the localisation of fatty acid synthesis multienzyme complex 19 . In addition, ME exists in a membrane-bound form in the oleaginous fungus Mcr circinelloides 20 Figure 1. A diagram to show the organization of a hypothesized lipogenic metabolon. The flux of carbon from the mitochondrion, via citrate efflux and acetyl-CoA formation in the cytosol, then into fatty acids and finally into long chain PUFAs (LCPUFA) occurring in the membranes of the endoplasmic reticulum, is shown by the continuous lines. The system uses pyruvate (from glycolysis) as the provider of intramitochondrial acetyl-CoA and for citric acid production as is shown in Fig. 1  possibly providing NADPH for the fatty acid desaturases to operate. However, to the best of the authors' knowledge, there has been no report regarding the isolation of all the lipogenic enzymes in the form of a complete multienzyme complex from any source.
We report here the isolation and identification of a lipogenic multienzyme complex (lipid metabolon) from an oleaginous fungus, Cunninghamella bainieri 2A1 isolated from Malaysian soil 21,22 , using Blue Native PAGE (BN-PAGE) coupled to liquid chromatography-tandem mass spectrometry (LC-MS/MS). We also provide evidence suggesting the involvement of the formation and dissociation of the multienzyme complex in the regulation of lipid biosynthesis. It was essential that minimal force was used to disrupt the cells in order to preserve the intracellular structures. This was achieved by preparing protoplasts that were then osmotically lysed. Other direct methods of cell disruption were too severe and always led to the loss of metabolon structures (unpublished work). Figure 2 shows the profiles of growth and lipid accumulation of C. bainieri 2A1 where phases of growth, lipid accumulation and cessation of lipid accumulation occurred between 0 to 8 h, 24 to 48 h and 96 to 120 h, respectively. Figure 3a-c represent BN-PAGE of the extracts obtained from the 8 h (balanced growth phase), 24 h (lipid accumulation phase) and 96 h (after cessation of lipid accumulation phase) cultures of C. bainieri 2A1. Proteins in a cell extract obtained by disruption of protoplasts prepared from these phases were separated by BN-PAGE gradient gel (3-12% providing resolutions of 15 kDa to 10 MDa). The approximate molecular weights of the proteins were estimated based on exponential plot of molecular weight of standard proteins against relative migration distance (R f ) of standard proteins.

Detection of lipogenic multienzyme complex using BN-PAGE coupled with LC-MS/MS.
After the electrophoresis, the gel was sliced into 10 groups with an approximate length of 8 mm each, and designated as group 1 to group 10. Group 1 represents the upper part of the BN gel and group 10 represents the lowest part of the gel. The estimated molecular weight of the proteins that could be resolved in each gel groups are as follows; group 1 (4.6 MDa to 10 MDa), group 2 (1.2 MDa to 4.6 MDa), group 3 (950 kDa to 1.2 MDa), group 4 (830 kDa to 950 kDa), group 5 (650 kDa to 830 kDa), group 6 (400 kDa to 650 kDa), group 7 (245 kDa to 400 kDa), group 8 (130 kDa to 245 kDa), group 9 (30 kDa to 130 kDa) and group 10 (0 to 30 kDa). The lipogenic multienzyme complex is postulated to consist of the key lipogenic enzymes: FAS, ME and ACL as well as ACC, PC and MDH with their approximate molecular weights of 2.6 MDa, 160 kDa, 375 kDa, 252 kDa, 330 kDa and 70 kDa, respectively 5 . Thus, the molecular weight of the complex is estimated to be more than 3.7 MDa assuming a complex of single enzymes. This should, therefore, be resolved in group 1 or group 2 of the gel.
