Thermal stabilization of immobilized lipase B from Candida antarctica on different supports: Effect of water activity on enzymatic activity in organic media

Thermal stabilization of immobilizedlipase B from Candida antarcticaondifferent supports: Effect of wateractivity on enzymatic activity inorganic mediaMiguel Arroyo, Jose Marı a Sanchez-Montero, and Jose´ Vicente Sinisterra

Department of Organic & Pharmaceutical Chemistry, Faculty of Pharmacy, UniversidadComplutense, Madrid, Spain

Covalent immobilization ofC. antarcticalipase B (CALB) on sepharose, alumina, and silica was undertaken. Thethermal stability of these covalently immobilized catalysts were studied and compared to adsorbed derivativesfrom Novo Nordisk at 50°C under wet conditions. Native enzyme and Novozym 435 follow a deactivation modelE3 E1 whereas covalently immobilized derivatives and SP435A follow the model E3 E13 E2. This differentbehavior is related to the nature of the support and the immobilization methodology. Water absorption isothermsof dry solid biocatalysts in air or isooctane were used to predict the optimum preequilibriumaw value to obtainthe highest rate in the esterification of(r,s)-ibuprofen. © 1998 Elsevier Science Inc.

Keywords: Lipase; immobilization; stability; ibuprofen; water activity

Introduction

Lipase B fromCandida antarctica(CALB)1 is an interest-ing lipase with potential application in a number of indus-trial processes such as the synthesis of triglycerides,2

esterification of terpenic alcohols,3 etc. Adsorbed CALB ondifferent supports has also proven to be very regioselectivein the esterification of sugars,4 nucleosides,5 and steroids,6

and very enantioselective in the resolution of secondaryalcohols via hydrolysis7 or esterification in organic sol-vents.8 One of these derivatives, called SP435A, has beenemployed successfully in the preparation of pureS(1)-2-arylpropionic acids with antiinflammatory effect.9

In the current paper, several immobilization methods andsupports have been tested for the covalent bonding of purelipase B fromC. antarctica. We have also compared the

thermal stability of our covalent immobilized derivativeswith those obtained by Novo Nordisk by absorption of thesame lipase on different polymers. All the biocatalysts weretested in the hydrolysis of triacetin due to the low activity ofthis lipase in the hydrolysis of triglycerides with long-chainfatty acids.10 Afterwards and at different initial wateractivity (aw), the best immobilized derivatives catalyzed thestereospecific esterification of racemic ibuprofen inisooctane in order to find a method to predict the bestpre-equilibrium aw of the whole system to perform thereaction.

Materials and methods

Materials

Native lipase B fromC. antarctica(SP525) and the same lipaseimmobilized on Lewatit OC 1600 (SP435A) and Lewatit E(Novozym 435) were kindly supplied by Novo Nordisk Bioindus-trias (Madrid, Spain). Tresylated sepharose 4B and epoxy-acti-vated sepharose 6B were purchased from Pharmacia (Uppsala,Sweden). Silica (Kiesegel 60, size 0.015–0.040 nm, average pore

Address reprint requests to Dr. J. M. Sanchez-Montero, UniversidadComplutense de Madrid, Facultad de Farmacia, Dept. Organic & Pharma-ceut. Chemist., 28040 Madrid, SpainReceived 22 January 1998; revised 7 April 1998; accepted 28 April 1998

Enzyme and Microbial Technology 24:3–12, 1999© 1998 Elsevier Science Inc. All rights reserved. 0141-0229/99/$–see front matter655 Avenue of the Americas, New York, NY 10010 PII S0141-0229(98)00067-2

diameter5 95 Å, surface area5 239 m2 g21), alumina (Alumi-num 60, size 0.063–0.200 nm, average pore diameter5 60 Å,surface area5 166 m2 g21) and isooctane (analytical grade) wereobtained from Merck (Darmstadt, Germany). 2,4,6-trichloro-1,3,5-triazine was supplied by Aldrich (Steinhiem, Germany). Puretriacetin was purchased from Sigma Chemical Co. (St. Louis,MO). (r,s)-2-((4-isobutyl)-phenyl)-propionic acid (ibuprofen) wasdonated by Boots Pharmaceuticals (Nottingham, UK).

Protein determination

The protein content of SP525 (0.1 mg protein mg21 derivative)was determined by the Biuret method.11 In a typical experiment,0.1 ml of a 125 mg ml21 SP525 solution was mixed with theBiuret reagent and the protein concentration was determinedspectrophotometrically at a wavelength of 545 nm using a calibra-tion curve of seroalbumin.

