Stability of Enzymes and Proteins in Dried Glassy Systems: Effectof Simulated Sunlight Conditions

Luis Espinosa,*,† Carolina Schebor,†,| Norma S. Nudelman,‡,§ and Jorge Chirife†

Departamento de Industrias and Departamento de Quımica Organica, Facultad de Ciencias Exactas y Naturales,Universidad de Buenos Aires, Ciudad Universitaria, 1428 Buenos Aires, Argentina

The purpose of the present work was to study the effects of simulated sunlightconditions on enzyme inactivation and structural damage in dehydrated glassysystems. Freeze-dried samples containing different enzymes (lactase, invertase,lysozyme and amyloglucosidase) were exposed to light using a medium-pressure metalhalide HPA 400 W lamp. After 1 h of light exposure, the samples showed a significantreduction (more than 50%) in the denaturation peak area as analyzed by DSC, andthis could be attributed to protein denaturation. For most of the pure enzymes, theloss of enzymic activity after 1 h of light exposure was around 50%. In the case ofenzymes included in anhydrous model systems (trehalose, raffinose, maltodextrin, anddextran), the remaining activity also decreased dramatically during the light treatment.We showed that the light exposure in dehydrated systems generated both the loss ofenzymic activity and structural changes such as denaturation (observed by DSC) andprotein fragmentation and aggregation (observed by electrophoresis). Overall, we canconclude that a short exposure to the light produces dramatic changes in the enzymicactivity in dehydrated systems with or without protective matrices.

Introduction

Previous work done in this laboratory analyzed thethermal inactivation of enzymes in dehydrated glassysystems. A glass is a liquid of such high viscosity that itmay cause a significant arrest of translational molecularmotion; thus, the formation of a glassy solid state mayseverely restrict the extent of chemical reactions. Thetemperature of transition from a glassy solid to asupercooled liquid (rubbery state) is labeled Tg, the glasstransition temperature (1). We have observed that amor-phous dried matrices of polymers (maltodextrin, PVP)and sugars (trehalose, raffinose, sucrose) kept below theirTg greatly protected invertase and lactase from thermalinactivation (2-5). As reported in those papers, somethermal inactivation of enzymes was observed in glassysystems, showing that Tg could not be considered anabsolute threshold of stability. However, in the glassystate inactivation was much lower than that observedat temperatures above Tg.

Most organic polymers undergo chemical change, orphotodegradation, when exposed to visible or ultravioletradiation. The biological activity of a protein depends onits native structure as determined by the three-dimen-sional arrangement of amino acids at the active site (6).Research about the photochemistry of proteins hasdemonstrated that physical and chemical alterations maybe observed following irradiation by ultraviolet (UV)

light (7). Most of the amino acids do not absorb UVA orvisible radiation. Only tryptophan and tyrosine and to alesser extent phenylalanine absorb UVB light. Theenzymes that are constituted exclusively by amino acidsand no other chemical groups are not affected by visiblelight irradiation in both their catalytic activity and theirstructure (8). The presence of a suitable photosensitizeris required to induce modifications (mostly oxidations)in the protein structure by visible light irradiation (9,10). Nevertheless, it is possible that visible radiationcould modify conjugated enzymes, whose prostheticgroups absorb visible light (8).

The purpose of the present work was to study theeffects of simulated sunlight conditions on enzyme inac-tivation and structural damage in dehydrated glassysystems.

Materials and Methods

The following (one or the other) materials were em-ployed to form the amorphous matrices: trehalose,raffinose (both from Sigma Chemical Co., St. Louis MO),maltodextrin (Maltrin M 040, Grain Processing Corp.Muscatine, IA), and dextran (average molecular weight473,000; Sigma).

One or the other of the following enzymes wereincluded in amorphous matrices: (1) lactase (â-galactosi-dase) from Kluyveromyces lactis, (Maxilact 5000L, fromGist Brocades Holland, The Netherlands), (2) invertase(â-d-fructofuranosidase) from Saccharomyces cerevisiae(Solvay, Bioproducts Div., Buenos Aires, Argentina), (3)amyloglucosidase (Diazyme L-200), from Milar S. A.Cordoba, Argentina), and (4) lysozyme (Sigma).

The enzymes were used without further purification.Lactase is a purified industrial liquid â-galactosidasepreparation, molecular weight 200,000, composed of two

* To whom correspondence should be addressed. Phone/Fax:(54) 11 4576-3366. E-mail: lespinos@di.fcen.uba.ar.