The LC-MS/MS analysis was done for each 8 h, 24 h and 96 h samples after running the BN-PAGE. The experiments were repeated three times and the results were shown to be reproducible. The data from one set of the experiment was used to represent the results. Results showed that all the targeted enzymes (ME, FAS, ACL, ACC, PC and MDH) were identified in gel group 2 from the sample of 24 h of cultivation (during lipid accumulation phase) (as shown in Table 1). The isolation of the six proteins in a single gel group indicates that the enzymes are physically associated during lipid biosynthesis phase. This observation could not have resulted from inadequate resolving time as the electrophoresis was run for 24 h. To the best of the author's knowledge, this is the first evidence confirming the existence of the proposed lipogenic multienzyme complex consisting of ME, FAS, ACL, ACC, PC and MDH. In addition, each of the enzymes were also resolved in different gel groups representing each individual molecular weight, representing the free forms of each of the enzymes. Results showed that PC, ACL and ACC were present in gel group 7 while ME and MDH were detected in gel group 8 and 9, respectively (as shown in Table 1). This suggests possible partial dissociation of the complex occurred although gentle disruption method was used. This also indicates some probable equilibrium occurring between the metabolon and the free form enzymes. However, this is uncertain as it is unknown how much unravelling of the metabolon might have occurred accidentally during the isolation procedures. Although metabolon is known to be fragile, until better methods have been developed to stabilize it, it is inconclusive whether the detached proteins are there in equilibrium with the metabolon or have arisen as artefacts. However, when extracts of cultures undergoing balanced growth (8 h) and in the phase after cessation of lipid biosynthesis (96 h) were analysed by LC-MS/MS, all the enzymes were detected in gel groups corresponding to their molecular weight (as shown in Table 2 and Table 3, respectively) and no evidence of possible physical association into any aggregate was observed. For both samples, only FAS (2.6 kDa) was detected in gel group 2 whereas PC, ACL and ACC were in gel group 7 while ME and MDH were in gel group 8 and 9, respectively. This suggests that the lipogenic enzymes were not physically associated during growth (8 h) as well as after the conclusion of the lipid accumulation phase (96 h) and that the formation of the lipogenic multienzyme complex occurred transiently, only during lipid accumulation phase.
With regards to other NADPH-generating enzymes (NADP + : ICDH, 6-PGDH and G-6-PDH), none of these were found to be associated to the complex detected in gel group 2 during lipid accumulation phase (24 h) where NADP + : ICDH and G-6-PDH were detected in gel group 8 and 6-PGDH was detected in gel group 9 (as shown in Table 1). This provides further evidence regarding the exclusive role of ME in the provision of NADPH to FAS. Furthermore, none of the enzymes that served as negative controls [phosphofructokinase (PFK) and pyruvate kinase (PK)] were detected in gel group 2 (as shown in Table 1). The complex that we observed is therefore not a spurious collection of proteins, some involved in lipid biosynthesis and some not, but is a dedicated group of enzymes all with established roles in fatty acid biosynthesis.
The presence of ME in the gel groups 2 and 8 was also confirmed by in-gel activity staining. Results showed that for the 24 h extract, two bands representing ME activity appeared in gel group 2 and 8 (as shown in Fig. 4a and b) whereas for 8 h and 96 h extract, only a single band appeared in gel group 8 thus confirming the tandem mass spectrometry data.

Discussion
Regulation of lipid accumulation in oleaginous yeasts and fungi has been well established where the extent of lipid accumulation has been reported to be determined by ME [3][4][5] . The lipid content accumulated by oleaginous fungi correlates with the presence and activity of ME. In Mcr circinelloides the specific activity of ME diminished shortly after N-limitation was achieved (18 h cultivation) coinciding with cessation of lipid accumulation with a lipid content of 30% (w/w) although in the presence of other lipogenic and NADPH-generating enzymes. Whereas ME activity of Mta alpina, which accumulated up to 45% (w/w) lipid lasted even more than 88 h of cultivation period 9 . A similar observation was also reported in C. bainieri 2A1 10 . This led to the suggestion that ME plays a specific and crucial role in generating NADPH for FAS catalysis that could be explained only by the existence  of physical association between ME and FAS. This would allow direct channelling of NADPH from ME to FAS. Hence a hypothesis was proposed by Ratledge 5 for the existence of a multienzyme complex involving ME and FAS as well as other essential lipogenic enzymes. Indeed, physical association of several of the enzymes has been reported such as the evidence of physical association between ACL, ACC and FAS in which these enzymes have been successfully isolated as a high molecular weight fraction from rat liver (ME was not included in the work) 19 .