Covalent immobilization on inorganic supports

The activation of silica and alumina was performed according tothe 2,4,6-trichloro-1,3,5-triazine (TCT) method previously de-scribed in the immobilization ofCandida rugosalipase.12 Theimmobilization of lipase B fromC. antarcticawas performed at4°C for 6 h with low stirring. Each support (1 g) was added todifferent concentrations of enzyme in 10 ml of standard buffer (0.1m Tris-HCl buffer pH 8.0). After the desired contact time, theinsoluble enzyme derivative was filtered and washed with standardbuffer. The percentage of immobilized enzyme was determined bythe difference between the initial activity of the native enzymesolution and the activity of the filtrate after the immobilizationprocess.

Covalent immobilization on organic supports

Epoxy-activated sepharose 6B (2 g) was mixed with 10 ml of theenzymatic solution of native CALB in 0.1m Tris-HCl pH 8.0buffer. The mixture was stirred at 4°C for 6 h and then filtered andwashed with 33 10 ml of buffer solution. The immobilizedbiocatalyst on tresyl-activated sepharose 4B was prepared by anexperimental methodology equivalent to that described above.

Electron microphotographs

Electron microphotographs were taken with a Zeiss DSM 940scanning electron microscope.

Hydrolysis assay

As standard assay, the hydrolysis of pure triacetin was performedin 1 mm Tris-HCl buffer pH 7.0 at 37°C. The acetic acid releasedwas continuously titrated to constant pH with the help of a pHstat(Crison model microTT 2022). Several NaOH solutions (1–10mm) were used as titrating agents. The catalytic efficiency ofimmobilized derivatives was determined as the ratio between theenzymatic activity of 3mg native lipase and the activity of theamount of immobilized derivative which contained 3mg enzyme,taking into account the percentage of immobilized enzyme(Table 1).

Thermal stability assays

The thermal stability assays were performed with the same amountof lipase: native or immobilized. The storage stability of native andinsolubilized enzymes was performed at 50°C in 0.1m Tris-HClbuffer pH 8.0. These experimental conditions were selected as theextreme conditions. After incubation for different times, theremaining activity was measured in the hydrolysis of triacetin asdescribed above.

Measurement of water absorption isotherms andpre-equilibration of the reaction media

Pure lipase or immobilized derivative (100–200 mg) were previ-ously predried with P2O5. Theaw values of solid preparations weremeasured at 25°C using a hygrometric sensor (Rotronic Hygro-scopic D.T.) precalibrated with two saturated salt solutions ataw 5 0.11 and 0.98. The isotherms in isooctane were measuredwith the same amount of solid plus 1 ml of dried solvent. Enzymepreparations (native or immobilized lipases) and reaction mediawere equilibrated at 25°C in separate containers with the help ofthe hygrometric sensor. Equilibration at differentaw was achievedby adding the required amount of water to reach the desiredaw.

General procedure for esterification

The reaction media was composed of isooctane (5 ml), racemicibuprofen (66 mm), and l-propanol (66 mm). The reaction wasstarted by mixing different amounts of pre-equilibrated immobi-lized lipase to the pre-equilibrated solution to the desiredaw of thesystem. The reactions were performed at a fixed temperature bystirring in 25 ml flasks for a specific time. The solution (100ml)

Table 1 Immobilization of lipase B from C. antarctica

Derivative Activated supportmg lipase added

g21 support

Immobilizatione

(mg lipaseg21 support) C.E.f (%) S.E.A.g(%)

CALB-ES-1 Epoxyactivated sepharosea 1,500 44 660 49 32CALB-ES-2 Epoxyactivated sepharosea 5,000 18.4 920 49 45CALB-TS-1 Tresylated speharoseb 1,500 42.5 640 72 47CALB-TS-2 Tresylated sepharoseb 5,000 27.4 1,370 72 100CALB-S-1 Silica-TCTc 2,000 90.2 1,810 38 62CALB-S-2 Silica-TCTc 6,700 48.8 3,250 31 86CALB-AL-1 Alumina-TCTd 6,700 9.3 620 27 14

a19–40 meq oxyrane groups ml21 wet gelbGrade of activation not supplied by Pharmacia LKBc0.24 g TCT g21 silicad0.19 g TCT g21 aluminaeReferred to the amount of CALB added in the immobilization processfCatalytic efficiency in the hydrolysis of triacetin (see MATERIALS AND METHODS)gSpecific enzymatic activity expressed as mmol acetic acid released min21 g21 dry derivative

