Biochimica et Biophysics Acta, 1123 (1992) 59-64 0 1992 Elsevier Science Publishers B.V. All rights reserved
BBALIP
59 OOOS-2760/92/$05.00
53806
Substrate specificities of lipases A and B from Geotrichum candidum CMICC 335426 Emmanuelle
Charton and Alasdair R. Macrae
lJniler,er Research, Colworth Laboratory, Colworth House, Sharnbrook, Bedford (U.K.)
(Revised
Key words:
Lipase A; Lipase
(Received manuscript
8 April 1991) received 29 July 1991)
B; Lipase structure:
Substrate
specificity;
(G. candidurn)
Ttie mould Geotrichum cundidum produces extracellular lipases with different substrate specificities according to strains. We purified two lipases - termed lipase A and lipase B - from Geotrichum candidum CMICC 335426. The specificity of the two lipases was investigated using hydrolysis assays on emulsions of pure acylglycerols and a wide range of fatty acid esters. Lipase B was very highly specific for hydrolysis of esters of cis-A9-fatty acids. Lipase A did not show such strict specificity, because it hydrolysed a wider variety of fatty acid esters, in particular those of palmitic acid and isomers of oleic acid. We think that differences in specificity previously observed for crude lipases from various strains of G. candidurn can be explained by the presence of different levels of specific (lipase B) and non-specific (lipase A) lipases. As lipases A and B are structurally related proteins, a minor variation in structure may be responsible for the differing specificities.
Introduction The mould Geotrichum candidum produces extracellular lipases whose substrate specificity have been widely investigated. Strains of the fungus were isolated for lipase production by several laboratories, the first being in the USA by Alford and Pierce [1,2]. The specificities of these extracellular lipases were examined using crude or partially purified enzymes [3-71. Although G. candidum lipases show a preference for long chain unsaturated substrates, some differences in specificity exist from one strain to the other [6,8,9]. In recent studies, G. candidum from the ATCC 34614 strain was reported to produce two lipases from two distinct genes [lO,ll]. The two cloned enzymes were shown to have similar substrate specificities. Mul-
Abbreviations: HPLC, high-performance liquid chromatography; FPLC, fast protein liquid chromatography; PPO, l(3)-oleoyl-2,3(l)dipalmitoylglycerol; POP, 1.3-dipalmitoyl-2-oleoylglycerol; SOS, 1,3distearoyl-2-oleoylglycerol. Correspondence: ratory, Colworth
A.R. Macrae, Unilever Research, Colworth LaboHouse, Sharnbrook, Bedford, MK44 ILQ. U.K.
tiple lipases were purified [13] or partially purified [14] from other strains, but no specificity studies of the isolates were carried out. Purification of lipases from a commercially available lipase (GC-20 Amano) has also been reported 193. One isoenzyme had different specificity from the others in term of its preference for the oleate ester of methyl umbelliferol over the palmitate ester. We have reported the purification of two lipases from the strain CMICC 335426, using high-performance liquid chromatography. They were termed ‘lipase A’ and ‘lipase B’ and their properties were investigated [15,161. Although structurally similar, the two lipases were shown to have very different substrate specificities, lipase B having an unprecedented specificity, releasing only unsaturated fatty acids from acylglycerols and other esters. We suggested that G. candidum may produce more than one lipase with different specificities, at different amounts according to the strain. This could explain the specificity differences reported for crude lipases in the literature. The aim of the work reported here was to study in detail the specificities of lipases A and B using a wider range of substrates. To obtain sufficient enzyme quantities for specificity studies, a larger scale purification procedure was developed.
60 Materials
and Methods
Materials
The methyl esters of cis-A7, -38, -dlO and -317 octadccenoic acids were provided by Prof. F.D. Gunstone (University of St. Andrews) and purified by preparative thin-layer chromatography. cis-A2, -33, -A 12 and -A 13 octadecenoic acids were also provided by Prof. F.D. Gunstone and the c&-J14 and -315 isomers were obtained from Mr. A.Visser (Unilever Research Laboratory, Vlaardingen). They were purified by preparative TLC and methylated by refluxing at X0 ’ C for 3 min with 14% boron trifluoride in methanol. The purities of the methyl esters determined by capillary gas chromatography were A2 95%, 33 890/o, A4 X7%, 37 91%, 3x 980/c, 310 91%, 1\12 X9%, 31s 94% and A17 78%. The major impurities of these samples consisted of esters of saturated fatty acids (mainly stearic acid) and trans isomers. The ester of the cis-39 isomer (oleic acid) was also present at very small amounts (less than 2%). Octyl oleate (99%) was provided by Mr. A. Peilow (Unilever Research Laboratory, Colworth). Dee-9-enoic acid (99%) was purchased from TCI, Tokyo and methylated using the boron trifluoride method. Isopropyl oleate was synthesised by refluxing for 7 h at SO “C an equimolar mixture of 99.7% pure isopropanol (BDHI and 99% pure oleic acid (Sigma) using Amberlyst 15 (BDH) as a catalyst; unreacted acid was removed by elution with diethyl ether through an alumina column (basic, activity II from Merck); unreacted alcohol was removed by distillation. The pure synthetic triacylglycerols l(3)-oleoyl-2,3(l)-dipalmitoylglycerol (PPO), 1,3-dipalmitoyl-2-oleoylglycerol (POP) and 1,3-distearoyl-2-oleoylglycerol (SOS) were obtained from Unilever Research Laboratory, Vlaardingen. Their purities determined by silver impregnated silica HPLC were POP 960/c, PPO 98% and SOS 98%. All other substrates were obtained from Sigma and were at the highest purity available. Diolein (Sigma) was a 99% pure mixture of the 1.3-t = 85%) and the 1,2(2,3)-(2 15%) isomers. Monoolein was 99% pure 1-monooleoyl-rat-glycerol.
