Prediction of penicillin V acylase stability in water-organicco-solvent monophasic systems as a function of

solvent composition

Miguel Arroyo*, Raquel Torres-Guzma´n, Isabel de la Mata, M. Pilar Castillo´n,Carmen Acebal

Departamento de Bioquı´mica y Biologı´a Molecular I, Facultad de Ciencias Biolo´gicas, Universidad Complutense de Madrid, Madrid, Spain

Received 27 October 1999; received in revised form 24 January 2000; accepted 15 February 2000

Abstract

Hydrolytic activity of penicillin V acylase (EC 3.5.1.11) can be improved by using organic cosolvents in monophasic systems. However,the addition of these solvents may result in loss of stability of the enzyme. The thermal stability of penicillin V acylase fromStreptomyceslavendulaein water–organic cosolvent monophasic systems depends on the nature of the organic solvent and its concentration in the media.The threshold solvent concentration (at which half enzymatic activity is displayed) is related to the denaturing capacity of the solvent. Wefound out linear correlations between the free energy of denaturation at 40°C and the concentration of the solvent in the media. On one hand,those solvents with logP values lower than21.8 have a protective effect that is enhanced when its concentration is increased in the medium.On the other hand, those solvents with logP values higher than21.8 have a denaturing effect: the higher this value and concentration, themore deleterious. Deactivation constants of PVA at 40°C can be predicted in any monophasic system containing a water-miscible solvent.© 2000 Elsevier Science Inc. All rights reserved.

Keywords:Penicillin V acylase;Streptomyces lavendulae; Organic solvents; logP; Enzyme stability; Monophasic systems

1. Introduction

The potential advantages of performing biocatalyticreactions in aqueous– organic solvent mixtures are welldocumented [1–3]. Although the logP value of a solventhas been a useful tool to predict the behavior of enzymesin systems containing low amounts of water [4], thisapproach has not been valid for monophasic systemscontaining water-miscible organic solvents. Enzymes areknown to be denaturated in monophasic systems contain-ing polar solvents, mainly because the hydratation stateof the enzyme molecules is distorted to the extent that thecatalytically active conformation is lost [5–7]. In addi-tion, the enzyme activity and stability depends on theconcentration of the water-miscible solvent. When theproportion of the latter exceeds a certain threshold, theessential bound water is stripped from the enzyme’s

surface, leading to deactivation [8]. The critical pointarrives when one has to determine the solvent concentra-tion before reaching its denaturating effect. In view ofthis, many authors have been looking for a correlationbetween any physicochemical property of the solvent andenzyme stability [9,10]. All of these reports give onlysome guiding rules for optimizing the enzyme-catalyzedreactions in nonaqueous media. In this work, we havesystematically studied the effect of increasing the con-centrations of several water-miscible organic solvents onthe stability of penicillin V acylase (PVA) fromStrepto-myces lavendulae. This enzyme catalyzes the hydrolysisof penicillin V to yield 6-amino penicillanic acid (6-APA), key intermediate for the production of semisyn-thetic penicillins [11]. In a previous work, we reportedthat the hydrolytic activity of PVA could be improved byusing organic cosolvents in monophasic systems [12].However, the addition of these solvents may result in lossof stability of the enzyme. We were able to distinguishbetween enzyme inactivation and enzyme denaturation inwater– organic cosolvent monophasic systems, and to

* Corresponding author. Tel.:134-91-394 41 50; fax:134-91-394 4672.

E-mail address:arroyo@solea.quim.ucm.es (M. Arroyo).

www.elsevier.com/locate/enzmictecEnzyme and Microbial Technology 27 (2000) 122–126

0141-0229/00/$ – see front matter © 2000 Elsevier Science Inc. All rights reserved.PII: S0141-0229(00)00183-6

predict the behavior of penicillin V acylase in thesenon-conventional environments.