Papers

4 Enzyme Microb. Technol., 1999, vol. 24, January/February 1

was added to 1.4 ml of isooctane and the ester conversion wasanalyzed by gas chromatography.9

Gas chromatographic analysis

Gas chromatography was performed in a Shimadzu GC-14A gaschromatograph equipped with a FID detector, a split injector (1:2),and a SPB-1 sulfur column (15 m3 0.32 mm; Supelco, Bellafonte,PA). Injector temperature was 200°C and the detector temperature350°C; the carrier gas was nitrogen. Conditions for quantitativeanalysis were: column temperature of 180°C and N2 stream of 12ml min21. An external standard method was used to quantify theremnant acid and the formed ester.

Results and discussion

Covalent immobilization of pure lipase Bfrom C. antarctica

We tested different activated organic and inorganic supportsfor the covalent immobilization of pure lipase B fromC.antarctica. As organic supports, we employed epoxy andtresylated sepharoses which differed in the length of thehydrocarbon chain attached to the matrix. As inorganicsupports, we employed silica and alumina activated by theTCT methodology. The results of covalent immobilizationof CALB on these activated supports are shown inTable 1.Comparing the results obtained with epoxy-activated(CALB-ES-1) and tresylated sepharose (CALB-TS-1), wemay conclude that the length of the spacer arm does notaffect the amount of immobilized lipase (640–660mgCALB g21 support, 42–44% immobilization) in contrastwith the results of Shawet al.13 who observed higherimmobilization values forC. rugosa lipase (CRL) withspacer arms longer than two carbons. Our results could berelated to the smaller size of CALB (Mw 5 33.3 kDa)1

compared to CRL (Mw 5 60.0 kDa);14 however, there is asignificant effect on the catalytic efficiency of the immobi-lized lipase. We expected higher catalytic efficiency inCALB-ES-1 (49%) where the enzyme should have largerconformational freedom due to the longer spacer arm of theepoxy-activated sepharose (12 carbons). Instead CALB-TS-1, where the enzyme is closely attached to the matrix,showed higher catalytic efficiency (72%). We tried toexplain this behavior taking into account the differencebetween the chemical bonds on different sites of theenzyme. Epoxy-activated sepharose is able to link proteinsthrough the hydroxyl, amino, and thiol groups of theiramino acids whereas tresylated sepharose only linksthrough the amino and thiol groups, then an interactionbetween the oxirane group of the sepharose and the hy-droxyl group ofserine 105of the catalytic triad could beinvolved in some extent during the immobilization. Thisfact could have diminished the catalytic efficiency ofCALB-ES derivatives with respect to the CALB-TS.

If the amount of added lipase is increased in theimmobilization on both activated sepharoses, themg en-zyme bonded to the support is increased as well, but thepercentage of immobilization decreases (Table 1) and thecatalytic efficiency remains the same (72% in CALB-TSderivatives and 49% in CALB-ES). Similar results have

also been observed in the immobilization ofC. rugosalipase on agarose-activated by tosylation methodology.15

The enzyme loading in the derivatives on silica isincreased with the amount of lipase added in the immobi-lization process (CALB-S-1 and CALB-S-2); nevertheless,lipase molecules form multilayers on the support surface(Figure 1) and, as a consequence, the catalytic efficiencydiminishes (31–38%,Table 1). Similar results were reportedin the immobilization of lipase fromC. rugosa12 on TCT-activated silica.

Finally, we can assume that alumina is not an adequatesupport for the immobilization of CALB according to thepoor catalytic efficiency value of CALB-AL-1 (Table 1).The hindering effect of the pore diameter of the support (95Å in silica, and 60 Å in alumina) can explain how the lipaseloading is higher in CALB-S-2 (3,250mg g21 support) thanin CALB-AL-1 (620 mg g21 support). This indicates thatthe enzyme (whose size is 303 40 3 50 Å)1 is located inthe external surface of alumina; therefore, a poor enzymeloading is observed. Lipase multilayers are also formed andthe catalytic efficiency of both derivatives is similar (31%and 27%, respectively,Table 1).

