IR and Raman Spectroscopic Studies of the Interaction of Trehalose with Hen Egg White Lysozyme

Transcription

1 R and Raman Spectroscopic Studies of the nteraction of Trehalose with Hen Egg White Lysozyme P. S. BELTON and A. M. G1 AFRC nstitute of Food Research, Norwich Laboratory, Norwich Research Park, Colney, Norwich NU4 7UA, UK SYNOPSS nfrared and Fourier transform Raman spectra are reported for dried mixtures of trehalose and lysozyme. The Raman spectra show effects on the protein amide band and some sugar hands that are not present when the components are dried separately. Comparison of ir spectra with those published previously show significant differences. t is concluded that these arise because of differences in the extent of drying of the moisture, and that, contrary to some claims, vibrational spectroscopy does not so far show any clear evidence of specific trehalose/protein interactions and that results may be interpreted in terms of entrapment of water within the mixture John Wiley & Sons, nc. NTRODUCTON The disaccharide commonly called trehalose is composed of two (1 + 1) -linked a,a units of glucopyranose and is one of the family of three trehaloses: a,a, a,& and P,p. This nonreducing disaccharide is widespread in nature, occurring in yeasts, higher fungi, insects,' bacteria, and some plants and algae.3 Trehalose is found in large quantities in the so-called resurrection plants Craterostigma plantagineum and Selaginella lepidophylla in which it is known to protect the plants against the effects of heat and dehydration. These plants grow in the deserts of South America and are subjected to high temperatures (up to 60 C) and extremely dry conditions without suffering damage. These plants can be maintained dry, apparently dead, for up to 50 years, during which time no metabolic processes are detected." Upon rehydration, however, the plants' metabolic activity restarts and its appearance changes rapidly. The presence of trehalose is believed to be responsible for protecting the cell structure at low water levels and the exact mechanism of interaction is the subject of many current investigat,i~ns.~-' Studies have been carried out into the phenomenon of anhydrobiosis (the ability to survive dehydration and rehydration processes) in the pres- Riopolymers, Vol. 34, (1994) 6' 1994 dohn Wiley & Sons, nc. CCC /94/ ence of the sugar trehalose, in systems such as lipid bilayer~,~~' micelles,6 and proteins? n this paper we report on a Raman and ir investigation of the interaction of lysozyme and trehalose. To the best of our knowledge, Raman studies have not been reported before. We compare our results with previously published data' and show some important differences in observation and interpretation. EXPERMENTAL The a,a-trehalose dihydrate from Sigma Chemical Co. was authenticated by solid state nmr," x-ray diffraction, lo Fourier transform ir (FTR) and FT Raman spectroscopy, and was used as received. n this work, this form will be denoted as form of trehalose. The anhydrous trehalose, form 111, was produced by heating up the dihydrate in the MAS rotor to 358 K, during variable temperature experiments, and its nature was checked by the resulting CPMAS spectrum. An intermediate trehalose form, form 11, was obtained by leaving form under vacuum for 48 h, at 323 K; the loss of water from the dihydrate sample was monitored gravimetrically, showing the loss of 2.0 water molecules per disaccharide molecule. Form 1 of trehalose is thus a second anhydrous form. 957

2 958 BELTON AND GL All dry lysozyme and trehalose powder samples were dried to constant weight under vacuum. The lysozyme / trehalose mixture samples were prepared by dissolving together both compounds in water and leaving to stir for 1-2 h. The solutions were freeze dried for approximately 12 h for all samples. Deuterated samples of lysozyme and lysozyme / trehalose for use in FTR studies were prepared by dissolution in D20 inside a "dry-box'' apparatus. KBr disks were produced from freeze-dried powders. Deuterated solutions, which were 20% w/v protein, were analyzed immediately after preparation in a BaFz liquid cell. Hydrated samples were prepared by leaving the dry powders over salt solutions of known relative humidity for 3-4 days. For Raman studies, the hydrated samples were prepared in the same way as for infrared. However, samples tended