CRYOBIOLOGY
29, 599-606 (1992)
Hydration
of Oligosaccharides: Anomalous of Trehalose
H. KAWAI, M. SAKURAI,
Hydration
Ability
Y. INOUE, R. CHUJ6, AND S. KOBAYASHI*
Department of Biomolecular Engineering, Tokyo Institute of Technology, 12-l 0-okayama 2-chome, Meguro-ku, Tokyo 152, Japan; and *National Food Research Institute, Ministry of Agriculture, Forestry Fisheries, Kan-nondai, Tukuba, Ibaragi 305, Japan
and
The dissacharide trehalose extensively exists in anhydrobiotic organism and is considered to play an important role in preserving the integrity of biomembrane. However, the preserving mechanism remains unclear. In this report, we examine the hydration abilities of trehalose and several oligosaccharides composed of a-o-glucopyranosyl residues. The unfrozen water fraction per molecule was determined from differential scanning calorimetry measurements of their aqueous solutions. Further, the NMR relaxation time of the natural abundance “0 of water is measured for several saccharide solutions. These results indicate that trehalose has the highest hydration ability among the saccharides studied. In other words, trehalose can effectively lower the mobility of water molecules hydrogenbonded with saccharides. It is thus reasonable that, among the disaccharides studied, trehalose exhibits the maximum stabilizing effect on the bilayer structure of lipid whose acyl chains are bonded with each other by the apolar interaction, because the apolar interaction is strengthened with the stabilization of the surrounding water structure. o 1992 Academic press, inc.
Anhydrobiotic organisms such as brine observed for dehydrated lipid bilayer in the shrimp cysts and certain nematodes can presence of trehalose. In addition, molecular mechanics calculation has been applied survive even after complete dehydration. This ability of these organisms is thought to to build molecular models for the interaccorrelate with the protective effect of disac- tion between bilayer surface and a few charide trehalose on the cell membranes kinds of saccharide including trehalose (22). However, no clear-cut conclusion has (18). In model systems, using phospholipid bilayers, it was shown that trehalose main- been reached for why trehalose is most actains the functional and structural integri- tive among the various mono- and disacties of bilayeres to a considerable extent charides studied so far. during dehydration and subsequent rehyIt is believed that the two kinds of stress dration processes (for reviews see (9-11)). vectors, drying and freezing, are fundamenMoreover, trehalose is known to be one of tally different, requiring different mechathe most effective cryoprotectants against a nisms to explain the protection of lipid bifreeze-thaw cycle, while a number of other layers by saccharides and other stabilizing saccharides such as sucrose and maltose solutes (11). For drying phospholipid bilayshare these protective properties (9-11,20). ers, one plausible adaptive mechanism inThere is extensive evidence for the interac- volves the hypothesized replacement of tion of trehalose with model lipid bilayers; lost water by a “compatible solvent” like for example, the depression of gel-liquid saccharides, leading to the direct interaccrystalline transition temperature (6-8, 24) tion between the saccharide and bilayer and some characteristic changes in IR (5) surface. On the other hand, such a direct and solid-state NMR (16, 17) spectra were interaction alone is insufficient to account for the various experimental data obtained from freezing phospholipid bilayers (11). In Received October 7, 1991; accepted May 27, 1992 599 001l-2240/92 $5.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
600
KAWAI
unfrozen solution, an entropically driven mechanism may play an important role in protecting lipid bilayers. Most saccharides are known to exhibit the structure-making effect on their surrounding water (25). Thus, when they are present in the aqueous dispersion of phospholipids, the apolar interaction between the acyl chains is strengthened, consequently contributing to the stabilization of bilayer structure. Our recent NMR study on unilamellar vesicles of dipalmitoylphosphatidylcholine (21) has demonstrated that trehalose increases the packing density of lipid bilayer, supporting the operation of the above mechanism in unfrozen solution. Similar results have been obtained from a study of the thermotropic phase transition of unilamellar vesicles in trehalose-containing solution (12). In such a case that the entropically driven mechanism is predominant, the hydration characteristics of saccharides, such as the so-called hydration number, may correlate with the preservation effect on the membrane integrity. However, little information has been obtained on the hydration of trehalose except for a recent study by Uedaira et al. (27). In the present study, we measured the spin-lattice relaxation time of the natural abundance “0 of water in aqueous solution and number of unfrozen water molecules in the presence of saccharides. On the basis of these data, we show that trehalose has a considerably high hydration ability. MATERIALS
AND
METHODS
Sample preparation. Trehalose, sucrose, maltose, and maltotriose were purchased from Nakarai Chemical Co., Tokyo, and maltotriose, maltohexaose, maltoheptaose, and maltopentaose were from Eusuikouseikou, Yokohama. All the reagents were used without further purification; nominal purities were over 99%. For the measurements of differential scanning calorimeter (DSC), all saccharides were hydrated with distilled water at room
ET AL.
temperature, then they were sealed in calorimeter pans. The saccharide concentrations were 0.1,0.2,0.3, and 0.4 w/w%. DSC measurements were taken without equilibration. For NMR measurements, all saccharides were hydrated in lo-mm-diameter NMR tubes at room temperature with pH 3.79 buffer solutions. Buffer solution was prepared with p-hydroxybenzoic acids. All the measurements were carried out in a wider range of saccharide concentrations from 0 to 3.0 mol kgg’. DSC measurements. DSC thermograms of all samples were recorded on the heat flux DSC (SEIKO DSC-20) in heating direction using a SEIKO SSC/580 thermal controller. All samples were cooled to -70°C (cooling rate was about 9”C/min) with liquid nitrogen and then were heated to 80°C (heating rate was S’C/min). Whenever the samples were cooled below o”C, the measurements were performed under an atmosphere of nitrogen. The purpose of this purge gas was to provide a stable environment for the analyzer head and to remove moisture around calorimetric pans. NMR measurements. The “0-NMR measurements were performed with a JEOL GSX-270 spectrometer operating at a “0 frequency of od2n = 36.6 MHz. “0 spin-lattice relaxation times (T,) were measured by the standard inversion-recovery method with the VT-T/~ pulse sequence. The 7r/2 pulse width for “0 nuclei was 21 (*.s. Acquisitions (2048) were accumulated for each of 10 different delay times T. All measurements were performed at 30°C. RESULTS
For all DSC samples prepared above, the enthalpy of fusion, Hr, was measured as a function of saccharide concentration. Figure 1 illustrates the data for the trehalosecontaining solutions. & decreases linearly with an increase in the weight fraction of trehalose, indicating that the number of unfrozen water molecules per trehalose mole-
HYDRATION
0
0.2
0.4
OF OLIGOSACCHARIDES
0.8
0.6
WEIGHT FRACTION
FIG. 1. Heat of fusion of ice vs. weight fraction of trehalose. The weight fraction (w/w) is the ratio of trehalose weight to the total amount of solution.
