Biochimica et Biophysica Acta, 1127 (1992) 19i - 198 © 1992 Elsevier Science Publishers B.V. All rights reserved 0005-2760/92/$05.00
191
BBALIP 53972
Use of specific polyclonal antibodies to detect heterogeneous lipases from Geotrichum candidum Emmanuelle Charton, Christine Davies and Aiasdair R. Macrae Unilever Research, Colwonh Laboratory, Cohvorth House, Shambrook, Bedford (UK) (Received 9 February 1992)
Key words: Lipase; Polyclonal antibody; G. candidum
Geotrichum candidum CMICC 335426 was previously shown to produce two lipases termed lipase A and lipase B, lipase B being highly specific for hydrolysis of esters of cis-A9 fatty acids. We now describe the isolation of polyelonal antibodies specific for lipase A and lipase B. These antibodies were used in Western blotting techniques to detect the appearance of the lipases during the course of the fermentation of G. candiaum CMICC 335426. Lipascs A and B were found to be produced simultaneously in the extracellular medium ~t the start of the growth phase. The two lipases were always present at similar levels in the medium. The specific antibodies were then used to detect the presence of A- and B-like lipases in crude lipase samples from other strains of G. candidum. The lipa~,es were found at different levels in all these samples, and the specificities of the crude lipases varied significantly from one strain to another. Differences in specificity could therefore be explained by different levels of specific (B-type) and non-specific (A-type) lipases in the m_edium. This was verified by purifying A- and B-type lipases from the G. candidum strain ATCC 34614.
Introduction Geotrichum candidum is a ~videspread mold in nature and many strains of the fungus have been isolated from various origins. G. candidum produces extracellular lipases whose specificities vary significantly from one strain to another, the crude enzymes showing more or less pronounced specificity far hydrolysis of esters of cis-Ag-unsaturated fatty acids [1-7]. Recently, distinct lipases have been purified from several crude G. candidum lipase preparations. The ATCC 34614 strain of the mold was reported to contain two distinct lipase genes, and two extracellular lipases have been purified from fermentation m e d i a [8,9]. The specit'icities of the enzymes were examined using triolein and triacylglycerols of saturated fatty acids of various chain lengths, and the lipases were shown to have similar specific!ties ,rl0~. More r~.~atly, four iipa~es were isolated from the same strain and the specificities of two of them were examined using the same substrates. One of them expressed an unusual specificity, releasing fatty acid preferentially from the 2-position of triolein [11]. Several iipases were purified from a commercial source of G. candidum lipase, and one lipase had a
Correspondence to: A.R. Macrae, Unilever Research, Colworth Lab..or~t~o.,,, Colworth House, Shambrook, Bedford MK44 1LQ, UK.
preference for the oleate ester of methylumbeiliferol over the palmitate ester [12]. We reported on the purification of two lipases from the CMICC 3354";6 strain using HPLC [13] or FPLC [14], this last method being particularly suitable for upscaling. The specificities of the two lipases, termed lipase A and lipase B, were tested on a wide range ot natural and synthetic triacylglycerols and fatty acid alkyl esters. Lipase B was very highly specific for hydrolysis of esters of cis-A9fatty 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 proposed that differet~ce~ in specificities previously observed for crude lipases from various strains of G. candidum could be explained by the presence of different levels of specific (B-type) and non.specific (A-type) lipases. The aim of the wor!presented here was to i.~olate polyclonal antibodic,~ ~zcific for each iipase , n d to use them in Western blotting techniques to detect A- and B-type lipases in crude G. candidum lipase preparations. Maierials and Methods Microorganisms and commercial lipase samples. G. candidum NRRL Y-552 attd NRRL Y-553 were obtained fi'om the Agricultural Research Service Culture Collection (Peoria, IL, USA). G.. candidum A~i'("
192 34614 was purchased from American Type Culture Collection (Rockville, MD, USA). G. candidum CMICC 335426 is deposited in Commonwealth Mycological Institute Culture Collection (Kew, Surrey, UK). Commercial samples of G. candidum lipases were obtained from Biocatalysts (Pontypridd, UK) and Amano (Nagoya, Japan). Growth of microorganisms. Spores of G. candidum were inoculated into a medium containing, per litre of distilled water: 1.5 g KH2PO 4, 1.0 g NH4CI, 1.2 g MgSO4.7H2G, 20.0 g Difco Beta Lab Yeast Extract, 25.0 g olive oil, 17 mg ZnSO4, 17 mg MnSO4 and 17 mg FeSO4. The strains NRRL Y-552 and NRRL Y-553 were grown in shake flasks at 30'C in 5011 ml of medium for 24 h. The strains ATCC 34614 and CMICC 335426 were grown in fermenters, as described before [13]. The growth of G. candidum CMICC 335426 was followed during its whole cycle. Samples w~re taken from 8 h after inoculation. Aliquots of fermentation broth were assayed for lipase activity. Culture filtrates were assayed for lipase activity and subsequently analysed by SDS-PAGE. Mycelia were imaged at a magnification of 100(), using interference contrast optics on a Reicl~ert Polyvar Research microscope and photographed using 4 × 5 inch polaroid film. Lipase.catalysed hydrolysis reactions. Lipase titre was determined using emulsions of 5% (w/v) olive oil in 2% gum arabic solution [13]. Digestions of palm oleine and methyl esters were performed as described previously [ 13].