zinc, HCI (pH 4.3). A pH gradient to 20 mM piperazinc HCI (pH 5.3) was devclopcd on the column for enzyme separation. Protein elution was detected by absorbance at 280 nm and fractions assayed for lipase activity. Fractions samples were analyscd by SDS-denaturing polyacrylamidc gel electrophoresis (SDS-PAGE) using 5% stacking and 6.5% running polyacrylamidc gels in a Bio-Rad Mini Protean II apparatus [ 171. Protein assay was done using bovine serum albumin as a protein standard for the kit provided by Bio-Rad [lg]. Hydrolysis reactions
Lipase titre was determined using emulsions of 5% (w/v) olive oil in 2%’ gum arabic solution as described previously [15]. Specificity studies were carried out using mono-, di-, triacylglycerols, methyl esters and various esters of oleic acid. Unless specified differently, 1.0 and 2.8 pg of lipases A and B, respectively, wcrc added to 20 ml of oil-in-water emulsions containing 0.5% (w/v) substrate. 2% (w/v) gum arabic and 0.5% (w/v) CaCl,. The pH of the reaction was maintained at 8.0 with 0.01 M NaOH using a pH-stat autotitrator (Radiometer, Copenhagen). The reaction was run at 30 “C. The rate of hydrolysis was deduced from the volume of alkali added during the 3 min following lipase addition. It was expressed in pmol fatty acids released per minute and per mg of protein. This rate was also calculated relative to that of methyl oleatc. The assays were run in duplicate and a mean taken. For each substrate, the rate of reaction obtained in the duplicate assays varied by less than 5%. The pure synthetic triacylglycerols l(3)-oleoyl2,3( l)-dipalmitoylglycerol (PPO), 1,3-dipalmitoyl-2oleoylglycerol (POP) and 1.3-distearoyl-2-oleoylglycerol (SOS) were hydrolysed at 40°C as 1% oil-in-water emulsions in 2% gum arabic containing 0.5% CaCl,. The pH of the medium was maintained at X.0 with 0.1 M NaOH. The reaction was catalysed with 9.0 pg of lipasc A and 12.0 pg of lipase B and was stopped after 15 min for analysis of the composition of the released fatty acids as described previously [lSl. Results
Production candidum
and purification
of lipases A and B from
G.
CMICC 335426 Production of crude lipase by growth of G. candidum in fermenters and concentration of the culture filtrate on Q Sepharose Fast Flow were achieved as described previously [15]. However, a new method for scaling up the subsequent purification was developed using a Pharmacia LKB FPLC system. The eluate from Q Sepharose was diluted 50: 50 (v/v) with water in order to reduce its conductivity and its pH adjusted to 4.3 with HCI. It was then loaded on to a 1.6 x 2Ocm column (Pharmacia code C16/20) packed with S Sepharose Fast Flow equilibrated in 20 mM pipera-
Large
scale purification
candidum
CMICC
qf’ lipases A and
B from
G.
335426
Fig. 1 shows the elution profile of lipase A and B from the FPLC S Sepharose column. Lipasc A was eluted between pH 4.2 and 4.6, lipasc B between 4.8 and 5.1. Both were obtained at a high level of purity. SDS-PAGE (Fig. 1) and IEF (data not shown) strongly indicate that both lipase fractions contain single protein species. The molecular masses of lipases A and B are 62 and 58 kDa, respectively [15]. Although lipase B (pl 4.5) has a more acidic pI than lipase A (PI 4.71) [lS], it was cluted at a higher pH than lipase A. This
61 TABLE
I
TABLE
Hydrolysis of acylglycerols and methyloleate by Geotrichum candidum CMICC 335426 lipases A and B
II
Composition of the fatty acids released by Geotrichum candidum CMICC 335426 iipases A and B from the pure synthetic triacylglycerols POP, PPO, SOS
Substrate
Monoolein Diolein Triolein POP PPO SOS Methyl oleate
Rate of hydrolysis
(pmol/min
lipase A
lipase B
1010 II20 1270 670 1000 hl0 720
1430 1540 2380 1890 IS30 1400 640
per mg) Triacylglycerol
POP PPO SOS
saturated
oleate
(P, S) 25.4 42.1 61.4
I.1 0.5 0.1
98.9 99.5 99.9
and triolein. For both enzymes, triolein was hydrolysed the fastest, followed by diolein and monoolein. Lipase B had a higher specific activity than lipase A for all these substrates. Lipases A and B readily hydrolysed POP, PPO and SOS, the rates of reaction varying between 50 and 80% those observed with triolein. Lipase A had a small preference for PPO and lipase B for POP (Table I>. The composition of the fatty acids released from POP, PPO, SOS are shown in Table II. Lipase A hydrolysed large amounts of palmitic, stearic and oleic acids from these substrates. In sharp contrast, the fatty acids released by lipase B contained more than 98.9% oleic acid. This confirms the results obtained previously with HPLC purified enzymes [El. Lipase B shows no significant regioselectivity with triacylglycerols, oleate being released from POP and PPO at a similar rate.
were efficient catalysts for the hydrolysis of mono-, di-
-_*
74.6 57.9 38.h
fatty acids ($6) lipase B
oleate
(P, S)
Specificity of lipases A and B from G. candidum CMtCC 335426 Hydrolysis of acylglycerols by G. candidum lipases A and B (Table I and II) G. candidum lipases A and B
Activity
of released
lipase A saturated
shows that the external charge and not the net charge of the molecule is important in its binding and separation on the ion exchange matrix. In a typical purification run, 100 000 units of crude lipase preparation were loaded on 10 ml of S Sepharose and eluted with a 1 1 gradient to give 80000 units of lipase B (spec. act. 2500 units/mg) and 20000 units of lipase A (spec. act. 1000 units/mgl. Lipases A and B were obtained at the following concentrations A: 100 kg/ml; B: 280 Fug/ml and used as such for specificity studies,
1
Composition
*,.*.rn 2
3
4
5
L
6
7
89
Wml)
600
6.4
600
6.2
&so
6
4.8
0.05
4.2 60 80 100 Elution volume (ml) Fig. 1. Purification by FPLC of G. candidum CMICC 335426 lipases A and B. 5 ml fractions were collected. n , lipase activity units/ml; +, pH; dotted line, absorbance at 280 nm. Top, SDS-PAGE of crude enzyme (L) and fractions 1 to 9.