† Departamento de Industrias.‡ Departamento de Quımica Organica.§ Present address: Institute for Materials Chemistry and

Engineering, Kyushu University, 6-10-1 Hakozaki, Fukuoka, 812-8581 Japan.

| Member of CONICET Argentina.

1220 Biotechnol. Prog. 2004, 20, 1220−1224

10.1021/bp034368m CCC: $27.50 © 2004 American Chemical Society and American Institute of Chemical EngineersPublished on Web 03/26/2004

identical subunits; it is a glycoprotein with 45% carbo-hydrate (w/w) (11). Invertase is a purified industrialliquid, molecular weight 270,000; it consists of twosubunits and contains up to 50% of its mass as carbohy-drate in the form of nine high-mannose oligosaccharidechains (12). Amyloglucosidase is a purified industrialliquid, and lysozyme is a dehydrated powder obtainedfrom chicken egg white, with a molecular weight of14,400 and 129 amino acids forming a single polypeptidechain stabilized by four disulfide bridges (13). Some pureenzymes and proteins were lyophilized and exposed tothe light without being included in a matrix: (1) inver-tase (â-D-fructofuranosidase) from Saccharomyces cer-evisiae (Sigma), (2) lactase (â-galactosidase) from As-pergillus oryzae (Sigma), (3) lysozyme from chicken eggwhite (Sigma), and (4) bovine serum albumin (BSA)(Pedro Zubizarreta Ward, Buenos Aires, Argentina).

Preparation of Model Systems. Amorphous modelsystems were obtained by freeze-drying solutions con-taining either 10% (w/w) trehalose or raffinose, 2% (w/w) dextran, or 10% (w/w) maltodextrin. The amount ofenzymes added was 1% (v/v) lactase, 0.005% (v/v) inver-tase, 0.01% (v/v) amyloglucosidase. Aliquots of 0.5 mLof the solution were placed in 5 mL glass vials, frozenfor 24 h at -26 °C, and immersed in liquid nitrogenbefore freeze-drying. The freeze-drying process lasted 48h. A Heto-Holten A/S cooling trap model CT110 freeze-drier (Heto Lab Equipment, Denmark) was used, whichoperated at -110 °C and at a chamber pressure of 4 ×10-4 mbar.

Thin layers of the sample to be freeze-dried wereprepared to ensure an effective and homogeneous lightpenetration. It is to be noted that properties of thedehydrated matrix affect light penetration and surfaceabsorption and for this reason samples of differentphysical structure may have some different behavior.

The freeze-dried samples were transferred to desicca-tors and kept at 26 °C over desiccant under vacuum untilmeasurements were performed.

Pure Enzymes and Proteins. Pure dehydrated en-zymes (invertase, lactase, and lysozyme) and BSA weredissolved in water (5% w/v), and 0.2 mL was loaded in 3mL vials. The samples were frozen and freeze-dried usingthe same procedure as described above.

Light Treatment. Dehydrated samples were sealedwith a cap covered with Teflon and an aluminum strapand exposed to the light. A control sample covered withaluminum foil was also exposed to the lamp. The lampemployed was a medium-pressure metal halide HPA 400W (Phillips). The distance between the lamp and thesamples was 18 cm. The ambient temperature wasmonitored and was always between 50 and 55 °C. Theradiation emitted by this lamp simulates sunlight condi-tions. The lamp light intensity was measured as reportedby Yoshioka et al. (14), using quinine monohydrochloridedihydrate aqueous solution as a chemical actinometer.The calibration was carried out following the recom-mendations of the International Conference on Harmo-nization (ICH); the unit of integrated intensity of radia-tion is expressed as the difference in absorbance at 400nm of the irradiated quinine solution (15). Under thepresent conditions, the integrated intensity of radiationafter 1 h exposure was of 0.13 quinine units.

Determination of Enzymatic Activity. After expos-ing the samples to light, the remaining enzymic activitywas determined as follows:

Lactase: 1 mL of phosphate buffer (pH 6.8, 0.1 M) wasadded to each sample. When total dissolution wasachieved, 2 mL of 15% (w/w) lactose aqueous solution was

added. The samples were incubated for 1 h at 37 °C andplaced in a boiling water bath for 5 min to inactivate theenzyme (16). Lactose hydrolysis was determined bymeasuring the amount of glucose formed using anenzymic method described in previous publications (17).