cule is constant at least in the range of about 0.1 to 0.4 w/w%. By extrapolating this straight line toward Hr = 0, the amount of unfrozen water was calculated. Similar analyses were carried out for the other saccharide solutions. Table 1 summarizes the number of unfrozen water molecules measured per saccharide molecule (N) and per its constituent glucose residue (N/n). The number of unfrozen water molecules for maltose, maltotriose, and maltopentaose examined by Miyajima et al. (20) was 19.5, 23, and 36, respectively, and these TABLE 1 Unfreezable Water and Equatorial OH for Various Saccharides Saccharide
n(e - OH)
n
N
Nin
Trehalose Sucrose Maltose Maltotriose Maltotetraose Maltopentaose Maltohexaose Maltoheptaose
8.0 6.3 7.6 10.6 13.6 16.6 19.6 22.6
2 2 2 3 4 5 6 7
7.95 6.33 6.50 1.67 12.01 16.16 20.86 17.01
3.98 3.17 3.25 2.56 3.00 3.23 3.48 2.43
Note. n(e - OH), numbers of equatorial OH groups per saccharide molecule; n, numbers of monosaccharide units; N, numbers of unfreezable water molecules per saccharide molecule.
601
values were quite different from our data. On the other hand, Bociek et al. (2) reported the number of unfrozen water molecules for maltose to be 6.5, showing agreement with our data of 6.5. It is unclear which factor, experimental conditions or methods for analysis, contributes to these differences. The unfrozen region of the solutions consists mainly of the water molecules directly hydrogen-bonded with the OH groups of saccharides. It is thus expected that the number of unfrozen water molecules is in proportion to the number of water molecules which were involved in the first solvation shell around each saccharide molecule. According to Uedaira et al. (26, 27), the hydration characteristics of saccharides correlate well with the number of equatorial OHS in the equilibrium state. In order to confirm this in the present case including longer oligosaccharides, the data for the equatorial OH are also included in Table 1. The number of equatorial OH was counted according to the following assumptions: (i) the conformation of the constituent saccharide is the same as that for the corresponding monosaccharide, and (ii) the anomeric OH groups contribute to the number of equatorial OH according to the proportion of its anomers; for example, the ratio of the l3 anomer is 0.6 for glucose (1). As shown in Table 1, among the disaccharides used, numbers of both unfrozen water molecules and equatorial OHS correlate well with each other. In the cases of the longer saccharide chains, similar results were obtained. However, in the cases of maltotriose and maltoheptaose, the number of unfrozen water molecules is less than predicted from the number of equatorial OHS. These disagreements will be discussed under Discussion. The relaxation time for pure water, Ty, was 7.8-8.2 ms at 30°C and 7.1 ms at 25°C which are consistent with the data reported by Uedaira et al. (26), Ty = 7.3 ms. The “0 relaxation times were measured
602
KAWAI ET AL.
as a function of saccharide concentrations (Figs. 2 and 3). Ty and T, represent the spin-lattice relaxation times of “0 nucleus in pure water and saccharide solutions, respectively. In a case where the so-called extremely narrowing condition is valid, l/T, is known to be linearly related to the correlation time 7,. As shown in Figs. 2 and 3, the value of ratio TYIT, increases almost linearly with saccharide concentration. This indicates that the saccharides decrease the mobility of water molecules. The strength of such an effect of each saccharide is measured as the slope of the corresponding straight line in Figs. 2 and 3. The results are plotted against the number of equatorial OHS in Fig. 4. The relative order of hydration ability is determined to be maltopentaose > trehalose > sucrose > maltotriose > maltose, although there remains some ambiguity in the order between sucrose, maltose, and maltotriose. In order to estimate the hydration ability of the saccharide in more detail, we calculated the correlation time of water molecules. In the extreme narrowing conditions, the relaxation rate l/T, for “0 nucleus is related to the correlation time T, of molecular reorientation as (26) l/T, = 3/125 (1 + 71~13) (e*qQlh)*7,
[II
0. 5
1. 5
1.0
MOLALITY
2. 0
OF SACCHARIDE.
2.5 kg-'
mol
FIG. 3. TYIT, ratio for “0 nuclei of H,“O in saccharide solutions vs. saccharide molalities. (0) trehalose, (@) maltotriose, (A) maltopentaose.
electrical field gradient and e*qQlh is the quadrupole coupling constant. The q value is usually small (26), so its relating term was omitted in the following calculation. The data for the relaxation time were analyzed according to the so-called two-site model (13). The “0 nuclei are assumed to be distributed between the two motional states, which correspond to the water of hydration sphere and the bulk water, respectively. If the exchange of water molecules between both states is sufficiently fast, the following relation is fulfilled; l/T, = (1 - x,)/T? + +,/T,,,.
121
Using Eq. [l] and [2], Uedaira et al. (26) derived the equations 55.5B = nh (K 7&p - 1) K = (e2qQ,lh)2/(e2qQolh)2 [3]
where q is an assymmetric parameter of the
0
70 2.5
60 ,2 ?. 0 c"
50 it 3 40 Lo 30
1. 5
0 0
0 0
20 0
1.0 1.5 2.0 0.5 MOLALITY OF SACCHAHIDE.
2.5 mol kg-'
FIG. 2. TYIT, ratio for “0 nuclei of H “0 in saccharide solutions vs. saccharide molkties. (0) trehalose, (A) maltose, (0) sucrose.
10 5
7
9
11
u
NUMBER OF EQUATORIAL
15
17 011
FIG. 4. Slopes of plots in Figs. 2 and 3 vs. number of equatorial OH groups.
