A Unique Mono- and Diacylglycerol Lipase from Penicillium cyclopium: Heterologous Expression, Biochemical Characterization and Molecular Basis for Its Substrate Selectivity

A cDNA gene encoding a mature peptide of the mono- and diacylglycerol lipase (abbreviated to PcMdl) from Penicillium cyclopium PG37 was cloned and expressed in Pichia pastoris GS115. The recombinant PcMdl (rePcMdl) with an apparent molecular weight of 39 kDa showed the highest activity (40.5 U/mL of culture supernatant) on 1,2-dibutyrin substrate at temperature 35°C and pH 7.5. The rePcMdl was stable at a pH range of 6.5–9.5 and temperatures below 35°C. The activity of rePcMdl was inhibited by Hg2+ and Fe3+, but not significantly affected by EDTA or the other metal ions such as Na+, K+, Li+, Mg2+, Zn2+, Ca2+, Mn2+, Cu2+, and Fe2+. PcMdl was identified to be strictly specific to mono- and diacylglycerol, but not triacylglycerol. Stereographic view of PcMdl docked with substrate (tri- or diacylglycerol) analogue indicated that the residue Phe256 plays an important role in conferring the substrate selectivity. Phe256 projects its side chain towards the substrate binding groove and makes the sn-1 moiety difficult to insert in. Furthermore, sn-1 moiety prevents the phosphorus atom (substitution of carboxyl carbon) from getting to the Oγ of Ser145, which results in the failure of triacylglycerol hydrolysis. These results should provide a basis for molecular engineering of PcMdl and expand its applications in industries.

The correlation between structure and function of Mdl has been studied widely. By site-directed mutagenesis, it was confirmed that the Mdl from P. camembertii has a similar catalytic center, Ser-Asp-His, to that of a triacylglycerol lipase family [3]. Analysis of the binding of diacylglycerol analogue to the Mdl from Malassezia globosa suggested that the substrate specificity of the enzyme mainly resulted from the shape and size of a narrow tunnel, in which there was no space for the settlement of the third chain of triacylglycerol [4,5]. The mdl from P. camembertii U-150 was heterologously expressed in Saccharomyces cerevisiae and Aspergillus oryzae [3,6]. However, there are few reports concerning the heterologous expression of mdl in Pichia pastoris and the molecular basis for substrate selectivity of Mdl from genus Penicillium.
In our previous studies, the genomic DNA encoding a monoand diacylglycerol lipase (PcMdl) was cloned (GenBank: AF288219) from Penicillium cyclopium PG37, which can produce PcMdl together with an alkaline lipase (Lip I). The lip I gene was efficiently expressed in P. pastoris GS115 by induction under optimized conditions [7]. In this work, the cDNA encoding PcMdl was cloned and functionally expressed in P. pastoris. The character of the recombinant PcMdl (rePcMdl), as well as the molecular basis for substrate selectivity of PcMdl was investigated.

Ethics Statement
No specific permissions were required for these locations/ activities. The field studies did not involve endangered or

Analysis of the Primary and 3-D Structures of PcMdl
The bioinformatics analysis of the signal peptide of PcMdl was carried out using the SignalP 3.0 program (http://www.cbs.dtu. dk/services/SignalP/). The physicochemical properties of PcMdl were identified using the Protparam program (http://au.expasy. org/tools/protparam.html). The homology search for lipase primary structures was performed using the BLAST server at NCBI website (http://www.ncbi.nlm.nih.gov/). The multiple homology alignment of the amino acid sequences among lipases was accomplished using the ClustalW2 program (http://www.ebi. ac.uk/Tools/msa/ClustalW2) and DNAMAN 6.0 software (http://dnaman.software.informer.com). The putative N-linked glycosylation site of PcMdl was located using the NetNGlyc program 1.0 (http://www.cbs.dtu.dk/services/NetNGlyc/). The 3-D structure of PcMdl was homologically modeled using the MODELLER 9.9 program (http://salilab.org/modeller/) based on the known crystal structure of a lipase from P. camembertii (PDB: 1TIA). The modeled 3-D structure of PcMdl was visualized using the PyMOL software (http://pymol.org).

