Reduced-lipid soy protein isolate (SPI), prepared from soy flour treated so that most of the polar lipids have been removed, exhibited an increase in protein solubility of 50% over that of the control SP1 prepared from hexane-defatted flour. Adding lipids from a commercial SPI during processing of reduced-lipid SPI decreased SPI solubility by 46%. The 19% decreased solubility caused by the lipids (primarily phospholipids) was largely recovered by treating the protein with a reducing agent (2-mercaptoethanol). The balance of protein insolubility, caused by the lipids, was attributed to a smaller lipid fraction (approximately 5% of the total lipids). Adding lipids during SPI processing contributed to both the formation of oxidized protein sulfhydryls, incapable of being reduced by 2-mercaptoethanol, and to oxidative deterioration of protein as determined by protein carbonyl contents.

Using gas chromatography/olfactometry (GCO), major odors from the headspace of aqueous solutions of soy protein isolates were evaluated. Many corresponding odorants were identified by correlating GCO with GC/mass spectrometry (MS) on two separate stationary phases followed by comparing retention times, mass spectra, odor descriptions and odor intensities with authentic standards. Based on aroma extract dilution analyses, the most powerful odorants (strongest and most volatile first) were (1) dimethyl trisulfide, (2) trans,trans-2,4decadienal, (3) an unidentified burnt soy sauce-like odor, (4) 2-pentyl pyridine, (5) trans,trans-2,4-nonadienal, (6) hexanal, (7) an unidentified charred sweaty feet-like odor, (8) acetophenone, and (9) 1-octen-3-one. This is the first reported occurrence of dimethyl trisulfide in soy protein isolates.

The effect of oxygen on the two separate one-electron reactions involved in the oxidation of ascorbic acid was investigated. The rate of ascorbate radical (Asc(-)) formation (and stability) was strongly dependent on the presence of oxygen. A product of ascorbic acid oxidation was measurable levels of hydrogen peroxide, as high as 32.5 μM from 100 μM ascorbic acid. Evidence for a feedback mechanism where hydrogen peroxide generated during the oxidation of ascorbic acid accelerates further oxidation of ascorbic acid is also presented. The second one-electron oxidation reaction of ascorbic acid leading to the disappearance of Asc(-) was also strongly inhibited in samples flushed with argon. In the range of 0.05-1.2 mM ascorbic acid, maximum levels of measurable hydrogen peroxide were achieved with an initial concentration of 0.2 mM ascorbic acid. Hydrogen peroxide generation was greatly diminished at ascorbic acid levels of 0.8 mM or above.

Addition of tert-butylhydroquinone or a mixture of butylated hydroxyanisole and tert-butylhydroquinone (200 ppm on a lipid basis) during SPI processing gave increased protein solibility over that of the con&l (55%. 56% and 34%. resuectivelv). These increased solubilities correspond to 32% and 18% decrease in oxidation of free sulthydryls and 20% and 12% reduction in protein oxidation, as determined by protein carbonyl content. Increased protein solubilities, due to added antioxidants, were accompanied by higher total protein surface hydrophobicity, as determined by the sodium dodecyl sulfate (SDS) binding method, and soluble protein hydrophobic&y, as determined by the fluorescence probe S-anilino-1-naphthalene sulfonate (ANS).

Volatile compounds from 2 samples of aqueous soy-protein isolates (SPI) (7%) were analyzed using both static and dynamic headspace methods. Based on dynamic headspace analyses, the most powerful odorants were (1) dimethyl trisulfide, (2) methanethiol, (3) hexanal, (4) an unidentified charred, sweaty feet-like odor, (5) 2-pentyl furan, (6) 2,3-butadione, and (7) an unknown burnt-like odor. The most powerful odorants by static headspace analyses were (1) dimethyl trisulfide, (2) hexanal, (3) methanethiol, and (4) 2-pentyl furan. Using deuterium labeled DMTS as an internal standard, DMTS was quantified at 60.1 and 45.5 ppb in the SPIs. This corresponds to odor values of 6014 and 4554, respectively. Using a cool, on-column technique, direct injection of concentrated-headspace volatiles and solvent-recovered volatiles with an internal standard of d 6 -DMTS detected both methanethiol and DMTS at similar levels as with the traditional injection methods.

Gas chromatography olfactometry/mass spectrometry (GCO/MS) studies on static and concentrated headspace of the aqueous slurries from soy protein concentrate (SPC) revealed acetaldehyde, methanethiol, hexanal, dimethyl trisulfide (DMTS), and 2-pentyl furan as the most odorous volatiles. Further aroma extract dilution analysis (AEDA) of the volatile extracts identified the following as the odorous substances: hexanal, 2-heptanone, octanal, 2-octanone, 1-octen-3-one, DMTS, 3-octen-2-one, 2-decanone, benzaldehyde, 2-pentyl pyridine and trans, trans-2,4-nonadienal, along with several unidentified odorants. Methanethiol and acetaldehyde, which have low boiling points, were not detected by AEDA, however. This is the first time that acetaldehyde, methanethiol, and dimethyl trisulfide have been identified as primary odorants in SPC.

Mean methanethiol headspace concentrations above aqueous slurries of isolated soy proteins (ISP) increased 17-to 36-fold over the controls with the addition of L-cysteine. Corresponding hydrogen sulfide levels were also greatly increased. Dithiothreitol, sodium sulfite, and glutathione increased headspace methanethiol from aqueous ISPs 23-to 44-fold, 8-to 9-fold, and 5-fold, respectively, but did not elevate hydrogen sulfide. These observations, along with the effects from the addition of dithiothreitol/O-acetyl-serine, the addition of a pyridoxial phosphate inhibitor and the intrinsic sulfite content of ISP samples (22 to 31 ppm), indicate that methanethiol from soy proteins is formed by way of components of a sulfite-to-cysteine pathway.

Solid-state electron paramagnetic resonance (EPR) spectroscopy of commercial samples of isolated soy proteins (ISP) revealed a symmetrical free-radical signal typical of carbon-centered radicals (g= 2.005) ranging from 2.96 x 10(14) to 6.42 x 10(14) spins/g. The level of free radicals in ISP was 14 times greater than similar radicals in sodium caseinate, 29 times greater than egg albumin, and about 100 times greater levels than casein. Nine soy protein powdered drink mixes contained similar types of free radicals up to 4.10 x 10(15) spins/g of drink mix, or up to 6.4 times greater than the highest free-radical content found in commercial ISP. ISP samples prepared in the laboratory contained trapped radicals similar to the levels in commercial ISP samples. When ISP was hydrated in 2.3 mM sodium erythorbate or 8.3 mM L-cysteine, frozen and dried, the level of trapped free radicals increased by about 17- and 19-fold, respectively. The ESR spectrum of defatted soybean flakes contained overlapping signals from the primary free-radical peak (g= 2.005) and a sextet pattern typical of manganese-II. The manganese signal was reduced in the laboratory ISP and very weak in the commercial ISP.