Thermal stability

The covalently immobilized derivatives can be stored at 4°Cfor at least two months without appreciable loss of catalyticactivity. The storage stability of native and covalentlyimmobilized lipase B fromC. antarcticawas studied underwet conditions at 50°C. The thermal deactivation curveshave been explained following the deactivation modelproposed by Henley and Sadana.16 This model involvesenzymatic states (E, E1, and E2) where k1 and k2 arefirst-order deactivation rate coefficients anda1 anda2 arethe ratios of specific activities E1/E and E2/E, respectively[Eq. (1)]. The experimental plots of residual activity versus

Figure 1 Electron microphotograph of CALB-S-2 (original mag-nitude 2,000X). Enzyme aggregates of lipase B from C. antarc-tica can be appreciated on the surface of silica

Immobilized lipase B from C. antarctica: M. Arroyo et al.

Enzyme Microb. Technol., 1999, vol. 24, January/February 1 5

storage time were adjusted to exponential decays [Eq. (2)](single or double, with or without offset) with the help of theSimfit program developed by Dr. Bardsley.17 From the dataof the adjusted equations and using Eq. (3), we couldcalculate all the parameters (k1, k2, a1, a2), the half-life ofthe biocatalyst, and the stabilization factor (F, considered asthe ratio between soluble and derivatives half-lives)(Table 2).

E™™™3k1

o1

E™™™3k2

a2

E2 (1)

A 5 A1e2k1t 1 A2e

2k2t 1 A3 (2)

A 5 F1001a1k1

k2 2 k12

a2k2

k2 2 k1G e2k1t

1F a2k1

k2 2 k12

a1k1

k2 2 k1G e2k2t 1 a2 (3)

At 50°C, native CALB deactivation followed a singleexponential decay which belongs to the classical first-orderdeactivation pattern (Table 2) in which a1 5 0, a2 5 0, andk2 5 0. Like native lipases A and B fromC. rugosa,18

CALB loses its activity in only one step, but just a littleslower. The difference between the half-life values ofCALB (0.5 h) andC. rugosapure lipases (approximately0.25 h) may be related to the number of SOS bonds in theirstructure which maintains the protein integrity. CALB hasthree disulfide bridges in the protein structure:1 cys22-cys64,cys216-cys258, and cys293-cys311 whereas both isoforms ofpure CRL have only two:14 cys60-cys97 and cys268-cys277.

Figure 2 Thermal stability of covalent immobilizedderivatives of lipase B from C. antarctica at 50°C

Table 2 Thermal deactivation of native and immobilized lipase B from C. antarctica at 50°C

Derivative A1a k1 (h21)a A2

a k2 (h21)a a1b a2

b (A3) t1/2 (h)c Fd

Native CALB 98.9 1.30 0 0 0 0 0.5 1CALB-ES-1 43.6 1.80 56.0 0.03 55 0 4 8CALB-ES-2 57.8 0.70 43.3 0.06 40 0 2.5 5CALB-TS-1 52.8 1.01 47.2 0.01 47 0 2.5 5CALB-TS-2 27.6 0.56 73.3 0.08 63 0 4.5 9CALB-S-1 12.5 0.85 87.8 0.02 86 0 24 48CALB-S-2 74.0 0.42 24.6 0.01 23 0 2.5 5SP435Ae 46.7 4.10 0 0 53 52.6 0.5 1Novozym 435f 101 0.05 0 0 0 0 14 28

aParameters (in %) from the fitted exponential decay equationba1 and a2 are the ratio in (in %) of specific activities E1/E and E2/E, respectivelycHalf-lifedStabilization factoreS.E.A. 5 147 mmol acetic acid min21 g21 SP435AfS.E.A. 5 238 mmol acetic acid min21 g21 Novozym 435

Papers


As a consequence, CALB would be more thermostable thanCRL isoforms with respect to the thermal deactivation byconformational change.

The deactivation of covalently bonded CALB followed adouble exponential decay when the derivatives were storedat 50°C (Figure 2). In all cases,k1 . k2, a1 , 100% anda2 5 0 (Table 2). The small values ofk2 mean a goodstabilization of enzymatic stateE1. The remaining activityof the intermediateE1, expressed asa1, is similar in allsepharose derivatives (ES or TS series). There is not astrong relationship between this value and the enzymaticloading. The hydrophilic nature of sepharose protects theenzyme independently from the length of the spacer arm,giving similara1, half-life values and deactivation profiles.In the derivatives prepared with silica, the activity of theintermediate state E1 may be affected by the lipase loading.The intermediate of CALB-S-1 has higher activity (a1 586%) than the intermediate of CALB-S-2 (a1 5 23%) dueto the formation of lipase multilayers in the second deriva-tive as mentioned above. As a consequence of this enzy-matic aggregation, the lipase is weakly linked in the outerlayers and it is quickly deactivated. We can conclude thatcovalent immobilization of pure lipase B fromC. antarcticaproduces an appreciable stabilization of the biocatalyst,changing its deactivation profile. This change from a singleexponential decay (native lipase) to a double one (immobi-lized lipase) was also observed in pure lipases fromC.rugosa.18