to lose water as a result of heating by the incident light. To avoid this, samples were placed in a capillary with a drop of water placed in a side that was used to maintain a high relative humidity. Under these circumstances there may have been a slight increase in the water contents of the samples during the experiment. n most studies of lysozyme /trehalose mixtures the weight ratio 1 : 10 was used since this ratio was reported to produce the biggest influence on the protein ir amide bands." FTR studies were carried out on a FTSGO Digilab spectrometer with a liquid RamanShift cu~l Figure 1. FT Raman spect,ra of (a) trehalose dihydrate, (b) dry lysozyme powder, arid ( c) dry lysozyme /trehalose 1 : 10 powder. a) Ramanshift cni' Figure 2. Expansions of the cm-' regions of the FT Raman spectra of (a) dry lysozyme powder and (b) dry lysozyme/trehalose 1 : 10 powder. nitrogen cooled MCT detector. A resolution of 4 cm-' was used and 1024 scans were acquired for all samples. FT Raman spectra were run on an Bruker FTSlO6 spectrometer, with a liquid nitrogen cooled Ge detector and a Neodymium YAG 1064 nm laser. Raman spectrum were acquired at 4 cm-' resolution and 500 scans were co-added unless indicated otherwise. A laser power of 95V was used. RESULTS The FT Raman spectra of crystalline a,a-trehalose dihydrate, dry lysozyme, and dry lysozyme 1 : 10 mixture are shown in Figure 1. ntensity from the trehalose signal does not overlap the conformationally sensitive amide and amide 1 of the protein in the cm-l region. Below 1550 cm-' the trehalose signal is intense, and is not significantly affected by the very weak protein bonds below 1200 crn-l. The sugar bands at 1000 to 800 cm-' are intense and there is no significant protein contribution to them. Figure 2 shows an expansion of the amide and 1 regions of dry lysozyme and dry lysozyme with trehalose. The signal from the lysozyme trehalose mixture is slightly narrower than that from the dry lysozyme and shows evidence of some structure. The narrowing is observed even in 1 : 1 lysozyme trehalose mixtures (Table ). On the addition of water to the mixture, the line splits and narrows. The line is significantly narrower than for lysozyme alone under similar hydration conditions. Although the signal-to-noise ratio in the hydrated spectra was worse than that in the dry spectra, the observations were reproducible and were still observed when acquisition conditions were altered to reduce signal to noise ratios (Figure 3 ). We therefore believe these are real effects.

3 NTERACTON OF TREHALOSE WTH HEN EGG WHTE LYSOZYME 959 Table Changes in the Amide Region of the Raman Spectrum of Lysozyme upon the Presence of Trehalose and upon Hydration Frequency Line Width Sample (cm-') (cm-') Dry lysozyme L:T~: L:T1: L: T 1 : Lysnzyrne, 20% water L : T 1 : 10, 18% water 1675, a L : T represents a mixture of lysozyme and trehalose The FT Raman spectra of various forms of trehalose and lysozyme /trehalose mixtures in the cm-' region are shown in Figure 4. This region arises solely from sugar bands. t is clear that the signal from the lysozyme/trehalose mixture differs from trehalose dihydrate and trehalose dried under similar conditions to the lysozyme mixture. The region cm-' has been assigned to bending vibrations of equatorial C ~ H of a anomers in the 4C1 conformation." n trehalose the monomeric hydrogens are equatorial so changes in this region reflect the occurrence of conformational changes affecting the anomeric hydrogens. The spectra, therefore, clearly show that the formation of the lysozyme/trehalose mixture causes conformational changes that are not observed in trehalose alone. The water content of each sample in Figure 4 was estimated by ratioing the intensity of the water bands at 3250 and 2665 cm-' to the CH peak at Ramanshift a-' Figure 4. FT Raman spectra of (a) trehalose form, (b) trehalose form dried under vacuum at room temperature, for 48 h, (c) trehalose form 11, and (d) lysozyme/ trehalose 1 : 10, dried under vacuum at room temperature for 48 h. cm-'. These calculations showed that sample (c), prepared by drying under vacuum for 48 h at 50 C, was the driest material. Samples (b) and (d), the trehalose and trehalose/protein mixtures dried at room temperature, were the next driest, and the dihydrate was the wettest. The ratio of the intensities of the peaks at 842 and 