Hydration and thermal denaturation of β-lactoglobulin. A calorimetric study

Biochimica et Biophysica Acta, 400 (1975) 334-342

© Elsevier Scientific Publishing Company, Amsterdam- Printed in The Netherlands BBA 37115 H Y D R A T I O N A N D T H E R M A L D E N A T U R A T I O N OF fl-LACTOGLOBUL1N. A C A L O R I M E T R I C STUDY

M. RQEGG, U. MOOR and B. BLANC Federal Dairy Research Institute, CH-3097 Liebefeld-Bern (Switzerland)

(Received February l lth, 1975)

SUMMARY The thermal properties of the fl-lactoglobulin-water system were investigated by differential scanning calorimetry in the temperature range from --50 to 130 °C. Determination of the heat and temperature of fusion of the absorbed water allowed resolution of the water into four different states. The amounts of water in these states were different for samples before and after heat denaturation. In the case of denatured fl-lactoglobulin, a smaller amount of water with thermal properties different from ordinary water was observed and its total water binding capacity was lower. The thermal stability of fl-lactoglobulin in the water content range from 0 to 0.75 g/g showed a strong dependence on the degree of hydration. A correlation was observed between the changes in the thermal stability of the protein and the changes in the state of the absorbed water. The results are compared with those obtained from similar measurements of other globular proteins and of fibrillar proteins.

INTRODUCTION The interaction of water with fl-lactoglobulin in aqueous solutions has been studied using various criteria and methods. X-ray scattering [1], sorption [2] and density measurements [3, 4] have been used to determine the amount of associated water but these techniques were not adequate to study its physical properties. Other techniques such as calorimetry [5] and dielectric measurements [6] have been used to determine both the amount and the physical properties of this strongly interacting water. Most of these investigations were carried out in dilute solution. No attempt has been made to study both the amount and the state of absorbed water over the whole range of water activity. The thermal denaturation of fl-lactoglobulin solutions has also been extensively studied by various techniques such as polarimetry [7], electrophoresis [8], light scattering [9] or chromatography [10], but there are no reports on the influence of hydration upon the thermal stability of this protein. Both the different states of associated water, and the changes of the thermal stability of the macromolecule upon hydration can be observed by measuring the heat effects accompanying the melting of the absorbed water and the thermal denaturation of the protein.

335 The present work was undertaken to study the mutual interactions between fl-lactoglobulin and water in the water activity range from 0.05 to 1, by differential scanning calorimetry and gravimetric sorption measurements. The results are compared with those obtained from similar measurements of other globular proteins and of fibrillar proteins. EXPERIMENTAL

Materials A commercial preparation of bovine fl-lactoglobulin (B.D.H. Chem. Ltd, lot No. 1545760)was purified by gel chromatography [l 1]. The purified solution was dialyzed against distilled water and lyophilized. The water content of the samples used for calorimetric measurements was adjusted either by equilibration in humidostats at appropriate relative humidities [13] or, for higher water contents, by directly adding water to the lyophilized preparations. The denatured form of fl-lactoglobulin used to obtain water sorption data was prepared by heating an aqueous solution of the protein for 1 h at 90 °C. The absence of a denaturation peak in a DSC scan indicated a complete and irreversible denaturation of the protein after this treatment. The following calorimetric standards were used: indium (Perkin-Elmer), naphtalene (Suchardt, 99.99 ~), benzoic acid (Merck, 99.98 ~o) and benzene (Merck, Uvasol).

Calorimetry A Perkin-Elmer Model DSC-2 differential scanning calorimeter, equipped with a subambiant kit and cryostate (Intracooler II), was used for the thermal measurements. The measurements of the fusion of associated water and the heat denaturation of the protein involved two different modes of operation. (A) In the case of water fusion studies, it is known that the thermal history of the sample under investigation affects the shape and area of the endothermal melting peak [14]. The following procedure was found to give reproducible results: 5-8 mg of wet fl-lactoglobulin were packed and hermetically sealed in volatile sample pans (Perkin-Elmer) and cooled in the calorimeter cell with a scan rate of --0.31 °C/min to --25 °C. At this temperature the samples were allowed to stand for 10-12 h in order to obtain an equilibrium state in the freezable water fractions. The samples were then cooled to --50 °C and measured at a programmed rate of 2.5 °C/min. Subsequent to the low temperature scan the samples were heated above ambient temperature at a rate of 5 °C/min until the thermal denaturation was complete. After heat denaturation the capsules were allowed to stand for 2 days in order to attain the new sorption equilibrium. The denatured samples were then investigated as described above. As judged from the absence of a denaturation peak in a DSC scan, no renaturation occurred during the equilibration period. Empty containers and covers were placed in the reference pan. (B) For denaturation studies, the calorimeter cell was cooled with tap water which allowed operation of the instrument at temperatures as low as 25 °C. 10-15-mg samples were packed in sealable pans and scanned at a rate of 5 °C/min until the denaturation process was complete. Compared with slower scan rates, this increase in temperature was found to be preferable in terms of noise level and baseline, thus