Purification of extracellular lipases frmn Geotrichum candidum CMICC 335426 and A TCC 34614. The purification of lipases from G. candidum CMICC 335426 and ATCC 34614 was performed using two steps of ion exchange chromatography, as described previously [14]. Polyacrylamide gel electrophoresis: (7. candidum lipaso samples were analysed by SDS-PAGE [15] using b.5 or 10% slab gels in a Bio-Rad mini Protean II apparatus at 180 V. Western blotting. Proteiras were separated by SDSPAGE and electroph,Jretically transferred to nitrocellulose (Bio-Rad) ia ~. Bio-Rad blotting cell at a constant voltage of 50 V for 3 h. After blocking in appropriate blocking solution, the nitrocellulose filters were incubated successively with rabbit anti-lipase pri. mary antibody and horseradish pcroxidase-labelled donkey anti-rabbit IgF(ab)~ fragment (Amersham) diluted 1:300 in Tris-buffc.red saline solution. Frocedures for membrane blocking and chemiluminescent detection were those given by Amersham. Anti4ipase antisera. Crude antisera against G. cand/d~t lipascs A a~d B were raised in rabbits [13]. i',,ffficrdou of specific anti.lipase antibodies. Specific antibodies were isolated by passing crude antisera through columns of pure immobilised lipases. For that purpose, pure lipases A and B were coupled to CNBr-
activated Sepharose 4B (Pharmacia) according to manufacturer's instructions to give final loadings of 3 mg lipase per ml of swollen gel. l~ltification of specific anti-lipase A antibody. Crude antiserum against lipase A was diluted 1:10 in 3 ml coupling buffer (0.5 M NaCl, 0.1 M NaHCO a, pH 8.3) and added to 1 ml of immobilised lipase B matrix equilibrated in the same buffer• Affinity binding was allowed to proceed for 2-3 h at room temperature with continuous end-over-end shaking. The mixture was then poured into a 10 ml column. The supernatant was collected in one fraction, diluted 1:25 in Tris buffered saline solution to give a specific anti-lipase A antibody used for Western blotting analysis. Purification of specific anti-lipase B antibody. Crude antiserum against lipase B was diluted 1:10 in coupling buffer, added to 1 ml of immobilised lipase A matrix, and processed for affinity binding as described above. The supernatant consisted of partially purified anti-lipase B antibody, and an additional purification step was required to obtain specific antibody. For this purpose, the solution was diluted 1 : 12 in Tris-buffered saline solution containing 1% hemoglobin, and incubated for at least 18 h at 4°C with 4 bands (4 cm × 0.4 cm) of nitrocellulose carrying 50 Izg denatured lipase A per cm 2 of membrane [16]. The resulting solution was the anti-lipase B specific antibody and was used as such for Western blot analysis. Results and Discussion
Specific anti.lipase antibodies The specificity of the crude antisera and purified antibodies was tested on pure denatured G. candidum iipases A and B using Western blot analysis (Fig. 1). The crude antiserum directed against lipase A showed
a B
b A
.
c
A
.
.
.
.
B
A
d B
A
B
A
B
.
Fig. I. Western blotting for the evaluation of tltc specifcity of crude antisera and specific antibodies. (a). crude anti-lipase A antiserum (!:250); (b), crude anti-lipase B antiserum (1:250); (c), CNBr Sepharose-purified specific anti-lipase A antibody (I :20); (d), CNBr Sepharose-partially purified anti-lipase B antibody (I :20); (e), CNBr Sepharose and nitrocellulose-purified specific anti-lipase B antibody (I :20). SDS-PAGE was done on 6.5,% polyaci~lamide running gels. A, 200 ng lipase A; B, 200 ng lipase B.
!93 slight specificity for lipase A (Fig. la), whereas the crude anti.B antiserum recognised both lipases A and B (Fig. lb). The finding that none of the antisera was specific showed that lipases A and lipase B are closely related and have at least one common antigenic determinant. The aim of the CNBr Sephar, Jse affinity chro-
mat0graphy step was to remove common antibodies from each antiserum. Western blots of the antibod~i¢s purified with the CNBr Sepharose columns are represented on Fig. lc and ld. Purifying crude anti-A antiserum with immobilised lipase B gave an antibody cross-reacting specifi-
x 1000 a
b
c
tl
Fig. 2. Phase-contrastmicrographsof morphological stages o!: ,'3eotriehum candidum CMICC 335426. (a), resting spores (inoculum):(b), growing myceiium(8 i~after inoculation);(.c),mature'mycelium(15 h after inoculation); (d), arthrospores(20 h after inoculation).
194 cally with lipase A (Fig. lc). It can be concluded that the antibodies specific for the common epitopes were removed in one step, leaving antibodies recognising epitopes specific of lipase A. Purifying crude anti-B antiserum with immobilised lipase A gave an almost specific antibody (Fig. ld). All the cross-reacting antibodies were not removed in this step. Specific anti-B a n t i ~ was finally obtained by incubating the partially purified antibody solution with nitrocellulose strips carrying SDS-denatureo iipase A (Fig. le). To summarize, antibody specific to lipase B had to not only be in contact with native lipase A, but also with denatured lipase A to make it specific for lipase B, A possible explanation for this is a common epitope which presents itself on the surface of native lipase B, but only presents itself in iipase A when denatured. The specific anti-A and anti-B antibodies were not only specific egainst denatured lipases, but also against native lipases, This was verified using ELISA assays against the pure native lipases (data not shown), and confirms that the antigenic determinants for which each population of antibodies is specific are present on the surface of the native iipases.
Lipase production by Geotrichum candidum CbtlCC 335426 in laboratory scale fermenters Growth of G. candidum and lipase production. Photomicrographs of the morphological states of G. candidum are given in Fig. 2. The inoculum (Fig. 2a) consists of spores with thick walls. 8 h after inoculation, th~ spores have germinated, the hyphae are short dnd ~tart branching (Fig. 2b). This is the beginning of
0
8
11
14
17
('l
(-) Activity (Ulml) 200 I
Biornass (g/I) 140
/
150
120
/
100 80
100 / J
60
i I I !
SO
tl
'
40
',j 20
0
5
10
15
20 Hours after inoculation
d!-
0 35
Fig, 3. Biomass and lipase productions during the fermentation of Geotrichum cm,.didum CMICC 335426. (*), Biomass; ( g ) , lipase activity ( - - , total activity; . . . . ,acdvity in culture filtrate).
the growth phase (Fig. 3). After 15 h, long mature hyphae start to fragment themselves (Fig. 2c) and after 20 h, at the end of the growth phase, the whole culture sporulates (Fig. 2d). Lipase activity (2 U/ml) is detected in the whole culture 8 h after inoculation, at the
20
37
A
B
Fig. 4. SDS-PAGE (6.5% polyaa~jlamide) of culture filtrate fractions (20 pl) from the fermentation of Geotrichum candidum CMICC 335426. Proteins were detected by silver staining. The figures above each lane correspond to the time (in h) following inoculation. A, 500 ng lipase A; B, 500 ng lipase B.