Finally, we compared the stability of our biocatalystswith Novozym 435 and SP435A, both prepared by theadsorption methodology using different polymeric resins.These adsorbed derivatives showed a very different deacti-vation pattern (Figure 3) compared to our covalently im-

mobilized derivatives (Figure 2). This fact may be related tothe support structure and the immobilization method. Thesupport of SP435A derivative is a polymeric resin (LewatitOC 1600) whereas Novozym 435 is prepared with amacroporous acrylic resin (Lewatit E). After a quick deac-tivation, SP435A keeps its residual activity (a1 5 53%) fora long time (Figure 3). The deactivation fits model 3

Figure 4 Electron microphotograph of SP435A (original mag-nification 200X). Lipase B from C. antarctica is located inside themicropores of the support (Lewatit OC 1600)

Figure 3 Thermal stability of SP435A and Novozym435 at 50°C



described by Henley and Sadana, with a very high-stabilizedintermediate E1 (k2 5 0). This stabilization could beexplained by the lipase location inside the micropores of thesupport where the enzyme is protected against alterations ofthe microenvironment (Figure 4). On the contrary, CALB iscompletely exposed to the medium in the surface of thesmall beads of Novozym 435 (Figure 5) so the deactivationmodel of the derivative is similar to the native lipase (a1 50, a2 5 0, andk2 5 0); nevertheless, CALB is stabilized

against temperature andk1 diminishes from 1.3 h231 (nativeenzyme) to 0.05 h21 (Novozym 435).

From the data ofTables 1and2, we can conclude thatSP435, Novozym 435, and CALB-S-1 are the most inter-esting biocatalysts due to their high activity and stability.

Water absorption isotherms and enzymatic activity

Water plays an important role in enzyme structure andfunction. If enzymes are used in organic media, little wateris necessary in the solvent to activate the enzyme.19 Whenthe enzyme is immobilized, the affinity of the support forwater may influence its catalytic activity. The amount ofwater absorbed by solid carriers can be predicted from theaquaphilicity of the support.20 This property may affect theenzymatic activity in the synthesis in organic media.21 Theavailability of water to the biocatalyst to maintain itsenzymatic activity varies depending on the water partition-ing among all the components of the system: the organicsolvent, the enzyme, and the solid support. The problem canbe simplified using thermodynamic water activity (aw). Atequilibrium,aw values will be the same for all the compo-nents in the system. In esterification reactions, lipasessometimes show an optimal initial activity at a certaincontent of water in the reaction medium but their behavioris different depending on the source of lipase,22 the support,and the solvent.23 Our aim was to find a general method topredict the optimum pre-equilibrium water activity of thewhole system to achieve the best catalytic activity of animmobilized lipase in organic media.

The adsorption isotherms of native CALB in air orisooctane are shown inFigure 6.The deviation between thewater absorption isotherms of the solid in the solvent and inair produces adivergence point(D.P.). This point shows the

Figure 5 Electron microphotograph of Novozym 435 (originalmagnification 2003). Lipase B from C. antarctica can be appre-ciated on the surface of the bead (Lewatit E)

Figure 6 Adsorption water isotherms of native CALB.Conditions for measurement: 100 mg of CALB (in air); 100mg of CALB included in 1 ml of isooctane (in solvent)

Papers


lowest water activity at which we must pre-equilibrate thereaction mixture to achieve active molecules of lipase in theorganic medium. The continuing similarity of the air andwater isotherm fromaw 5 0 to the divergence point (aw 50.45) suggest that this hydration range corresponds togradual solvation of the remaining polar groups of theenzyme; however, this does not mean that the water

monolayer coverage and the hydrophobic regions of thelipase will remain exposed to the solvent.24 As a conse-quence, pre-equilibration of the reaction mixture ataw ,D.P. gave low esterification rates (Figure 7). Up to aw 50.45, the divergence between the solvent and air isothermsbecomes increasingly large and a gradual increase in ester-ification rate was observed due to a general reorientation of

