Biochimica et Biophysica Acta, 400 (1975) 17-23

© Elsevier Scientific Publishing Company, Amsterdam -- Printed in The Netherlands BBA 37082 I N F R A R E D SPECTROSCOPY OF T H E W A T E R VAPOR SORPTION PROCESS OF CASEINS

MAX ROEGG and HEINZ H.~NI Federal Dairy Research Institute and Federal Research Institute for Agricultural Chemistry and Pollution Control, CH-3097 Liebefeld, Bern (Switzerland)

(Received October 15th, 1974)

SUMMARY Infrared spectra of as-, fl- and micellar casein were studied at relative water vapor pressures (P/Po) ranging from 0 to 0.98. The samples were prepared as selfsupporting films by evaporating concentrated aqueous suspensions of the caseins under study. An infrared cell and a vacuum apparatus were constructed which allowed exposure of the casein films either to vacuum or to sorbate vapor. Following the increase in intensities of the OH and O2H absorption bands during hydration, a sigmoid-shaped curve was observed, similar to the type II isotherm usually obtained by gravimetric sorption measurements. The pronounced frequency and intensity changes in the amide I, II and III bands in thep/po range from 0 to about 0.10 lead to the conclusion that water molecules are already attached to the peptide repeat unit at very low humidities. Based on calculations of the amount of polar groups per casein molecule it was shown that much less than one water molecule per polar group is needed to cause these significant spectral changes.

INTRODUCTION Polypeptides and proteins exhibit several characteristic absorption bands in the infrared region. These characteristic bands are commenly labeled amide A, amide B, amide I, II, III etc. [1]. Each one of these bands is associated with a particelar type of vibration of the peptide repeat unit [2]. A considerable amount of information can be deduced from the study of the splitting and positions of these amide bands. Thus, it was possible to characterize the conformation of several proteins [3-5]. In some favorable cases, specific interactions between protein and water molecules upon stepwise hydration were also observed [6, 7]. Until recently, hydration of only fibrillar proteins has been studied by infrared methods because of difficulties in preparing the thin protein films required for this purpose. The present communication describes a procedure for obtaining self-supporting films of proteins and reports an infrared study of the adsorption of normal and deuterated water vapor by various caseins.

18 EXPERIMENTAL Samples of bovine a~- and fl-casein were provided by Professor Ch. Alais, University of Nancy, France. Both preparations were dialyzed against distilled water and lyophilized. Micellar casein was prepared by high-speed centrifugation from skim milk as reported earlier [8]. A simple technique for the preparation of self-supporting films which has been used for inorganic colloidal suspensions was adapted for the casein solutions [9]: 0.2 ml (0.7 ~o, w/w) aqueous casein solutions were dried at room temperature on 0.1 mm thick polyethylene strips, held on glass slides by capillary forces from a water film. After evaporation of the solvent, the transparent protein film was removed by drawing the polyethylene strip over the edge of the glass plate. The film thickness normally used allowed 5 - 1 0 ~ transmitted radiation for the most intense infrared absorption band (amide I at 1650 cm-l). This film thickness was preferable for intensity studies of the bands which were of major interest (amide A, amide I, amide ll). The casein films were introduced in a cell fitted with AgCI windows. The cell was of a similar design to that described by Angell and Schaffer [10] and connected to a universal glass apparatus equipped with suitable accessories such as a trap (cooled in liquid nitrogen), a flask (containing degassed sorbates), a Pirani gauge head and various stopcocks allowing the cell to be connected either to vacuum or to sorbate vapor. A vacuum of approx, l0 -3 Tort was obtained using a rotatory oil pump. To study the adsorption of normal water vapor by casein, the free films were dried in the cell at 25 °C for 48 h in vacuo. Subsequently, the films were equilibrated with introduced water vapor, the vapor pressure of which was controlled by H2504 solutions of appropriate concentrations [l I]. An equilibriation time of 24 h was sufficient for all values of relative vapor pressure, as judged by the reproduction of the integrated intensity of the amide A band centered around 3300 cm -t. Prior to drying and stepwise hydration with ZHzO vapor, hydrogen-deuterium exchange was achieved by several sorption and desorption cycles of 2H20 vapor at P/Po - 0.90. The desired 2H20 vapor pressures were obtained by adjusting the sorbate temperature by means of a cryostat [12]. The degree of deuteration was judged by the intensity of the residual amide A band. As in the studies with water, an equilibration time of 24 h in the 2H20 atmosphere was sufficient to obtain constant integral band intensities of the amide A' band at 2400 c m - ~. (Absorption bands of deuterated proteins are conventionally labeled A', !', II' etc. [1].) The infrared spectra from 4000 to 600 cm -1 were obtained using a Beckman IR-20 grating spectrometer. A cell with AgCl windows was placed in the reference beam and filled with an appropriate amount of sorbate vapor. A time-dependent reproducible water desorption due to the heating effect of the infrared beam was only noted at the end of a scan, i.e. below approx. 1300 c m - ' . The relative shape of the intensity curves, however, did not change. Because of the complex overlapping of the protein absorption bands, the intensity measurements were based on semiquantitative methods similar to those proposed by other authors [l 3, 14]. Linear baselines for the estimation of the changes in intensity were drawn by connecting points of the spectrum which showed constant transmittance during the whole sorption experiment. In the case of the amide A and B bands around 3300cm -1, which are superimposed on CH-stretching modes at