336 giving greater accuracy in enthalpy determinations. The heat capacity of the reference cell was balanced using sealed pans which contained an appropriate amount of water. The water content and dry mass of the samples were determined by puncturing the lid of the capsule, and drying to constant weight over P205. The instrumental accuracy of temperature readings was ~- 0.2 °C. For broad m a x i m a in the thermograms, the accuracy of the reported temperatures is somewhat lower (0.5 °C). Determinations of heats of fusion of water and heats of denaturation were based on peak area measurements which account for the baseline shift due to changes in specific heat [15, 16]. Experimental precision for enthalpy determinations was ~ - 3 ~ .

Water sorption data Water sorption by /3-1actoglobulin before and after heat denaturation was determined gravimetrically. The equilibrium water contents were measured by the atmospheric isopiestic technique, using an apparatus developed by G~il and Bolliger [17, 18]. The values of equilibrium water content are based on a dry weight obtained over P20 5 under atmospheric pressure. Constant dry weights and equilibrium water contents were achieved after 7 and 3 days, respectively. Weighings were made using a Sartorius Model 2405 microbalance. The coefficient of variation of the reported values of equilibrium water content is 0.5 ~ . RESULTS

Heat and temperature of fusion of adsorbed water The melting peaks of water associated with fl-lactoglobulin were recorded in the temperature range --50 to 10 °C before and after thermal denaturation of the

7

c

,E, i~, s

I

r,

b

ii ii o

~

-30

2b

'

-~0

a

6

Fig. 1. Thermograms of hydrated fl-lactoglobulin in the temperature range - 30 to 5°C. - - - - - before; - - - - - , after thermal denaturation. Water contents (g/g) and sample weights (mg) are: (ai 0.398, 7.652; (b) 0.492, 10.010; (c) 0.859, 2.097•

337 protein. The water content of the samples ranged from 0.2 to about 4.5 g per g of dry protein. Fig. 1 shows typical thermograms obtained in three different water content ranges. The shape of the heat absorption curves shows a strong dependence on water content and is different before and after protein denaturation. Up to a threshold water content of 0.29 ± 0.02 g/g no measurable heat absorption was observed in the recorded temperature range, indicating that the sorbed water was unfreezable. Above this threshold value, in the range 0.29 to about 0.45 g/g, a broad endothermal peak was observed (Fig. la). As judged from the measurable departure from the baseline, this peak started in the vicinity of --27 °C and was complete near 0 °C. The temperature of the peak maximum shifted from --10 to --5 °C when the water content was raised from 0.3 to 0.45 g/g. On addition of more water a second water fusion peak was observed around 0 °C (Fig. lb). Further increase in the amount of water affected mainly the size and shape of the second peak at 0 °C. The peak of the low melting water fractions, now superimposed on the dominant peak at 0 °C, is recognized by a shoulder on the low temperature side of the latter (Fig. lc). The differences between thermograms obtained before and after protein denaturation were most pronounced at water contents below 0.7 g/g. Both the decrease of heat absorption below --3 °C and the significant increase of heat absorption at 0 °C indicate that bound water was released from the protein during the heat treatment. A smaller melting point depression is observed for samples containing the denatured form of the protein. The integral heats of fusion of water were calculated from the areas of the fusion peaks and plotted against water content in Fig. 2. The values obtained before denaturation and above a water content of 0.70 -t- 0.05 are well represented by a straight line. The line shown is the least squares regression for the experimental points. Its slope gave a value of 75.9 cal/g for the incremental heat of fusion of the corresponding water, which may be compared with 79.7 cal/g for the heat of fusion of

140120100-

~ "~

80-

60-

40-

20-

o

o'5

11o

1'5

210

2'~

WARE, CONTENT[0/0] Fig. 2. Measured heat of fusion of water associated with fl-lactoglobulin as a function of water content.