HYDRATION
603
OF OLIGOSACCHARIDES
where B is the slopes of Ty/T, vs m plots, n,, is the coordination number. The subscripts h and 0 represent the water of the hydration sphere and bulk water, respectively, Both values of e’qQ,/h and e*qQ,lh are assumed to be equal to those of pure water, and the value of K is taken to be 1. Uedaira et al. (26, 27) assumed that the value of n,, are 6, 10, and 14 for mono-, di-, and trisaccharides, respectively, according to Kabayama and Patterson’s model (15). However, this assumption is problematic, because this model does not consider the orientational difference of the individual OH groups of saccharides. As can be understood from Table 1, the equatorial OH is more preferably hydrogen-bonded with the surrounding water than with the axial OH. Thus, this qualitative difference between the equatorial and axial OHS should be reflected in nh in the stage of analyzing the NMR data. In this report, we substitute the number of unfrozen water molecules for &,. Table 2 shows the values for ~JT,o, .55SB, number of unfrozen water molecules, and the number of equatorial OH. In addition, the data obtained by Uedaira et al. (26, 27) are also shown in Table 2. The ratio ~&,o is a measure of the mobilities of the water molecules involved in the nearest neighbor of saccharide. The larger its value, the tighter binding of waters to saccharides in the dynamic process of solution. As shown in Table 2, the value of T,~/T,ofor trehalose was largest, suggesting that this saccharide has the most pow-
erful ability for lowering the mobility of the surrounding water molecules. DISCUSSION
The present results provide clear evidence for the anomalous characteristics of hydration of trehalose. This saccharide has the maximal number of unfrozen water molecules among the disaccharides studied. If the number of unfrozen water molecules is measured per constituent glucose unit, its value exhibits a maximum in trehalose among a series of saccharides studied. In addition to such thermodynamical changes of the aqueous solutions containing the saccharides, the molecular dynamics of water molecules, measured by “0 NMR, is a valuable indicator for estimating the hydration characteristics of the saccharides. The relative correlation time defined above is maximal in trehalose, although the measurements were not carried out for all types of saccharides. This finding indicates that the mobility of water molecules hydrogen-bonded with saccharide OH groups is most effectively lowered by an addition of trehalose in comparison with the other saccharides. On the basis of these results, the OH groups of trehalose are relatively easy to fit into the surrounding water structure. It is thus reasonable that trehalose exhibits the maximum stabilizing effect on the lipid bilayer structure at least among the disaccharides studied, because the apolar interaction is strengthened with the stabilization of the surrounding water structure.
TABLE 2 Several Hydration Characteristics of Saccharides Saccharide
n(e - OH)
55.SB”
55.5B”,’
r,h/r,o
r,h/r,oC
r,hlr,od
Trehalose Sucrose Maltose Maltotriose Maltopentaose
8.0 6.3 7.6 10.6 16.6
48.3 36.8 23.8 35.6 71.0
25.4 25.2 27.2 34.6 -
7.08 6.81 4.66 5.64 5.39
5.83 4.68 3.38 3.54 4.23
3.54 3.52 3.72 3.47 -
N
7.95 6.33 6.50 7.67 16.16
a See the Eq. [3] in the text. b+‘Data from Uedaira et al. (26, 27). ‘Values of n,, are 10, 14, and 22 for di-, tri-, and pentasaccharide, respectively.
604
KAWAI
ET AL.
Recently, Green and Angel1 (14) have re- mental lines for maltose and sucrose soluported glass transition temperatures (T,) tions cross at the concentration of 1.3 mol and phase diagrams for several saccharide- kg-‘. At present, the origin of the above water systems, including trehalose, mal- discrepancy remains unclear. tose, and sucrose. The trehalose-water sysHowever, it should be noted that our tem is distinguished from the others by a data are consistent with the available thersignificantly higher Tg at all water contents modynamic data mentioned above which with a particularly large advantage near the indicate the most powerful hydration ability stoichiometry of one water molecule per of trehalose. glucose unit. The relative order of Tg is The classical Kabayama and Patterson shown to be trehalose > maltose > sucrose model (15) suggests that saccharides have a in increasing temperature. In a sense, this structure-making effect on their surroundtrend is a measure of perturbation of the ing water and thereby stabilize the ice-like solutions and consistent with the above water structure. However, this model does data for the number of unfrozen water mol- not explicitly take into account some imecules. portant factors determining the hydration There are some discrepancies between characteristics of saccharides: namely, the our results and those given by Uedaira’s relative orientation of their OH groups, pioneering work (26, 27) in which the “0 apolar hydration arising from the partial hyrelaxation times and the correlation times drophobicity of saccharides (19, 23), and of water molecules were obtained for vari- conformational effects particularly seen in ous oligsaccharide-containing aqueous so- cases of longer saccharide chains. The rellutions. According to their results, the ative contributions of these effects may be slope of the T, ratio vs. saccharide concen- estimated from the present experimental tration plot as shown in Fig. 2 is smaller in data. As described under Results, the number the trehalose system than in the maltose of unfrozen water molecules correlates one, indicating the more stronger hydration with the number of equatorial OH despite to maltose. The experimental conditions are almost identical with each other except some exceptions such as maltotriose and for the prepared concentration range of sac- maltoheptaose (Table 1). It should be noted charides: the range in our NMR experi- that there exists about one unfrozen water ments for the disaccharides is about 1.5 molecule per equatorial OH. Uedaira et al. times wider. As can be seen from Fig. 2, the (25) have indicated so far that the equatorial T, ratio does not necessarily change lin- OH groups play an important roles in deearly with saccharide concentration, which termining the hydration ability of mainly is similar to the data by Uedaira et al. (26). mono- and disaccharides. The present reHowever, even if a small range, approxi- sults suggest the applicability of this empirmated by a straight line, is selected, our ical rule to longer saccharides. In the cases of maltotriose and maltohepdata always give the larger slope for the trehalose-containing solution than for the taose, the number of unfrozen water molecules was considerably smaller than the other disaccharide solutions. In particular, this is remarkable in the range of higher number of equatorial OH. There must other concentrations where the reliability of T, factors disturbing the formation of hydromeasurements are relatively higher. On the gen bonds between the saccharide and its other hand, the data for maltose and su- surrounding water molecules. The conforcrose obtained from both works are similar mational effect mentioned above may be to each other up to some detailed points; for most important, especially for maltohepexample, as shown in Fig. 2, the experi- taose. According to a recent theoretical cal-
HYDRATION
605
OF OLIGOSACCHARIDES
culation (3) based on Flory’s statistical mechanics for linear polymers, infinitely long linear saccharides composed of glucose units tend to take a helix-like conformation in wcuo with a period of 5-7 units. The formation of the helix would result in reducing the number of OH groups available for hydrogen bonding with the surrounding water molecules. The conformational analysis of oligosaccharides in aqueous solution must be performed to confirm this speculation, including the reason why such a effect occurs specifically in maltotriose and maltoheptaose. Next, we examine the data for the relaxation time T, in detail. The T, data represent the degree of mean mobility of all water molecules in the saccharide solution. This is different from the physical meaning of the correlation time mentioned above, which represents the mean mobility of hydrated waters as understood from the way of derivation (Eq. [3]). A characteristic feature is found in Figs. 2 and 3 and Table 2. Namely, the slope of the T, vs. concentration plot for maltopentaose solution is larger than that for trehalose solution. This means that the average mobility of water molecules in maltopentaose solution is lower than that in trehalose solution when the saccharide concentration is identical in each. The hydroxyl groups of maltopentaose effectively participate in the hydrogen-bonding network with the surrounding water molecules, without being shielded from the aqueous medium, which is consistent with the interpretation on the basis of the DSC data. In the case of trehalose, the number of hydroxyl groups per molecule is smaller than that of maltopentaose, leading to the smaller slope of the T, data. However, as understood from the data on the correlation time, the intrinsic ability to lower the mobility of hydrated water molecules is higher in the hydroxyl groups of trehalose than in those of maltopentaose. In conclusion, it is revealed that trehalose has a considerably higher hydration
ability. The present data alone are insufficient to explain the origin of such a characteristic property of trehalose and to describe fully the hydrogen-bonding network between trehalose and the surrounding water molecules. For this purpose, molecular dynamics and Monte Carlo simulations may be useful. Such an investigation is in progress now. ACKNOWLEDGMENT
We thank Ensuikou-seitou Co., Ltd., Yokohama, for the sample (maltopentaose for the use of NMR measurement). REFERENCES
I. Angyal, S. J. The composition and conformation of sugars in solution. Angew. Chem. Int. Edit. 8, 157-166 (1969). 2. Bociek, S., and Franks, F. Proton exchange in aqueous solutions of glucose. .I. Chem. Sot. Faraday
Trans. I. IS, 262-270 (1979).