Invertase: 1 mL of citrate buffer (pH 5.0, 0.1 M) wasadded to each sample. When total dissolution wasachieved, 1 mL of 40% (w/w) sucrose aqueous solutionwas added. Vials were incubated for 1 h in a bath at 37°C, and the enzyme was inactivated with 1 mL of 0.33M solution of Na2CO3 (3). Sucrose hydrolysis was followedby glucose determination using the same method asdescribed for lactase activity.

Amyloglucosidase: 1 mL of acetate buffer (pH 5.0,0.1 M) containing 1% (w/v) pregelatinized wheat starchand 0.186% EDTA. Samples were incubated for 1 h in abath at 50 °C, and the enzyme was inactivated with 1mL of 0.33 M solution of Na2CO3. Starch hydrolysis wasfollowed by glucose determination using the same methodcited for lactase activity.

The remaining activity of the enzymes was calculatedas follows. The amount of substrate (lactose, sucrose, orstarch) hydrolyzed by enzymes without light treatment(L0) was considered as 100% enzyme activity; the amountof substrate hydrolyzed after light treatment (Lt) wasreferred to (L0), and the residual activity (RA) wasexpressed as RA ) (Lt/L0) × 100.

Lysozyme: 1.5 mg of dehydrated Micrococcus lysodeik-ticus (Sigma) was resuspended in 1.5 mL of phosphatebuffer (pH 6.8, 0.1 M), and 0.5 mL of NaCl (0.3 M) wasadded. The reaction was started by the addition of 1 mLof the enzyme solution (10 µg/mL). The activity of theenzyme was that corresponding to the transmittancevalue at 540 nm after 10 min of enzymatic reaction.

An average value of two replicate samples was re-ported, both for treated and control samples, and thestandard deviations were about (3%.

Calorimetric Measurements. DSC was used todetermine the endothermal peak areas related to theprotein denaturation (the results were expressed as J/gof dry protein). The calorimetric analyses were performedin a DSC 822e Mettler Toledo calorimeter (Schwerzen-bach, Switzerland). The instrument was calibrated withindium (156.6 °C), lead (327.5 °C), and zinc (419.6 °C).All measurements were performed at a heating rate of10 °C/min in the temperature range 25-120 °C. Her-metically sealed 40 µL medium-pressure pans were used(an empty pan served as reference). Thermograms wereevaluated using Mettler Stare program. An average valueof two replicates was reported, both for treated andcontrol samples.

The calorimetric measurements were performed onpreviously humidified (2 days at 93% RH) freeze-driedenzymes and proteins, because it is known that in nearlydry samples the denaturation temperature may be veryhigh and the sample may decompose. Elkordy et al. (18)recently showed that the denaturation peak for dehy-drated (8% w/w water content) lysozyme was observedat 200 °C.

Electrophoresis Analysis. SDS-PAGE (15% acryl-amide and 0,2% bisacrylamide) was conducted accordingto the procedure described by Laemmli (19) in 1 mm thickgels in a Mini-PROTEAN 3 Cell apparatus (BIO-RAD).Protein fractions were dissolved in (1) 0.5 M Tris-HClbuffer, pH 6.8, containing 1% sodium dodecyl sulfate and(2) 0.5 M Tris-HCl buffer, pH 6.8, containing 1% w/vsodium dodecyl sulfate and 3% v/v 2-mercapto-ethanol,both heated at 95 °C for 5 min prior to loading.

Biotechnol. Prog., 2004, Vol. 20, No. 4 1221

The densitometric analysis of the gels was carried outusing the Molecular Analyst/PC Molecular Imager ver-sion 1.4.1.

Results and DiscussionThe effect of light treatment on protein denaturation

in various freeze-dried enzymes and protein was analyzedby DSC (differential scanning calorimetry). Becauseduring the light treatment the temperature in thephotoreactor was about 50 °C, a control sample coveredwith aluminum foil was also put in the photoreactor atthe same time as the samples, and analyzed to check thethermal effects. Figure 1 shows the DSC thermogramsof lysozyme (Figure 1a) and lactase (Figure 1b) comparedto nontreated samples (labeled “control”) and to samplescovered with an aluminum foil (labeled “covered”). En-thalpy values for lysozyme denaturation were 14.83,14.60, 11.60, and 3.40 J/g for control, covered, and light-treated (1 and 3.5 h) samples, respectively. For lactase,enthalpy values were 1.66, 1.38, and 0.69 J/g for “control”,“covered”, and 1 h treatment samples, respectively.

