Proton Relaxation of Starch and Gluten by Solid-State Nuclear Magnetic Resonance Spectroscopy

Cereal Chem. 73(6):736-743 Proton rotating frame relaxation times [Tlp(H)] were used to characof gluten decreased the Tjp(H), which was also dependent on moisture. terize the molecular dynamics and structural homogeneity in waxy corn When mixed at a 1:1 starch-to-gluten ratio and heated, the Tjp(H) starch, wheat gluten, and mixtures of both. Single-phase relaxation of associated with the gluten were similar to those for pure gluten at 20% Tjp(H) was found in native starch, indicating a relatively small dimension moisture content (mc). However, when dried to 2% mc, the gluten Tip(H) of structural heterogeneity in terms of spin-diffusion. Heating of the increased to 9.3-9.6 msec. The Tjp(H) values for starch in the mixture starch samples decreased the Tjp(H) to 3.2-3.4 msec, as compared to raw were slightly increased to 5.7 msec. The different Tjp(H) values for starch samples at 5.3-6.2 msec, possibly due to the presence of more starch and gluten suggested a limited miscibility of the two components. amorphous domains. The native wheat gluten displayed a slightly inhoCompared to starch, gluten Tjp(H) was far more sensitive to moisture mogeneous Tjp(H) of 4.9-6.3 msec, suggesting the presence of a struccontent. tural inhomogeneity, different from that of native waxy corn starch. Heating Water, starch and protein are important constituents in foodstuffs. In many aspects, the quality and properties of food products are determined by the molecular dynamics and interactions in the systems. For instance, the retrogradation and crystallization of starch molecules during storage depends, to a large extent, on molecular mobility. This was previously determined by the glass transition temperature (Tg) (Levine and Slade 1988, Kalichevsky and Blanshard 1992). The mobilities and thermodynamic properties of starch and gluten in foods, on the other hand, depend on the moisture content in the system. Water serves as a plasticizer for starch and gluten molecules to lower the glass transition temperature in the food complex (Slade and Levine 1984). The inherent effects of water on starch and gluten are complicated. If water molecules form hydrogen bonds with starch and gluten molecules, the water relations in food may be interpreted in terms of such interactions. "Unfreezable" water refers to the fraction of water in starch that does not crystallize, even at temperatures far below the freezing point of pure water, within reasonable experimental timeframes (Hally and Snaith 1968, 1971; Kuntz and Kauzmann 1974; Cooke and Kuntz 1974). Many properties of food polymers are related to the unfreezable water. However, the molecular dynamics of starch and proteins in the presence of unfreezable water is unclear. Techniques such as nuclear magnetic resonance (NMR) spectroscopy could provide such information. Unlike small molecules or synthetic polymers, characterization of molecular dynamics in starch and gluten systems is difficult due to the complexity of the molecular structures. Solid-state NMR has been used to characterize the structure and dynamics of such systems (Veregin et al 1986, Belton et al 1987, Blanshard et al 1990, Grad and Bryant 1990, Hills 1991, Morgan et al 1992, Wu and Eads 1993). With the use of cross-polarization and magic angle spinning (CP-MAS), high-resolution spectra can be readily obtained. The molecular mobilities in the scale of milliseconds to seconds can be characterized using spin-relaxation of both protons and carbons in the laboratory frame and rotating frame (Schaefer et al 1977). t Department of Food Science University of Massachusetts, Amherst, MA 01003. Department of Polymer Science and Engineering, University of Massachusetts, Amherst, MA01003. Corresponding author. E-mail: [email protected] Publication no. C-1996-1004-01R. C 1996 American Association of Cereal Chemists, Inc. 736 CEREAL CHEMISTRY In this study, we examined the proton relaxation in the rotating frame for waxy corn starch and wheat gluten. The effects of water were studied at various moisture contents for both the starch and gluten samples. Effects of heating (gelatinization for starch and denaturation for gluten) were also investigated for starch and gluten samples and for samples mixed 1:1. MATERIALS AND METHODS Materials Waxy corn starch samples (Waxy No. 1, A.E. Staley, Decatur, IL) had a 10.5% mc (total basis), as measured by thermogravimetric analysis (TGA). Native wheat gluten samples (G-5004, Sigma Chemical, St. Louis, MO) with 80% protein content and 7% fat had a 14.9% mc as measured by TGA. Three kinds of samples were prepared for the NMR measurements: 1) native, 2) heated and 3) gluten heated and cooled stepwise. Sample Preparation Native waxy starch (10.5% mc) was used as is. Native starch of 2.0% and native gluten of 2.0 and 7.5% mc were prepared by drying the native samples at 60°C in vacuum oven. Heated samples were prepared by hydrating starch and gluten samples to '50% mc, and then spreading samples evenly onto a flat-bottomed disk covered with a ground-joint lid (to minimize water loss). Samples were heated for 15 min in a preheated oven (150°C). During heating, the sample temperature rose to 100°C in =10 min, and remained relatively unchanged thereafter. The heated samples were then cooled, dried in open air at ambient temperature to '30-40% mc, and then further dried at 60°C in a vacuum oven. Heated