in olive oil
Methyl lipasc A Methanol Prop;u-
I-01
Propan-I-01 Butan-I-ol Oct;ln-
I -(II
Tctradwan-l-01
reaction
lipase A Satur;cted
rate liparc
6:O
0.0
0.0
IS?
x:0
7.4
23
IO:0
0.0
0.X 0.0
I71
I?:0
0.0
0.0
40
I‘$:0
0.0
71
Ih:O
4.4 04.2
IX:0
13.1)
I.1
I I.6
3.7
I -(II
-16
Octadecan-
I-01
-1-l
Octadeccn-
I -(II
42 0
ci.dO IO:
Monounsaturated
13.6
l3:l 16: I IX:
Hydrolysis of methyl oleate by G. candidum iipases A und B (Tuble I). The results show that the specific activity of lipase A on methyl oleate was slightly higher than that of lipase B. Lipases A and B hydrolysed triolein 2 and 4-times faster than methyl ester of oleic acid, respectively. Hydrolysis of other esters of oleic ucid by G. candidum lipuses A nnd B (Table 111). Esters of oleic acid and linear saturated alcohols from 1 to 18 carbons were all digested by G. cundidum lipases A and B at significant rates. However, some differences between the two lipascs were observed. Lipase A had a prefercncc for esters of short chain alcohols (Cl, C2, C3, C4) with a peak of activity for propyl oleate. For esters of medium to long chain alcohols CC8 to Cl8), reduced reaction rates were observed. In contrast, the activity of lipase B was high for esters short chain alcohols ((3, C4) as well as those of medium (C8, Cl2) and long (ClO, ClX) chain alcohols with a peak for octyl oleate. The two enzymes hydrolysed an ester of the secondary alcohol isopropyl oleate. However. the rates of hydrolysis of this alcohol were about X-times lower than those of the corresponding primary alcohol. Interestingly enough, lipase B hydrolysed oleyl oleate twice as fast as stearyl oleatc showing that cis-J%unsaturation of the alcohol was preferred to saturation. Finally. lipase B hydrolyscd cholestcryl oleate at a significant rate whereas lipase A did not catalyse the hydrolysis of this substrate. Hydrolysis of methyl esters by G. cundidum lipuses A und H (Tables IV und V). Methyl esters of suturated fatty ucids (Table IV). G. cundidum lipases A and B were tested on methyl esters of the 6: 0, 8: 0, 10: 0, 12: 0. 14:0, 16: 0 saturated fatty acids. Both enzymes were inactive on 6 : 0. 10: 0 and 12: 0 methyl esters but gave a slow rate with the 8 : 0 ester. Appreciable hydrolysis of trioctanoin by lipases A and B was also measured
0.0
FFA
I
0.5
XX
I
12.OH
B
FFA
II7
Hcxadccan-
C’holerterol
Relative
lipase B
IO0
Ethanol
cstcr
35.11
IO0 IX: I (ricinoleic
Polyunsaturated Linoleic
acid)
5.0
FFA
acid
ru-Linolenic
acid
y-Linolenic
acid
Arachidonic
100
17.1
I94
01.0
I66
703
acid
3.‘)
I.1
0.0
Il.0
(unpublished data). Higher activities with lipase A were observed when longer chain fatty acid esters were used (14: 0, 16:O and 18:O). Methyl palmitate was significantly the best substrate for lipase A, with a rate of 64% that of methyl olcate. In contrast, lipase B had no activity on 14: 0 and 16 : 0 methyl esters and very small activity on 18 : 0 methyl ester. - Methyl ester,s of cis-J 9 monounsutur~lter1~~~ firtty wids (Table IV). The effect of fatty acid chain length on
Isomer
Relative lipw
A
reaction
rate lipase B
2.0
0.0
16.4
0.3
0.X
0.0
5.7
0.0
25.0
()..I
7.3
100
0.x I00
lb.7
I.?
13.1
0.0
14.2
0.0
Il.7
0.0
IX.3
I .8
14.6
0.7
23.9
0.7
63 lipase activity was further investigated with methyl esters of the cis-A9 monounsaturated fatty acids 10: 1, 14: 1, 16: 1 and 18: 1. For both enzymes the rates of hydrolysis of these cis-A9 unsaturated substrates were higher than their saturated equivalents. Moreover, longer chain fatty acid esters were hydrolysed more rapidly than shorter chain fatty acid esters. The 16: 1 methyl ester was the best substrate for lipase A, followed by 18:1, 14: 1 and 10: 1. However, lipase B hydrolysed I8 : 1 methyl ester the fastest; it also hydrolysed 16: 1 significantly but 14: 1 and 10: 1 were digested very slowly. Finally, the presence of a hydroxyl group at the 12-position of the carbon chain of 18: 1 considerably reduced the activity of both enzymes. - Methyl ester;v of polyunsaturated fatty acids (Table IV). Lipase B hydrolysed methyl linoleate and and cu-linolenate at rates slightly lower than that of methyl oleate. But the activity of lipase A on these substrates was almost twice the activity measured on methyl oleate. Methyl y-linolenatc was a poor substrate for both enzymes. This was due to the presence of a double bond in the h-position. Methyl arachidonate (C20), which has double bonds in the 5, 8, 11 and 14-positions was not hydrolysed at all. - Methyl esters of isomers of oleic acid (Table V). G. candidum lipases A and B hydrolysed methyl oleate at a maximum rate. Although lipase A showed a preference for the cis-A9 isomer, it did not have the strong specificity of lipasc B which was inactive on the 42, 44, 46, All, Al2 and Al3 isomers. Very small activities were measured for the hydrolysis by lipase B of the remaining isomers, but these very slow rates could be due to the presence in the samples of small amounts of impurities (especially the cis-A9 isomer). The methyl ester of trans-A9 18: 1 was hydrolysed by lipase A but very slow reaction was observed with lipase B.