19

'°°t "t

~6040-

:4.

~

"

I

20,

i

,

x./

,

II

i

,

3000

i

,

,

I

i

i

1

1

2000 WAVE NUMI~R (cm'l)

i

,

,

i

i

i

,

i

1000

Fig. I. Infrared spectrum of fl-casein as a function of relative water vapor pressure. Curves of increasing intensities correspond to P/Po values of O, 0.11, 0.46 and 0.98. Only the curves for P/Po = 0 and 0.98 are shown in the 2000-600cm-~ region. about 2900 cm -~ [16], the baseline was drawn as a tangent to the 2400 and 3800 cm -1 region. Relative intensity changes in the amide A band region (amide A + amide B) were obtained by integrating the peak areas in steps of 20 cm- 1 using the approximation Z" log lo/l.A*,. The intensity changes of the CH-stretching bands were considered to be extremely small during the sorption experiment and were therefore included in this integration. The baseline for the determination of absorbance changes in the lower wavenumber range was drawn as tangent to the 1300 and 1800 cm -~ regions. RESULTS

Water sorption studies Infrared spectra of a~-,/3- and micellar casein from 4000 to 600 cm- 1 were obtained at relative water vapor pressures of approx. 0.025, 0.05, 0.20, 0.30, 0.45, 0.60, 0.75, 0.85, 0.90 and 0.98, the exact values being determined after each experiment. Fig. 1 shows the spectra of fl-casein equilibrated at various relative humidities. The spectra of a~- and micellar casein differed only in the spectral range from 1000 to about 1200 cm -1. In the spectra of caseins, this range is dominated by absorptions due to phosphate groups [15]. It is evident from Fig. 1 that water sorption causes spectral changes over almost the whole region of the recorded spectrum. The most pronounced and significant frequency shifts were observed in the amide A, amide I and amide II regions. Fig. 2 shows the frequency shifts ofthese bands as a function of water vapor pressure. Consistent with their main vibration nature, the amide A and amide I bands shifted to lower and the amide II band to higher frequencies upon hydration. However, the amide A band was shifted at relative vapor pressures higher than about 0.6 to higher frequencies, because of the increasing intensity of the OH absorption on the high frequency side of the amide A band. OHstretching bands appeared at lower vapor pressures (P/Po < 0.3) as a shoulder on the high frequency side of the amide A band near 3460 cm -t. At higher humidities this shoulder moved to about 3380 cm -1 and a second shoulder was recognized in the vicinity of 3540 c m - L Smaller, but significant, frequency shifts of 8-10 cm -t were observed for the amide III band at about 1240 cm -~, the band centering near 1070

20 a)

b)

•

•

c) •

O [ 0

3280"

~ 200

-~ , ~

1520

.o,- .

.

o

.

.010.

41"- •

~

coo" } ,oo

~

r,

-'= • •

1511~

0.5

P/e,

0

P/~

P/e,

Fig. 2. Spectral changes of main infrared absorption bands of fl-casein as a function of relative water vapor pressure. (a) Frequency shifts of amide A, amide I and amide I1 bands. (b) Changes in absorbance of amide I, II, II1 and COO- bands. (c) Integral band intensity of amide A band region (amide A -!- amide B); increase calculated as percentage of area at P/Po= 0. The diameter of the circles represents the estimated standard deviation. cm -t (mainly phosphate-stretching modes [15]), and the band at 1445 cm -1 (which is probably associated with CH-bending modes [6]). The integrated intensity of the absorption centering around 3 3 2 0 c m -~ (amide A -- B bands) was plotted against the relative water vapor pressure in Fig. 2c. The resulting sigmoid-shaped curve may be compared with the type II isotherm obtained by gravimetric sorption measurements [8, 17]. The intensity increase of the amide A band upon hydration was markedly asymmetric. From dryness to about P/Po = 0.05, the broadening of the band occured mainly on the low frequency side. Increasing the vapor pressure from P/Po = 0.05 to about 0.6 affected mainly the high frequency side. At higher humidities, the band broadened symmetrically probably because o f the dominating role of the intense O H bands which arise from the large a m o u n t of attached water molecules. The relative absorbance changes o f the band at 1395 c m - t which is associated with side chain C O 0 - groups [18, 19], and the bands associated with vibrations of the peptide repeat unit, amide I, I1 and 111, are shown in Fig. 2b. The four curves show similar shapes at low and medium P/Po values. A pronounced increase in absorbance in the range of P/Po = 0 to about 0.1 is followed by a plateau region.

