Cloning, expression, purification and characterization of patatin, a novel phospholipase A


  • Definition: phospholipids are named according to the IUBMB recommmendations as follows: PCho, phosphocholine; diC6PCho, 1,2-dihexanoyl-sn-glycero-3-phosphocholine; diC7PCho, 1,2-diheptanoyl-sn-glycero-3-phosphocholine; diC8PCho, 1,2-dioctanoyl-sn-glycero-3-phosphocholine; diC9PCho, 1,2-dinonanoyl-sn-glycero-3-phosphocholine; diC10PCho, 1,2-didecanoyl-sn-glycero-3-phosphocholine; diC12PCho, 1,2-didodecanoyl-sn-glycero-3-phosphocholine; diC16PCho, 1,2-dihexadecanoyl-sn-glycero-3-phosphocholine; C18lysoPCho, 1-octadecanoyl-2-lyso-sn-glycero-3-phosphocholine; C8thioglycolPCho, 2-octanoylthio-ethane-1-phosphocholine; C12thioglycolPCho, 2-dodecanoylthio-ethane-1-phosphocholine; C16thioglycolPCho, 2-hexadecanoylthio-ethane-1-phosphocholine; diC8dithioPCho, rac-1,2-dioctanoyldithiopropyl-3-phosphocholine

M. R. Egmond, Department of Enzymology and Protein Engineering, Padualaan 8, De Uithof, 3584 CH Utrecht, the Netherlands


Patatin is the major protein constituent of potato tubers and displays broad esterase activity. The native enzyme actually belongs to a highly homologous multigene family of vacuolar glycoproteins. From these, the patB2 patatin gene was selected and cloned into pUC19 without its signal sequence but with an N-terminal histidine-tag. This patatin was overexpressed under the control of the lac promotor in Escherichia coli strain DH5α. The protein was recovered as inclusion bodies, folded into its native state by solubilization in urea and purified to homogeneity. Starting with one gram of inclusion bodies, 19 mg of pure and active recombinant patatin was isolated, with even higher specific activity than the glycosylated wild-type patatin purified from potato tubers. The purified enzyme showed esterolytic activity with p-nitrophenylesters dissolved in Triton X-100 micelles. The activity of patatin on p-nitrophenylesters with different carbon chain lengths showed an optimum for p-nitrophenylesters with 10 carbon atoms. Besides general esterolytic activity, the pure enzyme was found to display high phospholipase A activity in particular with the substrates 1,2-dioctanoyl-sn-glycero-3-phosphocholine (diC8PCho) (127 U·mg−1) and 1,2-dinonanoyl-sn-glycero-3-phosphocholine (diC9PCho) (109 U·mg−1).

Recently, the structure of human cytosolic PLA2 (cPLA2) was solved, showing a novel Ser-Asp active site dyad [1]. Based on a partial sequence alignment of patatin with human cPLA2, we propose that patatin contains a similar active site dyad. To verify this assumption, conserved Ser, Asp and His residues in the family of patatins have been modified in patatin B2. Identification of active site residues was based on the observation of correctly folded but inactive variants. This led to the assignment of Ser54 and Asp192 as the active site serine and aspartate residues in patatin B2, respectively.


5,5’-dithio-bis(2-nitrobenzoic acid)


cytosolic calcium-dependent phospholipase A2


Patatin subclass B2.

Patatin is a member of a multigene family of vacuolar glycoproteins with a molecular mass of ≈ 40 kDa. It represents 40% of total soluble potato tuber protein and is considered to be a storage protein. Unlike other storage proteins patatin displays lipid acyl hydrolase and acyl transferase activities [2–4]. The enzyme is N-glycosylated at Asn60 and Asn90 [5]. This glycosylation has no influence on proteolytic stability or on enzymatic properties [6].

Patatin exhibits extensive charge heterogeneity [2,7] that may be due, in part, to its glycosylation pattern and post-translational modifications [8]. Several cDNA and genomic clones of patatin have already been isolated and a number of nucleotide sequences have been determined [9–11]. Most patatin isoforms have nearly identical amino-acid sequences and are immunologically identical within a given cultivar as well as among cultivars [8].