338 0,5-

~04c:3

~

J

0.3

o2 %

;:z=

'='0,1

A , ~ - ~

0.2 '-

~_.~A_. ,6....~.._A ~A "~' "~t~ '

0.4 '

'

06'

'

0'8

'

P/Po

Fig. 3. Water sorption isotherm for/4-1actoglobulin at 25 °C. O, before; II, after thermal denatur ation; A, difference between the two sorption isotherms. bulk water. At water contents below 0.7 g/g, however, a change of slope indicates an amount of water different from the bulk liquid with respect to its heat of fusion. The experimental points in the water content range 0.3 to 0.7 also lie close to a straight line. The slope of the line corresponds to an incremental heat of fusion of 56 cal/g. The values obtained after protein denaturation were not significantly different from those obtained before denaturation.

Sorption isotherms Fig. 3 shows sorption isotherms of/%lactoglobulin obtained at 25 °C in the relative water vapor pressure range P/Po ~ 0 to 0.98. The difference between the sorption capacities before and after heat denaturation is significant above P/Po = 0.4 and is most pronounced in the P/Po range from 0.93 to 1. The threshold value of water content for the freezable water fractions, 0.29 g/g, corresponds to the amount of water absorbed at a relative water vapor pressure of approx. 0.93. In other words, the water attached to fl-lactoglobulin in the P/Po range from 0 to 0.93 was found to be unfreezable. Thermal denaturation of [3-lactoglobulin The thermal denaturation of fl-lactoglobulin was accompanied by an endothermal heat effect. Fig. 4 shows typical thermograms obtained with the hydrated protein. The following quantities were measured in the thermograms and plotted as a function of water content in Fig. 5: the total heat involved in the denaturation process (A Qd), the temperature of maximum heat absorption (T~), and the half width of the denaturation peak (AT1/2). All of the three functions show a marked dependence on the degree of hydration of the protein. The resulting curve for the T~ values indicates that the denaturation temperatures for/3-1actoglobulin follow an exponential equation, similar to that found by

339

T

-g F.

,E, a:

~ 6'0

a

7'o

eb 9'0 1~0 TE.,,,,T0,, ;°el Fig. 4. Thermograms of hydrated fl-lactoglobulin in the temperature range 60 to 105 °C. Water contents (g/g) and sample weights (mg) are: (a) 0.176, 6.275; (b) 0.439, 8.562; (c) 4.591, 17.710.

Flory and Garrett [19] for other proteins. According to these authors the Td values depend mainly on the volume fraction of solvent in the protein and reflect no detailed solvent-protein interactions. The dependence on water content of the Td values becomes very small at water contents higher than about 0.7 g/g. The mean value above 0.7 g/g is 80.3 -6 0.2 °C. In the same water content region, the incremental heats of fusion of the absorbed water reaches a constant value (see Fig. 2). The heat of denaturation (A Qd), similarly to the denaturation temperatures (Ta), remains constant within the experimental error at water contents higher than approx. 0.7 g/g. The mean value in the water content range from 0.7 to 3.0 g/g is 3.44 -/- 0.02 cal/g. At water contents lower than 0.7 g/g, A Qd decreases gradually to a value as low as 1.8 g/g. A Qd at water contents lower than about 0.15 g/g could not be measured, because the denaturation reaction occurs in the temperature region of thermal decomposition.

12, 10 .~TI h =

=

.

_-..-o

2-

,120

i,oo -

~

..

L i Jit

2-

0

80

"A~id

'

'

'

'

0.5

'

'

'

'

|.0'

'

'

'

'

1.5'

'

'

'

WATEA C0,TE.T E,lo]

Fig. 5. Heat of denaturation (AQd), temperature of denaturation (Td) and half width of the denaturation peak (A 7",) of fl-lactog]obulin as a function of water content.