195 start of the growth (Fig. 3). It gradually increases to reach the maximum value of 120 U / m l (or more) at sporulation and usually stays constant for at least 20 h. Lipase activity partitions between the biomass and the culture filtrate. Early on in the growth phase most of the activity is associated with the biomass, whereas at the end of the growth phase the activity is found predominantly in the culture filtrate. During sporulation, a proportion of the soluble lipase activity appears to reassociate with the biomass, before being released back into the culture filtrate on prolonged incubation of the arthrospores. The reasons for the fluctuations during fermentation of the relative proportions of bound and soluble lipase are not understood.
Protein composition of the extracellular medium during the fermentation of G. candidum. SDS-PAGE of supernatant samples show that very few proteins are present in the extracellular medium (Fig. 4). Two major components have been identified as lipases A and B by Western blotting techniques (Fig. 5). Lipases A and B are always present at approximately equivalent levels in the supernatant and appear at the same time in the culture supernatant, 8 to 9 h after inoculation. Interestingly, a protein cross-reacting with the specific anti-B antibody, and having slightly higher molecular weight than lipase A, seems to be produced by G. candidum (Fig. 5b). This protein is present at the same level as lipase B until 11 h, but becomes a minor component at the peak of lipase production. It is not visible on silver-stained polyacrylamide gel. The protein could be lipase B with a higher degree of glycosylation, or a third lipase produced by G. candidum. No work has been carried out yet to verify whether any of these assumptions is correct. Lastly, lipases A and B are almost free from any other proteins at the final stage of lipase production (20 h after inoculation). It is only after 37 h that p~otein contaminants appear in the supernatant. This phenomenon is illustrated on Fig. 4, and is probably due to the release of proteinases in the extracellular medium. It may also be due to the accumulation of cell debris after senescence of the mycelia. Specificity studies and Western blotting analysis of crude Geotrichum candidum lipases from various origins The objective of this section was to compare the properties of G. candidum CMICC lipases A and B with those of other strains of G. candidum. The lipases were from the G. candidum strains ATCC 34614, NRRL Y-552, NRLL Y-553 and from commercial sources (Amano and Biocatalysts). Specificity studies. Crude G. candidum lipases were tested on emulsions of palm oleine and of methyl esters of 16:0, cis.A9-18: 1, cis.All-18:1 and transA9-18:1 fatty acids. Table I shows the composition of the fatty acids released from palm oleine by the crude lipases. Table II shows the realtive rates of hydrolysis
A
9
13
16
19
B
9
13
16
19
i/~~: .....
~
~
i/iil//! iiii ~,i'~ i~
Fig. 5. Western blotting of culture filtrate fractions (20 #1) from the fermentation of Geotrichmn candidum CMICC 335426 using specific anti-lipase A (a) and anti-lipase B (b) antibodies. SDS-PAGE was performed on 6.5% polyac~lamide gels. The figures above each lane correspond to the time (in h) following inoculation.
of the methyl esters by the same lipases. It appears from these results that the specificities of the various crude G. candidum lipases are different. The composition of the fatty acids released from palm oleine by the various lipases did not differ much in terms of stearic acid ( < 1%) or linoleic acid content (10.7% to 15.7%). However, the levels of palmitic and oleic acids analysed in the hydrolysates differed markedly from one lipase to another. The fatty acids released from palm oleine by Amano, NRRL and CMICC lipases contained relatively low levels of palmitic acid (10=20%), while in contrast, the fatty acids released by Biocatalysts and ATCC lipases contained 35-40% palmitic acid. Differences in relative rates of hydrolysis of the
196 TABLE !
Composin'on (%) of the fatty acids released from palm oleine by t,arious crime preparations of Geotrichum candidum lipases
a 1
Palm oleine was hydrolysed for 15 rain at 40°C and pH 8, as a 1% oibin-v,'ater emulsiol~ ~i 20 ml 2% gum arabic, containing 0.5% CaCI2. Reactions were eatalysed with 30 units of crude lipase solutions. Analysis of fatty acids was done by GLC after extraction, preparative T I C and methylation. I: Amano; II: Biocatalysts; III: ATCC 34614; IV: NRRL Y-552; V: NRRL Y-553; VI: CMICC 335426~ Fatty acid
Lipase i
!1
!I!
IV
V
VI
16:0 18:0 18:1 18:2
21,0 0,7 6%6 10.7
34,7 0,5 52.8 12,0
41 1,0 44.8 13.2
7,8 0~0 77,3 14,9
18.9 0,0 69.5 ! 1,6
20,5 0.2 66,6 12,7
2
3
~ i!:
4
5
6
i • :•i:i•¸
TABLE !!
Relati~,e rates of hydrolysis of fatty acid methyl esters by t,arious crude preparations of Geotrichum candidum lipases The substrates were hydrolysed at 30°C and pH 8, as 0.5% oil-in-water emulsions ia 20 ml 2% gum arabic containing 0.5% CaCI 2. Reactions were catalysed with 5 units of crude lipase solutions, i: Amano; !i: Biocatalysts; llh ATCC 34614: IV: NRRL Y-552; V: NRRL Y-553; Vh CMICC 335426.
Methyl ester 16: 0 cb-.49-18: ! cis.~! 1-18: I trans-~18: I
1
2
3
4
5
Lipase !