Both enzymes, lysozyme and lactase, exposed to lightshowed a significant reduction in the denaturation peakarea. This could be attributed to protein denaturationupon exposure to light, so that the proportion of nativeprotein decreased, leading to a reduced denaturationpeak. It can be observed that the denaturing effect causedby the heating of the sample (up to 50-55 °C) duringthe light treatment was minimal, because the aluminumfoil covered samples showed only a slight decrease in thepeak area with respect to the control sample. However,a synergism between temperature and light cannot beruled out.

According to Branchu et al. (20), the endothermictransition observed for lysozyme is explained by the

breaking of the H-bonded structure and the loss ofhydrophobic interactions.

Figure 2 compares the percent denaturation (i.e., %reduction in endothermic peak area) observed after 1 hof light treatment for lactase, invertase, lysozyme, andBSA. It can be seen that 1 h of light exposure producessignificant changes in the degree of protein denaturation.The percent denaturation was calculated considering itwas 0% for the covered samples (in all cases these valueswere almost zero; except for lactase, which showed 10%thermal denaturation and was the more sensitive to thelight irradiation).

The effect of light treatment on the activity of theenzymes was also analyzed. Figure 3 shows the relation-ship between the percent denaturation and the remainingactivity for dehydrated lysozyme. Irradiation of driedlysozyme with light (at around 50 °C) leads to a progres-sive loss of enzymic activity, which could be related tothe percent denaturation. The loss of enzymic activityafter 1 h of light exposure was also evaluated for lactaseand invertase, being 50% for both enzymes. Very earlystudies, with no description of photochemical conditions,showed that after a few minutes of exposure to ultravioletlight, the enzyme lysozyme (in solution) lost 55% activity(21); however, no literature data is known to us on theeffect of light in the dried state. Isutzu et al. (22) reporteda good correlation between the denaturation peak andthe remaining activity in freeze-dried lactase samples,after thermal treatment; however, no literature data isknown to us on the effect of light for enzymes in the driedstate.

The effect of light on the protein structure (i.e.,aggregation and/or fractures in the molecule) was alsostudied by electrophoresis analysis. Protein aggregationinvolves the formation of higher molecular weight com-plexes from the denatured protein, which then cross-link

Figure 1. DSC thermograms of (a) pure freeze-dried lysozymeand (b) lactase exposed to the light compared to control andcovered samples.

Figure 2. Percent denaturation observed after 1 h of lighttreatment for different proteins. The error bars indicate thestandard deviation.

Figure 3. Relationship between the percent denaturation andthe remaining activity for dehydrated lysozyme after exposureto the light. The error bars indicate the standard deviation.

1222 Biotechnol. Prog., 2004, Vol. 20, No. 4

by specific bonding (23). Both disulfide and hydrogenbonding, as well as ionic interactions, are involved in thecross-linking of aggregates from denatured protein (24,25).

Figure 4 shows densitograms corresponding to thedifferent bands of an SDS-PAGE gel for lysozyme. Thenontreated sample (control) shows the presence of a mainband corresponding to the enzyme and a minor band athigher molecular weight (Figure 4, line 1). In the densi-tograms corresponding to samples treated with SDS afterlight exposure, an increase in high molecular weight

protein fractions was observed (Figure 4, lines 3 and 4).Meanwhile, the peak corresponding to the enzyme wasreduced, probably as a result of the aggregation of proteinmolecules. In the samples treated with â-mercaptoetha-nol, after light exposure the main peak was not signifi-cantly reduced, and the peak profile did not showimportant changes comparing to the control sample(Figure 4, lines 5-8). It is possible that the aggregationobserved for samples exposed to the light (Figure 4, lines3 and 4) was due to disulfide bond formation betweenprotein molecules.