3. Brant, D. A., and Christ, M. D. Realistic conformational modelling of carbohydrates. In “ACS Symposium Series 430: Computer Modelling Carbohydrate Molecules” (A. D. French and J. W. Brady, Eds.), pp. 42-68. American Chemical Society, Washington, DC, 1990. 4. Clegg, J. S., Seitz, P., and Hazlewood, F. Cellular responses to extreme water loss: The waterreplacement hypothesis, Cryobiology 19, 306 316 (1982). 5. Crowe, J. H., Crowe, L. M., and Chapman, D. Infrared spectroscopic studies on interaction of water and carbohydrates with a biological membrane. Arch. B&hem. Biophys. 232, 4OMO7 (1984). 6. Crowe, L. M., Womersley, C., Crowe, J. H., Reid, D., Appel, L., and Rudolph, A. Prevention of fusion and leakage in freeze-dried liposomes by carbohydrates. B&him. Biophys. Acta 861, 131-140 (1986). 7. Crowe, L. M., Crowe, J. H., and Chapman, D. Interaction of carbohydrates with dry dipalmitoylphosphatidylcholine. Arch. B&hem. Biophys. 236, 289-296 (1985). 8. Crowe, L. M., Crowe, J. H., Rudolph, A., Womersley, C., and Appel, L. Preservation of freeze-dried liposomes by trehalose. Arch. Biothem. Biophys. 242, 240-247 (1985). 9. Crowe, J. H., Crowe, L. M., Carpenter, J. F., and Aurell Wistrom, C. Stabilization of dry phospholipid bilayers and proteins by sugars. Biochem. J. 242, l-10 (1987).
606
KAWAI
IO. Crowe, J. H., Crowe, L. M., Carpenter, J. F., Rudolph, A. S., Wistrom, C. A., Spargo, B. J., and Anchordoguy, T. J. Interactions of sugars with membranes. Biochim. Biophys. Acta 947, 367-384 (1988). Il. Crowe, J. H., Carpenter, J. F., Crowe, L. M., and Anchordoguy, T. J. Are freezing and dehydration similar stress vectors? A comparison of modes of interaction of stabilizing solutes with biomolecules. Cryobiology 27, 219-231 (1990). 12. Crowe, L. M., and Crowe, J. H. Solution effects on the thermotropic phase transition of unilamellar liposomes. Biochim. Biophys. Acta 1064, 267-274 (1991). 13. Fullerton, G. D., Potter, J. L., and Dornbluth, N. L. NMR relaxation of protons in tissues and other macromolecular water solutions. Magn. Reson. fmaging
1, 209-228 (1982).
14. Green, J. L., and Angel], C. A. Phase relations and vitrification in saccharide-water solutions and the trehalose anomaly. /. Phys. Chem. 93, 2880-2882 (1989). 15. Kabayama, M. A., and Patterson, D. The thermodynamics of mutarotation of some sugars. Can. J. Chem. 36, 563-573 (1958). 16. Lee, C. W. B., Waugh, J. S., and Griffin, R. G. Solid-state NMR study of trehalose/l,2dipalmitoyl-sn-phosphatidylcholine Interactions. Biochemistry 25, 3737-3742 (1986). 17. Lee, C. W. B., Dus Gupta, S. K., Mattai, J., Shipley, G. G., Abdel-Mageed, 0. H., Makriyannis, A., and Griffin, R. G. Characterization of the L, phase in trehalose-stabilized dry membranes by solid-state NMR and X-ray diffraction. Biochemistry 28, 5000-5009 (1989). 18. Madin, K. A. C., and Crowe, J. H. Anhydrobiosis in nematodes: Carbohydrate and lipid me-
ET AL.
19.
20.
21.
22.
23.
24.
25.
26.
27.
tabolism during dehydration. J. Exp. 2001. 193, 335-342 (1975). Miyajima, K., Machida, K., and Nakagaki, M. Hydrophobic indexes for various monosaccharides. Bull. Chem. Jpn. 58, 2595-2599 (1985). Miyajima, K., Tomita, K., and Nakagaki, M. Effect of saccharides on the freezing and thawing of liposome dispersion. Chem. Parm. Bull. 34, 2689-2697 (1986). Nishiwaki, T., Sakurai, M., Inoue, Y., Chujo, R., and Kobayashi, S. Increasing packing density of hydrated dipalmitoylphosphatidylcholine unilamellar vesicles induced by trehalose. Chem. Left., 1841-1844 (1990). Rudolph, B. R., Chandrasekhar, I., Garber B. P., and Nagumo, M. Molecular modelling of saccharide-lipid interactions. Chem. Phys. Lipids 53, 243-261 (1990). Sundari, C. S., Raman, B., and Balasubramanian, D. Hydophobic surfaces in oligosaccharides: Linear dextrins are amhiphilic chains. Biochim. Eiophys. Acfa 1065, 35111 (1991). Tsvetkov, T. D., Tsonev, L. I., Tsvetkova, N. M., Koynava, R. D., and Tenchov, B. G. Effect of trehalose on the phase properties of hydrated and lyophilized dipalmitoylphosphatidylcholine multilayers. Cryobiology 26, 162169 (1989). Uedaira, H. The effect of sugars on the stability of protein structure. Sen-i Gakkaishi 37, 436-443 (1981). Uedaira, H., Ikura, M., and Uedaira, H. Naturalabundance oxygen-17 magnetic relaxation in aqueous solutions of carbohydrates. Bull. Chem. Sot. Jpn. 62, 14 (1989). Uedaira, H., Ishimura, M., Tsuda, S., and Uedaira, H. Hydration of oligosaccharides. Bull. Chem. Sot. Jpn. 63, 3376-3379 (1990).