ll
Ill
IV
V
VI
45,7 100 10,6 20,2
66.7 100 15,9 30,2
42,9 100 21,0 15,2
54,4 1(10 22,2 41,1
34,4 100 26,1 34,8
26,8 100 8,5 15.9
fatty acid methyl esters by the various crude lipases were also observed, For example, the rates of hydrolysis of methylpalmitate by the Biocatalysts and ATCC lipases were high (66,7% ant~ 42,9% of the rates obtained with methyloleate, respectively), while hydrolysis of methylpalmitate by CMICC lipase was comparatively low, The variations observed from one strain to another suggested that these preparations may contain different levels of specific (B-type) and non-specific (Aotyp¢) lipases, Specific antibodies were used to verify this, Western blotting. Coomassie-stained SDS-PAGE gels of lipases from various strains of (3, candidum are shown on Fig, 6a, Corresponding Western blots using anti.A and anti.B specific antibodies are presented on
Fig, 6. Cmss-reactivities between anti-A or anti~B specific antibodies and crude tipases from various strains of Geotrichum candidm~n,(a), Coomassk-stained SDS gel; (b), Anti-lipas~ A specific antibody; (c), Anti-lipase B specific antibody, SDS-PAGE was performed on 6.5% IxJlyacwlamide gels, ], Amano; 2, BiocataJysts; 3, N R R L Y-552; 4, NRRL Y-5~3; 5, AT~.~ 34614; 6, CMICC 335426,
C 1
4
6
197 TABLE III Relative rates of hydrolysis of methyl esters from pure lipases from Geotrichum candidum A TCC 34614 and CMICC 335426 Experimental conditions are those described in Table II. Methyl ester
Lipase CMICC
16:0 cis-A9-18:1 trans-A9-18:1 cis-All-18:1
ATCC
A-type
B-type
A-type
B-type
64.2 100 23.9 13.1
0 100 0.7 0
67.0 100 37.6 32.1
0 100 2.4 0
Fig. 6b and Fig. 6c, respectively. A protein cross-reacting with specific anti-A antibody was the major component of all G. candidum lipases preparations. A lesser component cross-reacting with specific anti-B antibody was also present in these preparations. The (7. candidum CMICC was the strain which produced the highest levels of lipase B-type component.
Specificity of purified lipases from Geotrichum candidum ATCC 34614 In an earlier paper we reported the purification of two lipases fiom (7. candidum ATCC 34614 [13]. The.se enzymes, named lipases I and II, have similar specificities and are beleived to be variants of the same protein, possibly differing only in glycosylation [17]. We have now isolated two lipases frorn the ATCC strain by a modified purification procedure, and examined their specificities using fatty acid methyl esters as substrates (Table III). One lipase, which is identical to lipase I, cross-reacted with anti-lipase A antibody and hydrolysed methylpalmitate, methylelaidate and methylvaccenate at significant rates. The specificity of this lipase was similar to that of CMICC lipase A. The second lipase, which is distinct from both lipases I and If, cross-reacted with anti-lipase B. This enzyme was inactive on methylpalmitate and methylvaccenate, and expressed very low activity with methylelaidate, showing a specificity profile similar to that of CMICC lipase B. It was concluded that both A-type and B-type lipases are produced by G. candidum ATCC 34614 and, probably, other strains of G. candidum. Conclusion G. candidum CMICC produces two lipases, lipase A and lipase B, having markedly different substrate specificities. Although G. candidum CMICC 335426 lipases A and B are very similar proteins, it is possible to obtain specific anti-lipase A and anti-lipase B antibodies. These antibodies revealed the presence of different levels of lipase A- and B-like proteins in crude lipase solutions from other strains of G. candidum. We
examined the specificities of these preparations using methyl esters and palm oleine as substrates. Significant variations were observed from one strain to another, supporting the results from Baillargeon et al. [7] and strongly suggesting that the preparations contained different levels of specific (B-type) and non-specific (Atype) lipases." The two lipases we purified from culture filtrates of G. candidum ATCC 34614 were found to have the specificity profiles of CMICC lipases A and B. These findings contrast with those of Sugihara et ai., who also isolated two lipases from cultures of the ATCC strain and described them as products of two different genes [10]. The specificity profiles of their lipases were similar when examined using triolein and triacylglycerols of various saturated fatty acids, and neither lipase displayed the pronounced specificity of our iipase B. A possible explanation of these conflicting results is that the ATCC strain, in common with other G. candidum strains, produces several lipases from multiple lipase genes, the relative amounts of the various lipases produced being dependent on the culture conditions. Under our conditions, significant amounts of the specific B-type lipase are produced, whereas under the culture conditions used by Sugihara et al. production of this lipase is suppressed.
Acknowledgements We are very grateful to Gary Mycock for growing Geotrichum candidum in fermenters. We also thank Nigel Lindner for taking microphotographs of Geotrichum candidum mycelia.
References ! Jensen, G. (1973) Lipids 9, 149-157. 2 Franzke, C., Kroll, J. and Petzold, R. (1973) Nahrung 17, 171-174. 3 Tsujisaka, Y., lwai, M. and Tominaga, Y. (1973) Agric. Biol. Chem. 37, 1457-1464. 4 Aneja, R. and Hollis, W.H. (1993) Papers of the American Chem. Soc. 186, Abstract 129. 5 Jacobsen, T., Olsen, J. and Allerman, K. (1990) Biotechnol, Lett. 12, 121-126. 6 Aneja, R. (1987) J. Am. Oil Chem. Soc. 64, 645. 7 Baillargeon, M.W,, Bistline, R.G. and Sonnet, P.E. (1989) Appl. MicrobioL Biotechnol. 30, 92-96. 8 Shimada, Y., Sugihara, A., Tominaga, Y. and lizumi, T. (1989) J. Biochem. 106, 383-388. 9 Shimada, Y., Sugihara, A., lizumi. "'.. attd Tominaga, Y. (1990) J. Biochem. 107, 703-707. 10 Sugihara, A., Shimada, Y. and Tominaga, Y. (1990) J. Biochem. 107, 426-430. 11 Sugihara, A,, Shimada, Y. and Tominaga, Y. (1991) Appl. Microbiol. Biotechnol. 35, 738-740. 12 Baillargeon, M.W, (1990) Lipids 25, 841-848, 13 Sidebottom, C., Charton, E,, Dunn, P.P.J., Mycock, G., Davies, C., Sutton, J., Macrae, A. and Siabas, A.R. (1991) Eur. J. Biochem, 202, 485-491.
198 14 Charton, E, and Macrae, A.R. (1992) Biochim. Biophys. Acta 1123, 59-64. 15 Laemmli, U.K. (1970) Nature 22% 680-685. 16 Robinson, P.A., Anderton, B.H. and Loviny, T.L.F. (1988) J. lmmunol. Methods 108, 115-122.
17 Charton, E., Davies, C., Sidebottom, C,M., Sutton, J,L,, Dunn, P.PJ., Slabas, A.R. and Macra¢, A.R, (1991) in Lipases: Structure, Mechanism and Genetic Engineering (AIberghina, L. and Schmid, R.D., eds.) pp. 335-338, VCH Press, Weinheim.