Two multigene families with different expression patterns have been identified: class I patatin genes are mainly expressed in tubers whereas class II genes are found both in tubers and roots though at a much lower level as compared to class I genes [12].

In 1970, Galliard reported the partial purification of a lipolytic acyl hydrolase that was active on phospholipids, mono-acylglycerols and p-nitrophenyl esters, moderately active on galactolipids and inactive towards di- and tri-acylglycerols [13,14]. This lipid acyl hydrolase was later identified as patatin [3]. The role of patatin in potato tubers remains unclear. It has been speculated that, besides being the main storage protein of potato tubers, patatin might also be involved in the resistance reaction induced by attack by a pathogen. The lipid acyl hydrolase activity of patatin could be important for the rapid degradation of cell membranes and thus rapid degradation of certain metabolites. Recently, patatin was found to be identical to a cytosolic phospholipase A2 (cPLA2) from potato [15], shedding new light on its possible physiological function. As a storage protein, patatin is mainly localized in the vacuoles where it is inactive. Upon wounding or attack by pathogens, patatin is translocated to the cytosol where the enzyme becomes active under basic conditions. cPLA2 is a member of a diverse superfamily of phospholipase-A2-type serine hydrolases [15]. In animal cells, cPLA2 cleaves phospholipid membranes to release arachidonic acid which in turn is metabolized to prostaglandins and leukotrienes, two classes of inflammatory mediators. Membrane phospholipids in plant cells contain linoleic acid instead of arachidonic acid at the sn-2 position of glycerol. The liberation of linoleic acid might be an important step in the production of derivatives such as endogeneous elicitors. Recently the crystal structure of human cPLA2 has been solved [1] showing a novel active site comprising a dyad of Ser228 and Asp549. Based on a partial sequence alignment of patatin with cPLA2 together with the previous finding that patatin is identical to a cytosolic phospholipase A2, we propose in this paper a similar catalytic dyad for patatin.

Materials and methods

General DNA techniques

All enzymes for DNA manipulations were purchased from New England Biolabs and applied according to manufacturers instructions.

Escherichia coli DH5α strains were used for all plasmid constructions. Point mutations were introduced according to the Quick Change site directed mutagenesis method (Stratagene). Temperature cycling was performed in 0.5-mL Eppendorf tubes using a Pharmacia LKB Gene ATAQ controller.

Plasmid DNA was isolated using the Qiaprep Spin Miniprep Kit (Qiagen, Dusseldorf, Germany). DNA fragments were isolated from an agarose gel using the QiaxII Gel extraction Kit (Qiagen).

Construction of expression plasmid pHH3

E. coli DH5α strains were used as cloning and expression hosts.

A synthetic duplex of 5′-GCTCT(CACCAT)3GCATG-3′ and 5′-C(ATGGTG)3AG-3′ was cloned into HindIII/SphI-digested pUC 19. The resulting vector pHH1 contains a hexa-histidine tag and the original HindIII site was deleted to facilitate restriction analysis.

The 1107-bp NspI/SmaI fragment of plasmid pPATB2 [9] encoding mature patatin was cloned into the SphI/SmaI sites of vector pHH1 to obtain plasmid pHH2. This pHH2 plasmid contains the patatin gene preceded by a His-tag under control of the lac promoter. This pHH2 plasmid contained an introduced cysteine residue due to the filling of the SphI overhang. This cysteine residue was replaced by an alanine residue to avoid possible formation of undesired disulfide bridges. To accomplish this, the following oligonucleotides were used: 5′-CACCATCACCATGCAGCTGCTAAGTTGGAAG-3′ and 5′-CTTCCAACTTAGCAGCTGCATGGTGATGGTG-3′. The DNA sequence of the resulting pHH3 construct was verified by DNA sequencing.


E. coli DH5α cells transformed with plasmid pHH3 were grown overnight at 37 °C in Luria–Bertani medium (10 g bacto-tryptone, 5 g bacto yeast extract, 10 g NaCl) supplemented with ampicillin (100 mg·L−1). The overnight cultures (2 × 200 mL) were added to 9.6 L Luria–Bertani medium supplemented with 100 mg ampicillin per litre. Cultivation took place in a fermentor (New Brunswick) at 37 °C under vigorous stirring and aeration. The cells were grown until the absorbance at 600 nm did not increase further (typically a D600 of 4 was reached after 8 h). Cells were collected by centrifugation (20 min 5000 g, 4 °C) and stored at −20 °C.