It has been reported that protein unfolding is mechani-cally inhibited by entrapment of protein molecules in anamorphous matrix (26, 27). The protective effect ofsugars, particularly trehalose, and polymers on enzymesduring drying and heat treatment has been proved inprevious works. Carpenter and Crowe (28), concludedthat during drying the maintenance of the native stateof lysozyme included in a trehalose matrix was due tohydrogen bonding between sugar and protein. Previousresults from our laboratory showed that certain enzymeswere protected when immersed in glassy matrices andexposed to high temperatures. Lactase and invertaseretained almost 100% activity after a few hours of heattreatment at 75 °C (2-4). To test if the presence ofmatrices could protect the enzymes from the damagecaused by light exposure, freeze-dried model systemscontaining the enzymes included in different matriceswere analyzed. Figure 5 shows the remaining activity ofthe different enzymes in anhydrous model systems oftrehalose, raffinose, maltodextrin, and dextran as afunction of time of exposure to light. In all systems, theenzymic activity decreased dramatically during the lighttreatment. None of the matrices employed could protectthe enzymes from being inactivated, even considering theshort exposure time employed (max 1 h). The remainingactivity of the enzymes was related to a sample coveredwith aluminum film to take into account the effect oftemperature (50-55 °C). Thus, light rather than tem-

Figure 4. Densitograms corresponding to the different bandsof an SDS-PAGE gel of dried lysozyme irradiated with lightfor 0 (1 and 5), 1 (3 and 7), and 3.5 (4-8) h and 3.5 h coveredwith aluminum foil (2 and 6).

Figure 5. Remaining activity of (a) invertase, (b) lactase, (c) amyloglucosidase, and (d) lysozyme in freeze-dried matrices of “zero”moisture content as a function of exposure time to the light at 50 °C: (/) trehalose, (9) raffinose, (b) maltodextrin, and (2) dextran.Theerror bars indicate the standard deviation.

Biotechnol. Prog., 2004, Vol. 20, No. 4 1223

perature induced the dramatic activity loss found in thepresent experiments. The lack of protective effect ofglassy sugar matrices with respect to light exposure isnot necessarily in conflict with the idea that the protec-tive effects are due to hydrogen bonding between sugarand protein (28). Light exposure could damage theproteins without determining the unfolding in the amor-phous matrix. We have to consider that protein dena-turation, causing the observed decrease of enzymicactivity, could occur upon resuspension of the glassymatrix, meaning that damage in the protein moleculecaused during the exposure to light can be manifestedupon rehydration.

We showed that the light exposure in dehydratedsystems generated both loss of enzymic activity andstructural changes such as denaturation (observed byDSC) and protein fragmentation and aggregation (ob-served by electrophoresis). Overall, we can conclude thata short exposure to the light produces dramatic changesin the enzymic activity in dehydrated systems with orwithout protective matrices.

Acknowledgment

The authors acknowledge financial support from theAgencia Nacional de Promocion Cientıfica y Tecnologica(Proyecto 06251, prestamo BID 1201/OC-AR).

References and Notes(1) Roos, Y. H. The physical state of amorphous materials. In

Phase Transitions in Foods; Taylor, S., Eds.; AcademicPress: New York, 1995; pp 26-34.

(2) Mazzobre, F.; Buera, M. P.; Chirife, J. Glass transition andthermal stability of lactase in low-moisture amorphouspolymeric systems. Biotechnol. Prog. 1997, 13, 195-199.

(3) Schebor, C.; Buera, M. P.; Chirife, J. Glassy state in relationto the thermal inactivation of the enzyme invertase inamorphous dried matrices of trehalose, maltodextrin andPVP. J. Food Eng. 1996, 30, 269-282.

(4) Schebor, C.; Burın, L.; Buera, M. P.; Aguilera, J. M.; Chirife,J. Glassy state and thermal inactivation of invertase andlactase in dried amorphous matrices. Biotechnol. Prog. 1997,13, 857-863.

(5) Cardona, S.; Schebor, C.; Buera, M. P.; Karel, M.; Chirife,J.; Thermal stability of invertase in reduced moisture amor-phous matrices in relation to glassy state and trehalosecrystallization. J. Food Sci. 1997, 62, 105.

(6) Rosenberger, F.; Vekilov, P. G.; Muschol, M.; Thomas, B.R. Nucleation and crystallization of globular proteinsswhatwe know and what is missing. J. Cryst. Growth 1996, 168,1-27.

(7) Brooks, S. C. In Biological Effects of Radiation; Dugar, B.M., Eds.; McGraw-Hill: New York, 1936; pp 341.

(8) Silva, E. Sensitized photooxidation of amino acids in pro-teins. In Chemical Modification of Enzymes; Ezaguirre, J.,Eds.; Ellis Horwood Limited: Chichester, 1987; pp 63-73.