2HzO sorption studies The results obtained by studies of the adsorption of 2H20 vapor on deuterated casein films were consistent with those described above. Again, the differences between the spectra and the spectral changes o f the three caseins under investigation were not significant. Fig. 3 shows infrared spectra of as-casein at three different 2H20 vapor pressures. The band at 3300 cm -~ indicated a small a m o u n t o f unexchanged hydrogen atoms. The intensity of this residual amide A band remained constant d u r i n g t h e w h o l e sorption experiment. It was thus not possible under the experimental conditions (25 °C, P/Po = 0.98) to achieve complete deuteration.

21 100-

~0"

!

..../"'-'

80-

:.... ".....,. 20-

.

.

.

.

I 3000

.

.

.

.

~ 2000

.

.

WAVE N U M 8 ~

.

.

.

.

.

.

.

10~00

(cm-t)

Fig. 3. Infrared spectrum of deuterated as-casein at different ~H~O vapor pressures. Curves of increasing intensities correspond to p/Po values of 0, 0.30 and 0.95. Feature at 2000cm -~ is of instrumental origin.

In agreement with the isotope effect, two r-O2H shoulders around 2460 and 2570 cm -~, corresponding to the r-OH shoulders at 3380 and 3540 cm -1, were observed upon sorption of 2H20 on the casein films. Because of the absence of an amide B' component and other overlapping bands in the 2200 c m - ~ region, the asymmetric intensity change of the amide A' band on hydration was more pronounced than that of the amide A band in the corresponding experiments using normal water. The amide !' band which is observable in the "window region" of the sorbate spectrum, showed a frequency shift of about 10 cm -~ to lower frequencies, and a peak absorbance curve with a plateau between P/Po ----0.12 and 0.9. Thus, in the case of the sorption of normal water, the overlapping of the OH absorption bands with the amide I band had only a small influence on the spectral behaviour of the latter band as shown in Fig. 2. The amide II' band, however, showed a different behaviour when compared with the corresponding band of the non-deuterated samples. The frequency shift to higher wavenumbers was only 10 cm -1, compared with about 20 cm -1 for the amide 1I band. In addition, the changes in intensity were less pronounced. This can be explained by the change in the coupling nature of the band upon deuteration [20]. Increased intensity of the band around 1200 cm -1 (6-O2H [21]) and a double band centering near 1040 and 1070 cm-1 (phosphate groups [15]) are the most significant spectral changes in the region below 1300 c m - L The absorbance values of the &O2H band followed a sigmoid-shaped curve when plotted against increasing P/Po values. The two other bands showed nearly a linear dependence on the 2H20 vapor pressure. DISCUSSION Following the increase in intensities of the amide A and amide A' infrared absorption bands of caseins, a sigmoid-shaped curve was observed, similar to the type II isotherm obtained by gravimetric sorption measurements. This result is in contrast with that reported by other authors who investigated the water binding of collagen

22 as a function of relative humidity [6]. Although collagen exhibits a similar type I1 isotherm [22], these authors found no significant changes in the spectra of collagen films when the relative humidity was raised from 75 to 9 5 ~ . The failure to observe an intensity increase for the OH-absorption bands might be caused by the different experimental conditions (the films were deposited on windows and the equilibration periods were shorter). On the basis of the frequency shifts and intensity changes of the casein absorption bands due to polar groups, it can be shown by the infrared method that all these groups, including the peptide repeat unit, participate simultaneously in primary water sorption at low humidities. Moreover, the strongest hydrogen bonds between the polar sorption sites and water molecules are already formed at about P/Po - 0.1. This is clearly recognized by the pronounced spectral changes of the bands associated with NH or CO vibrations at low vapor pressures, followed by weaker changes at medium and high P/Po values. Other authors calculated the water sorption capacities of proteins assuming that water molecules are attached to polar side chain groups only [23, 24]. Although they found reasonable correlations between their predicted and observed waterbinding capacities, infrared spectroscopy revealed evidence for a participation of the peptide groups in water sorption at low humidities. The P/Po value of about 0.1, where the strongest hydrogen bonds between polar sorption sites and water molecules are formed, corresponds to a water content of about 4-5~o (w/w) [8, 17]. According to the known primary structures, there are various polar groups in addition to the peptide groups, located in the side chains, such as charged or uncharged NH, C O 0 and PO4 groups [25]. A calculation based on the assumption that only one water molecule is attached to each polar side chain group leads to a water content of approx. 7.5 ~ . Provided that one water molecule is also held by each peptide group, a total water content of 22.5 ~o is obtained. These calculations lead to the conclusion that considerably less than one water molecule per polar group is needed to cause the pronounced frequency and intensity changes of the infrared absorption bands which are observed in the P/Porange from 0 to about 0.1. The changes observed in the absorption bands associated with CH vibrations are consistent with those reported by Susi et al. [6]. These authors concluded from the high frequency shift of the band at 1440 c m - ' that CH to O hydrogen bonds are formed at high water vapor pressures. However, the overlapping by bands associated with carboxyl stretching around 1395 cm -1 [18, 19], which show a strong intensity increase at high humidities, might be the origin of the spectral changes rather than C H . . . O hydrogen bonds. This intensity increase similar to that of the amide A band at high water vapor pressures may be a result of swelling of the casein which opens up previously bonded, hence inaccessible, sites. ACKNOWLEDGMENTS The authors wish to thank Professor Ch. Alais, University of Nancy, for providing samples of a~- and fl-casein. Valuable suggestions by Dr H. Susi, Eastern Regional Research Center, Philadelphia, and linguistic assistance by Dr P. J. Wood, Food Research Institute, Ottawa, are gratefully acknowledged.