The proposed hypothesis involves all lipogenic enzymes catalysing sequential reactions i.e. ACC, ACL and FAS plus ME and MDH and PC which operates as a "transhydrogenase cycle" for NADPH generation. However, up to now, no evidence regarding the existence of the hypothesized complex has been reported. We report the first evidence validating the hypothesis using an oleaginous fungus, C. bainieri, as a model. In this work, a lipogenic multienzyme complex consisting of ME, FAS and ACL as well as PC, ACC and MDH in C. bainieri has been isolated using protoplasts for the initial step followed by their gentle breaking. As we have tried but failed on numerous occasions to isolate this complex starting with whole cells coupled with a variety of cell disruption systems (unpublished work), we argued that harsh cell disintegration was also leading to disruption of any liposome complex. Hence, by using protoplasts, we were able to preserve the internal structure of the cell and this was then kept intact during their gentle lysis. As shown in Table 1, during lipid accumulation (24 h) the enzymes were present in gel group 2 despite significant differences in molecular weight of each individual enzymes indicating positive physical association. The absence of two enzymes (PFK and PK) that serve as negative controls in group 2 indicates the unlikely probability that the observed association could have resulted from non-specific interactions. The formation of this complex is shown to be transient, as the complex was only detected during lipid accumulation phase (24 h) and was absent during growth (8 h) and after cessation of lipid accumulation (96 h). A similar observation has been reported in Bacillus subtilis where a multienzyme complex involving malate dehydrogenase and phosphoenolpyruvate carboxykinase forms only during growth on gluconeogenic carbon sources 17 . In addition, the transient nature of multienzyme complexes was also demonstrated in relation to the association of glycolytic enzymes on the surface of mitochondria in Arabidopsis thaliana, where the association was dependent on the rate of respiration. These findings show that the formation of multienzyme complexes can be attributed to the change of metabolic flux where the complex formation would support direct channeling of metabolites for efficient operation of the pathways 23 .
Despite the detection of the lipogenic enzymes as clusters in gel group 2, each of the enzymes was also detected in separate bands corresponding to their molecular weights. This could be due to partial dissociation of the complex during the final disruption of the protoplasts and subsequent handling and possibly during electrophoresis itself. Most multienzyme complexes are fragile and held together by weak interactions that will be strengthened by attachment to appropriate membranes within the cell 24,25 . In this case, we suggest that attachment of the complex to membranes of the endoplasmic reticulum and mitochrondrion (as shown in Fig. 1) might be the key stabilizing factor. Attachment to the membrane of a (growing) lipid droplet is also a possibility. The stability and integrity of multienzyme complexes will also be affected by other factors such as pH, temperature and metal ions concentration 14 and we make no claims that we have optimized the conditions for the isolation of this liposome complex.
No evidence of involvement of other NADPH-generating enzymes, besides ME, in the lipogenic multienzyme complex was observed. G-6-PDH, 6-PGDH and NADP + : ICDH all were not found to be resolved in gel group 2 (as shown in Table 1). This supports the previous suggestion where ME functions exclusively as the NADPH provider for FAS.

Conclusions
This work provides evidence of the existence of a lipogenic multienzyme complex consisting of ME, FAS, ACL, ACC, PC and MDH. The multienzyme complex exists transiently and was only detected during lipid accumulation phase whilst dissociating during growth and cessation of lipid accumulation. No evidence of probable involvement of G-6-PDH, 6-PGDH and NADP + : ICDH was observed indicating the vital role of ME as NADPH provider in lipid biosynthesis in C. bainieri 2A1.   sterilized at 121 °C for 40 min. Glucose (to give 30 g/L) was added separately after sterilization. Seed culture was prepared by transferring a spore suspension into 500 mL shake-flasks containing 200 mL nitrogen-limited medium to give 10 5 spores/mL. The cultures were grown at 30 °C and shaken at 200 rpm for 48 h. Ten percent (v/v) of the culture was then used for subsequent inoculations. All experiments were carried out using 500 mL conical flasks containing 200 mL nitrogen-limited medium as described above. Cultivation was carried out at 30 °C, with agitation at 200 rpm and cultures were harvested at 8 h (growth phase), 24 h (lipid accumulation phase) and 96 h (cessation of lipid accumulation).    (20 kDa) was loaded into the well of the same gradient gel. All steps were performed at 4 °C. The molecular weight of the proteins that resolved in the gradient gel were estimated based on exponential plot of molecular weight of standard proteins against relative migration distance (R f ) of standard proteins.

Preparation of protoplasts and cell extract.