29, 599-606 (1992)
Hydration
of Oligosaccharides: Anomalous of Trehalose
H. KAWAI, M. SAKURAI,
Hydration
Ability
Y. INOUE, R. CHUJ6, AND S. KOBAYASHI*
Department of Biomolecular Engineering, Tokyo Institute of Technology, 12-l 0-okayama 2-chome, Meguro-ku, Tokyo 152, Japan; and *National Food Research Institute, Ministry of Agriculture, Forestry Fisheries, Kan-nondai, Tukuba, Ibaragi 305, Japan
and
The dissacharide trehalose extensively exists in anhydrobiotic organism and is considered to play an important role in preserving the integrity of biomembrane. However, the preserving mechanism remains unclear. In this report, we examine the hydration abilities of trehalose and several oligosaccharides composed of a-o-glucopyranosyl residues. The unfrozen water fraction per molecule was determined from differential scanning calorimetry measurements of their aqueous solutions. Further, the NMR relaxation time of the natural abundance “0 of water is measured for several saccharide solutions. These results indicate that trehalose has the highest hydration ability among the saccharides studied. In other words, trehalose can effectively lower the mobility of water molecules hydrogenbonded with saccharides. It is thus reasonable that, among the disaccharides studied, trehalose exhibits the maximum stabilizing effect on the bilayer structure of lipid whose acyl chains are bonded with each other by the apolar interaction, because the apolar interaction is strengthened with the stabilization of the surrounding water structure. o 1992 Academic press, inc.
Anhydrobiotic organisms such as brine observed for dehydrated lipid bilayer in the shrimp cysts and certain nematodes can presence of trehalose. In addition, molecular mechanics calculation has been applied survive even after complete dehydration. This ability of these organisms is thought to to build molecular models for the interaccorrelate with the protective effect of disac- tion between bilayer surface and a few charide trehalose on the cell membranes kinds of saccharide including trehalose (22). However, no clear-cut conclusion has (18). In model systems, using phospholipid bilayers, it was shown that trehalose main- been reached for why trehalose is most actains the functional and structural integri- tive among the various mono- and disacties of bilayeres to a considerable extent charides studied so far. during dehydration and subsequent rehyIt is believed that the two kinds of stress dration processes (for reviews see (9-11)). vectors, drying and freezing, are fundamenMoreover, trehalose is known to be one of tally different, requiring different mechathe most effective cryoprotectants against a nisms to explain the protection of lipid bifreeze-thaw cycle, while a number of other layers by saccharides and other stabilizing saccharides such as sucrose and maltose solutes (11). For drying phospholipid bilayshare these protective properties (9-11,20). ers, one plausible adaptive mechanism inThere is extensive evidence for the interac- volves the hypothesized replacement of tion of trehalose with model lipid bilayers; lost water by a “compatible solvent” like for example, the depression of gel-liquid saccharides, leading to the direct interaccrystalline transition temperature (6-8, 24) tion between the saccharide and bilayer and some characteristic changes in IR (5) surface. On the other hand, such a direct and solid-state NMR (16, 17) spectra were interaction alone is insufficient to account for the various experimental data obtained from freezing phospholipid bilayers (11). In Received October 7, 1991; accepted May 27, 1992 599 001l-2240/92 $5.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
600
KAWAI
unfrozen solution, an entropically driven mechanism may play an important role in protecting lipid bilayers. Most saccharides are known to exhibit the structure-making effect on their surrounding water (25). Thus, when they are present in the aqueous dispersion of phospholipids, the apolar interaction between the acyl chains is strengthened, consequently contributing to the stabilization of bilayer structure. Our recent NMR study on unilamellar vesicles of dipalmitoylphosphatidylcholine (21) has demonstrated that trehalose increases the packing density of lipid bilayer, supporting the operation of the above mechanism in unfrozen solution. Similar results have been obtained from a study of the thermotropic phase transition of unilamellar vesicles in trehalose-containing solution (12). In such a case that the entropically driven mechanism is predominant, the hydration characteristics of saccharides, such as the so-called hydration number, may correlate with the preservation effect on the membrane integrity. However, little information has been obtained on the hydration of trehalose except for a recent study by Uedaira et al. (27). In the present study, we measured the spin-lattice relaxation time of the natural abundance “0 of water in aqueous solution and number of unfrozen water molecules in the presence of saccharides. On the basis of these data, we show that trehalose has a considerably high hydration ability. MATERIALS
AND
METHODS
Sample preparation. Trehalose, sucrose, maltose, and maltotriose were purchased from Nakarai Chemical Co., Tokyo, and maltotriose, maltohexaose, maltoheptaose, and maltopentaose were from Eusuikouseikou, Yokohama. All the reagents were used without further purification; nominal purities were over 99%. For the measurements of differential scanning calorimeter (DSC), all saccharides were hydrated with distilled water at room
ET AL.
temperature, then they were sealed in calorimeter pans. The saccharide concentrations were 0.1,0.2,0.3, and 0.4 w/w%. DSC measurements were taken without equilibration. For NMR measurements, all saccharides were hydrated in lo-mm-diameter NMR tubes at room temperature with pH 3.79 buffer solutions. Buffer solution was prepared with p-hydroxybenzoic acids. All the measurements were carried out in a wider range of saccharide concentrations from 0 to 3.0 mol kgg’. DSC measurements. DSC thermograms of all samples were recorded on the heat flux DSC (SEIKO DSC-20) in heating direction using a SEIKO SSC/580 thermal controller. All samples were cooled to -70°C (cooling rate was about 9”C/min) with liquid nitrogen and then were heated to 80°C (heating rate was S’C/min). Whenever the samples were cooled below o”C, the measurements were performed under an atmosphere of nitrogen. The purpose of this purge gas was to provide a stable environment for the analyzer head and to remove moisture around calorimetric pans. NMR measurements. The “0-NMR measurements were performed with a JEOL GSX-270 spectrometer operating at a “0 frequency of od2n = 36.6 MHz. “0 spin-lattice relaxation times (T,) were measured by the standard inversion-recovery method with the VT-T/~ pulse sequence. The 7r/2 pulse width for “0 nuclei was 21 (*.s. Acquisitions (2048) were accumulated for each of 10 different delay times T. All measurements were performed at 30°C. RESULTS
For all DSC samples prepared above, the enthalpy of fusion, Hr, was measured as a function of saccharide concentration. Figure 1 illustrates the data for the trehalosecontaining solutions. & decreases linearly with an increase in the weight fraction of trehalose, indicating that the number of unfrozen water molecules per trehalose mole-
HYDRATION
0
0.2
0.4
OF OLIGOSACCHARIDES
0.8
0.6
WEIGHT FRACTION
FIG. 1. Heat of fusion of ice vs. weight fraction of trehalose. The weight fraction (w/w) is the ratio of trehalose weight to the total amount of solution.