and B. c&-A9 polyunsaturated esters are also good substrates for the enzymes. However, the presence of a double bond between the 9-position and the carboxylic extremity of the acid reduces considerably the activity of the lipases. Moreover, monocne positional isomers of oleic acid are all digested by lipase A at significant rates but lipase B does not hydrolyse any of these isomers at a rate higher than 1.8% that of methyl oleate. These results confirm the discovery of a G. candidum lipase which is very highly specific for esters of fatty acids containing a cis-double bond in the 9-position. Previous studies have reported the specificity of crude G. candidum lipase preparations using esters of saturated fatty acids [4,6,8], cis-A9 monounsaturated fatty acids [4,6,8,19,201, polyunsaturated fatty acids [19,20], isomers of oleic acid [3,20]. The significant variations of specificities reported in these publications can be explained by the presence in the crude preparations of different levels of specific (lipase B) and nonspecific (lipase A) lipases. We are currently investigating this using specific polyclonal antibodies. Although they are structurally related 1151 and appear to possess similar partial amino acid sequences (unpublished data), lipases A and B from G. candidum CMICC 335426 show striking differences in their specificities. Lipases from other strains of G. candidum have been crystallised [5,20-221 and three dimensional structures will hopefully become available soon. Crystallisation of lipases A and B are under way. It will be of considerable interest to determine the structural differences between lipases A and B which result in their differing specificity. Further work with lipases A and B will generate information about the influence of protein structure on lipase specificity and will eventually enable one to use protein engineering for the modification of lipase specificity.
Discussion Acknowledgements G. candidum lipases A and B are effective catalysts for the hydrolysis of acylglycerols and alkyl esters of oleic acid. Lipase A has a preference for oleic acid esters of short primary alcohols, whereas lipase B is very active on esters of short, medium and long chain alcohols as well as cholesterol. But for both enzymes, an ester of a primary alcohol is preferred to the secondary alcohol. Lipase A hydrolyses methyl esters of long saturated fatty acids at significant rates, with an unexpected peak for methyl palmitate. Interestingly, 8 : 0 is the only saturated fatty acid hydrolysed by lipase B from a methyl ester or a triacylglycerol to an appreciable extent. This suggests that caprylic acid may have a configuration particularly adapted to the active site of the enzyme. Among the methyl esters of cis-A9 monounsaturated fatty acids, methyl oleate and methyl palmitoleate are hydrolysed the fastest by lipases A
We are grateful to Mr G. Mycock for growing G. candidum in fermenters and wish to thank Prof. F.D. Gunstone (University of St. Andrews) and Mr. A. Visser (Unilever Research Laboratory, Vlaardingen) for kindly providing us with isomers of oleic acid. We also thank Prof. A.R. Slabas, Mr. C. Sidebottom and Dr. S. Doyle for their advice on protein purification procedures. References I Alford, J.A. and Pierce, D.A. (1961) J. Food Sci., 26, 518-524. 2 Alford, J.A., Pierce, D.A. and Suggs, P.G. (1964) J. Lip. Res. 5, 390-394. 3 Jensen, G. (1973) Lipids 9 (3), 149-157. 4 Franzke, C., Kroll, J. and Petzold, R. (1973) Die Nahrung 17 (2), 171-174.
64 5 Tsujisaka, Y., Iwai, M. and Tominaga, Y. (1973) Agr. Bio. Chem. 37 fh), 1457- 1764. 6 Aneja, R. and Hollis, W.H. (1983) Abstracts No. 120 of papers of the Am. Chem. Sot.. Vol. 1Xh. 7 Jacobsen. T., Olsen, J. and Allerman, K. (1900) Biotechnol. Lett. 12 (2), 121-126. X Aneja. R. (1987) J. Am. Oil Chem. Sot. 64 (5). 645. 9 Baillargeon, M.W., Bistline, R.G. and Sonnet, P.E. (1989) Appl. Microhiol. Biotechnol. 30, Y2-96. 10 Shimada. Y., Sugihara, A.. Tominiga. Y. and lizumi. T. (1989) J. Biochem. 106, 383-388. I I Shimada. Y.. Sugihara. A.. lizumi, T. and Tominaga. Y. (1990) J. B&hem. 107, 703-707. 12 Sugihara, A., Shimada, Y. and Tominaga, Y. (1990) J. Biochem. 107. 4X-430. 13 Veeraragavan, K.. Colpitts. T. and Gibbs, B.F. (19YO) Biochem. Biophys. Acta 1044, 26-33. 14 Jacobsen, T.. Olsen, J., Allerman, K., Poulsen. O.M. and Hau, J. (1989) Enzyme Microbial. Technol. 11, 90-95. IS Sidebottom. C., Charton, E., Dunn. P.P.J., Mycock, G.. Davies. C.. Sutton, J.. Macrae, A.. Slabas. A.R. (1991) Eur. J. Biochem.. in press. 16 Charton, E., Davies, C.. Sidebottom. C., Sutton, J.. Dunn, P.P.J..
17 18 IY 20 21
22
23
Slabas, A.R. and Macrae, A.R. (1991) in Lipases: Structure. Mechanism and Genetic Engineering (Alberghina, L., Verger, R. and Schmid, R.D.. eds.), pp. 335-338, GBF Monographs 16. VCH Weinheim lY91. Laemmli, U. flY70) Nature 227, h80-685. Bradford. M.M. (1976) Anal. Biochem. 72. 248-254. Marks, T.A., Quinn. J.G.. Sampugna. J. and Jensen, R.G. (196X) Lipids. 3 (2). 142-146. Jensen, G. and Pitas, E. (1976) Lipids I. 141-146. Menge. U., Hecht. H-J., Schomburg. D., Schmid, R.D., Hedrich. H.C. and Spener. F. (IYY I) in Lipases: Structure. Mechanism and Genetic Engineering (Alherghina, L., Verger. R. and Schmid. R.D., eds.). pp. 59-62. GBF Monographs 16, V
BBALIP
59 OOOS-2760/92/$05.00
53806
Substrate specificities of lipases A and B from Geotrichum candidum CMICC 335426 Emmanuelle
Charton and Alasdair R. Macrae
lJniler,er Research, Colworth Laboratory, Colworth House, Sharnbrook, Bedford (U.K.)