Isolation and refolding of inclusion bodies

All isolation steps were performed at 4 °C. The cell pellet of the 10 L culture was homogenized in 300 mL Tris/EDTA buffer (50 mm Tris/HCl, pH 8 and 40 mm EDTA). Sucrose (75 g) and 60 mg lysozyme were added and the solution was incubated for 30 min. An osmotic shock was applied by the addition of 300 mL ice-cold Tris/EDTA buffer. The solution was again homogenized and incubated for 30 min. The cells were sonicated (Branson sonifier-450 using a 1-cm diameter tip, duty control 9, duty cycle 50%) in 100 mL portions, each 3 × 1.5 min with intervals of 3.5 min to prevent overheating. Subsequently, Brij-35 was added to a final concentration of 0.1% (w/v) and the solution was sonicated for 3 min. The inclusion bodies were collected by centrifugation (20 min, 6726 g, 4 °C) and washed with 400 mL Tris/EDTA buffer. After centrifugation, the pellet was resuspended in 100 mL 10 mm Tris/HCl (pH 8) by sonication and divided into small portions. After centrifugation, the inclusion bodies were stored at −20 °C.

The amount of patatin was determined by SDS/PAGE and by absorbance measurements of a solution of the inclusion bodies in 8 m urea using A280 of 0.748 (1 mg·mL−1).

One gram of inclusion bodies was dissolved in 1 L of 8 m urea and 5 mm glycine (pH 9) at room temperature. After centrifugation of insoluble material, the unfolded protein was refolded at 4 °C by dialysis against 10 vol. 20 mm glycine (pH 9), 20 mm 2-mercaptoethanol and 5 mm EDTA, three times for 24 h.


All purification steps were performed at 4 °C. The pH of the dialysed protein solution was adjusted to pH 7 with 20 mm acetic acid, before loading onto a 100-mL DEAE–Sepharose anion-exchange column equilibrated with 300 mL buffer A: 20 mm imidazole (pH 7). The column was washed with 10 column vol. of buffer A and the bound protein was eluted with 1 m NaCl. The eluted protein was dialysed twice against 3 L buffer A and loaded on a 15-mL Ni-nitrilotriacetic acid column, equilibrated with 20 mm buffer A. The column was subsequently washed with 200 mL buffer A, 200 mL Tween-20 (10 mm), 200 mL NaCl (0.5 m) and 100 mL buffer A. The bound protein was eluted with 500 mm imidazole and dialysed against 3 L 20 mm Tris/HCl (pH 8). The dialysed protein was concentrated to a volume of 10 mL by centrifugation using Amicon YM-30 filters. The concentrated protein was loaded on a 200-mL gel-filtration column (Sephadex G100), which was equilibrated with 20 mm Tris/HCl (pH 8). The purified protein was stored at ≈ − 20 °C.