Electrophoresis. Electrophoresis was performed at 4 °C, 100 V for 3 h and subsequently run overnight at 150 V.
Slicing of the BN-PAGE gel. After the electrophoresis, the gel was sliced into 10 groups with an approximate length of 8 mm each, and designated as group 1 (upper part of the gel) to group 10 (lowest part of the gel).
In-gel trypsin digestion. In-gel trypsin digestion 29 was performed using trypsin profile IGD kit for in-gel digests (Sigma-Aldrich). Each of the gel bands (1 to 10) was further sliced separately into smaller pieces and transferred into silicone tube. Two hundred µL of destaining solution (400 mM NH 4 HCO 3 in 40% acetonitrile) was added into the tubes and incubated at 37 °C for 30 min. This step was repeated twice. Reduction of disulphite bonds of the proteins was done by the addition of 200 µL 10 mM dithiothreitol (DTT) with incubation at room temperature for 30 min. The DTT solution was then removed. Two hundred µL 55 mM iodoacetamide was then added into the silicone tubes for alkylation of the proteins and incubated in the dark at room temperature. The solution was then removed and replaced with 400 µL of washing solution (50 mM NH 4 HCO 3 ) and incubated at room temperature for 15 min. This step was repeated twice. Then, 400 µl acetonitrile (100%) was added into the silicone tubes to dehydrate the gel pieces and incubated for 10 min at room temperature. The supernatant was removed and the gel pieces were dried by incubation at 50 °C for 5 min.
Next, 20 µL (0.4 µg trypsin) trypsin (1 mM HCl and 40 mM NH 4 HCO 3 in 9% of acetonitrile) was added into the silicone tube and incubated at room temperature for 5 min. Then 50 µL of 40 mM NH 4 HCO 3 in 9% acetonitrile was added in the tubes and incubated overnight at 37 °C. The digested peptides were then transferred into new silicone tubes and dried again in vacuum centrifuge. The dried peptides were stored at −20 °C before tandem mass spectrometry analysis was performed.
Identification of protein using high performance liquid chromatography-tandem mass spectrometry (LC-MS/MS). Ten uL 0.1% formic acid was added into the dried peptides tubes and centrifuged at 13,000 x g for 5 min. Five µL of the supernatant was transferred into new silicone tubes. Samples were analysed by Agilent 6520 Accurate Mass Q-TOF LC/MS-Nano ESI (ChipCube) (Agilent Technologies). The parameters of analysis were as follows: two columns were used to separate the peptides; 1. large capacity chip, C18, 300 Å, 160 nL enrichment column, 2. analytical column, 75 µm × 150 mm (Agilent Part no: G4240-62010), flow rate: 4 uL/min from Agilent 1200 Series capillary pump and 0.3 uL/min from Agilent 1200 Series nano pump, solvents: 0.1% formic acid in water (A); 90% acetonitrile in water with 0.1% formic acid (B), injection volume: 2 uL. Mass spectrometry was operated in positive ion mode with acquired 4 MS spectra s −1 from 300 until 3000 m/z. Auto MS/MS mode applied from 50 until 3000 m/z with total maximum precursor of 4 in a cycle and exception of 2 spectra for 1 minute. Identification of peptides was performed by Spectrum Mill MS Proteomics Workbench (Rev B.04.00.127; Agilent Technologies). Cysteine carbamidomethylation was set as fix and variable modification. Search was done online for Mucorales at UniProtKB/Swiss-Prot database. Standard criteria for actual proteins selection were as follows; peptides filtered with score ≥ 8, score peak intensity percentage (SPI%) ≥ 60 and protein filtered with score ≥ 10, mass MH + Error ≤ 10 ppm, local false discovery rate ≤ 0.1% and database Fwd-Rev score ≥ 2.
In-gel activity staining of ME. Activity staining for ME in gradient gel was performed by immersing the gel in a phosphate buffer (pH 7.4) containing 0.47 mM NADP + , 17.2 mM L-malate, 0.1 M MgSO 4 , 0.55 mg/mL nitroblue tetrazolium and 0.097 mg/mL phenazine methosulphate 30 . After 3 h of incubation, the reaction was stopped by replacing the staining solution with 5% (v/v) acetic acid.
Data availability. All data generated or analysed during this study are included in this published article.