(Revised
Key words:
Lipase A; Lipase
(Received manuscript
8 April 1991) received 29 July 1991)
B; Lipase structure:
Substrate
specificity;
(G. candidurn)
Ttie mould Geotrichum cundidum produces extracellular lipases with different substrate specificities according to strains. We purified two lipases - termed lipase A and lipase B - from Geotrichum candidum CMICC 335426. The specificity of the two lipases was investigated using hydrolysis assays on emulsions of pure acylglycerols and a wide range of fatty acid esters. Lipase B was very highly specific for hydrolysis of esters of cis-A9-fatty acids. Lipase A did not show such strict specificity, because it hydrolysed a wider variety of fatty acid esters, in particular those of palmitic acid and isomers of oleic acid. We think that differences in specificity previously observed for crude lipases from various strains of G. candidurn can be explained by the presence of different levels of specific (lipase B) and non-specific (lipase A) lipases. As lipases A and B are structurally related proteins, a minor variation in structure may be responsible for the differing specificities.
Introduction The mould Geotrichum candidum produces extracellular lipases whose substrate specificity have been widely investigated. Strains of the fungus were isolated for lipase production by several laboratories, the first being in the USA by Alford and Pierce [1,2]. The specificities of these extracellular lipases were examined using crude or partially purified enzymes [3-71. Although G. candidum lipases show a preference for long chain unsaturated substrates, some differences in specificity exist from one strain to the other [6,8,9]. In recent studies, G. candidum from the ATCC 34614 strain was reported to produce two lipases from two distinct genes [lO,ll]. The two cloned enzymes were shown to have similar substrate specificities. Mul-
Abbreviations: HPLC, high-performance liquid chromatography; FPLC, fast protein liquid chromatography; PPO, l(3)-oleoyl-2,3(l)dipalmitoylglycerol; POP, 1.3-dipalmitoyl-2-oleoylglycerol; SOS, 1,3distearoyl-2-oleoylglycerol. Correspondence: ratory, Colworth
A.R. Macrae, Unilever Research, Colworth LaboHouse, Sharnbrook, Bedford, MK44 ILQ. U.K.
tiple lipases were purified [13] or partially purified [14] from other strains, but no specificity studies of the isolates were carried out. Purification of lipases from a commercially available lipase (GC-20 Amano) has also been reported 193. One isoenzyme had different specificity from the others in term of its preference for the oleate ester of methyl umbelliferol over the palmitate ester. We have reported the purification of two lipases from the strain CMICC 335426, using high-performance liquid chromatography. They were termed ‘lipase A’ and ‘lipase B’ and their properties were investigated [15,161. Although structurally similar, the two lipases were shown to have very different substrate specificities, lipase B having an unprecedented specificity, releasing only unsaturated fatty acids from acylglycerols and other esters. We suggested that G. candidum may produce more than one lipase with different specificities, at different amounts according to the strain. This could explain the specificity differences reported for crude lipases in the literature. The aim of the work reported here was to study in detail the specificities of lipases A and B using a wider range of substrates. To obtain sufficient enzyme quantities for specificity studies, a larger scale purification procedure was developed.
60 Materials
and Methods
Materials
The methyl esters of cis-A7, -38, -dlO and -317 octadccenoic acids were provided by Prof. F.D. Gunstone (University of St. Andrews) and purified by preparative thin-layer chromatography. cis-A2, -33, -A 12 and -A 13 octadecenoic acids were also provided by Prof. F.D. Gunstone and the c&-J14 and -315 isomers were obtained from Mr. A.Visser (Unilever Research Laboratory, Vlaardingen). They were purified by preparative TLC and methylated by refluxing at X0 ’ C for 3 min with 14% boron trifluoride in methanol. The purities of the methyl esters determined by capillary gas chromatography were A2 95%, 33 890/o, A4 X7%, 37 91%, 3x 980/c, 310 91%, 1\12 X9%, 31s 94% and A17 78%. The major impurities of these samples consisted of esters of saturated fatty acids (mainly stearic acid) and trans isomers. The ester of the cis-39 isomer (oleic acid) was also present at very small amounts (less than 2%). Octyl oleate (99%) was provided by Mr. A. Peilow (Unilever Research Laboratory, Colworth). Dee-9-enoic acid (99%) was purchased from TCI, Tokyo and methylated using the boron trifluoride method. Isopropyl oleate was synthesised by refluxing for 7 h at SO “C an equimolar mixture of 99.7% pure isopropanol (BDHI and 99% pure oleic acid (Sigma) using Amberlyst 15 (BDH) as a catalyst; unreacted acid was removed by elution with diethyl ether through an alumina column (basic, activity II from Merck); unreacted alcohol was removed by distillation. The pure synthetic triacylglycerols l(3)-oleoyl-2,3(l)-dipalmitoylglycerol (PPO), 1,3-dipalmitoyl-2-oleoylglycerol (POP) and 1,3-distearoyl-2-oleoylglycerol (SOS) were obtained from Unilever Research Laboratory, Vlaardingen. Their purities determined by silver impregnated silica HPLC were POP 960/c, PPO 98% and SOS 98%. All other substrates were obtained from Sigma and were at the highest purity available. Diolein (Sigma) was a 99% pure mixture of the 1.3-t = 85%) and the 1,2(2,3)-(2 15%) isomers. Monoolein was 99% pure 1-monooleoyl-rat-glycerol.