863 cm-' declines slightly from sample (b) to sample (c), indicating that water loss causes a decrease in the ratio of intensities. n sample (d) the ratio of intensities is much lower. This cannot arise from water loss from the mixture since the sample is wetter than sample (c). The effect must therefore be due to the presence of the protein. These results together with those from the amide regions suggest that the formation of the dry lysozyme/trehalose mixture causes conformational changes in both the lysozyme and trehalose. A report by other workersg has suggested that the effect of trehalose is to establish a hydrogen-bonding ndtwork similar to that in hydrated lysozyme. These conclusions were made on the basis of three observations of the ir spectrum of a dry bovine serum albumin / trehalose mixtures and lysozyme /trehalose mixtures. The observations were as follows: 1x RRHRN SHFT CH-1 Figure 3. F'1' Haman spectra of hydrated samples of (a) lysozyme, 20% water and (b) lysozyme/trehalose 1 : 10, 18% water at 4 cm-', and at (c) 8 cm-' resolution. a loss of intensity in the trehalose bands at cm-' in trehalose/protein mixtures compared to trehalose freeze dried alone, the features of the lysozyme/trehalose mixtures in the amide 1 region are identical to those in hydrated lysozyme, and a band at 1583 cm-' assigned to hydrated

4 960 BELTON AND GL ZSE wavenumber* Figure 5. FTR spectra of (a) crystalline and (b) freezedried trehalose. carbonyl groups was observed in the hydrated proteins and in lysozyme/ trehalose but not in the lysozyme dried alone. t was also observed that the amide bands of the dry protein and the trehalose protein mixtures were almost identical. Our results are not consistent with these observations. Figure 5a shows the spectrum of authenticated trehalose dihydrate. The cm-' region is very similar to that described in Ref. 9 as arising from a trehalose sample freeze dried alone. The freeze-dried material used here shows no such bands (Figure 5b). We conclude, therefore, that the freeze-drying process used in the previous report resulted in the formation of trehalose dihydrate and that freeze drying the protein /trehalose mixture resulted simply in the dehydration of the trehalose. The spectrum obtained by us of the dry lysozyme/ trehalose mixture is identical to that observed for trehalose dried alone. We conclude, therefore, that spectra from this region do not yield evidence of trehalose /protein interactions. The spectra obtained by us in the amide and 1 regions (Table 11) differ from those reported in Ref. 9. n our spectra the amide band of the hydrated protein is similar to that observed by Carpenter and Crowe,' but has a maximum at cm-' as opposed to their value of' Our value is closer to that observed by Poole and Finney, '* who measured a peak maximum at 1634 cm-' in a 30% w/v solution. When the lysozyme/trehalose mixture is dried, this peak shifts to cm-l, in agreement with the reported value and broadens to the same width as reported. When lysozyme is freeze dried alone, we observe an increased broadening to 67 cm-l, as opposed to the reported value of 58 cm-', and a peak position of cm-', substantially in agreement with the reported value. The amide 1 region of the dried-alone protein is substantially different in intensity and position to the hydrated protein and the mixture. The mixture and hydrated material show weaker bands and have some similarity in shape. However, it is not possible, because of the problem of the rolling baseline, to be conclusive about the degree of similarity. n Ref. 9 the bands are shown to be very similar indeed. n the hydrated protein we observe a substantial peak at cm-', in agreement with reported value of 1583 cm-'. However, in contradistinction to the observation of Carpenter and Crowe' we do not observe a band in the mixture at 1583 cm-' or there abouts. DSCUSSON The Raman spectra of dry trehalose/lysozyme mixtures show clear evidence of effects that are not observed in either component when dried alone. There is a narrowing of the amide band of the protein relative to that in the dried alone material and distinctly different effects are observed when these are partially hydrated. Similarly, the trehalose bands in the region cm-' show distinct behavior for the mixed material. On the basis of their evidence, Carpenter and Crowe' concluded that