Expression of Pcmdl in P. pastoris
The gene Pcmdl was excised from pUCm-T-Pcmdl and inserted into the pPIC9K vector at the EcoR I and Not I sites. The resulting pPIC9K-Pcmdl was linearized with Sal I, and transformed into P. pastoris GS115 by electroporation using a Gene Pulser apparatus (Bio-Rad, Hercules, CA, USA) according to the manufacturer's instruction. The P. pastoris transformants were screened by an MD plate and then a set of increasing concentrations of G418-containing (up to 4.0 mg/mL) YPD plates in turn. A negative control strain (P. pastoris GSC) was constructed by introducing only the vector pPIC9K by electroporation into P. pastoris GS115. The heterologous expression of Pcmdl in P. pastoris GS115 was performed followed the instruction of Multi-Copy Pichia Expression Kit (Invitrogen, USA) with slight modifications. Each single colony of the P. pastoris transformants was inoculated into a 30 mL of BMGY medium in a 250 mL flask, and cultured at 30uC on a rotary incubator (220 rpm) until the OD 600 reached 2-4. Then, the yeast cells were harvested by centrifugation, resuspended in a 30 mL of BMMY medium, and induced for rePcMdl expression by adding methanol to a final concentration of 1.0% (v/v) at 24 h intervals at 30uC for 72 h.

Enzyme Activity and Protein Assays
A quantitative assay for rePcMdl activity was performed according to the method described previously with modifications [8]. In brief, 1 vol of 1,2-dibutyrin (Sangon, China) and 7 vol of 1.0% (w/v) polyvinyl alcohol in 50 mM Na 2 HPO 4 -NaH 2 PO 4 buffer (pH 7.5) were mixed with a blender (10,000 rpm) till homogeneity was achieved. Then, 9 mL of this emulsified 1,2dibutyrin was incubated with 1 mL of suitably diluted enzyme at 30uC for 10 min, and terminated by the addition of 15 mL ethanol. The amount of released free butyric acid was measured by titration using 50 mM NaOH. One unit (U) of lipase activity was defined as the amount of enzyme liberating 1 mmol of fatty acid per minute under the standard assay conditions (at pH 7.5 and 30uC for 10 min). Corresponding to U/mg, it was defined as the active amount (U) per milligram of purified lipase protein. To confirm the substrate selectivity of PcMdl, 1-monobutyrin (Sigma, St. Louis, MO, USA) and tributyrin (Sangon, China) were pretreated and used as substrate like 1,2-dibutyrin in quantitative assay for rePcMdl activity. All data of enzyme activities or relative ones were expressed as the mean 6 standard deviation (SD) from three independent experiments or parallel measurements.
Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) was carried out as reported previously on a 12.5% gel [11]. Coomassie Brilliant Blue R-250 (Sigma, USA) and Quantity One software were used for visualizing and estimating the apparent molecular weights of the separated proteins, respectively. The BCA-200 Protein Assay Kit (Pierce, Rockford, IL, USA) was applied for protein concentration measuring. The expressed rePcMdl was purified by a simple combination of ammonium sulfate precipitation, ultrafiltration and Sephadex G-75 gel filtration [12].

Determination of Temperature Optimum and Stability of rePcMdl
The temperature optimum of rePcMdl was measured under the standard lipase activity assay conditions, except reaction temperatures of 20-50uC. To evaluate its temperature stability, aliquots of rePcMdl were incubated in the absence of substrate at different temperatures (20-45uC), respectively, for 1.0 h. To avoid effect of temperature on the pH value, the 50 mM Na 2 HPO 4 -NaH 2 PO 4 buffer had been accurately adjusted to pH 7.50 at different temperatures (20-50uC). The residual activity of rePcMdl was measured under the standard assay conditions. In this work, the thermostability was defined as the temperature, at or below which the residual activity of rePcMdl was over 85% of its initial activity.

Determination of pH Optimum and Stability of rePcMdl
The pH optimum of rePcMdl was determined using the emulsified 1,2-dibutyrin prepared with 50 mM Na 2 HPO 4 -NaH 2 PO 4 buffer (pH 6.5-8.0) and glycine-NaOH buffer (pH 8.5-9.0), at the optimum temperature for 10 min. To estimate its pH stability, aliquots of rePcMdl were incubated at 30uC and at different pH values (50 mM Na 2 HPO 4 -NaH 2 PO 4 buffer, pH 6.0-8.0; 50 mM glycine-NaOH buffer, pH 8.5-10.0) for 1.0 h. The pH range over which the retained activity of rePcMdl was more than 85% of its initial activity was defined as the range of its pH stability.