2. Materials and methods

2.1. Enzyme production and partial purification

Penicillin V acylase (EC 3.5.1.11) was produced fromS.lavendulae(ATCC 13664) by fermentation [13] and it waspurified according to Torres et al. [14]. Protein was mea-sured by the Coomassie Blue binding method using bovineserum albumin as the standard [15].

2.2. Enzyme activity

The potassium salt of phenoxy methyl penicillin (PVK)from Sigma (St. Louis, MO, USA) was used as substrate inthe enzyme assays. A total of 15ml of the purified PVAsolution was mixed with 120ml of water, 15ml of 1 Mpotassium phosphate buffer, pH 8.0, and 150ml of a PVKsolution (45 mg/ml) in water. The reaction mixture wasincubated for 30 minutes at 40°C under gentle shaking. Thereaction was stopped by addition of 0.9 ml of 20% (v/v)acetic acid solution. The reaction mixture was centrifuged,and an aliquot of 0.9 ml was processed for estimation of6-APA by addition of 300ml of a 5% (p/v) solution ofp-dimethyl-amino benzaldehyde (Sigma) in methanol [16].One activity unit (U) is defined as the amount of enzymeproducing 1mmol/min of 6-APA under the assay conditions(PDAB method). An enzyme solution with an activity of 1.0U/mg protein was used in all the experiments.

2.3. Enzyme activity in water–organic solvent mixtures

The hydrolytic activity of PVA was assayed in the pres-ence of different concentrations of triethylene glycol, dieth-ylene glycol, ethylene glycol, dimethylsulfoxide, dimethyl-formamide, methanol, acetonitrile and acetone (Merck,Darmstadt Germany). The solvents were incorporated into50ml of enzyme solution before the addition of penicillin V.The volume of water was corrected for the added solvent sothat the volume remained 150ml. The enzyme activity, incontact with the solvent, was measured after the addition of150 ml of PVK solution, so that the solvent concentrationwas referred to a final reaction volume of 300ml. All of theexperiments were carried out in triplicate.

2.4. Effect of organic cosolvents on enzyme stability

The stability of PVA was determined in 100 mM phos-phate buffer, pH 8.0, and in the mixtures containing differ-ent concentration of the organic cosolvents. The sampleswere incubated at 40°C, and aliquots of 150ml were with-drawn at different storage times to check the residual activ-

ity. The residual enzyme activity was determined by usingthe PDAB method.

3. Results and discussion

3.1. Predicting the threshold concentration of organicsolvent for PVA denaturation

We reported that penicillin V acylase (EC 3.5.1.11) fromS. lavendulaeshowed enhanced activity in water–organicsolvent mixtures [12]. In the case of water-miscible sol-vents, the catalytic activity was increased up to a criticalconcentration of these cosolvents, but further addition leadto a gradual protein deactivation. As shown in Fig. 1, itappears that the catalytic activity of PVA change after acertain threshold, as reported for other enzymes [5,9,17–19]. We calculated the threshold concentration for eachsolvent (C50), defined as the concentration (in molarity) ofthe organic solvent at which one-half of the activity of theenzyme is observed (Table 1). In view of this, the denatur-ation capacity of the solvent seemed to be a valuable tool forpredicting the threshold concentration for PVA. The dena-turation capacity is a parameter which characterizes thedenaturing strength of organic solvents. Scaling of co-sol-vents according to their denaturing capacity may have ageneral character as it has been proved correct in experi-ments with 8 proteins in more than 30 solvents [9]. Therelationship with other physicochemical parameters such asdielectric constant [20,21], logP [22], Dimroth-Reichardtparameter [23], and polarity index [10] was not reliable forpredicting the threshold concentration. However, we couldfind a linear correlation between the DC values of thesolvents and the log(C50) values of each solvent (Fig. 2).Our results are in accordance with the statement that co-solvents that are both hydrophobic and have a high solva-tion capacity (1,4 dioxane, tetrahydrofurane and pyridine)

Fig. 1. Calculation of the threshold concentration of water-miscible organicsolvents.