the interaction between trehalose and protein was a hydrogen bonding one with features in common with the interaction of water with protein. Our results do not substantiate this conclusion. Our spectra differ substantially from those of Carpenter and Crowe and suggest that the effects they observe in trehalose are not due to the presence of protein, but to the presence of water. Our spectra show no effect of protein on trehalose in the region described by those authors. We observe effects of trehalose on the amide region in the FTR spectra consistent with the effects we observe in the Raman spectra and in contrast to those that Carpenter and Crowe observed in dry protein and dry mixture. A major difference arises from the band at 1583 cm-'. n common with the previous report, we do observe this band in the hydrated material, but in contrast we do not see it in the mixture. Our results in the amide 1 region of the ir specta are not as clearcut as those reported, but we do see a shift toward the behavior observed in the hydrated material. Taken together our results from Raman and ir do clearly indicate that trehalose has an effect

5 NTERACTON OF TREHALOSE WTH HEN EGG WHTE LYSOZYME 961 Table 1 Effects of Hydration and of Trehalose in the R Amide and 1 Bands of Lysozyme (Line Widths Are ndicated Between Brackets) Frequency (cm-') From Ref Dry lysozyme (58) (broader than in 2. below) 2. Hydrated lysozyme (44) " Dry L : T 1 : (57) v.wa From this Work (67) (38) a (55) , v.wb a Assigned to hydrated carboxyl groups. v.w = very weak. Unassigned. on the protein and that the protein in turn affects trehalose. However, the spectral changes observed are not by themselves sufficient to support the conclusions of Carpenter and Cr~we.~ n interpreting the spectral changes it is important to take into consideration the fact that water is still present in the system. Some, at least, of the observations reported previously, such as the spectra of the freezedried trehalose, may arise from variable water contents in the system. Even if water content is not reproducible, it is not zero, and an explanation of the spectra consistent with the suggestions of Timasheff l3 may explain the observed phenomena. The basis of this hypothesis is that in solution the sugar is excluded from the proximity of the protein by water. We propose that this relationship remains in the dried state. The effect of trehalose is to concentrate what water remains close to the protein, thus causing its spectra to resemble more the hydrated than the dehydrated protein. The converse is that the trehalose is in a dehydrated form. As the system dries, water concentrates at the protein sugar interface and may be trapped there by glass formation. f the water content is low enough there will be some sugar protein contacts, which will affect conformation and thus sugar spectra; in any case it is reasonable to assume that at the protein /sugar interface there will be some conformational changes. These may account for the changes seen in the trehalose spectra. f the trehalose has a role in trapping water at the protein interface, the degree of hydration may depend upon the detailed history of the sample. This may explain some of the differences between this work and that previously reported. AMG thanks the Calouste Gulbenkian Foundation, Portugal, for provision of a grant. REFERENCES 1. Elbein, A. D. (1974) Adu. Carbohydr. Chem. Biochem. 30, Barnett, J. A. (1981) Adu. Carbohydr. Chem. Biochem. 39, Roser, B. (1991) Trends Food Sci. Technol., Lee, C. W. B., Waugh, J. S. & Griffin, R. G. (1986) Biochemistry 25, Park, Y. S. & Huang, L. (1992) Biochim. Biophys. Acta 1124, Ram, P., Mazzola, L. & Prestegard, J. H. (1989) J. Am. Chem. SOC. 111, Sharon, N. & Lis, H. (1982) Mol. Cell Biochem. 42, Lee, C. W., Das Gupta, S. K., Mattai, J., Shipley, G. G., Abdel-Mageed, 0. H., Makrtyannis, A. & Griffin, R. G. (1989) Biochemistry 28, Carpenter, J. F. & Crowe, J. H. (1989) Biochemistry 28, Belton, P. S. & Gil, A. M., Carbohydrate Res., submitt e d. 11. Parker, F. S. (1983) Applications of nfrared, Raman and Resonance Raman Spectroscopy in Biochemistry, Plenum Press, New York, London. 12. Poole, P. L. & Finney, J. L. (1984) Biopolymers 23, Timasheff, S. N. (1982) in Biophysics of Water, Franks, F. & Mathias, S., Eds., Wiley, New York, pp Received July 20, 1993 Accepted December 13, 1993