Construction of patatin variants

The substitutions S10C, S10T, S54A, S54C, S54T, S234C, S234T, K17A, D192A, H175N and H288N were introduced into the patatin gene of plasmid pHH3 according to the Quick Change site-directed mutagenesis method. The following oligonucleotides were used: 5′-GTTACTGTTCTATGCATTGATGGAGGTGGA-3′ and 3′-CAATGACAAGATACGTAACTACCTCCACCT-5′ for S10C; 5′-GTTACTGTGTTAACTATTGATGGAGGTGGA-3′ and 3′-CAATGACACAATTGATAACTACCTCCACCT-5′ for S10T hereby introducing a HpaI site; 5′-TGGAGGAACGGCCACAGGAGGTTTATTG-3′ and 3′-ACCTCCTTGCCGGTGTCCTCCAAATAAC-5′ for S54A hereby introducing an EaeI site; 5′-GGAGGAACATGTACAGGAGG-3′ and 3′-CCTCCTTGTACATGTCCTCC-5′ for S54C hereby introducing an NspI site; 5′-GGAGGAACAACTACAGGAGG-3′ and 3′-CCTCCTTGTTGATGTCCTCC-5′ for S54T; 5′-GTTGTTGCTCTGCTTAGGCACTGG-3′ and 3′-CAACAACGAGACGAATCCGTGACC-5′ for S234C; 5′-GCAAATGTTGTTGTTAACATTAGGCACTGG-3′ and 3′-CGTTTACAACAACAATTGTAATCCGTGACC-5′ for S234T hereby introducing a HpaI site; 5′-CATCAATCTTGTTGCCGGCGCTGTTGCTAC-3′ and 3′-GTAGTTAGAACAACGGCCGCGACAACGATG-5′ for D192A hereby introducing a NaeI site; 5′-CTGGAATGATGCCGGCAATTCCACCTC-3′ and 3′-GACCTTACTACGGCCGTTAAGGTGGAG-5′ for K17A hereby introducing a NaeI site; 5′-GTAACAAAGTGATTCGGAGGAAAATATATTGG-3′ and 3′-CATTGTTTCACTAAGCCTCCTTTTATATAACC-5′ for H175N hereby introducing a HinfI site; 5′-GTTTTTCAAGCTCGGAATTCACAAAACAATTACC-3′ and 3′-CAAAAAGTTCGAGCCTTAAGTGTTTTGTTAATGG-5′ for H288N hereby introducing an EcoRI site.

For the subcloning into plasmid pHH4 (plasmid pHH3 with an EcoRI deletion) the different mutants were digested with the following enzymes: mutants S10C, S10T and S54T with AflIII/BamHI, mutant S54A with BstBI/SapI, mutants S234C and S234T with BamHI/AvaI, mutant K17A with BstBI/AflIII, and mutant D192A with SpeI/AvaI. The fragments were subcloned into plasmid pHH4. All final constructs were verified by DNA sequencing.

Expression and purification of patatin variants

E. coli DH5α cells transformed with the different plasmids were grown at 37 °C in 10 mL Luria–Bertani medium supplemented with 0.75 mg ampicillin. These 10-mL cultures were diluted to 200 mL cultures and subsequently grown overnight. Inclusion bodies were isolated as described above. The inclusion bodies were unfolded in 8 m urea and 5 mm glycine (pH 9) and refolded at 4 °C by dialysis against three changes of 5 vol. of 20 mm glycine (pH 9), 20 mm 2-mercaptoethanol and 5 mm EDTA. The refolded variants were tested with respect to activity. Inactive variants were purified by chromatography using a 5-mL DEAE column and a 1-mL Ni-nitrilotriacetic acid column as described above.

Inhibition reactions

Methyl-p-nitrophenyl-octylphosphonate inhibitor

A stock solution of 1 mm methyl-p-nitrophenyloctylphosphonate was prepared in acetonitrile. Residual activity of 0.1 µm patatin after addition of 1 µm methyl-p-nitrophenyl-octylphosphonate, was measured spectrophotometrically using 0.1 mmp-nitrophenylcaprate in 1 mm Triton X-100 and 50 mm Tris/HCl (pH 8).

Diethylpyrocarbonate inhibitor

A stock solution of 100 mm diethylpyrocarbonate was prepared in acetonitrile. Patatin was incubated for 15 min with a 100-fold molar excess of diethylpyrocarbonate. The acetonitrile concentration did not exceed 5% (v/v) and was found to have negligible effects on the activity and stability of the enzyme. Residual activity was measured spectrophotometrically using 0.1 mmp-nitrophenylcaprate in 1 mm Triton X-100 and 50 mm Tris/HCl (pH 8) after addition of the treated enzyme.

Standard enzymatic activities

Standard enzymatic activities were measured spectrophotometrically at 400 nm. The assay contained 1 mL of 10 mm Tris/HCl (pH 8), 1 mm Triton X-100 and 0.1 mmp-nitrophenylcaprate. Activities, given in U, where 1 U = 1 µmol·min−1, were determined by measuring the release of p-nitrophenoxide ions (using an absorption coefficient of 16 888 m−1·cm−1).