191
BBALIP 53972
Use of specific polyclonal antibodies to detect heterogeneous lipases from Geotrichum candidum Emmanuelle Charton, Christine Davies and Aiasdair R. Macrae Unilever Research, Colwonh Laboratory, Cohvorth House, Shambrook, Bedford (UK) (Received 9 February 1992)
Key words: Lipase; Polyclonal antibody; G. candidum
Geotrichum candidum CMICC 335426 was previously shown to produce two lipases termed lipase A and lipase B, lipase B being highly specific for hydrolysis of esters of cis-A9 fatty acids. We now describe the isolation of polyelonal antibodies specific for lipase A and lipase B. These antibodies were used in Western blotting techniques to detect the appearance of the lipases during the course of the fermentation of G. candiaum CMICC 335426. Lipascs A and B were found to be produced simultaneously in the extracellular medium ~t the start of the growth phase. The two lipases were always present at similar levels in the medium. The specific antibodies were then used to detect the presence of A- and B-like lipases in crude lipase samples from other strains of G. candidum. The lipa~,es were found at different levels in all these samples, and the specificities of the crude lipases varied significantly from one strain to another. Differences in specificity could therefore be explained by different levels of specific (B-type) and non-specific (A-type) lipases in the m_edium. This was verified by purifying A- and B-type lipases from the G. candidum strain ATCC 34614.
Introduction Geotrichum candidum is a ~videspread mold in nature and many strains of the fungus have been isolated from various origins. G. candidum produces extracellular lipases whose specificities vary significantly from one strain to another, the crude enzymes showing more or less pronounced specificity far hydrolysis of esters of cis-Ag-unsaturated fatty acids [1-7]. Recently, distinct lipases have been purified from several crude G. candidum lipase preparations. The ATCC 34614 strain of the mold was reported to contain two distinct lipase genes, and two extracellular lipases have been purified from fermentation m e d i a [8,9]. The specit'icities of the enzymes were examined using triolein and triacylglycerols of saturated fatty acids of various chain lengths, and the lipases were shown to have similar specific!ties ,rl0~. More r~.~atly, four iipa~es were isolated from the same strain and the specificities of two of them were examined using the same substrates. One of them expressed an unusual specificity, releasing fatty acid preferentially from the 2-position of triolein [11]. Several iipases were purified from a commercial source of G. candidum lipase, and one lipase had a
Correspondence to: A.R. Macrae, Unilever Research, Colworth Lab..or~t~o.,,, Colworth House, Shambrook, Bedford MK44 1LQ, UK.
preference for the oleate ester of methylumbeiliferol over the palmitate ester [12]. We reported on the purification of two lipases from the CMICC 3354";6 strain using HPLC [13] or FPLC [14], this last method being particularly suitable for upscaling. The specificities of the two lipases, termed lipase A and lipase B, were tested on a wide range ot natural and synthetic triacylglycerols and fatty acid alkyl esters. Lipase B was very highly specific for hydrolysis of esters of cis-A9fatty 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 proposed that differet~ce~ in specificities previously observed for crude lipases from various strains of G. candidum could be explained by the presence of different levels of specific (B-type) and non.specific (A-type) lipases. The aim of the wor!presented here was to i.~olate polyclonal antibodic,~ ~zcific for each iipase , n d to use them in Western blotting techniques to detect A- and B-type lipases in crude G. candidum lipase preparations. Maierials and Methods Microorganisms and commercial lipase samples. G. candidum NRRL Y-552 attd NRRL Y-553 were obtained fi'om the Agricultural Research Service Culture Collection (Peoria, IL, USA). G.. candidum A~i'("
192 34614 was purchased from American Type Culture Collection (Rockville, MD, USA). G. candidum CMICC 335426 is deposited in Commonwealth Mycological Institute Culture Collection (Kew, Surrey, UK). Commercial samples of G. candidum lipases were obtained from Biocatalysts (Pontypridd, UK) and Amano (Nagoya, Japan). Growth of microorganisms. Spores of G. candidum were inoculated into a medium containing, per litre of distilled water: 1.5 g KH2PO 4, 1.0 g NH4CI, 1.2 g MgSO4.7H2G, 20.0 g Difco Beta Lab Yeast Extract, 25.0 g olive oil, 17 mg ZnSO4, 17 mg MnSO4 and 17 mg FeSO4. The strains NRRL Y-552 and NRRL Y-553 were grown in shake flasks at 30'C in 5011 ml of medium for 24 h. The strains ATCC 34614 and CMICC 335426 were grown in fermenters, as described before [13]. The growth of G. candidum CMICC 335426 was followed during its whole cycle. Samples w~re taken from 8 h after inoculation. Aliquots of fermentation broth were assayed for lipase activity. Culture filtrates were assayed for lipase activity and subsequently analysed by SDS-PAGE. Mycelia were imaged at a magnification of 100(), using interference contrast optics on a Reicl~ert Polyvar Research microscope and photographed using 4 × 5 inch polaroid film. Lipase.catalysed hydrolysis reactions. Lipase titre was determined using emulsions of 5% (w/v) olive oil in 2% gum arabic solution [13]. Digestions of palm oleine and methyl esters were performed as described previously [ 13].