cule is constant at least in the range of about 0.1 to 0.4 w/w%. By extrapolating this straight line toward Hr = 0, the amount of unfrozen water was calculated. Similar analyses were carried out for the other saccharide solutions. Table 1 summarizes the number of unfrozen water molecules measured per saccharide molecule (N) and per its constituent glucose residue (N/n). The number of unfrozen water molecules for maltose, maltotriose, and maltopentaose examined by Miyajima et al. (20) was 19.5, 23, and 36, respectively, and these TABLE 1 Unfreezable Water and Equatorial OH for Various Saccharides Saccharide
n(e - OH)
n
N
Nin
Trehalose Sucrose Maltose Maltotriose Maltotetraose Maltopentaose Maltohexaose Maltoheptaose
8.0 6.3 7.6 10.6 13.6 16.6 19.6 22.6
2 2 2 3 4 5 6 7
7.95 6.33 6.50 1.67 12.01 16.16 20.86 17.01
3.98 3.17 3.25 2.56 3.00 3.23 3.48 2.43
Note. n(e - OH), numbers of equatorial OH groups per saccharide molecule; n, numbers of monosaccharide units; N, numbers of unfreezable water molecules per saccharide molecule.
601
values were quite different from our data. On the other hand, Bociek et al. (2) reported the number of unfrozen water molecules for maltose to be 6.5, showing agreement with our data of 6.5. It is unclear which factor, experimental conditions or methods for analysis, contributes to these differences. The unfrozen region of the solutions consists mainly of the water molecules directly hydrogen-bonded with the OH groups of saccharides. It is thus expected that the number of unfrozen water molecules is in proportion to the number of water molecules which were involved in the first solvation shell around each saccharide molecule. According to Uedaira et al. (26, 27), the hydration characteristics of saccharides correlate well with the number of equatorial OHS in the equilibrium state. In order to confirm this in the present case including longer oligosaccharides, the data for the equatorial OH are also included in Table 1. The number of equatorial OH was counted according to the following assumptions: (i) the conformation of the constituent saccharide is the same as that for the corresponding monosaccharide, and (ii) the anomeric OH groups contribute to the number of equatorial OH according to the proportion of its anomers; for example, the ratio of the l3 anomer is 0.6 for glucose (1). As shown in Table 1, among the disaccharides used, numbers of both unfrozen water molecules and equatorial OHS correlate well with each other. In the cases of the longer saccharide chains, similar results were obtained. However, in the cases of maltotriose and maltoheptaose, the number of unfrozen water molecules is less than predicted from the number of equatorial OHS. These disagreements will be discussed under Discussion. The relaxation time for pure water, Ty, was 7.8-8.2 ms at 30°C and 7.1 ms at 25°C which are consistent with the data reported by Uedaira et al. (26), Ty = 7.3 ms. The “0 relaxation times were measured
602
KAWAI ET AL.
as a function of saccharide concentrations (Figs. 2 and 3). Ty and T, represent the spin-lattice relaxation times of “0 nucleus in pure water and saccharide solutions, respectively. In a case where the so-called extremely narrowing condition is valid, l/T, is known to be linearly related to the correlation time 7,. As shown in Figs. 2 and 3, the value of ratio TYIT, increases almost linearly with saccharide concentration. This indicates that the saccharides decrease the mobility of water molecules. The strength of such an effect of each saccharide is measured as the slope of the corresponding straight line in Figs. 2 and 3. The results are plotted against the number of equatorial OHS in Fig. 4. The relative order of hydration ability is determined to be maltopentaose > trehalose > sucrose > maltotriose > maltose, although there remains some ambiguity in the order between sucrose, maltose, and maltotriose. In order to estimate the hydration ability of the saccharide in more detail, we calculated the correlation time of water molecules. In the extreme narrowing conditions, the relaxation rate l/T, for “0 nucleus is related to the correlation time T, of molecular reorientation as (26) l/T, = 3/125 (1 + 71~13) (e*qQlh)*7,
[II
0. 5
1. 5
1.0
MOLALITY
2. 0
OF SACCHARIDE.
2.5 kg-'
mol
FIG. 3. TYIT, ratio for “0 nuclei of H,“O in saccharide solutions vs. saccharide molalities. (0) trehalose, (@) maltotriose, (A) maltopentaose.
electrical field gradient and e*qQlh is the quadrupole coupling constant. The q value is usually small (26), so its relating term was omitted in the following calculation. The data for the relaxation time were analyzed according to the so-called two-site model (13). The “0 nuclei are assumed to be distributed between the two motional states, which correspond to the water of hydration sphere and the bulk water, respectively. If the exchange of water molecules between both states is sufficiently fast, the following relation is fulfilled; l/T, = (1 - x,)/T? + +,/T,,,.
121
Using Eq. [l] and [2], Uedaira et al. (26) derived the equations 55.5B = nh (K 7&p - 1) K = (e2qQ,lh)2/(e2qQolh)2 [3]
where q is an assymmetric parameter of the
0
70 2.5
60 ,2 ?. 0 c"
50 it 3 40 Lo 30
1. 5
0 0
0 0
20 0
1.0 1.5 2.0 0.5 MOLALITY OF SACCHAHIDE.
2.5 mol kg-'
FIG. 2. TYIT, ratio for “0 nuclei of H “0 in saccharide solutions vs. saccharide molkties. (0) trehalose, (A) maltose, (0) sucrose.
10 5
7
9
11
u
NUMBER OF EQUATORIAL
15
17 011
FIG. 4. Slopes of plots in Figs. 2 and 3 vs. number of equatorial OH groups.