Substrate specificity p-nitrophenylesters

Activity towards acyl p-nitrophenylesters was determined in mixed micelles with Triton X-100 in a buffer of 10 mm Tris/HCl (pH 8). Substrate stocks of p-nitrophenyl-butyrate, -caproate, -caprylate, -caprate and -palmitate were prepared in acetonitrile. The enzyme kinetics were studied either by varying the amount of detergent containing a fixed mole fraction of substrate, or by varying the mole fraction of substrate at saturating levels of detergent.


Activity towards mono-acylglycolphosphocholines was assayed in 50 mm Tris/HCl (pH 8) on pure micelles of C8thioglycolPCho, C12thioglycolPCho and C16thioglycolPCho at substrate concentrations of 10 mm. 5,5′-Dithio-bis(2-nitrobenzoic acid) (DTNB) was included in a concentration of 0.1 mm and the increase in absorption was followed at 412 nm. The molar extinction coefficient used for the 2-nitro-5-thiobenzoic acid anion was 13 600 m−1·cm−1.

Titrimetric assays

Activity on pure synthetic phospholipids (diC8PCho, C18lysoPCho), mono-acylglycerols (myverol 18–92) and triacylglycerols (tributyrin and olive oil) were measured using a titrimetric assay. Emulsions of 0.5% triglycerides or phospholipid in 5 mm Tris/HCl pH 8, 2% gum arabic and 0.5% NaCl were prepared by three times sonication for 1 min. Emulsions (3 mL) were stirred in a reaction vessel of an autotitrator (TTT80, Radiometer Copenhagen) maintained at 40 °C, and the pH of the emulsion was kept at pH 8 by addition of a sodium hydroxide solution. Enzyme was added and the release of fatty acid was measured at pH 8 by autotitration with 10 mm sodium hydroxide under nitrogen atmosphere.


Expression of pHH3 construct and purification

From a 10-L culture, 3 g inclusion bodies were isolated. As estimated from SDS/PAGE gel (Fig. 1, lane 2) the inclusion bodies were about 60% pure. Inclusion bodies (1 g) were unfolded and refolded, giving rise to 800 U active patatin. This amounts to 53 mg patatin (refolding efficiency 5% using a specific activity of 15 U·mg−1 for pure patatin, as discussed below).

Figure 1.

 SDS/PAGE analysis of patatin at different stages of purification. All samples were boiled before loading. Lane 1, molecular mass marker; lane 2, inclusion bodies; lane 3, in vitro refolding; lane 4, fraction loaded on DEAE–Sepharose column, lane 5, fraction loaded on Ni-nitrilotriacetic acid; lane 6, eluted fraction after dialysis.

The refolded protein was purified on DEAE ion-exchange chromatography. While patatin remained bound, impurities could be eluted from the column (data not shown). The eluted protein was dialysed and had an activity of 612 U (Table 1), which is a recovery of 76%. At this stage the patatin had a specific activity of 5 U·mg−1.

Table 1. Purification overview of recombinant patatin. Activities were measured on mixed micelles of 0.1 mmp-nitrophenylcaprate and 1 mm Triton X-100.
Inclusion bodies   ≈1000  
In vitro folding8001000.61333534
After DEAE6127651224033
After Ni-nitrilotriacetic acid4505610453067
After gel filtration2953715201997

All recovered patatin was loaded on a Ni-nitrilotriacetic acid column. A small amount of active enzyme (5%) was found in the flow-through (30 U), whereas the different washing steps did not release further active enzyme. D280 measurements showed that impurities were eluted during the washing steps (data not shown). The bound protein was eluted as a single band on SDS/PAGE. After dialysis, 450 U was recovered with a specific activity of 10 U·mg−1. After the gel-filtration column, the eluted fraction contained 295 U with a specific activity of 15 U·mg−1. This corresponds to 19 mg protein with an estimated purity of above 95% (SDS/PAGE analysis, see Fig. 2, lane 3). The purification scheme and results are summarized in Table 1.

Figure 2.

 SDS/PAGE analysis of patatin: influence of boiling and trypsin treatment. The 65-kDa band represents the correctly folded protein which is absent when the sample is boiled. Lane 4 shows that the correctly folded protein is not digested by trypsin. Lane 1, marker; lane 2, unboiled sample; lane 3, boiled sample; lane 4, unboiled sample after limited digestion with trypsin; lane 5, boiled sample after limited digestion with trypsin.