zinc, HCI (pH 4.3). A pH gradient to 20 mM piperazinc HCI (pH 5.3) was devclopcd on the column for enzyme separation. Protein elution was detected by absorbance at 280 nm and fractions assayed for lipase activity. Fractions samples were analyscd by SDS-denaturing polyacrylamidc gel electrophoresis (SDS-PAGE) using 5% stacking and 6.5% running polyacrylamidc gels in a Bio-Rad Mini Protean II apparatus [ 171. Protein assay was done using bovine serum albumin as a protein standard for the kit provided by Bio-Rad [lg]. Hydrolysis reactions
Lipase titre was determined using emulsions of 5% (w/v) olive oil in 2%’ gum arabic solution as described previously [15]. Specificity studies were carried out using mono-, di-, triacylglycerols, methyl esters and various esters of oleic acid. Unless specified differently, 1.0 and 2.8 pg of lipases A and B, respectively, wcrc added to 20 ml of oil-in-water emulsions containing 0.5% (w/v) substrate. 2% (w/v) gum arabic and 0.5% (w/v) CaCl,. The pH of the reaction was maintained at 8.0 with 0.01 M NaOH using a pH-stat autotitrator (Radiometer, Copenhagen). The reaction was run at 30 “C. The rate of hydrolysis was deduced from the volume of alkali added during the 3 min following lipase addition. It was expressed in pmol fatty acids released per minute and per mg of protein. This rate was also calculated relative to that of methyl oleatc. The assays were run in duplicate and a mean taken. For each substrate, the rate of reaction obtained in the duplicate assays varied by less than 5%. The pure synthetic triacylglycerols l(3)-oleoyl2,3( l)-dipalmitoylglycerol (PPO), 1,3-dipalmitoyl-2oleoylglycerol (POP) and 1.3-distearoyl-2-oleoylglycerol (SOS) were hydrolysed at 40°C as 1% oil-in-water emulsions in 2% gum arabic containing 0.5% CaCl,. The pH of the medium was maintained at X.0 with 0.1 M NaOH. The reaction was catalysed with 9.0 pg of lipasc A and 12.0 pg of lipase B and was stopped after 15 min for analysis of the composition of the released fatty acids as described previously [lSl. Results
Production candidum
and purification
of lipases A and B from
G.
CMICC 335426 Production of crude lipase by growth of G. candidum in fermenters and concentration of the culture filtrate on Q Sepharose Fast Flow were achieved as described previously [15]. However, a new method for scaling up the subsequent purification was developed using a Pharmacia LKB FPLC system. The eluate from Q Sepharose was diluted 50: 50 (v/v) with water in order to reduce its conductivity and its pH adjusted to 4.3 with HCI. It was then loaded on to a 1.6 x 2Ocm column (Pharmacia code C16/20) packed with S Sepharose Fast Flow equilibrated in 20 mM pipera-
Large
scale purification
candidum
CMICC
qf’ lipases A and
B from
G.
335426
Fig. 1 shows the elution profile of lipase A and B from the FPLC S Sepharose column. Lipasc A was eluted between pH 4.2 and 4.6, lipasc B between 4.8 and 5.1. Both were obtained at a high level of purity. SDS-PAGE (Fig. 1) and IEF (data not shown) strongly indicate that both lipase fractions contain single protein species. The molecular masses of lipases A and B are 62 and 58 kDa, respectively [15]. Although lipase B (pl 4.5) has a more acidic pI than lipase A (PI 4.71) [lS], it was cluted at a higher pH than lipase A. This
61 TABLE
I
TABLE
Hydrolysis of acylglycerols and methyloleate by Geotrichum candidum CMICC 335426 lipases A and B
II
Composition of the fatty acids released by Geotrichum candidum CMICC 335426 iipases A and B from the pure synthetic triacylglycerols POP, PPO, SOS
Substrate
Monoolein Diolein Triolein POP PPO SOS Methyl oleate
Rate of hydrolysis
(pmol/min
lipase A
lipase B
1010 II20 1270 670 1000 hl0 720
1430 1540 2380 1890 IS30 1400 640
per mg) Triacylglycerol
POP PPO SOS
saturated
oleate
(P, S) 25.4 42.1 61.4
I.1 0.5 0.1
98.9 99.5 99.9
and triolein. For both enzymes, triolein was hydrolysed the fastest, followed by diolein and monoolein. Lipase B had a higher specific activity than lipase A for all these substrates. Lipases A and B readily hydrolysed POP, PPO and SOS, the rates of reaction varying between 50 and 80% those observed with triolein. Lipase A had a small preference for PPO and lipase B for POP (Table I>. The composition of the fatty acids released from POP, PPO, SOS are shown in Table II. Lipase A hydrolysed large amounts of palmitic, stearic and oleic acids from these substrates. In sharp contrast, the fatty acids released by lipase B contained more than 98.9% oleic acid. This confirms the results obtained previously with HPLC purified enzymes [El. Lipase B shows no significant regioselectivity with triacylglycerols, oleate being released from POP and PPO at a similar rate.
were efficient catalysts for the hydrolysis of mono-, di-
-_*
74.6 57.9 38.h
fatty acids ($6) lipase B
oleate
(P, S)
Specificity of lipases A and B from G. candidum CMtCC 335426 Hydrolysis of acylglycerols by G. candidum lipases A and B (Table I and II) G. candidum lipases A and B
Activity
of released
lipase A saturated
shows that the external charge and not the net charge of the molecule is important in its binding and separation on the ion exchange matrix. In a typical purification run, 100 000 units of crude lipase preparation were loaded on 10 ml of S Sepharose and eluted with a 1 1 gradient to give 80000 units of lipase B (spec. act. 2500 units/mg) and 20000 units of lipase A (spec. act. 1000 units/mgl. Lipases A and B were obtained at the following concentrations A: 100 kg/ml; B: 280 Fug/ml and used as such for specificity studies,
1
Composition
*,.*.rn 2
3
4
5
L
6
7
89
Wml)
600
6.4
600
6.2
&so
6
4.8
0.05
4.2 60 80 100 Elution volume (ml) Fig. 1. Purification by FPLC of G. candidum CMICC 335426 lipases A and B. 5 ml fractions were collected. n , lipase activity units/ml; +, pH; dotted line, absorbance at 280 nm. Top, SDS-PAGE of crude enzyme (L) and fractions 1 to 9.