Lipid Hydrolysis by rePcMdl and Triacylglycerol Lipase
Two triacylglycerol lipases, a neutral Candida antarctica lipase B (CALB) and an alkaline P. cyclopium lipase (Lip I ) which were purchased from Sigma (USA) and produced by the method reported previously [7] in turn, were used for tests of olive oil hydrolysis with rePcMdl. The hydrolysis by CALB and rePcMdl was carried out at pH 7.5 and 30uC, while by Lip I and rePcMdl was at pH 8.5 and 25uC. An emulsified olive oil was used as the substrate in the hydrolysis experiments. Briefly, 1 vol of olive oil and 3 vol of 3% (w/v) polyvinyl alcohol were mixed well with a blender (10,000 rpm) to form the emulsified substrate. Aliquots of 20 mL 50 mM Na 2 HPO 4 -NaH 2 PO 4 (pH 7.5) or glycine-NaOH (pH 8.5) buffer were mixed well with 8 mL emulsified substrate and 1 mL diluted rePcMdl, triacylglycerol lipase or two lipases mixture with appropriate activity units, respectively. The mixture of substrate and lipase(s) was incubated at 25 or 30uC for 2 h on a magnetism mixer with 200 rpm agitation, meanwhile titration was performed with 50 mM NaOH in a pH-stat to keep pH 7.5 or 8.5. The heat inactivated lipase instead of the active one was used as the control. The total volume of the used NaOH measured at 5 min intervals for first hour and at 10 min intervals for the next hour had a positive correlation with the amount of free fatty acids released from olive oil.

3-D Structure of PcMdl in an Open Conformation
The primary sequence of PcMdl shares a 98.6% identity with the MdlA from P. camembertii (PDB: 1TIA), which is in a closed conformation. In order to provide insight into the responsibility for substrate selectivity of PcMdl, its open conformation was homologically modeled using the crystal structure of T. lanuginosa lipase (PDB: 1GT6) as the template on the MODELLER 9.9 program. The modeled 3-D structure of PcMdl was visualized using the PyMOL software.

Molecular Dynamics Simulations
In order to fully relax the steric clashes in PcMdl-analogue docked complexes, molecular dynamics (MD) simulations were carried out by the Discovery Studio ver. 3.5 (http://accelrys.com/). The CHARMm force field was used for MD simulations. All structures were inserted into an octahedral box with explicit solvent and simulated with periodic boundary conditions. Then the systems were neutralized by insertion of counter ions. First energy minimization was performed by 2,000 steps of steepest descent algorithm with a RMS gradient of 0.1 kcal/mol. Second energy minimization was carried out by 20,000 steps of conjugate gradient algorithm with a RMS gradient of 0.0001 kcal/mol. The following simulations were overall conducted using Particle Mesh Ewald electrostatics. Heating and cooling simulations were run at an initial and a target temperature, 50 and 300 K, respectively, for 4,000 steps (0.001 ps per step). Equilibration simulations were undergone at a target temperature 300 K for 4,000 steps (0.001 ps per step). Production simulations were worked at a target temperature 300 K for 30,000 steps (0.001 ps per step). The optimized 3-D structure of PcMdl-substrate analogue was visualized using the PyMOL software.

Cloning of the cDNA Gene Pcmdl
Using the first-strand cDNA transcribed reversely from the P. cyclopium PG37 total RNA as the template, an about 1.1-kb clear band and several faint bands were amplified by the first-round PCR. On the basis of the principle of nested PCR technique, each band was gel-purified, and subjected to the second-round PCR. As a result, an about 850-bp specific band was amplified only using a 1.1-kb band as the template. The nucleotide sequence of the cDNA gene Pcmdl encoding PcMdl has been deposited in the GenBank database under accession no. HM135194.