Active site variants

As shown by SDS/PAGE (results not shown), the expression levels of all variants were similar to wild-type pHH3. From a 200-mL overnight culture, 40 mg inclusion bodies were isolated. After refolding, only the H288N variant was as active as wild-type (see Table 2). All other variants, S54A, S54C, S54T, S10C, S10T, S234C, S234T, H175N, D192A and K17A, which were inactive after refolding were further purified on DEAE and Ni-nitrilotriacetic acid columns. After purification, the variants S10C, S10T displayed activities of 29% and 25%, respectively, relative to wild-type patatin. The variant S54C was found to be slightly active (1% relative to wild-type). The other histidine variant H175N was also found to be active, but displayed low activity (10%), whereas the remaining variants K17A, S54A, S54T, S234C, S234T and D192A were all inactive.

Table 2. Activities (%) on mixed micelles of p-nitrophenylcaprate and Triton X-100 of different variants compared to wild-type. Activities were measured with the purified mutants, with the exception of variant H288N. This variant showed already 100% activity after refolding.
Variants% Activity

On SDS/PAGE (see Fig. 3) all inactive variants migrated with an apparent mass of about 65 kDa, except for the variants S234C and S234T. It should be noted that data for one variant, S54T, are not included in Fig. 3. For this variant, the band at 65 kDa was found to be present, but very vague indeed. In contrast, the 65-kDa band of the other Ser54 variant, S54A, can be clearly observed (lane 7, Fig. 3). On SDS/PAGE, the 65-kDa band corresponds with the correctly folded active form of patatin as can be seen in Fig. 2. Here it is shown that the correctly folded protein (65-kDa band) is resistant towards limited trypsin treatment, whereas the misfolded protein (40-kDa band) is rapidly degraded. When patatin samples are boiled prior to SDS/PAGE analysis, only the 40-kDa band is observed, representing unfolded protein.

Figure 3.

 SDS/PAGE analysis of purified patatin variants. The variants were purified on a DEAE and a Ni-nitrilotriacetic acid column. Samples were not boiled before loading. The 65-kDa band represents correctly folded patatin, whereas unfolded protein runs at the correct mass of 40 kDa. Lane 1, marker; lane 2, wild-type; lane 3, D192; lane 4, K17A; lane 5, S234C; lane 6, S234T; lane 7, S54A.

Kinetic characterization of patatin

Substrate specificity

Substrate specificity for p-nitrophenylesters was investigated for varying carbon chain lengths. To avoid physico-chemical effects on kinetics for these substrates, reactions were studied using 20 mm mixed micelles of Triton X-100 and 10 mol % substrate. As shown in Fig. 4, patatin displays highest activity towards p-nitrophenylcaprate (C10).

Figure 4.

 Substrate specificity for p-nitrophenyl esters with varying carbon chain lengths. Activities were measured in mixed micelles with Triton X-100 in 10 mm Tris/HCl (pH 8). The activity of p-nitrophenyl caprate was set at 100%.

Besides its activity on p-nitrophenyl esters, patatin is also found to be active towards mono-acylglycolphosphocholines and diacylphospholipids (see Table 3). These results show that chain length preference as noted above (maximum for C10) is not absolute, as activity increases with chain length for the series of mono-acyl-glycolphosphocholines from C12 to C16. The highest specific activities were obtained with the synthetic phospholipids diC8PCho and diC9PCho.

Table 3. Specific activities of recombinant patatin towards phospholipids and neutral lipids.
SubstrateSpecific activity (U·mg−1)
Olive oil0

Activities towards neutral lipids were measured on triacyl- and mono-acylglycerols. Activities were only observed for mono-olein and myverol 18–92, a mixture of 96% mono-acylglycerols and 2% diacylglycerols. In agreement with previous findings patatin is inactive towards triacylglycerols.

Active site directed inhibitors

Several compounds were tested for their ability to affect enzyme activity with the aim of identifying active site residues. A likely active site serine was found on the basis of the inhibitory action of phosphonate substrate analogs. After the addition of already a 100-fold molar excess of methyl-p-nitrophenyl-octylphosphonate, patatin was found to be inactivated rapidly and completely (data not shown). Subsequently the presence of an active site histidine was tested using diethylpyrocarbonate. Incubation of patatin with this inhibitor for 15 min did not lead to any substantial decrease in enzyme activity, indicating that an active site histidine is lacking.