Purification of extracellular lipases frmn Geotrichum candidum CMICC 335426 and A TCC 34614. The purification of lipases from G. candidum CMICC 335426 and ATCC 34614 was performed using two steps of ion exchange chromatography, as described previously [14]. Polyacrylamide gel electrophoresis: (7. candidum lipaso samples were analysed by SDS-PAGE [15] using b.5 or 10% slab gels in a Bio-Rad mini Protean II apparatus at 180 V. Western blotting. Proteiras were separated by SDSPAGE and electroph,Jretically transferred to nitrocellulose (Bio-Rad) ia ~. Bio-Rad blotting cell at a constant voltage of 50 V for 3 h. After blocking in appropriate blocking solution, the nitrocellulose filters were incubated successively with rabbit anti-lipase pri. mary antibody and horseradish pcroxidase-labelled donkey anti-rabbit IgF(ab)~ fragment (Amersham) diluted 1:300 in Tris-buffc.red saline solution. Frocedures for membrane blocking and chemiluminescent detection were those given by Amersham. Anti4ipase antisera. Crude antisera against G. cand/d~t lipascs A a~d B were raised in rabbits [13]. i',,ffficrdou of specific anti.lipase antibodies. Specific antibodies were isolated by passing crude antisera through columns of pure immobilised lipases. For that purpose, pure lipases A and B were coupled to CNBr-
activated Sepharose 4B (Pharmacia) according to manufacturer's instructions to give final loadings of 3 mg lipase per ml of swollen gel. l~ltification of specific anti-lipase A antibody. Crude antiserum against lipase A was diluted 1:10 in 3 ml coupling buffer (0.5 M NaCl, 0.1 M NaHCO a, pH 8.3) and added to 1 ml of immobilised lipase B matrix equilibrated in the same buffer• Affinity binding was allowed to proceed for 2-3 h at room temperature with continuous end-over-end shaking. The mixture was then poured into a 10 ml column. The supernatant was collected in one fraction, diluted 1:25 in Tris buffered saline solution to give a specific anti-lipase A antibody used for Western blotting analysis. Purification of specific anti-lipase B antibody. Crude antiserum against lipase B was diluted 1:10 in coupling buffer, added to 1 ml of immobilised lipase A matrix, and processed for affinity binding as described above. The supernatant consisted of partially purified anti-lipase B antibody, and an additional purification step was required to obtain specific antibody. For this purpose, the solution was diluted 1 : 12 in Tris-buffered saline solution containing 1% hemoglobin, and incubated for at least 18 h at 4°C with 4 bands (4 cm × 0.4 cm) of nitrocellulose carrying 50 Izg denatured lipase A per cm 2 of membrane [16]. The resulting solution was the anti-lipase B specific antibody and was used as such for Western blot analysis. Results and Discussion
Specific anti.lipase antibodies The specificity of the crude antisera and purified antibodies was tested on pure denatured G. candidum iipases A and B using Western blot analysis (Fig. 1). The crude antiserum directed against lipase A showed
a B
b A
.
c
A
.
.
.
.
B
A
d B
A
B
A
B
.
Fig. I. Western blotting for the evaluation of tltc specifcity of crude antisera and specific antibodies. (a). crude anti-lipase A antiserum (!:250); (b), crude anti-lipase B antiserum (1:250); (c), CNBr Sepharose-purified specific anti-lipase A antibody (I :20); (d), CNBr Sepharose-partially purified anti-lipase B antibody (I :20); (e), CNBr Sepharose and nitrocellulose-purified specific anti-lipase B antibody (I :20). SDS-PAGE was done on 6.5,% polyaci~lamide running gels. A, 200 ng lipase A; B, 200 ng lipase B.
!93 slight specificity for lipase A (Fig. la), whereas the crude anti.B antiserum recognised both lipases A and B (Fig. lb). The finding that none of the antisera was specific showed that lipases A and lipase B are closely related and have at least one common antigenic determinant. The aim of the CNBr Sephar, Jse affinity chro-
mat0graphy step was to remove common antibodies from each antiserum. Western blots of the antibod~i¢s purified with the CNBr Sepharose columns are represented on Fig. lc and ld. Purifying crude anti-A antiserum with immobilised lipase B gave an antibody cross-reacting specifi-
x 1000 a
b
c
tl
Fig. 2. Phase-contrastmicrographsof morphological stages o!: ,'3eotriehum candidum CMICC 335426. (a), resting spores (inoculum):(b), growing myceiium(8 i~after inoculation);(.c),mature'mycelium(15 h after inoculation); (d), arthrospores(20 h after inoculation).
194 cally with lipase A (Fig. lc). It can be concluded that the antibodies specific for the common epitopes were removed in one step, leaving antibodies recognising epitopes specific of lipase A. Purifying crude anti-B antiserum with immobilised lipase A gave an almost specific antibody (Fig. ld). All the cross-reacting antibodies were not removed in this step. Specific anti-B a n t i ~ was finally obtained by incubating the partially purified antibody solution with nitrocellulose strips carrying SDS-denatureo iipase A (Fig. le). To summarize, antibody specific to lipase B had to not only be in contact with native lipase A, but also with denatured lipase A to make it specific for lipase B, A possible explanation for this is a common epitope which presents itself on the surface of native lipase B, but only presents itself in iipase A when denatured. The specific anti-A and anti-B antibodies were not only specific egainst denatured lipases, but also against native lipases, This was verified using ELISA assays against the pure native lipases (data not shown), and confirms that the antigenic determinants for which each population of antibodies is specific are present on the surface of the native iipases.
Lipase production by Geotrichum candidum CbtlCC 335426 in laboratory scale fermenters Growth of G. candidum and lipase production. Photomicrographs of the morphological states of G. candidum are given in Fig. 2. The inoculum (Fig. 2a) consists of spores with thick walls. 8 h after inoculation, th~ spores have germinated, the hyphae are short dnd ~tart branching (Fig. 2b). This is the beginning of
0
8
11
14
17
('l
(-) Activity (Ulml) 200 I
Biornass (g/I) 140
/
150
120
/
100 80
100 / J
60
i I I !
SO
tl
'
40
',j 20
0
5
10
15
20 Hours after inoculation
d!-
0 35
Fig, 3. Biomass and lipase productions during the fermentation of Geotrichum cm,.didum CMICC 335426. (*), Biomass; ( g ) , lipase activity ( - - , total activity; . . . . ,acdvity in culture filtrate).
the growth phase (Fig. 3). After 15 h, long mature hyphae start to fragment themselves (Fig. 2c) and after 20 h, at the end of the growth phase, the whole culture sporulates (Fig. 2d). Lipase activity (2 U/ml) is detected in the whole culture 8 h after inoculation, at the
20
37
A
B
Fig. 4. SDS-PAGE (6.5% polyaa~jlamide) of culture filtrate fractions (20 pl) from the fermentation of Geotrichum candidum CMICC 335426. Proteins were detected by silver staining. The figures above each lane correspond to the time (in h) following inoculation. A, 500 ng lipase A; B, 500 ng lipase B.