Figure 8 Adsorption isotherms of CALB-S-1. Condi-tions for measurement: 200 mg of CALB-S-1 (in air); 200mg of CALB included in 1 ml of isooctane (in solvent)

Figure 7 Esterification of (R,S)-ibuprofen catalyzedby native CALB in isooctane pre-equilibrated at differ-ent aw values. Conditions: 66 mM acid, 66 mM l-propanol in 5 ml of isooctane; 37°C; 20 mg of SP525ml21; 300 rpm



the polar side chains to the most favorable conformation forcatalytic function. Effectively, pure lipase displayed highactivity at aw values. D.P. If the initial aw of the systemis higher than 0.8, a monolayer of water is placed around theenzyme and the esterification rate is far higher (Figure 7).Similar results were observed for pure lipases fromC.rugosa25 in the same reaction. Because water is formed inthe esterification, the water activity is not controlledthroughout the reaction. After 20% conversion, the progress

curve of the reaction pre-equilibrated at an initialaw 5 0.35changes its shape and there is a sudden increase in theesterification rate and yield. This “activation” effect isexplained by the water production during the reaction,26

which increases the water activity of the system to anawvalue. D.P. which in turn increases the reaction rate.

The immobilized derivative CALB-S-1 adsorbs a loweramount of water g21 sample than pure lipase as could beexpected from the hydrophobic nature of silica (Figures 6and8). The continuing similarity of the CALB-S-1 absorp-tion isotherms (in air or isooctane) up to very high values ofaw (Figure 8) is related to the low affinity of silica for waterwhose hydrophobicity is not strongly reduced by the pres-ence of isooctane. In fact, we could check that the behaviorof silica in air or isooctane was very similar (data notshown). Evidently, the lipase immobilized on silica in-cluded in isooctane is slightly more hydrophobic than thederivative itself and, as a consequence, the adsorptionisotherm with the solvent was lower than the isotherm in air.Like native lipase, after pre-equilibration of the reactionmixture ataw values higher than the divergence point (aw 50.6), we obtained active lipase molecules to start the esterifi-cation reaction. The highest esterification rate and yield wereachieved with the system pre-equilibrated ataw 5 0.85 and 0.9(Table 3). These values are very close to theP point which isconsidered as the cross point of the extension of the secondslope of the absorption isotherm in isooctane and the x-axis.Native lipase also displayed its best catatalyic activity at itsPpoint (Figures 6and7). Up to aw . P, the esterification ratedecreases which may be a result of water acting as a substratein hydrolysis of the acylintermediate.27

The absorption water isotherms of lipase immobilized ontresylated sepharose (CALB-TS-1) are shown inFigure 9.In this case, lyophilized CALB-TS-1 was used to make the

Table 3 Effect of aw on the activity of immobilized lipase Bfrom C. antarctica

Derivative

Wateractivity

(aw)

Initial esterification rate(mmol ibuprofen propyl ester

formed h21)

CALB-S-1a 0.10 9.20.40 14.60.60 13.70.75 16.30.85 151.20.90 181.51.00 99.0

CALB-TS-1b 0.25 3.90.45 9.80.80 11.00.90 89.41.00 17.9

P435Ac 0.65 1,5620.85 2,2001.00 3,806

Conditions: 66 mM ibuprofen, 66 mM 1-propanol in 5 mlisooctane; 37°C, 300 rpma200 mg of CALB-S-1 ml21 isooctaneb30 g of CALB-TS-1 ml21 isooctanec50 mg of SP435A ml21 isooctane

Figure 9 Adsorption isotherms of CALB-TS-1. Condi-tions for measurement: 100 mg of CALB-TS-1 (in air);100 mg of CALB-TS-1 included in 1 ml of isooctane (insolvent)

Papers


isotherm. Obviously, the affinity of sepharose for water isvery high compared to silica (Figures 9and8, respectively).The absorption isotherm of CALB-TS-1 included inisooctane was lower than in air but no divergence point wasobserved; thus, the shape of the water isotherm depends onthe nature of the support of the biocatalyst. The highestesterification rate was achieved at anaw 5 0.9 (Table 3).When higher initial systemaw values were used (aw 5 1),the matrix was strongly hydrated and so is the microenvi-ronment of the immobilized lipase. As a consequence, theesterification reaction was thermodynamically disfavoredwith respect to the hydrolysis reaction.