Patatin provides an interesting opportunity for application of its catalytic power as a mono-acyl esterase, or ester synthase at low water activities [16]. Its relative abundance in potato is an additional advantage for application purposes. A drawback for a detailed investigation of its catalytic properties, however, is the complexity of the wild-type protein. Several homologous genes have been described coding for patatins [9–11], while different glycosylation patterns add to the complexity of the native enzyme mixture.

In our studies, we have decided to focus on one patatin type only (patB2) and express the enzyme in E. coli, thereby avoiding glycosylation. For the expression of patatin a pUC19 plasmid was used. High yields of the protein were obtained as inclusion bodies requiring unfolding in urea and subsequent refolding to obtain fully active enzyme. Optimal refolding was found at protein concentrations of 1 mg·mL−1, at pH 9 under reducing conditions (20 mm 2-mercaptoethanol). Although the refolding efficiency is rather low (50 mg of active patatin obtained from 1 g inclusion bodies, see Table 1), sufficient protein is recovered for purification and functional analysis. Three purification steps were found to be sufficient for yielding pure patatin. The specific activity of the purified recombinant protein of 15 U·mg−1 for the substrate pNPCho10 dissolved in Triton X-100 micelles is higher than that of wild-type patatin purified from potato (10 U·mg−1). This difference in specific activity may be due to the heterogeneity of the wild-type patatin. Our experiments confirm the previous findings [5] that glycosylation is not essential for esterolytic action of patatin.

Besides its esterolytic activity towards p-nitrophenylesters, potato tuber patatin is also active on mono- and di-acyl phospholipids and mono-acylglycerols, but inactive towards diacyl- and triacyl-glycerols [3]. These results were confirmed by us using the recombinant protein.

Studies of substrate specificity of lipolytic enzymes such as patatin may give rise to artefacts when changes in physico-chemical properties of the substrates are not controlled. To avoid undesired effects, the substrates were dissolved at low molar fractions in the inert detergent Triton X-100. Kinetic data obtained at varying detergent concentrations but at fixed molar fractions of substrate yield a set of apparent affinities for the detergent and maximal rates at high detergent levels. Replotting these maximal rates as a function of the mole fraction of substrate gives reliable information about substrate preference. As a control for the absence of interfacial artefacts, the apparent affinity of patatin for the detergent should not change as a function of the mole fraction of substrate.

Summarizing these kinetic studies, we find that acyl chain length preference is not absolute, and depends on the identity of the headgroup of the substrate used. This points to a delicate balance between hydrophobicity and hydrophilicity affecting patatin activity. Interestingly, highest activities were obtained for diacyl phospholipids such as diC8PCho and diC9PCho. We also tested several substrates, C16thioglycolPCho, diC8PCho, myverol and mono-olein, with patatin isolated from the potato tuber. For these substrates, no major differences were observed as compared to our recombinant patatin.

Using pure mono-olein, we found a specific activity of 15 U·mg−1 for the recombinant enzyme and 20 U·mg−1 for the enzyme isolated from potato tubers. The latter activity is three times lower than published in literature [16]. A surprising observation is the five times higher specific activity of native and recombinant patatins on myverol, an impure fraction of mono-olein (see Table 3). The difference between our measurements (20 U·mg−1) and the literature value (60 U·mg−1) can therefore be explained by differences in purity of the mono-olein substrates used.

The high phospholipase A activity besides the action of patatin as a mono-acyl esterase prompted us to investigate amino-acid sequence identity with several phospholipases.

As was shown in Fig. 5, only limited homology was found for short stretches of sequence of patatin B2 and the catalytic domains of the large multidomain phospholipase, cPLA2[1]. Strikingly, the overall sequence homology of patatin with this phospholipase is below 10% indicating that the overall fold of these proteins must be very different indeed. Recently, the X-ray structure of cytosolic PLA2 from human has been solved [1]. While this enzyme belongs to the large family of serine hydrolases, it is interesting to note that a novel active site architecture is present. A Ser-Asp dyad, rather than the more common Ser/His/(Asp or Glu), triad was identified. The limited homology with patatin prompted us to suggest that also patatin contains such an active site dyad. Experimental evidence for an active site serine in patB2 was derived from the inhibitory action of methyl-p-nitrophenyl-octylphosphonate, which specifically blocks a nucleophilic serine residue by covalent attachment of the methyl-octylphosphonate group.