195 start of the growth (Fig. 3). It gradually increases to reach the maximum value of 120 U / m l (or more) at sporulation and usually stays constant for at least 20 h. Lipase activity partitions between the biomass and the culture filtrate. Early on in the growth phase most of the activity is associated with the biomass, whereas at the end of the growth phase the activity is found predominantly in the culture filtrate. During sporulation, a proportion of the soluble lipase activity appears to reassociate with the biomass, before being released back into the culture filtrate on prolonged incubation of the arthrospores. The reasons for the fluctuations during fermentation of the relative proportions of bound and soluble lipase are not understood.
Protein composition of the extracellular medium during the fermentation of G. candidum. SDS-PAGE of supernatant samples show that very few proteins are present in the extracellular medium (Fig. 4). Two major components have been identified as lipases A and B by Western blotting techniques (Fig. 5). Lipases A and B are always present at approximately equivalent levels in the supernatant and appear at the same time in the culture supernatant, 8 to 9 h after inoculation. Interestingly, a protein cross-reacting with the specific anti-B antibody, and having slightly higher molecular weight than lipase A, seems to be produced by G. candidum (Fig. 5b). This protein is present at the same level as lipase B until 11 h, but becomes a minor component at the peak of lipase production. It is not visible on silver-stained polyacrylamide gel. The protein could be lipase B with a higher degree of glycosylation, or a third lipase produced by G. candidum. No work has been carried out yet to verify whether any of these assumptions is correct. Lastly, lipases A and B are almost free from any other proteins at the final stage of lipase production (20 h after inoculation). It is only after 37 h that p~otein contaminants appear in the supernatant. This phenomenon is illustrated on Fig. 4, and is probably due to the release of proteinases in the extracellular medium. It may also be due to the accumulation of cell debris after senescence of the mycelia. Specificity studies and Western blotting analysis of crude Geotrichum candidum lipases from various origins The objective of this section was to compare the properties of G. candidum CMICC lipases A and B with those of other strains of G. candidum. The lipases were from the G. candidum strains ATCC 34614, NRRL Y-552, NRLL Y-553 and from commercial sources (Amano and Biocatalysts). Specificity studies. Crude G. candidum lipases were tested on emulsions of palm oleine and of methyl esters of 16:0, cis.A9-18: 1, cis.All-18:1 and transA9-18:1 fatty acids. Table I shows the composition of the fatty acids released from palm oleine by the crude lipases. Table II shows the realtive rates of hydrolysis
A
9
13
16
19
B
9
13
16
19
i/~~: .....
~
~
i/iil//! iiii ~,i'~ i~
Fig. 5. Western blotting of culture filtrate fractions (20 #1) from the fermentation of Geotrichmn candidum CMICC 335426 using specific anti-lipase A (a) and anti-lipase B (b) antibodies. SDS-PAGE was performed on 6.5% polyac~lamide gels. The figures above each lane correspond to the time (in h) following inoculation.
of the methyl esters by the same lipases. It appears from these results that the specificities of the various crude G. candidum lipases are different. The composition of the fatty acids released from palm oleine by the various lipases did not differ much in terms of stearic acid ( < 1%) or linoleic acid content (10.7% to 15.7%). However, the levels of palmitic and oleic acids analysed in the hydrolysates differed markedly from one lipase to another. The fatty acids released from palm oleine by Amano, NRRL and CMICC lipases contained relatively low levels of palmitic acid (10=20%), while in contrast, the fatty acids released by Biocatalysts and ATCC lipases contained 35-40% palmitic acid. Differences in relative rates of hydrolysis of the
196 TABLE !
Composin'on (%) of the fatty acids released from palm oleine by t,arious crime preparations of Geotrichum candidum lipases
a 1
Palm oleine was hydrolysed for 15 rain at 40°C and pH 8, as a 1% oibin-v,'ater emulsiol~ ~i 20 ml 2% gum arabic, containing 0.5% CaCI2. Reactions were eatalysed with 30 units of crude lipase solutions. Analysis of fatty acids was done by GLC after extraction, preparative T I C and methylation. I: Amano; II: Biocatalysts; III: ATCC 34614; IV: NRRL Y-552; V: NRRL Y-553; VI: CMICC 335426~ Fatty acid
Lipase i
!1
!I!
IV
V
VI
16:0 18:0 18:1 18:2
21,0 0,7 6%6 10.7
34,7 0,5 52.8 12,0
41 1,0 44.8 13.2
7,8 0~0 77,3 14,9
18.9 0,0 69.5 ! 1,6
20,5 0.2 66,6 12,7
2
3
~ i!:
4
5
6
i • :•i:i•¸
TABLE !!
Relati~,e rates of hydrolysis of fatty acid methyl esters by t,arious crude preparations of Geotrichum candidum lipases The substrates were hydrolysed at 30°C and pH 8, as 0.5% oil-in-water emulsions ia 20 ml 2% gum arabic containing 0.5% CaCI 2. Reactions were catalysed with 5 units of crude lipase solutions, i: Amano; !i: Biocatalysts; llh ATCC 34614: IV: NRRL Y-552; V: NRRL Y-553; Vh CMICC 335426.
Methyl ester 16: 0 cb-.49-18: ! cis.~! 1-18: I trans-~18: I
1
2
3
4
5
Lipase !