starch at 2.0 and 12.0% mc and heated gluten at 2.0 and 20.0% mc were prepared in this fashion. TI~ypically, it took =4 hr to reach a 20% mc and three days to reach a 2% mc. Heated sample mixtures (1:1, starch and gluten) were prepared similarly. Approximately 5 g of waxy corn starch and 5 g of wheat gluten were mixed and hydrated with 10 g of water. The mixture was spread onto a flat-bottomed disk with ground-joint lid and heated as described above. The heated mixtures were then cooled and dried to 2.0 and 20.0% mc. Samples were stored in sealed containers at ambient temperature. Polarized optical microscopy showed no sign of crystalline structure (lack of maltese cross) of waxy corn starch during storage. Dried samples were ground into fine powder in liquid nitrogen using a Spex Freezer Mill (Spex Industries, Edison, NJ) for NMR measurements. This practice was necessary to facilitate solid state CP-MAS NMR experiments. Gluten heated and cooled stepwise samples were prepared to study the effect of temperature and water on the molecular dynamics of gluten. Heated gluten at 2.0 and 20.0% mc (same heating method as above) was cooled from =60 to =30 0C, stepwise, with -30 min at each step. Cooling was done in the NMR instrument and Tlp(H) was measured at each temperature. NMR Analysis '3C spectra, 1H spectra and the Tlp(H) were obtained using a Bruker ASX 300 NMR spectrometer. Samples were packed into 7-mm ceramic rotors and spun at =3,500 Hz at MAS to eliminate chemical shift anisotropy. The 13C spectra and Tlp(H) were obtained with CP pulse sequence to enhance the sensitivity. Figure 1 shows the pulse sequence for the Tlp(H) measurements. The Tlp(H) values were obtained at ambient temperature by spin-locking the proton spins in the rotating frame for a variable delay (spin-lock time) before cross-polarization (Li et al 1994a). The proton spins were flipped to the rotating frame by a 5 jisec 90° pulse and were immediately locked with a spin-lock field strength of 50 kHz. After a variable spin-lock time (X) in the rotating frame, the proton spins were brought in contact with ' 3C spins with a fixed contact time of 2 msec for the cross-polarization and followed by acquisition of the ' 3C signal. Thus, the residual proton magnetization after relaxation in the rotating frame for a delay (X) was measured using the '3C spins. The details of the experiments were described elsewhere (Li et al 1994a). Typically the Tlp(H) was determined by fitting about 10-12 data points (X) using a one-component or two-component model. Values were considered valid when the standard error of the fitting was within the order of 102. NMR analysis of starch, gluten, and starch-gluten mixtures were performed at ambient temperature. Additionally, some wheat gluten was subjected to measurement at various temperatures from =60 to =300C (stepwise cooling done using a Nestlab chiller with air with a -73°C dewpoint). 1H NMR spectra for gluten and starch were also obtained using a 3.5 ,usec 900 pulse and a 101,800 Hz spectral width. The large spectral width was necessary to observe both the wideline resonance of the solid protons as well as the narrow resonance of the liquid ones (Wu et al 1992). Some hydrated gluten was spiked with excess water by injecting 410 gl of water into 250 mg of dried sample (2.0% mc) immediately before data acquisition. The 'H NMR spectrum collected in this fashion might lead to one resembling a superposition of spectra of gluten protons and free water protons. 90 spin lock cp decoupling 1H FT Moisture Analysis The water content of the samples was measured by TGA. Samples were heated at a rate of 20C/min from 30 to 250 0C with nitrogen gas purge (Fig. 2). The weight loss due to the evaporation of water was continuously monitored, and the water content of the samples was measured by the weight loss at 180°C. Further heating to a higher temperature was necessary to assure a smooth baseline, but this might have involved thermal decomposition. The weight loss measured by TGA at 180°C was consistent with values obtained from the standard vacuum oven drying method (600C at 736 mm Hg vacuum for two days). T\wo weight loss curves (heating rates of 2°C/min and 20'C/min) are shown in Figure 2. The results indicated that the weight loss from water evaporation was independent of heating rate. The water content was determined immediately after the NMR measurements. RESULTS AND DISCUSSION CP-MAS 'C Spectra Native waxy corn starch is a semicrystalline powder that contains both crystalline and amorphous structures. The multiphase structure in the sample can be detected by solid-state NMR. Figure 3A shows the CP-MAS ' 3C spectra for native (10.5% mc) and heated starch (12.5% mc) samples in comparison with A-type starch a-(1-4)-glucan (Blanshard et al 1990). The resonance at 95-103 ppm represented the C1 carbon, the resonance at 60 ppm corresponded to the C6 carbon, and the strong signals at around 71 ppm represented the resonance from C2-C5 carbons in starch. The amorphous C4 resonance appeared as a weak peak at 79 ppm (Veregin et al 1986). The crystalline structure in starch has been identified by the line shape in the C1 resonance (Veregin et al 1986). This and other groups (Belton et al 1987, Blanshard et al 1990) characterized a triplet splitting in the C1 carbon resonance due to the A form crystalline structure (Fig. 3A, top spectrum) and a doublet splitting due to the B form crystalline structure of starches.