Figure 5.

 Alignment of partial sequences around conserved residues of patatin, and cPLA2. Active site residues of cPLA2: Ser228, Asp549 and oxyanion ligand Arg200 are indicated in bold as well as the corresponding residues in the aligned proteins. The structurally important Ser234 in patB2 is also shown aligned with Ser573 of cPLA2. Sequence data are obtained for cPLA2 from the structural data (RCSB PDB code 1CJY).

In our search for the identity of active site residues in patatin, several mutants were constructed. All serine residues conserved in the family of patatins (Ser10, Ser54, Ser234) were mutated. Based on the partial homology with cPLA2 (Fig. 5) we expect Ser54 to be the active site serine. This residue was mutated to an alanine, a cysteine and a threonine residue. The kinetic properties of the variants obtained demonstrated the catalytic importance of Ser54, as Ser54→Ala and Ser54→Thr were completely inactive, whereas only Ser54→Cys displayed some residual activity. Based on SDS/PAGE analysis, all of these variants were found to fold correctly, albeit to different extents. The yield of correctly folded S54A was highest, whereas only a small fraction of correctly folded S54T was found. Overall, for the Ser54 variants it is concluded that the lack of activity is not due to protein misfolding. Based on the activity of the Ser10 variants, this residue was discarded as a possible active site residue. Variants of the third conserved serine residue (Ser234) were also found to be inactive. For these variants, lack of activity is explained by misfolding of the proteins (see Fig. 3). Thus we conclude that Ser234 is important for the correct folding of the protein. Most likely it is present in a structurally conserved area. Partial homology with cPLA2 (Fig. 5) strengthens our conclusion that Ser234 is structurally important, as homology was found around Ser573, which is not, however, the active site serine of cPLA2.

The folding or misfolding of the protein can be followed on by SDS/PAGE when the samples are not boiled prior to loading. On SDS/PAGE, incorrectly folded proteins migrate at an apparent mass of 40 kDa, whereas correctly folded patatin runs at 65 kDa [2,6]. These two masses have been observed previously [17] and were attributed to the monomeric form (40 kDa) or the oligomeric form (65 kDa) of patatin. Only the oligomeric form of patatin possesses esterase activity. The reason why the supposedly oligomeric form runs faster than expected (65 kDa and not 80 kDa) is not clear. An alternative explanation may be the fact that patatin acts at lipid–water interfaces. Only correctly folded protein displays this property leading to slower migration due to interaction with detergent present in the gel.

Patatin remained active when incubated with diethylpyrocarbonate, which reacts mainly with histidine residues in proteins [18]. This is an indication that histidine residues are absent in the active site. To confirm this hypothesis, we mutated several conserved histidine residues (H175N, H288N). As expected, both variants displayed activity in our assay. The reduced activity of the H175N may also be due to an effect on protein folding or slight structural rearrangements.

Based on partial homology with cPLA2 we identified Asp192 as a likely active site residue forming a catalytic dyad with the proposed Ser54. On SDS/PAGE (see Fig. 3) the D192A variant was found to partially fold into the correct 65-kDa species. However, this variant lacked activity in our assays, in agreement with our hypothesis.

Finally, it was reported that the Ser-Asp dyad phospholipases may operate effectively through a nearby basic residue acting as oxyanion hole ligand [1]. Again based on partial homology, we identified Lys17 in patatin as a possible candidate. Indeed we found that Lys17 patatin variants displayed strongly reduced activity.

In conclusion, we propose that patatin is a serine hydrolase with an active site Ser-Asp dyad. Our directed mutagenesis studies have indicated that Ser54 and Asp192 form a catalytic dyad in patatins where Lys17 probably acts as an oxyanion hole ligand.


We would like to thank Mr Ruud Cox for synthesis of the substrates and the phosphonate inhibitor. We also would like to thank Nourdin Ghalit and Tom Wijnhoven for their work on several mutants.