(9) Silva, E.; De Landea, C.; Edwards, A. M.; A. M.; Lissi, A.M. Lysozyme photooxidation by singlet oxygen: propertiesof the partially inactivated enzyme. J. Photochem. Photobiol.,B 2000, 55, 196-200.

(10) Edwards, A. M.; Silva, E. Effect of visible light on selectedenzymes, vitamins and amino acids. J. Photochem. Photobiol.,B 2001, 63, 126-131.

(11) Cavaille, D.; Combes, D. Characterization of â-galactosi-dase from Kluyveromyces lactis. Biotechnol. Appl. Biochem.1995, 22, 55-64.

(12) Chu, F. K.; Watorek, W. W.; Maley, F. Factors affectingthe oligomeric structure of yeast external invertase. Arch.Biochem. Biophys. 1983, 223, 543-555.

(13) Canfield, R. E. The amino acid sequence of egg whitelysozyme. J. Biol. Chem. 1963, 238, 2698.

(14) Yoshioka, S.; Ishihara, Y.; Terazono, T.; Tsunakawa, N.;Murai, M.; Yasuda, T.; Kitamura, Y.; Kunihiro, K.; Sakai, Y.;Hirose, K.; Tonooka, K.; Takayama, F.; Imai, M.; Godo, M.;Matsuo, K.; Nakamura, Y.; Aso, S.; Kojima, Y.; Takeda Terao,T. Quinine actinometry as a method for calibrating ultravioletradiation intensity in light stability testing of pharmaceuti-cals. Drug Dev. Ind. Pharm. 1994, 20 (13), 2049-2062.

(15) Harth, Y.; Rapoport, M. Drug Saf. 1996, 14, 252-259.(16) Mazzobre, M. F.; Buera, M. P. Combined effects of trehalose

and cations on the thermal resistance of â-galactosidase infreeze-dried systems. Biochim. Biophys. Acta 1999, 1473,337-344.

(17) Buera, M. P.; Chirife, J.; Karel, K. A study of acid-catalyzedsucrose hydrolysis in an amorphous polymeric matrix atreduced moisture contents. Food Res. Int. 1995, 28, 359-365.

(18) Elkordy, A.; Forbes, R. T.; Barry, B. W. Integrity ofcrystalline lysozyme exceed that of a spray-dried form. Int.J. Pharm. 2002, 247, 79-90.

(19) Laemmli, U.K. Cleavage of structural proteins during theassembly of the head of bacteriophage T4. Nature 1970, 227,680-685.

(20) Branchu, S.; Forbes, R. T.; York, P.; Nyquist, H. A centralcomposite design to investigate the thermal stabilization oflysozyme. Pharm. Res. 1999, 16, 702-708.

(21) Shugar, D. The Measurement of lysozyme activity and theultra-violet inactivation of lysozyme. Biochim. Biophys. Acta1952, 8, 302-309.

(22) Isutzu, K.; Yoshioka, S.; Takeda, Y. Protein denaturationin dosage forms measured by differential scanning calorim-etry. Chem. Pharm. Bull. 1999, 38, 800-803.

(23) Schmidt, R. H. Gelation and coagulation. In ProteinFunctionality in Foods; Pour-El, A., Eds.; American ChemicalSociety: Washington, DC, 1981; pp 131

(24) Catsimpoolas, N.; Meyer, E. W. Gelation phenomena ofsoybean globulins: Protein-protein interactions. Cereal Chem.1970, 47, 559.

(25) Morrissey, P. A.; Mulvihill, D. M.; O’Weill, E. M. Functionalproperties of muscle proteins. In Development in Food Pro-teins; Hudson B. J. F., Eds.; Elsevier Applied Science:London, 1987; Vol. 5, p 237.

(26) Carpenter, J. F.; Crowe, J. H. An infrared spectroscopicstudy of the interactions of carbohydrates with dried proteins.Biochemistry 1989, 28, 3916-3922.

(27) Allison, S. D.; Chang, B.; Randolph, T. W.; Carpenter, J.F. Hydrogen bonding between sugar and protein is respon-sible for inhibition of dehydration-induced protein unfolding.Arch. Biochem. Biophys. 1999, 365, 289-298.

(28) Crowe, L. M.; Crowe, J. H.; Chapman, D. Preservation ofmembranes of anhydrobiotic organism: the role of trehalose.Science 1984, 233, 701-703.

Accepted for publication February 19, 2004.

BP034368M

1224 Biotechnol. Prog., 2004, Vol. 20, No. 4