23 REFERENCES 1 Miyazawa, T. (1962) in Polyamino Acid, Polypeptides, and Proteins (Stahmann, M. A., ed.), pp. 15-43, University of Wisconsin Press, Madison 2 Miyazawa, T., Shimanouchi, T. and Mizushima, S. (1958) J. Chem. Phys. 29, 611-616 3 Krimm, S. (1962) J. Mol. Biol. 4, 528-535 4 Susi, H., Timasheff, S. N. and Stevens, L. (1967) J. Biol. Chem. 242, 5460-5466 5 Timasheff, S. N., Susi, H. and Stevens, L. (1967) J. Biol. Chem. 242, 5467-5473 6 Susi, H., Ard, J. S. and Carroll, R. J. (1971) Biopolymers 10, 1597-1604 7 Chirgadze, Y. N., Veriyaminov, S. Y. and Zimont, S. L. (1969) in Water in Biological Systems (Kayushin, L. P., ed.), pp. 51-53, Consultants Bureau, New York 8 R/.iegg, M., Liischer, M. and Blanc, B. (1974) J. Dairy Sci. 57, 387-393 9 Farmer, V. C., Russel, J. D. and Ahlrichs, J. L. (1968) Int. Congr. Soil Sci. Trans. 3, 101-108 10 Angell, C. L. and Schaffer, P. C. (1965) J. Phys. Chem. 69, 3463-3465 11 Gb,I, S. (1967) Die Methodik der Wasserdampf-Sorptionsmessungen, pp. 33-38, Springer-Verlag, Berlin 12 Landolt-B6rnstein (1960) Gleichgewichte Dampf-Kondensat und osmotische Ph/inomene (Sch~ifer, K. and Lax, E., eds), II Band, 2 Tell, Bandteil a, b. 63, Springer-Verlag, Berlin 13 Blout, E. R., de Loz6, C. and Asadourian, A. (1961) J. Am. Chem. Soc. 83, 1895-1900 14 Ehrlich, S. H. and Bettelheim, F. A. (1963) J. Phys. Chem. 67, 1954-1960 15 Susi, H. (1969) in Structure and Stability of Biological Macromolecules (Timasheff, S. N. and Fasman, D., eds), p. 647, Marcel Dekker, New York 16 Susi, H., Ard, J. S. and Carroll, R. J. (1971) J. Am. Leather Chem. Assoc. 66, 508-554 17 Berlin, E., Anderson, B. A. and Pallansch, M. J. (1969) J. Phys. Chem. 73, 303-307 18 Ehrlich, G. and Sutherland, G. B. B. (1954) J. Am. Chem. Soc. 76, 5268-5273 19 Lenormant, H. and Blout, E. R. (1953) Nature 172, 770-773 20 Hallam, H. E. (1969) Spectrochim. Acta 25, 1785-1789 21 Bellamy, L. J. (1958) The Infrared Spectra of Complex Molecules, pp. 55-62, John Wiley, New York 22 Bull, H. B. (1944) J. Am. Chem. Soc. 66, 1499-1507 23 Pauling, L. (1945) J. Am. Chem. Soc. 67, 555-557 24 Bull, H. B. and Breese, K. (1968) Arch. Biochem. Biophys. 128, 488-496 25 Mercier, J. C., Grosclaude, F. and Ribadeau Dumas, B. (1972) Milchwissenschaft 27, 403-408