ll
Ill
IV
V
VI
45,7 100 10,6 20,2
66.7 100 15,9 30,2
42,9 100 21,0 15,2
54,4 1(10 22,2 41,1
34,4 100 26,1 34,8
26,8 100 8,5 15.9
fatty acid methyl esters by the various crude lipases were also observed, For example, the rates of hydrolysis of methylpalmitate by the Biocatalysts and ATCC lipases were high (66,7% ant~ 42,9% of the rates obtained with methyloleate, respectively), while hydrolysis of methylpalmitate by CMICC lipase was comparatively low, The variations observed from one strain to another suggested that these preparations may contain different levels of specific (B-type) and non-specific (Aotyp¢) lipases, Specific antibodies were used to verify this, Western blotting. Coomassie-stained SDS-PAGE gels of lipases from various strains of (3, candidum are shown on Fig, 6a, Corresponding Western blots using anti.A and anti.B specific antibodies are presented on
Fig, 6. Cmss-reactivities between anti-A or anti~B specific antibodies and crude tipases from various strains of Geotrichum candidm~n,(a), Coomassk-stained SDS gel; (b), Anti-lipas~ A specific antibody; (c), Anti-lipase B specific antibody, SDS-PAGE was performed on 6.5% IxJlyacwlamide gels, ], Amano; 2, BiocataJysts; 3, N R R L Y-552; 4, NRRL Y-5~3; 5, AT~.~ 34614; 6, CMICC 335426,
C 1
4
6
197 TABLE III Relative rates of hydrolysis of methyl esters from pure lipases from Geotrichum candidum A TCC 34614 and CMICC 335426 Experimental conditions are those described in Table II. Methyl ester
Lipase CMICC
16:0 cis-A9-18:1 trans-A9-18:1 cis-All-18:1
ATCC
A-type
B-type
A-type
B-type
64.2 100 23.9 13.1
0 100 0.7 0
67.0 100 37.6 32.1
0 100 2.4 0
Fig. 6b and Fig. 6c, respectively. A protein cross-reacting with specific anti-A antibody was the major component of all G. candidum lipases preparations. A lesser component cross-reacting with specific anti-B antibody was also present in these preparations. The (7. candidum CMICC was the strain which produced the highest levels of lipase B-type component.
Specificity of purified lipases from Geotrichum candidum ATCC 34614 In an earlier paper we reported the purification of two lipases fiom (7. candidum ATCC 34614 [13]. The.se enzymes, named lipases I and II, have similar specificities and are beleived to be variants of the same protein, possibly differing only in glycosylation [17]. We have now isolated two lipases frorn the ATCC strain by a modified purification procedure, and examined their specificities using fatty acid methyl esters as substrates (Table III). One lipase, which is identical to lipase I, cross-reacted with anti-lipase A antibody and hydrolysed methylpalmitate, methylelaidate and methylvaccenate at significant rates. The specificity of this lipase was similar to that of CMICC lipase A. The second lipase, which is distinct from both lipases I and If, cross-reacted with anti-lipase B. This enzyme was inactive on methylpalmitate and methylvaccenate, and expressed very low activity with methylelaidate, showing a specificity profile similar to that of CMICC lipase B. It was concluded that both A-type and B-type lipases are produced by G. candidum ATCC 34614 and, probably, other strains of G. candidum. Conclusion G. candidum CMICC produces two lipases, lipase A and lipase B, having markedly different substrate specificities. Although G. candidum CMICC 335426 lipases A and B are very similar proteins, it is possible to obtain specific anti-lipase A and anti-lipase B antibodies. These antibodies revealed the presence of different levels of lipase A- and B-like proteins in crude lipase solutions from other strains of G. candidum. We
examined the specificities of these preparations using methyl esters and palm oleine as substrates. Significant variations were observed from one strain to another, supporting the results from Baillargeon et al. [7] and strongly suggesting that the preparations contained different levels of specific (B-type) and non-specific (Atype) lipases." The two lipases we purified from culture filtrates of G. candidum ATCC 34614 were found to have the specificity profiles of CMICC lipases A and B. These findings contrast with those of Sugihara et ai., who also isolated two lipases from cultures of the ATCC strain and described them as products of two different genes [10]. The specificity profiles of their lipases were similar when examined using triolein and triacylglycerols of various saturated fatty acids, and neither lipase displayed the pronounced specificity of our iipase B. A possible explanation of these conflicting results is that the ATCC strain, in common with other G. candidum strains, produces several lipases from multiple lipase genes, the relative amounts of the various lipases produced being dependent on the culture conditions. Under our conditions, significant amounts of the specific B-type lipase are produced, whereas under the culture conditions used by Sugihara et al. production of this lipase is suppressed.
Acknowledgements We are very grateful to Gary Mycock for growing Geotrichum candidum in fermenters. We also thank Nigel Lindner for taking microphotographs of Geotrichum candidum mycelia.
References ! Jensen, G. (1973) Lipids 9, 149-157. 2 Franzke, C., Kroll, J. and Petzold, R. (1973) Nahrung 17, 171-174. 3 Tsujisaka, Y., lwai, M. and Tominaga, Y. (1973) Agric. Biol. Chem. 37, 1457-1464. 4 Aneja, R. and Hollis, W.H. (1993) Papers of the American Chem. Soc. 186, Abstract 129. 5 Jacobsen, T., Olsen, J. and Allerman, K. (1990) Biotechnol, Lett. 12, 121-126. 6 Aneja, R. (1987) J. Am. Oil Chem. Soc. 64, 645. 7 Baillargeon, M.W,, Bistline, R.G. and Sonnet, P.E. (1989) Appl. MicrobioL Biotechnol. 30, 92-96. 8 Shimada, Y., Sugihara, A., Tominaga, Y. and lizumi, T. (1989) J. Biochem. 106, 383-388. 9 Shimada, Y., Sugihara, A., lizumi. "'.. attd Tominaga, Y. (1990) J. Biochem. 107, 703-707. 10 Sugihara, A., Shimada, Y. and Tominaga, Y. (1990) J. Biochem. 107, 426-430. 11 Sugihara, A,, Shimada, Y. and Tominaga, Y. (1991) Appl. Microbiol. Biotechnol. 35, 738-740. 12 Baillargeon, M.W, (1990) Lipids 25, 841-848, 13 Sidebottom, C., Charton, E,, Dunn, P.P.J., Mycock, G., Davies, C., Sutton, J., Macrae, A. and Siabas, A.R. (1991) Eur. J. Biochem, 202, 485-491.
198 14 Charton, E, and Macrae, A.R. (1992) Biochim. Biophys. Acta 1123, 59-64. 15 Laemmli, U.K. (1970) Nature 22% 680-685. 16 Robinson, P.A., Anderton, B.H. and Loviny, T.L.F. (1988) J. lmmunol. Methods 108, 115-122.
17 Charton, E., Davies, C., Sidebottom, C,M., Sutton, J,L,, Dunn, P.PJ., Slabas, A.R. and Macra¢, A.R, (1991) in Lipases: Structure, Mechanism and Genetic Engineering (AIberghina, L. and Schmid, R.D., eds.) pp. 335-338, VCH Press, Weinheim.