123M. Arroyo et al. / Enzyme and Microbial Technology 27 (2000) 122–126

are strong denaturants and cause PVA deactivation at lowconcentrations. On the other hand, concentrated solutions ofsolvents with low denaturing capacity (ethylene glycol andglycerol) are not so deleterious for penicillin V acylase, asreported by Khmelnitsky for other enzymes [9].

3.2. Predicting the enzyme deactivation constants in themonophasic systems

The irreversible deactivation mechanism of PVA obeyedfirst-order deactivation kinetics in buffer and in the presenceof all the organic cosolvents tested. First-order exponentialdecay in activity has also been found in other enzymes inthe presence of different water-miscible organic solvents

[24–26]. According to this mechanism, the enzyme under-goes an irreversible deactivation in one step as indicatedbelow:

Ekd

3 Ed

where E and Ed are the initial and the deactivated forms ofPVA respectively, andkd is the apparent deactivation con-stant of the enzyme (h21). The time course of enzymedeactivation can be obtained from a linear equation ln (Ed/E) 5 2kd z t. Deactivation plots of PVA, stored in bufferand in the presence of different concentrations of organicsolvent at 40°C, let us calculate the apparent deactivationrate constants. These values (Table 2) have been calculatedfrom the slopes of the best-fit curves obtained by linearregression when ln (Ed/E) was plotted against time. On theother hand, the free energy (DG#) of the deactivation pro-cess could be calculated from the following equation:

2DG# 5 RT ln @~kB z T!/~kd z h)]

wherekd is the first-order deactivation constant (h21), kB isthe Boltzmann’s constant (1.383 10223 J z °K21), h is thePlanck’s constant (1.843 10237 J.h), R is the gas constant(8.314 Jz mol21 z °K21) and T is the temperature (°K). Asit is shown in Table 2, triethylene and diethylene glycolexerted a protective effect, supported by an increase in thefree energy of denaturation [27]. However, the other sol-vents decreased the free energy of denaturation. We couldalso state that the free energy of denaturation was increasedwhen the concentration of triethylene and ethylene glycolwas increased in the media. On the other hand, there was a

Table 1Threshold concentration (C50) of water-miscible organic solvents for penicillin V acylase deactivation.

Solvent DCa Ib «c LogPd ET(30)e

(KJ/mol)C50 (M) logC50

Ethylene glycol 18.7 5.4 37.7 21.80 236 10.8 1.03Glycerol 20.2 — 42.5 23.03 238 7.5 0.86Methanol 30.5 5.1 32.7 20.76 232 4.9 0.69Ethanol 54.4 4.3 24.6 20.24 213 5.5 0.73Dimethylsulfoxide 60.3 7.2 46.5 21.30 188 3.5 0.54Dimethylformamide 63.3 6.4 36.7 21.00 183 3.2 0.50Acetonitrile 64.3 5.8 35.9 20.30 192 3.5 0.481-propanol 69.2 3.9 20.3 0.28 212 2.7 0.432-propanol 70.2 3.9 18.3 0.05 203 3.3 0.52Acetone 78.2 5.1 20.7 20.24 177 3.4 0.481,4 Dioxane 92.1 4.8 2.2 21.10 151 1.3 0.13Tetrahydrofurane 100.0 4.0 7.6 0.49 156 1.2 0.08Diethylene glycol — — 31.7 22.30 225 3.7 0.57Triethylene glycol — — 23.7 22.80 224 2.25 0.35Pyridine — — 12.9 0.71 160 0.86 20.06

a The denaturing capacities of the solvents were taken from Khmelnitsky et al. [9].b The polarity indexes of the solvents were taken from Gupta et al. [10].c The dielectric constant values at 20°C were taken from the monographs by Reichardt [20] and Riddick et al. [21].d The logP values were taken from, or calculated on the basis of Rekker [22].e The ET(30) parameter describes the solvent polarity and the values for the solvents were taken from Reichardt [23].

Fig. 2. Relationship between the denaturing capacity (DC) of the solventand the threshold concentration (C50) for penicillin V acylase deactivation.