HYDRATION
603
OF OLIGOSACCHARIDES
where B is the slopes of Ty/T, vs m plots, n,, is the coordination number. The subscripts h and 0 represent the water of the hydration sphere and bulk water, respectively, Both values of e’qQ,/h and e*qQ,lh are assumed to be equal to those of pure water, and the value of K is taken to be 1. Uedaira et al. (26, 27) assumed that the value of n,, are 6, 10, and 14 for mono-, di-, and trisaccharides, respectively, according to Kabayama and Patterson’s model (15). However, this assumption is problematic, because this model does not consider the orientational difference of the individual OH groups of saccharides. As can be understood from Table 1, the equatorial OH is more preferably hydrogen-bonded with the surrounding water than with the axial OH. Thus, this qualitative difference between the equatorial and axial OHS should be reflected in nh in the stage of analyzing the NMR data. In this report, we substitute the number of unfrozen water molecules for &,. Table 2 shows the values for ~JT,o, .55SB, number of unfrozen water molecules, and the number of equatorial OH. In addition, the data obtained by Uedaira et al. (26, 27) are also shown in Table 2. The ratio ~&,o is a measure of the mobilities of the water molecules involved in the nearest neighbor of saccharide. The larger its value, the tighter binding of waters to saccharides in the dynamic process of solution. As shown in Table 2, the value of T,~/T,ofor trehalose was largest, suggesting that this saccharide has the most pow-
erful ability for lowering the mobility of the surrounding water molecules. DISCUSSION
The present results provide clear evidence for the anomalous characteristics of hydration of trehalose. This saccharide has the maximal number of unfrozen water molecules among the disaccharides studied. If the number of unfrozen water molecules is measured per constituent glucose unit, its value exhibits a maximum in trehalose among a series of saccharides studied. In addition to such thermodynamical changes of the aqueous solutions containing the saccharides, the molecular dynamics of water molecules, measured by “0 NMR, is a valuable indicator for estimating the hydration characteristics of the saccharides. The relative correlation time defined above is maximal in trehalose, although the measurements were not carried out for all types of saccharides. This finding indicates that the mobility of water molecules hydrogen-bonded with saccharide OH groups is most effectively lowered by an addition of trehalose in comparison with the other saccharides. On the basis of these results, the OH groups of trehalose are relatively easy to fit into the surrounding water structure. It is thus reasonable that trehalose exhibits the maximum stabilizing effect on the lipid bilayer structure at least among the disaccharides studied, because the apolar interaction is strengthened with the stabilization of the surrounding water structure.
TABLE 2 Several Hydration Characteristics of Saccharides Saccharide
n(e - OH)
55.SB”
55.5B”,’
r,h/r,o
r,h/r,oC
r,hlr,od
Trehalose Sucrose Maltose Maltotriose Maltopentaose
8.0 6.3 7.6 10.6 16.6
48.3 36.8 23.8 35.6 71.0
25.4 25.2 27.2 34.6 -
7.08 6.81 4.66 5.64 5.39
5.83 4.68 3.38 3.54 4.23
3.54 3.52 3.72 3.47 -
N
7.95 6.33 6.50 7.67 16.16
a See the Eq. [3] in the text. b+‘Data from Uedaira et al. (26, 27). ‘Values of n,, are 10, 14, and 22 for di-, tri-, and pentasaccharide, respectively.
604
KAWAI
ET AL.
Recently, Green and Angel1 (14) have re- mental lines for maltose and sucrose soluported glass transition temperatures (T,) tions cross at the concentration of 1.3 mol and phase diagrams for several saccharide- kg-‘. At present, the origin of the above water systems, including trehalose, mal- discrepancy remains unclear. tose, and sucrose. The trehalose-water sysHowever, it should be noted that our tem is distinguished from the others by a data are consistent with the available thersignificantly higher Tg at all water contents modynamic data mentioned above which with a particularly large advantage near the indicate the most powerful hydration ability stoichiometry of one water molecule per of trehalose. glucose unit. The relative order of Tg is The classical Kabayama and Patterson shown to be trehalose > maltose > sucrose model (15) suggests that saccharides have a in increasing temperature. In a sense, this structure-making effect on their surroundtrend is a measure of perturbation of the ing water and thereby stabilize the ice-like solutions and consistent with the above water structure. However, this model does data for the number of unfrozen water mol- not explicitly take into account some imecules. portant factors determining the hydration There are some discrepancies between characteristics of saccharides: namely, the our results and those given by Uedaira’s relative orientation of their OH groups, pioneering work (26, 27) in which the “0 apolar hydration arising from the partial hyrelaxation times and the correlation times drophobicity of saccharides (19, 23), and of water molecules were obtained for vari- conformational effects particularly seen in ous oligsaccharide-containing aqueous so- cases of longer saccharide chains. The rellutions. According to their results, the ative contributions of these effects may be slope of the T, ratio vs. saccharide concen- estimated from the present experimental tration plot as shown in Fig. 2 is smaller in data. As described under Results, the number the trehalose system than in the maltose of unfrozen water molecules correlates one, indicating the more stronger hydration with the number of equatorial OH despite to maltose. The experimental conditions are almost identical with each other except some exceptions such as maltotriose and for the prepared concentration range of sac- maltoheptaose (Table 1). It should be noted charides: the range in our NMR experi- that there exists about one unfrozen water ments for the disaccharides is about 1.5 molecule per equatorial OH. Uedaira et al. times wider. As can be seen from Fig. 2, the (25) have indicated so far that the equatorial T, ratio does not necessarily change lin- OH groups play an important roles in deearly with saccharide concentration, which termining the hydration ability of mainly is similar to the data by Uedaira et al. (26). mono- and disaccharides. The present reHowever, even if a small range, approxi- sults suggest the applicability of this empirmated by a straight line, is selected, our ical rule to longer saccharides. In the cases of maltotriose and maltohepdata always give the larger slope for the trehalose-containing solution than for the taose, the number of unfrozen water molecules was considerably smaller than the other disaccharide solutions. In particular, this is remarkable in the range of higher number of equatorial OH. There must other concentrations where the reliability of T, factors disturbing the formation of hydromeasurements are relatively higher. On the gen bonds between the saccharide and its other hand, the data for maltose and su- surrounding water molecules. The conforcrose obtained from both works are similar mational effect mentioned above may be to each other up to some detailed points; for most important, especially for maltohepexample, as shown in Fig. 2, the experi- taose. According to a recent theoretical cal-
HYDRATION
605
OF OLIGOSACCHARIDES
culation (3) based on Flory’s statistical mechanics for linear polymers, infinitely long linear saccharides composed of glucose units tend to take a helix-like conformation in wcuo with a period of 5-7 units. The formation of the helix would result in reducing the number of OH groups available for hydrogen bonding with the surrounding water molecules. The conformational analysis of oligosaccharides in aqueous solution must be performed to confirm this speculation, including the reason why such a effect occurs specifically in maltotriose and maltoheptaose. Next, we examine the data for the relaxation time T, in detail. The T, data represent the degree of mean mobility of all water molecules in the saccharide solution. This is different from the physical meaning of the correlation time mentioned above, which represents the mean mobility of hydrated waters as understood from the way of derivation (Eq. [3]). A characteristic feature is found in Figs. 2 and 3 and Table 2. Namely, the slope of the T, vs. concentration plot for maltopentaose solution is larger than that for trehalose solution. This means that the average mobility of water molecules in maltopentaose solution is lower than that in trehalose solution when the saccharide concentration is identical in each. The hydroxyl groups of maltopentaose effectively participate in the hydrogen-bonding network with the surrounding water molecules, without being shielded from the aqueous medium, which is consistent with the interpretation on the basis of the DSC data. In the case of trehalose, the number of hydroxyl groups per molecule is smaller than that of maltopentaose, leading to the smaller slope of the T, data. However, as understood from the data on the correlation time, the intrinsic ability to lower the mobility of hydrated water molecules is higher in the hydroxyl groups of trehalose than in those of maltopentaose. In conclusion, it is revealed that trehalose has a considerably higher hydration
ability. The present data alone are insufficient to explain the origin of such a characteristic property of trehalose and to describe fully the hydrogen-bonding network between trehalose and the surrounding water molecules. For this purpose, molecular dynamics and Monte Carlo simulations may be useful. Such an investigation is in progress now. ACKNOWLEDGMENT
We thank Ensuikou-seitou Co., Ltd., Yokohama, for the sample (maltopentaose for the use of NMR measurement). REFERENCES
I. Angyal, S. J. The composition and conformation of sugars in solution. Angew. Chem. Int. Edit. 8, 157-166 (1969). 2. Bociek, S., and Franks, F. Proton exchange in aqueous solutions of glucose. .I. Chem. Sot. Faraday
Trans. I. IS, 262-270 (1979).