340 The curve obtained for the/JTll 2 values also reaches a plateau around a water content of 0.7 g/g (5.32 -5_ 0.07 °C). At lower water contents, however, the variations of AT1~2 with water content were basically different when compared with those observed for Td and A Qd. The dT1/2 function has a minimum around a water content of 0.45 g/g. Presumably this minimum reflects a change in the mechanism of the denaturation process [20-22]. DISCUSSION The calorimetric determination of the heat and temperature of fusion of water associated with fl-lactoglobulin allowed resolution of the water into four states. These states were similar to those observed in other protein-aqueous systems [13, 23, 24] and may be described as (1) non-freezable water, (2) freezable water with both heat and temperature of fusion different from ordinary water, (3) freezable water with heat of fusion lower than that of bulk water, and (4) water indistinguishable by differential scanning calorimetric methods from bulk water. It is interesting to note that the threshold water content for the freezable water fractions, 0.29 g/g, does not differ significantly from that observed for various other water-protein systems. Albumins [25], collagen [14], keratin [23] and caseins [24] all showed a similar amount of non-freezable water. Furthermore, a recent N M R study of frozen protein solutions [26] and various other investigations using different criteria and methods, revealed a change of the physical properties of the attached water around a water content of 0.3 g/g [2, 6, 26]. Haly and Snaith [14, 23] when studying water-collagen and water-keratin systems concluded that this change is caused by the formation of clusters of water. Gurney [27] put forward the concept of different shells of water about solutes and Lumry [28, 29] has recently applied this concept with certain restrictions to water-protein systems. Others have put forward the concept of primary and secondary hydration water as opposed to the bulk water [30]. In these interpretations the water existing near the protein surface is characterized by different degrees of interactions with the solute. The so-called A shell or primary hydration water is strongly bound to hydrogen-bonding sites of the protein molecule. The B shell or secondary hydration water contains water molecules which ambivalently interact with the A shell and the normal bulk water which constitutes the C shell. The unfreezable (state 1) and the freezable water fractions (states 2 and 3) with physical properties different from the bulk liquid may well correspond to such A- and B-type domains, respectively. Upon thermal denaturation a decrease in the quantity of water in states 2 and 3 and an increase of the amount of water in state 4 was observed (see Fig. 1). The same phenomenon has been reported for egg albumin [25]. This leads one to the conclusion that changes in protein conformation involve the water in states 2 and 3. This is consistent with Lumry's [28, 29] view of water-protein interactions, where he would expect that the most interesting part of protein hydration to involve the B shell. It seems likely that unfolding of proteins affects mainly the water in states 2 and 3. Further evidence for this conclusion arises from the results obtained with caseins [24]. Caseins which are known to have an insignificant amount of ordered structures [31 ], revealed an accordingly small amount of water in these states. The data available on the change in water binding capacity on protein de-

341 naturation are contradictory. Privalov and Mrevlishvili [25] for globular and fibrillar proteins found a small positive change in hydration on denaturing. However, the sorption isotherms for fl-lactoglobulin as shown in Fig. 3 clearly indicate a decrease in water absorption after denaturation. The apparent decrease in water binding observed in sorption measurements after heat denaturation, was explained by Lewin [32] as being due to changes other than deconformations, since unfolding of proteins should result in water uptake. Turning to the thermostability of fl-lactoglobulin, it was shown in Fig. 5 that A Qd, Td and AT1~2 remained constant within experimental errors above a water content of approx. 0.7 g/g. This water content was also found to be the threshold value for the water in state 4, i.e. water indistinguishable by calorimetry from ordinary water. This indicates that hydration which is essential for stabilizing the spacial structure of fl-lactoglobulin in water is complete is approx. 75 % (w/w) of water is associated with the protein. Removal of this structural water leads to a decrease in the heat of denaturation (AQd), presumably because of the decreasing number of intramolecular hydrogen bonds [20]. Privalov et al. [20] compared the denaturation enthalpy of a number of globular proteins with the number of intramolecular hydrogen bonds and found a good correlation. In the case of fl-lactoglobulin, unfortunately, such detailed calculations are not possible at present, because the contribution of the different reactions involved in denaturing to the measured heat effect, as well as the spacial structure of the protein, are unknown. The minimum in the tiT1~ 2 function, which reflects a change in the mechanism of the denaturation process, was found around a water content of 0.45 g/g. The same water content was found to be the upper threshold value for the water in state 2, e.g. water with both heat of fusion and temperature of fusion different from the bulk liquid. It has been shown that dT1/z values may be used in favorable cases to obtain quantitative information on the mechanism and the kinetics of the process [20-22]. In the case of the thermal denaturation of fl-lactoglobulin, such detailed calculations are not meaningful, because of the irreversibility of the reactions involved [12]. Summarizing the results we may conclude that hydration of fl-lactoglobulin involves basically three steps: (1) Sorption of water molecules to hydrophilic sorption sites (water content range from dryness to about 0.29 g/g). These water molecules (which would correspond to Gurney's A shell) cannot participate in a normal ice lattice without rupture of hydrogen bonds with the protein and are thus unfreezable. (2) In the water content range from 0.29 to about 0.45 g/g a second shell of water is absorbed, which freezes in the temperature range from 0 °C to approx. --27 °C and which has a heat of fusion of about 56 cal/g. (3) From 0.45 g/g to about 0.7 g/g, a third layer of water is absorbed which freezes in the vicinity of 0 °C but interacts strongly enough with the protein or the adjacent layers of water to have a lower heat of fusion than bulk water (approx. 56 cal/g). Steps 2 and 3 might well correspond to the absorption of Gurney's B-type shell. At water contents higher than 0.7 g/g, specific hydration is complete and bulk water is absorbed. The division of the hydration process into different steps does not imply discontinuous changes. It is evident from Figs 2 and 5 that the hydration-dependent changes occur gradually. We are aware of the fact that the relationship between protein stability and the physical state of absorbed water is mainly of an empirical nature in the case of fl-lactoglobulin. A quantitative interpretation of the calorimetric results must await