124 M. Arroyo et al. / Enzyme and Microbial Technology 27 (2000) 122–126

linear decay in the free energy of denaturation when theconcentration of DMSO, DMF, methanol, acetonitrile, andacetone was increased in the media (Fig. 3). A similar trendwas found when we calculated theDG# from the deactiva-tion constants that appear in the literature for other enzymessuch as aminoacylase from pig kidney in water–DMF mix-tures [26] and cytochrome c oxidase in water–THF mixtures[24]. The slopes of these curves may be considered as aparameter that could characterize the nature of the solventand its effect on enzyme stability. If the slope has a positivevalue, the solvent would always have a stabilizing effect onthe enzyme, whereas a negative value would mean that thesolvent is a denaturant. The magnitude of this deleterious orstabilizing effect would depend on a physicochemical pa-rameter of the solvent. When we plotted the values of theslopes from Fig. 3 versus the logP values of the solvents,they fitted to a linear equation (Fig. 4). Triethylene anddiethylene glycol protect the enzyme against thermal inac-tivation. These solvents would behave like glycerol inwhich enzymes are more thermostable than in water [27–29]. In addition, this protective effect is enhanced if the

concentration of the solvent is increased in the medium.Those solvents with logP values higher than21.8 are strongdenaturants and disrupt the tertiary structure of the enzymewhen its concentration is increased. The denaturation isquicker at low solvent concentration when the logP value ofthe solvent is higher. Supporting these facts, it has beenreported that thermal transition temperature for enzymes inwater–alcohol mixtures [30,31] dropped regularly with in-creasing alcohol chain length at equal alcohol concentra-tion. As it can be calculated by Rekker method, the logPvalue of an alcohol increases when the number of methylenegroups in its chain increases. Therefore, the higher the logPvalue, the more denaturing effect of the alcohol on enzymestructure. In contrast to the effects of increasing hydrocar-bon content, increasing hydroxyl content tends to decreasethe denaturing ability of the alcohols [30]. These effectshave proved to occur for penicillin V acylase in the presenceof ethylene, diethylene, and triethylene glycols. The appli-cation of thermodynamics for the evaluation of the dena-turing effects of solvents on enzymes have been also re-ported by other authors [28,31]. A plot of free energy of

Table 2First-order deactivation constants and free energies of denaturation at 40°C for PVA in the presence of different concentrations of water-miscibleorganic solvents.

Concentration (v/v) 3% 5% 8% 10% 20%

Solvent LogP kd

(h21)DG#

(KJ/mol)kd

(h21)DG#

(KJ/mol)kd

(h21)DG#

(KJ/mol)kd

(h21)DG#

(KJ/mol)kd

(h21)DG#

(KJ/mol)

Acetone 20.24 0.0058 111.5 0.0077 110.8 0.0077 110.8 0.0129 109.4 0.066 105.2Acetonitrile 20.3 0.0058 111.5 0.0134 109.3 0.0231 107.9 0.038 106.6 — —Methanol 20.76 0.0039 112.5 0.0046 112.1 0.0067 111.1 — — — —DMF 21.0 0.0039 112.5 0.0042 112.3 0.0076 110.7 0.0086 110.5 0.0039 110.6DMSO 21.3 0.0039 112.5 0.0041 112.4 0.006 111.4 0.0088 110.4 0.023 107.9Ethylene glycol 21.8 0.0038 112.6 0.0029 113.3 0.0028 113.4 0.003 113.2 0.0025 113.7Diethylene glycol 22.3 0.0024 113.8 0.0019 114.4 0.0018 114.6 0.0013 115.3 0.0006 117.2Triethylene glycol 22.8 0.0014 115.3 0.0013 115.4 0.0011 115.9 0.0009 116.2 0.0007 116.9

For native PVA in buffer,kd 5 0.003 h21 andDG# 5 113.2 KJ/mol at 40°C

Fig. 3. Effect of the solvent concentration on the free energy of thedeactivation process at 40°C.