3. Brant, D. A., and Christ, M. D. Realistic conformational modelling of carbohydrates. In “ACS Symposium Series 430: Computer Modelling Carbohydrate Molecules” (A. D. French and J. W. Brady, Eds.), pp. 42-68. American Chemical Society, Washington, DC, 1990. 4. Clegg, J. S., Seitz, P., and Hazlewood, F. Cellular responses to extreme water loss: The waterreplacement hypothesis, Cryobiology 19, 306 316 (1982). 5. Crowe, J. H., Crowe, L. M., and Chapman, D. Infrared spectroscopic studies on interaction of water and carbohydrates with a biological membrane. Arch. B&hem. Biophys. 232, 4OMO7 (1984). 6. Crowe, L. M., Womersley, C., Crowe, J. H., Reid, D., Appel, L., and Rudolph, A. Prevention of fusion and leakage in freeze-dried liposomes by carbohydrates. B&him. Biophys. Acta 861, 131-140 (1986). 7. Crowe, L. M., Crowe, J. H., and Chapman, D. Interaction of carbohydrates with dry dipalmitoylphosphatidylcholine. Arch. B&hem. Biophys. 236, 289-296 (1985). 8. Crowe, L. M., Crowe, J. H., Rudolph, A., Womersley, C., and Appel, L. Preservation of freeze-dried liposomes by trehalose. Arch. Biothem. Biophys. 242, 240-247 (1985). 9. Crowe, J. H., Crowe, L. M., Carpenter, J. F., and Aurell Wistrom, C. Stabilization of dry phospholipid bilayers and proteins by sugars. Biochem. J. 242, l-10 (1987).
606
KAWAI
IO. Crowe, J. H., Crowe, L. M., Carpenter, J. F., Rudolph, A. S., Wistrom, C. A., Spargo, B. J., and Anchordoguy, T. J. Interactions of sugars with membranes. Biochim. Biophys. Acta 947, 367-384 (1988). Il. Crowe, J. H., Carpenter, J. F., Crowe, L. M., and Anchordoguy, T. J. Are freezing and dehydration similar stress vectors? A comparison of modes of interaction of stabilizing solutes with biomolecules. Cryobiology 27, 219-231 (1990). 12. Crowe, L. M., and Crowe, J. H. Solution effects on the thermotropic phase transition of unilamellar liposomes. Biochim. Biophys. Acta 1064, 267-274 (1991). 13. Fullerton, G. D., Potter, J. L., and Dornbluth, N. L. NMR relaxation of protons in tissues and other macromolecular water solutions. Magn. Reson. fmaging
1, 209-228 (1982).
14. Green, J. L., and Angel], C. A. Phase relations and vitrification in saccharide-water solutions and the trehalose anomaly. /. Phys. Chem. 93, 2880-2882 (1989). 15. Kabayama, M. A., and Patterson, D. The thermodynamics of mutarotation of some sugars. Can. J. Chem. 36, 563-573 (1958). 16. Lee, C. W. B., Waugh, J. S., and Griffin, R. G. Solid-state NMR study of trehalose/l,2dipalmitoyl-sn-phosphatidylcholine Interactions. Biochemistry 25, 3737-3742 (1986). 17. Lee, C. W. B., Dus Gupta, S. K., Mattai, J., Shipley, G. G., Abdel-Mageed, 0. H., Makriyannis, A., and Griffin, R. G. Characterization of the L, phase in trehalose-stabilized dry membranes by solid-state NMR and X-ray diffraction. Biochemistry 28, 5000-5009 (1989). 18. Madin, K. A. C., and Crowe, J. H. Anhydrobiosis in nematodes: Carbohydrate and lipid me-
ET AL.
19.
20.
21.
22.
23.
24.
25.
26.
27.
tabolism during dehydration. J. Exp. 2001. 193, 335-342 (1975). Miyajima, K., Machida, K., and Nakagaki, M. Hydrophobic indexes for various monosaccharides. Bull. Chem. Jpn. 58, 2595-2599 (1985). Miyajima, K., Tomita, K., and Nakagaki, M. Effect of saccharides on the freezing and thawing of liposome dispersion. Chem. Parm. Bull. 34, 2689-2697 (1986). Nishiwaki, T., Sakurai, M., Inoue, Y., Chujo, R., and Kobayashi, S. Increasing packing density of hydrated dipalmitoylphosphatidylcholine unilamellar vesicles induced by trehalose. Chem. Left., 1841-1844 (1990). Rudolph, B. R., Chandrasekhar, I., Garber B. P., and Nagumo, M. Molecular modelling of saccharide-lipid interactions. Chem. Phys. Lipids 53, 243-261 (1990). Sundari, C. S., Raman, B., and Balasubramanian, D. Hydophobic surfaces in oligosaccharides: Linear dextrins are amhiphilic chains. Biochim. Eiophys. Acfa 1065, 35111 (1991). Tsvetkov, T. D., Tsonev, L. I., Tsvetkova, N. M., Koynava, R. D., and Tenchov, B. G. Effect of trehalose on the phase properties of hydrated and lyophilized dipalmitoylphosphatidylcholine multilayers. Cryobiology 26, 162169 (1989). Uedaira, H. The effect of sugars on the stability of protein structure. Sen-i Gakkaishi 37, 436-443 (1981). Uedaira, H., Ikura, M., and Uedaira, H. Naturalabundance oxygen-17 magnetic relaxation in aqueous solutions of carbohydrates. Bull. Chem. Sot. Jpn. 62, 14 (1989). Uedaira, H., Ishimura, M., Tsuda, S., and Uedaira, H. Hydration of oligosaccharides. Bull. Chem. Sot. Jpn. 63, 3376-3379 (1990).