in olive oil
Methyl lipasc A Methanol Prop;u-
I-01
Propan-I-01 Butan-I-ol Oct;ln-
I -(II
Tctradwan-l-01
reaction
lipase A Satur;cted
rate liparc
6:O
0.0
0.0
IS?
x:0
7.4
23
IO:0
0.0
0.X 0.0
I71
I?:0
0.0
0.0
40
I‘$:0
0.0
71
Ih:O
4.4 04.2
IX:0
13.1)
I.1
I I.6
3.7
I -(II
-16
Octadecan-
I-01
-1-l
Octadeccn-
I -(II
42 0
ci.dO IO:
Monounsaturated
13.6
l3:l 16: I IX:
Hydrolysis of methyl oleate by G. candidum iipases A und B (Tuble I). The results show that the specific activity of lipase A on methyl oleate was slightly higher than that of lipase B. Lipases A and B hydrolysed triolein 2 and 4-times faster than methyl ester of oleic acid, respectively. Hydrolysis of other esters of oleic ucid by G. candidum lipuses A nnd B (Table 111). Esters of oleic acid and linear saturated alcohols from 1 to 18 carbons were all digested by G. cundidum lipases A and B at significant rates. However, some differences between the two lipascs were observed. Lipase A had a prefercncc for esters of short chain alcohols (Cl, C2, C3, C4) with a peak of activity for propyl oleate. For esters of medium to long chain alcohols CC8 to Cl8), reduced reaction rates were observed. In contrast, the activity of lipase B was high for esters short chain alcohols ((3, C4) as well as those of medium (C8, Cl2) and long (ClO, ClX) chain alcohols with a peak for octyl oleate. The two enzymes hydrolysed an ester of the secondary alcohol isopropyl oleate. However. the rates of hydrolysis of this alcohol were about X-times lower than those of the corresponding primary alcohol. Interestingly enough, lipase B hydrolysed oleyl oleate twice as fast as stearyl oleatc showing that cis-J%unsaturation of the alcohol was preferred to saturation. Finally. lipase B hydrolyscd cholestcryl oleate at a significant rate whereas lipase A did not catalyse the hydrolysis of this substrate. Hydrolysis of methyl esters by G. cundidum lipuses A und H (Tables IV und V). Methyl esters of suturated fatty ucids (Table IV). G. cundidum lipases A and B were tested on methyl esters of the 6: 0, 8: 0, 10: 0, 12: 0. 14:0, 16: 0 saturated fatty acids. Both enzymes were inactive on 6 : 0. 10: 0 and 12: 0 methyl esters but gave a slow rate with the 8 : 0 ester. Appreciable hydrolysis of trioctanoin by lipases A and B was also measured
0.0
FFA
I
0.5
XX
I
12.OH
B
FFA
II7
Hcxadccan-
C’holerterol
Relative
lipase B
IO0
Ethanol
cstcr
35.11
IO0 IX: I (ricinoleic
Polyunsaturated Linoleic
acid)
5.0
FFA
acid
ru-Linolenic
acid
y-Linolenic
acid
Arachidonic
100
17.1
I94
01.0
I66
703
acid
3.‘)
I.1
0.0
Il.0
(unpublished data). Higher activities with lipase A were observed when longer chain fatty acid esters were used (14: 0, 16:O and 18:O). Methyl palmitate was significantly the best substrate for lipase A, with a rate of 64% that of methyl olcate. In contrast, lipase B had no activity on 14: 0 and 16 : 0 methyl esters and very small activity on 18 : 0 methyl ester. - Methyl ester,s of cis-J 9 monounsutur~lter1~~~ firtty wids (Table IV). The effect of fatty acid chain length on
Isomer
Relative lipw
A
reaction
rate lipase B
2.0
0.0
16.4
0.3
0.X
0.0
5.7
0.0
25.0
()..I
7.3
100
0.x I00
lb.7
I.?
13.1
0.0
14.2
0.0
Il.7
0.0
IX.3
I .8
14.6
0.7
23.9
0.7
63 lipase activity was further investigated with methyl esters of the cis-A9 monounsaturated fatty acids 10: 1, 14: 1, 16: 1 and 18: 1. For both enzymes the rates of hydrolysis of these cis-A9 unsaturated substrates were higher than their saturated equivalents. Moreover, longer chain fatty acid esters were hydrolysed more rapidly than shorter chain fatty acid esters. The 16: 1 methyl ester was the best substrate for lipase A, followed by 18:1, 14: 1 and 10: 1. However, lipase B hydrolysed I8 : 1 methyl ester the fastest; it also hydrolysed 16: 1 significantly but 14: 1 and 10: 1 were digested very slowly. Finally, the presence of a hydroxyl group at the 12-position of the carbon chain of 18: 1 considerably reduced the activity of both enzymes. - Methyl ester;v of polyunsaturated fatty acids (Table IV). Lipase B hydrolysed methyl linoleate and and cu-linolenate at rates slightly lower than that of methyl oleate. But the activity of lipase A on these substrates was almost twice the activity measured on methyl oleate. Methyl y-linolenatc was a poor substrate for both enzymes. This was due to the presence of a double bond in the h-position. Methyl arachidonate (C20), which has double bonds in the 5, 8, 11 and 14-positions was not hydrolysed at all. - Methyl esters of isomers of oleic acid (Table V). G. candidum lipases A and B hydrolysed methyl oleate at a maximum rate. Although lipase A showed a preference for the cis-A9 isomer, it did not have the strong specificity of lipasc B which was inactive on the 42, 44, 46, All, Al2 and Al3 isomers. Very small activities were measured for the hydrolysis by lipase B of the remaining isomers, but these very slow rates could be due to the presence in the samples of small amounts of impurities (especially the cis-A9 isomer). The methyl ester of trans-A9 18: 1 was hydrolysed by lipase A but very slow reaction was observed with lipase B.