Finally, the high hydrophibicity of the polymeric resin(Lewatit OC 1600) explains how SP435A reaches anaw 51 with very few milligrams of water (Figure 10) comparedto the other supported biocatalysts. In addition, the wateradsorption isotherms in air or isooctane are very closetogether so neither the support nor the solvent are supposedto strip water off from the enzyme. Under these conditions,lipase B fromC. antarcticais always hydrated and activewith a very small amount of water. Once again, we achievedthe highest esterification rate pre-equilibrating the system atan aw near theP point (Table 3).

The low esterification rates achieved with CALB-S-1and CALB-TS-1 with respect to native lipase and SP435Ashown inTable 3could be expected from the low specificenzymatic activity of these biocatalysts (Tables 1and2).

Conclusions

Finally, we can conclude that SP435A is the best biocatalystfor many reasons: higher activity than the covalent immo-bilized lipase (in hydrolysis and synthesis), an interestingdeactivation profile with a very stable intermediate state at

high temperature under wet conditions (Figure 3), and avery easy methodology to be prepared. In addition, ourmethodology based on theP point to predict the optimumaw value to achieve the highest esterification rate in organicmedia is valid for different enzymes (crude and pure lipasesA and B from C. rugosa25,28 and lipase B fromC.antarctica) in native or immobilized form (adsorbed orcovalently bonded on different supports).

Acknowledgments

Financial support from the Ministerio de Educacioń yCiencia of Spain (PB93-0469) and from the ComunidadAutonoma de Madrid (AE 00232/94) are appreciated.

List of symbols

aw Water activitya1, a2 Ratios of specific activities E1/E and

E2/E, respectively [Eq. (1)]CALB Lipase B fromC. antarcticaCALB-AL-1 CALB immobilized on TCT-acti-

vated aluminaCALB-ES-1 and 2 CALB immobilized on epoxy-acti-

vated sepharoseCALB-S-1 and 2 CALB immobilized on TCT-acti-

vated silicaCALB-TS-1 and 2 CALB immobilized on tresyl-acti-

vated sepharoseCRL Lipase fromC. rugosaE, E1, E2 Enzymatic states during thermal de-

activation [Eq. (1)]F Ratio between soluble and immobi-

lized lipase half-lives

Figure 10 Adsorption isotherms of SP435A. Conditionsfor measurement: 100 mg of SP435A (in air); 100 mg ofSP435A included in 1 ml of isooctane (in solvent)



k1, k2 First-order deactivation rate coeffi-cients

TCT Trichlorotriazine

References1. Uppenberg, J., Hansen, M. T., Patkar, S., and Jones, T. A. The

sequence, crystal structure determination and refinement of twocrystal forms of lipase B fromCandida antarctica. Structure1994,2, 293–308

2. Akoh, C. C. Lipase catalyzed synthesis of partial glyceride.Bio-technol. Lett. 1993,15, 949–954

3. Claon, P. A. and Akoh, C. C. Effect of reaction parameters onSP435 lipase catalyzed synthesis of citronellyl acetate in organicsolvent.Enzyme Microb. Technol. 1994,16, 835–838

4. Kirk, O., Bjorkling, F., Godtfredsen, S. E., and Larsen, T. O. Fattyacid specificity in lipase-catalyzed synthesis of glucoside esters.Biocatalysis1992,6, 127–134

5. Moris, F. and Gotor, V. A useful and versatile procedure for theacylation of nucleosides through an enzymatic reaction.J. Org.Chem.1993,58, 653–660

6. Bertinotti, A., Carrea, G., Ottolina, G., and Riva, S. Regioselectiveesterification of polyhydroxilated steroids byCandida antarcticalipase B.Tetrahedron1994,50, 13165–13172

7. Hansen, T. V., Waagen, V., Partali, V., Anthosen, H. W., andAnthosen, T. Co-solvent enhancement of enantioselectivity inlipase-catalysed hydrolysis of racemic esters. A process for produc-tion of homochiral C-3 building blocks using lipase B fromCandidaantarctica. Tetrahedr. Assym.1995,6, 499–504

8. Frykman, H., Ohrner, N., Norin, T., and Hult, K.S-ethyl thiooctano-ate as acyl donor in lipase catalysed resolution of secondaryalcohols.Tetrahedr. Lett.1993,34, 1367–1370