Fig. 4. Relationship between the logP of the solvents and the increment inthe free energy of the deactivation of penicillin V acylase at 40°C.

125M. Arroyo et al. / Enzyme and Microbial Technology 27 (2000) 122–126

denaturation versus solvent concentration showed a linearresponse with different slope values depending on the na-ture of the solvent. At high alcohol concentration (.2M) theslope value was lower in the order propanol. ethanol.methanol, indicating that the former alcohol was the moredeleterious. This order would correspond to the logP valueof these short-chain alcohols (0.28;20.25, and20.76,respectively), proving again the relationship between thehydrophobicity of the solvent and its influence on enzymestability. Supporting this fact, the addition of increasingamounts of glycerol brought about a positive change in thefree energy of denaturation of chymotrypsinogen and ribo-nuclease at 45°C [28], indicating its stabilizing effect.

Taking into account all of these facts, we can predict thedeactivation constants of PVA at 40°C in any monophasicsystem containing a water-miscible solvent, by using thefollowing equation:

kd(SOLV) 5 kd(BUFFER)z e2(DGSOLV2DGBUFFER)/RzT

5 kd(BUFFER)z e2DDG#/RzT

5 kd(BUFFER)z e2(A1BzlogP)CSOLV/RzT

wherekd(SOLV) is the deactivation constant of the enzyme inthe presence of a water-miscible solvent at a concentrationof c molar (Csolv) at 40°C; andkd(BUFFER) is the deactivationconstant of the native enzyme at 40°C in buffer.

Acknowledgment

Financial support from the CICYT (QUI97-00490) isgratefully acknowledged.

References

[1] Zaks A, Klibanov AM. Enzymatic catalysis in nonaqueous solvents.J Biol Chem 1988;263:3194–201.

[2] Dordick JS. Enzyme catalysis in monophasic organic solvents. En-zyme Microb Technol 1989;11:194–211.

[3] Gupta MN. Enzyme function in organic solvents. Eur J Biochem1992;203:25–32.

[4] Laane C, Boeren S, Vos K, Veeger C. Rules for optimization ofbiocatalysis in organic solvents. Biotechnol Bioeng 1987;30:81–7.

[5] Mozhaev VV, Khmelnitsky YL, Sergeeva MV, Belova NL, KlyachkoNL, Levashov AV, Martinek K. Catalytic activity and denaturation ofenzymes in water/organic cosolvent mixtures:a-chymotrypsin andlaccase in mixed water/alcohol; water/glycol and water/formamidesolvents. Eur J Biochem 1989;184:597–602.

[6] Lozano P, de Diego T, Iborra JL. Influence of water-miscible aproticsolvents ona-chymotrypsin stability. Biotechnol Prog 1996;12:488–93.

[7] Griebenow K, Klibanov AM. On protein denaturation in aqueous-organic mixtures but not in pure organic solvents. J Am Chem Soc1996;118:11695–700.

[8] Gorman LAS, Dordick JS. Organic solvents strip water off enzymes.Biotechnol Bioeng 1992;39:392–7.

[9] Khmelnitsky YL, Mozhaev VV, Belova AB, Sergeeva MV, MartinekK. Denaturation capacity: a new quantitative criterion for the selec-

tion of organic solvents as reaction media in biocatalysis. Eur J Bio-chem 1991;198:31–41.

[10] Gupta MN, Batra R, Tyagi R, Sharma A. Polarity index: the guidingsolvent parameter for enzyme stability in aquous-organic cosolventmixtures. Biotechnol Prog 1997;13:284–8.

[11] Shewale JG, Sudakaran VK. Penicillin V acylase: its potential in theproduction of 6-amino penicillanic acid. Enzyme Microb Technol1997;20:402–10.

[12] Arroyo M, Torres R, de la Mata I, Castillo´n MP, Acebal C. Interac-tion of penicillin V acylase with organic solvents. Catalytic activitymodulation on the hydrolysis of penicillin V. Enzyme Microb Tech-nol 1999;25:378–83.