Analysis of the Primary and 3-D Structures of PcMdl
The cloned cDNA gene Pcmdl is exactly 840 bp in length, encoding a mature peptide (PcMdl) of 279 amino acids (Fig. 1, line  1). The theoretical molecular weight and isoelectric point (pI) of PcMdl were calculated to be 30,181 Da and 4.81, respectively. The homology alignment between two amino acid sequences using the DNAMAN 6.0 software displayed that the identities of PcMdl from P. cyclopium (in this work) with other four mono-and diacylglycerol lipases from P. camembertii (GenBank: AAB26004), Penicillium chrysogenum (GenBank: XP_002567712), A. oryzae (GenBank: XP_001823459), and Aspergillus kawachii (GenBank: GAA92588), were 98.6%, 89.0%, 63.7%, and 58.3%, respectively. The temperature optima and thermostability of rePcMdl. The temperature optimum (100% relative activity) of rePcMdl, measured at pH 7.5, was 35uC. Incubated at different temperatures for 1.0 h, the residual activity of rePcMdl retained over 85% of its original activity at 35uC or below, but declined rapidly over 35uC. (B) The optimal pH and pH stability of rePcMdl. The purified rePcMdl exhibited higher relative activities over a pH range of 7.0-8.0 (measured at 35uC). Its pH optimum (100% relative activity) was at pH 7.5. Incubated at different pH values (6.0-10.0) for 1.0 h, rePcMdl was highly stable (more than 85%) over a broad pH range of 6.5-9.5. doi:10.1371/journal.pone.0102040.g004 Figure 5. Effects of various metal ions and EDTA on rePcMdl activity. The rePcMdl activity was not significantly affected by metal ions that were tested or EDTA, but was inhibited by Hg 2+ and Fe 3+ . doi:10.1371/journal.pone.0102040.g005 Moreover, the multiple homology alignment among the five lipase sequences showed that a pentapeptide G-H-S-L-G (located at a site from Gly 143 to Gly 147 ) was strictly conserved (Fig. 1), corresponding with the classical motif G-X-S-X-G (X: arbitrary amino acid residue) of almost all lipases or esterases reported thus far [13,14]. While, the most conserved three amino acid residues, Ser 145 , Asp 199 , and His 259 (numbered by PcMdl), constitute a typical catalytic triad of lipases. All these primary structure features implied that PcMdl is a member of the lipase family.
Based on the known crystal structure of the P. camembertii (PDB: 1TIA), which shares a 98.6% identity with PcMdl in the primary structure, the 3-D structure of PcMdl was predicted by homology modeling in the closed conformation as shown in Fig. 2. The modeled 3-D structure displays a globular shape, which is composed of one eight-stranded b-sheet, three long and several short a-helices, and the catalytic triad Ser 145 -Asp 199 -His 259 . Although there are wide differences in their primary structures and biochemical properties, PcMdl and the majority of lipases share a similar core topology of the 3-D structure, well known as the a/b hydrolase fold [15].

Expression of Pcmdl in P. pastoris
The protein expression level was not directly proportional to the concentration of geneticin G418 [11]. For this reason, 45 transformants separately resistant to 1.0, 2.0, and 4.0 mg/mL of G418, numbered as P. pastoris GSM1-1 to GSM1-15, GSM2-1 to GSM2-15, and GSM4-1 to GSM4-15, were picked out for flask expression tests. P. pastoris GS115 transformed with pPIC9K vector instead of pPIC9K-Pcmdl, numbered as P. pastoris GSC, was used as the negative control. After P. pastoris transformants were induced by 1.0% methanol for 72 h, activities and proteins of the expressed rePcMdl in the cultured supernatants were assayed. Among all transformants tested, one transformant expressing the highest rePcMdl activity of 40.5 U/mL, P. pastoris GSM4-2, was selected and used for subsequent studies. As a negative control

Protein Assays of the Expressed rePcMdl
It was reported that the purities of the recombinant Aspergillus sulphureus b-mannanase and Aspergillus usamii xylanase AuX-yn11D expressed in P. pastoris X-33 and GS115 were 97% and 90%, respectively [16,17]. In this work, by protein band-scanning, more than 86% of the total supernatant protein was the secreted rePcMdl from P. pastoris GSM4-2 (Fig. 3, lane 1). Therefore, the expressed rePcMdl was purified to homogeneity by a simple combination of ammonium sulfate precipitation, ultrafiltration and Sephadex G-75 gel filtration (Fig. 3, lane 3).
SDS-PAGE analysis of the purified rePcMdl displayed a distinctive protein band (Fig. 3, lane 3). The apparent molecular weight (approximate 39 kDa) of rePcMdl was similar to that of the mono-and diacylglycerol lipase purified from P. camembertii [2], but much larger than its theoretical weight (30,181 Da). P. pastoris enables expressed proteins to undergo some post-translational modifications, such as the exclusion of signal peptide [18], assembly of disulfide bridges and glycosylation of the mature peptide. The bioinformatics analysis of PcMdl amino acid sequence using the NetNGlyc program displayed that there is one putative N-linked glycosylation site (N 225 -A 226 -T 227 ). Simultaneously, the carbohydrate content of the purified rePcMdl was assayed to be 12.3% using the phenol-sulfuric acid method [19]. These analytical results implied that the increased molecular weight of rePcMdl was caused by N-glycosylation [12].