9. Arroyo, M. and Sinisterra, J. V. High enantioselective esterificationof 2-arylpropionic acids catalyzed by immobilized lipase B fromC.antarctica: A mechanistic approach.J. Org. Chem.1994, 59,4410–4417

10. Rogalska, K., Cudrey, C., Ferrato, F., and Verger, R. Stereoselectivehydrolysis of triglycerides by animal and microbial lipases.Chiral-ity 1993,5, 24–30

11. Robson, R. M., Goll, D. E., and Temple, M. J. Determination ofproteins in “Tris” buffer by the biuret reaction.Anal. Biochem.1968,24, 339–341

12. Moreno, J. M. and Sinisterra, J. V. Immobilization of lipase fromCandida cylindraceaon inorganic supports.J. Mol. Catal.1994,93,357–369

13. Shaw, J. F., Chang, R. C., Wang, F. F., and Wang, Y. J. Lipolyticactivities of a lipase immobilized on six selected supporting mate-rials. Biotechnol. Bioeng.1990,35, 132–137

14. Lotti, M., Grandori, R., Fusetti, F., Longhi, S., Brocca, S., Tramon-

tano, A., and Alberghina, L. Cloning and analysis ofCandidacylindracealipase sequences.Gene1993,124,45–55

15. Arroyo, M., Moreno, J. M., and Sinisterra, J. V. Immobilization/stabilization, on different hydroxylic supports, of lipase fromCandida rugosa. J. Mol. Catal.1993,83, 261–271

16. Henley, J. P. and Sadana, A. Categorization of enzyme deactivationsusing a series-type mechanism.Enzyme Microb. Technol.1985,7,50–60

17. Bardsley, W. G.SIMFIT Program: A Computer Package forSimulation, Curve-fitting, Graph-plotting and Statistical AnalysisUsing Life Science Models. School of Biological Sciences, Univer-sity of Manchester, U.K., 1995

18. Moreno, J. M., Arroyo, M., Hernaíz, M. J., and Sinisterra, J. V.Covalent immobilization of pure isoenzymes from lipase ofCan-dida rugosa. Enzyme Microb. Technol.1997,21, 552–558

19. Zaks, A. and Klibanov. A. M. The effect of water on enzyme actionin organic media.J. Biol. Chem.1988,263,8017–8021

20. Reslow, M., Adlercreutz, P., and Mattiasson, B. On the importanceof the support material for bioorganic synthesis. Influence of waterpartition between solvent, enzyme, and solid support in water-poorreaction media.Eur. J. Biochem.1988,172,573–578

21. Adlercreutz, P. On the importance of the support material forenzymatic synthesis in organic media. Support effects at controlledwater activity.Eur. J. Biochem.1991,199,609–614

22. Valivety, R. H., Halling, P. J., Peilow, A. P., and Macrae, A. R.Lipases from different sources vary widely in dependence ofcatalytic activity on water activity.Biochim. Biophys. Acta1992,1122,143–146

23. Valivety, R. H., Halling, P. J., Peilow, A. P., and Macrae, A. R.Relationship between water activity and catalytic activity of lipasesin organic media. Effects of supports, loading and enzyme prepa-ration.Eur. J. Biochem.1994,222,461–466

24. Parker, M. C., Moore, B. D., and Blacker, A. J. In situ hydratationof enzymes in non-polar organic media can increase the catalyticrate.Biocatalysis1994,10, 269–277

25. Hernaíz, M. J., Sanchez-Montero, J. M., and Sinisterra, J. V. Newdifferences between isoenzymes A and B fromCandida rugosalipase.Biotechnol. Lett.1997,19, 303–306

26. Kvittingen, L., Sjursnes, B., and Anthosen, T. Use of salt hydratesto buffer optimal water level during lipase catalysed synthesis inorganic media: A practical procedure for organic chemists.Tetra-hedron1992,48, 2793–2802

27. Yamane, T., Kojima, Y., Ichiryu, T., Nagata, M., and Shimizu, S.Intramolecular esterification by lipase powder in microaqueousbenzene.Biotechnol. Bioeng.1989,34, 838–843

28. De la Casa, R. M., Sańchez-Montero, J. M., and Sinisterra, J. V.Water adsorption isotherm as a tool to predict the pre-equilibriumwater amount in preparative esterification.Biotechnol. Lett.1996,18, 13–18

Papers


Documents

Thermal stabilization of immobilized lipase B from Candida antarctica on different supports: Effect of water activity on enzymatic activity in organic media