[13] Torres R, Ramo´n F, de la Mata I, Castillo´n MP, Acebal C. Enhancedproduction of penicillin V acylase fromStreptomyces lavendulae.Appl Microbiol Biotechnol 1999;53:81–4.

[14] Torres R, de la Mata I, Castillo´n MP, Arroyo M, Torres J, Acebal C.Purification and characterization of penicillin V acylase fromStrep-tomyces lavendulae. In: Ballesteros A, Plou FJ, Iborra JL, Halling PJ,editors. Stability and stabilization of biocatalysts. Amsterdam:Elsevier Science B.V., 1998. p. 719–24.

[15] Bradford MM. A rapid and sensitive method for quantitation ofmicrogram quantities of protein utilizing the principle of protein-dyebinding. Anal Biochem 1976;72:248–54.

[16] Shewale JG, Kumar KK, Ambedkar GR. Evaluation of determinationof 6-amino penicillanic acid by p-dimethylamino benzaldehyde. Bio-technol Techniques 1987;1:69–72.

[17] Torres E, Tinoco R, Va´zquez–Duhalt R. Solvent hydrophobicitypredicts behaviour of lignin peroxidase and cytochrome c in aqueoussolution of water miscible organic solvents. J Biotechnol 1996;49:59–67.

[18] Torres E, Siminovich B, Barzana E, Va´zquez–Duhalt R. Thermody-namic hydrophobicity of aqueous mixtures of water-miscible solventspredicts peroxidase activity. J Mol Catal B: Enzym 1998;4:155–159.

[19] Simon LM, Laszlo K, Vertesi A, Bagi K, Sza´jani B. Stability ofhydrolytic enzymes in water-organic solvent systems. J Mol Catal B:Enzym 1998;4:41–45.

[20] Reichardt C. Solvents and solvent effects in organic chemistry. 2nded. Weinheim: VCH, 1988.

[21] Riddick JA, Bunger WB, Sakano TK. Organic solvents. Physicalproperties and methods of purification, 4th ed. New York: Wiley,1986.

[22] Rekker RF. The hydrophobic fragmental constant. New York:Elsevier, 1977.

[23] Reichardt C. Empirical parameters of solvent polarity as linear free-energy relationships. Angew Chem Int Ed 1979;18:98–110.

[24] Vazquez–Duhalt R, Semple KM, Westlake DWS, Fedorak PM. Ef-fect of water miscible solvents on the catalytic activity of cytochromec. Enzyme Microb Technol 1993;15:936–43.

[25] Moreno JM, Fagain CO. Activity and stability of native and modifiedalanine aminotransferase in cosolvent systems and denaturants. J MolCatal B: Enzym 1997;2:271–9.

[26] Boross L, Kosa´ry J, Stefanovits–Ba´nyai E, Sisak C, Sza´jani B. Stud-ies on kinetic parameters and stability of aminoacylase in non-con-ventional media. J Biotechnol 1998;66:69–73.

[27] Gekko K, Timasheff SN. Mechanism of protein stabilization byglycerol: preferential hydration in glycerol-water mixtures. Biochem-istry 1981;20:4667–76.

[28] Gekko K, Timasheff SN. Thermodynamic and kinetic examination ofprotein stabilization by glycerol. Biochemistry 1981;20:4677–86.

[29] Knubovets T, Osterhout J, Connolly PJ, Klibanov AM. Structure,thermostability and conformational flexibility of hen egg-white ly-sozyme dissolved in glycerol. Proc Natl Acad Sci USA 1999;96:1262–7.

[30] Herskovits TT, Gadegbeku B, Jaillet H. On the structural stability andsolvent denaturation of proteins. J Biol Chem 1970;245:2588–8.

[31] Velicelebi G., Sturtevant JM. Thermodynamics of the denaturation oflysozyme in alcohol-water mixtures. Biochemistry 1979;18:1180–6.

126 M. Arroyo et al. / Enzyme and Microbial Technology 27 (2000) 122–126