and B. c&-A9 polyunsaturated esters are also good substrates for the enzymes. However, the presence of a double bond between the 9-position and the carboxylic extremity of the acid reduces considerably the activity of the lipases. Moreover, monocne positional isomers of oleic acid are all digested by lipase A at significant rates but lipase B does not hydrolyse any of these isomers at a rate higher than 1.8% that of methyl oleate. These results confirm the discovery of a G. candidum lipase which is very highly specific for esters of fatty acids containing a cis-double bond in the 9-position. Previous studies have reported the specificity of crude G. candidum lipase preparations using esters of saturated fatty acids [4,6,8], cis-A9 monounsaturated fatty acids [4,6,8,19,201, polyunsaturated fatty acids [19,20], isomers of oleic acid [3,20]. The significant variations of specificities reported in these publications can be explained by the presence in the crude preparations of different levels of specific (lipase B) and nonspecific (lipase A) lipases. We are currently investigating this using specific polyclonal antibodies. Although they are structurally related 1151 and appear to possess similar partial amino acid sequences (unpublished data), lipases A and B from G. candidum CMICC 335426 show striking differences in their specificities. Lipases from other strains of G. candidum have been crystallised [5,20-221 and three dimensional structures will hopefully become available soon. Crystallisation of lipases A and B are under way. It will be of considerable interest to determine the structural differences between lipases A and B which result in their differing specificity. Further work with lipases A and B will generate information about the influence of protein structure on lipase specificity and will eventually enable one to use protein engineering for the modification of lipase specificity.
Discussion Acknowledgements G. candidum lipases A and B are effective catalysts for the hydrolysis of acylglycerols and alkyl esters of oleic acid. Lipase A has a preference for oleic acid esters of short primary alcohols, whereas lipase B is very active on esters of short, medium and long chain alcohols as well as cholesterol. But for both enzymes, an ester of a primary alcohol is preferred to the secondary alcohol. Lipase A hydrolyses methyl esters of long saturated fatty acids at significant rates, with an unexpected peak for methyl palmitate. Interestingly, 8 : 0 is the only saturated fatty acid hydrolysed by lipase B from a methyl ester or a triacylglycerol to an appreciable extent. This suggests that caprylic acid may have a configuration particularly adapted to the active site of the enzyme. Among the methyl esters of cis-A9 monounsaturated fatty acids, methyl oleate and methyl palmitoleate are hydrolysed the fastest by lipases A
We are grateful to Mr G. Mycock for growing G. candidum in fermenters and wish to thank Prof. F.D. Gunstone (University of St. Andrews) and Mr. A. Visser (Unilever Research Laboratory, Vlaardingen) for kindly providing us with isomers of oleic acid. We also thank Prof. A.R. Slabas, Mr. C. Sidebottom and Dr. S. Doyle for their advice on protein purification procedures. References I Alford, J.A. and Pierce, D.A. (1961) J. Food Sci., 26, 518-524. 2 Alford, J.A., Pierce, D.A. and Suggs, P.G. (1964) J. Lip. Res. 5, 390-394. 3 Jensen, G. (1973) Lipids 9 (3), 149-157. 4 Franzke, C., Kroll, J. and Petzold, R. (1973) Die Nahrung 17 (2), 171-174.
64 5 Tsujisaka, Y., Iwai, M. and Tominaga, Y. (1973) Agr. Bio. Chem. 37 fh), 1457- 1764. 6 Aneja, R. and Hollis, W.H. (1983) Abstracts No. 120 of papers of the Am. Chem. Sot.. Vol. 1Xh. 7 Jacobsen. T., Olsen, J. and Allerman, K. (1900) Biotechnol. Lett. 12 (2), 121-126. X Aneja. R. (1987) J. Am. Oil Chem. Sot. 64 (5). 645. 9 Baillargeon, M.W., Bistline, R.G. and Sonnet, P.E. (1989) Appl. Microhiol. Biotechnol. 30, Y2-96. 10 Shimada. Y., Sugihara, A.. Tominiga. Y. and lizumi. T. (1989) J. Biochem. 106, 383-388. I I Shimada. Y.. Sugihara. A.. lizumi, T. and Tominaga. Y. (1990) J. B&hem. 107, 703-707. 12 Sugihara, A., Shimada, Y. and Tominaga, Y. (1990) J. Biochem. 107. 4X-430. 13 Veeraragavan, K.. Colpitts. T. and Gibbs, B.F. (19YO) Biochem. Biophys. Acta 1044, 26-33. 14 Jacobsen, T.. Olsen, J., Allerman, K., Poulsen. O.M. and Hau, J. (1989) Enzyme Microbial. Technol. 11, 90-95. IS Sidebottom. C., Charton, E., Dunn. P.P.J., Mycock, G.. Davies. C.. Sutton, J.. Macrae, A.. Slabas. A.R. (1991) Eur. J. Biochem.. in press. 16 Charton, E., Davies, C.. Sidebottom. C., Sutton, J.. Dunn, P.P.J..
17 18 IY 20 21
22
23
Slabas, A.R. and Macrae, A.R. (1991) in Lipases: Structure. Mechanism and Genetic Engineering (Alberghina, L., Verger, R. and Schmid, R.D.. eds.), pp. 335-338, GBF Monographs 16. VCH Weinheim lY91. Laemmli, U. flY70) Nature 227, h80-685. Bradford. M.M. (1976) Anal. Biochem. 72. 248-254. Marks, T.A., Quinn. J.G.. Sampugna. J. and Jensen, R.G. (196X) Lipids. 3 (2). 142-146. Jensen, G. and Pitas, E. (1976) Lipids I. 141-146. Menge. U., Hecht. H-J., Schomburg. D., Schmid, R.D., Hedrich. H.C. and Spener. F. (IYY I) in Lipases: Structure. Mechanism and Genetic Engineering (Alherghina, L., Verger. R. and Schmid. R.D., eds.). pp. 59-62. GBF Monographs 16, V