342 exact knowledge of the spatial structure of this protein a n d of the forces c o n t r i b u t i n g to its stability. ACKNOWLEDGMENTS The authors wish to t h a n k Dr A. Lukesch for providing samples of purified /4-1actoglobulin. Valuable suggestions by Dr Madeleine Lfischer, University of Berne, a n d linguistic assistance by D r H. Susi, Eastern Regional Research Center, Philadelphia, are gratefully acknowledged. REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

Pessen, H., Kumosinski, T. F. and Timasheff, S. N. (1971) J. Agric. Food Chem. 19, 698-702 Bull, H. B. and Breese, K. (1968) Arch. Biochem. Biophys. 128, 488-502 McMeekin, T. L., Groves, M. L. and Hipp, N. J. (1954) J. Am. Chem. Soc. 76, 407-413 Lee, J. C. and Timasheff, S. N. (1974) Biochemistry 13, 257-265 Berlin, E., Kliman, P. G. and Pallansch, M. J. (1970) J. Coll. Interf. Sci. 34, 488-494 Buchanan, T. J., Haggis, G. H., Hasted, J. B. and Robinson, B. G. (1952) Proc. R. Soc. London, A 213, 379-391 Sawyer, W. H., Norton, R. S., Nichol, L. W. and McKenzie, G. H. (1971) Biochim. Biophys. Acta 243, 19-30 Sawyer, W. H. (1968) J. Dairy Sci. 51, 323-329 Kresheck, G. C., Van Winkle, Q. and Gould, I. A. (1964) J. Dairy Sci. 47, 126-130 Nakanishi, T. and Wada, Y. (1972) J. Dairy Sci. 21, 108-112 McKenzie, H. A. (1971) Milk Prot. Vol. 2, p. 267, Academic Press, New York McKenzie, H. A. (1971) Milk Prot. Vol. 2, pp. 316-318, Academic Press, New York Gal, S. (1967) Die Methodik der Wasserdampf Sorptionsmessungen, pp. 35-40, Springer, Berlin Haly, A. R. and Snaith, J. W. (1971) Biopolymers 10, 1681-1699 Adam, G. and Miiller, F. H. (1967) Kolloid-Z. 192, 29-33 Privalov, P. L. and Khechinashvili, N. N. (1974) J. Mol. Biol. 86, 665-684 Robinson, R. A. and Sinclair, D. A. (1934) J. Am. Chem. Soc. 56, 1830-1835 Gal, S. (1971) J. Food Sci. 36, 800-803 Flory, P. J. and Garrett, R. R. (1958) J. Am. Chem. Soc. 80, 4836-4845 Privalov, P. L., Khechinashvili, N. N. and Atanasov, B. P. (1971) Biopolymers 10, 1865-1890 Koch, E. (1973) Angew. Chem. 85, 381-390 Hanbaba, P., Jiintgen, H. and Peters, W. (1968) Ber. Bunsenges. Phys. Chem. 72, 554-562 Haly, A. R. and Snaith, J. W. (1969) Biopolymers 7, 459-474 Riiegg, M., L~scher, M. and Blanc, B. (1974) J. Dairy Sci. 59, 387-393 Privalov, P. L. and Mrevlishvili, G. M. (1967) Biofizika 12, 22-29 Ramirez, J. E., Cavanaugh, J. R. and Purcell, J. M. (1974) J. Phys. Chem. 78, 807-810 Gurney, R. (1962) Ionic Processes in Solutions, Dover Publications, New York Lumry, R. (1974) Ann. N.Y. Acad. Sci. 227, 471-485 Lumry, R. (1973) J. Food Sci. 38, 744-755 L~scher, M., R/iegg, M. and Schindler, P. (1974) Biopolymers 13, 2489-2503 Herskovits, T. T. (1966) Biochemistry 5, 1018-1026 Lewin, S. (1974) Displacement of Water and its Control of Biochemical Reactions, pp. 186-191, Academic Press, London