Partial Enzymatic Properties of rePcMdl
The specific activity of the purified rePcMdl, towards 1,2dibutyrin under the standard assay conditions, was 344.9 U/mg. The temperature optimum (100% relative activity) of rePcMdl at pH 7.5 was 35uC (Fig. 4A). It was similar to that of the mono-and diacylglycerol lipase (MDL) of P. camembertii [6], but 10uC higher than that of the triacylglycerol lipase (Lip I ) from P. cyclopium PG37 [7]. Incubated at different temperatures for 1.0 h, the residual activity of rePcMdl retained over 85% of its original activity at 35uC or below, but declined rapidly over 35uC (Fig. 4A). The purified rePcMdl exhibited higher relative activities over a pH range of 7.0-8.0 (measured at 35uC). Its pH optimum (100% relative activity) was at pH 7.5 (Fig. 4B), which was similar to that of the MDL of P. camembertii and much lower than that (pH 10.5) of Lip I. Incubated at different pH values (6.0-10.0) for 1.0 h, rePcMdl was highly stable (more than 85%) over a broad pH range of 6.5-9.5 (Fig. 4B). Incubated with an array of metal ions and EDTA, respectively, the activity of rePcMdl was not significantly affected by most metal ions tested and EDTA, but inhibited by Hg 2+ and Fe 3+ (Fig. 5). The rePcMdl was confirmed to be strictly specific to 1-monobutyrin (189.3 U/mg) and 1,2dibutyrin (344.9 U/mg), but not tributyrin (0 U/mg).

Lipid Hydrolysis by rePcMdl and Triacylglycerol Lipase
The rePcMdl could not hydrolyze olive oil without help from CALB or Lip I (Fig. 6). CALB or Lip I together with rePcMdl had bigger effect on olive oil hydrolysis than CALB or Lip I alone. By lipase catalyzing, olive oil was hydrolyzed and mainly releasing oleic acids [20]. The shape of the hydrolysis curves mainly depended on the characteristic of the triacylglycerol lipase, CALB or Lip I .

Molecular Basis for Substrate Selectivity of PcMdl
The active PcMdl in the open conformation was modeled with the lid in a-helix form, uncovered the catalytic center and led to the catalytic triad Ser 145 -Asp 199 -His 259 exposing to the solvent (Fig. 7). The difference between the open (modeled structure) and closed (crystal structure) conformations of PcMdls mainly existed in lid orient. In the open conformation, the residue Ser 83 together with the Ser 145 (Fig. 7), constituted the oxyanion hole which stabilizes the tetrahedral intermediates. The catalytic pocket, organized by residues Tyr 21 , Phe 112 , Leu 146 , Pro 174 , Val 201 , Phe 256 , Val 266 , and Asp 267 , exposed to the solvent with the lid opening (Fig. 8). All of the above made the modeled open conformation be reliable [4,[21][22][23].
After MD simulations, PcMdl-analogue docked complexes fully relaxed the steric clashes. In the optimized PcMdl-diacylglycerol analogue complex (Fig. 9A and C), the distance between phosphorus atom (substitution of carboxyl carbon) and O c of Ser 145 is 3.1 Å , while in the optimized PcMdl-triacylglycerol analogue complex (Fig. 9B and D) it is 5.9 Å . A smooth reaction of lipid hydrolysis requires that a distance between O c of serine in catalytic triad and carboxyl carbon of substrate is no more than 4.0 Å [24,25]. The Phe 256 projects its side chain towards the substrate binding groove and makes the sn-1 moiety difficult to insert in. Furthermore, sn-1 moiety prevents the phosphorus atom from getting to the O c of Ser 145 . The substrate selectivity mechanism of PcMdl differs from this of Mdl from M. globosa despite the steric hindrance effect [4]. In the M. globosa Mdl, the Phe 278 and residue at the opposite side, Leu 103 , forms a bridge-like obstacle and restricts two fatty acid chains fitted in tunnel or groove. In PcMdl, only one residue, Phe 256 , plays an important role in substrate selectivity.

Conclusions
In this work, Pcmdl which encodes the mono-and diacylglycerol lipase from P. cyclopium PG37 was cloned and expressed in P. pastoris GS115. The recombinant PcMdl was characterized and confirmed to be strictly specific to mono-and diacylglycerol, but not triacylglycerol. Stereographic view of PcMdl docked with substrate analogue indicated that the residue Phe 256 plays an important role in conferring the substrate selectivity. These results should provide a basis for molecular engineering of PcMdl and expand its applications in industries.