Rapid wheat cultivar differentiation by capillary electrophoresis (CE). CE is one of the newest instrumental methods for analytical separations of proteins. Commercial instruments have been available only during the past 5 years. The various parameters (capillary length and inside diameter, buffer, temperature, voltage, extraction conditions, etc.) were optimized to reduce analysis time and increase resolution of wheat gliadin proteins. Gliadins were extracted from cultivars representative of hard red winter (HRW), hard red spring (HRS), and soft wheat classes and separated by CE. Cultivars that were not differentiable by polyacrylamide gel electrophoresis (PAGE) at pH 3.1 were segregated in less than 10 min by CE. Cultivars that were related closely (sister lines or intercrossings) were differentiated readily and cultivars that were not close genetically exhibited quite different CE patterns.

High performance capillary electrophoresis (HPCE) optimization for wheat gliadin analyses and comparison of individual protein peak positions to high performance liquid chromatography (HPLC) peaks. Conditions to optimize speed and resolution of wheat gliadins by HPCE included a 20 micro m (i.d.) untreated fused capillary 27 cm long (20 cm to detector) at 45 C and 22 kV, using a pH 2.5 phosphate buffer containing a polymer modifier from Bio Rad (#148-5011). Gliadins from the wheat cultivar `Shawnee` that were previously analyzed by HPLC, acid (A-)PAGE, and sodium dodecyl sulfate (SDS)-PAGE were extracted by 30 % ethanol, separated by HPLC; then each HPLC peak was collected and analyzed on HPCE. In general, the gliadins separate in HPCE as they do in PAGE; FONT SIZE=2 FACE="WP Greek Century"" gliadins moved fastest, followed by FONT SIZE=2 FACE="WP MultinationalA Roman"8, FONT SIZE=2 FACE="WP Greek Century"(,and then the slowest, FONT SIZE=2 FACE="WP Greek Century"o gliadins.

Wheat varietal identification by capillary electrophoresis: an interlaboratory comparison of methods. Optimal CE conditions have not been established yet, nor methods compared between laboratories. Therefore, J.A. Bietz at Peoria, IL, and G.L. Lookhart at Manhattan, KS, analyzed 30 % ethanol-soluble gliadins from several wheat varieties by two methods in two laboratories using Beckman 2100 P/ACE systems. Both methods used a commercial 0.1 M phosphate buffer, pH 2.5, containing a linear hydrophilic polymer, and 27 cm uncoated silica columns (20 cm inlet to detector). One method used a 50 micro m i.d. capillary at 7 kV; separations required 30-40 min. The other method used a 20 micro m column at 22 kV; separations took about 10 min. Within each laboratory, both methods gave excellent separations, with comparable selectivity and resolution. Results from the two laboratories for short analyses were equivalent, with peak elution times differing by an average of FONT SIZE=2 FACE="WP MathA"" 0.030 min. For the longer separations, resolution and elution times were also equivalent at both laboratories. However, elution times and operating currents varied slightly for different batches of buffer, showing the need for careful standardization of buffer composition. CE is a rapid, reliable, and reproducible method for wheat varietal identification and a useful alternative or replacement to slab gel electrophoresis or HPLC.

Rapid identification of some U.S. wheat lines by near infrared diffuse reflectance (NIR) spectroscopy. Current methods of wheat varietal identification are visual, HPLC, PAGE, and the most recent, capillary zone electrophoresis (CZE) determination. Although HPLC and PAGE methods are quite accurate, they are not applicable to fast, routine, quality control in the grain industry. The visual method is fast, but it is not as accurate as it used to be because of the increasing number of wheat lines entering the market that represent crosses between distinct varieties. The feasibility of rapid varietal identification of some U.S. wheats by NIR analysis was investigated. Wheat samples of 30 known HRW, hard white winter (HWW), and soft red winter (SRW) commercial wheat lines, which consisted of a total of 1,544 samples, were collected over 3 crop years (1992-1994) from two sources (Kansas Assoc. of Wheat Growers, and the Kansas Winter Wheat Performance Trials), and their visible and NIR spectra were recorded. Twelve varieties (10 HRW, 1 HWW, and 1 SRW), comprising 796 samples, then were selected for multivariate discriminant analysis (MDA). To reduce the effect of nonlinear spectral effects such as granularity, the spectral data were corrected mathematically by removing the linear and quadratic curvatures of each spectrum and forcing the standard normal variate of each spectrum to equal 1.0. Principal component analysis and MDA were applied to the corrected data. MDA allowed an overall 81.5 % efficiency of identification of the genetic origin of an unknown set of samples based upon the 12 wheat lines selected. A comparison of the MDA results obtained by different spectral math pretreatments was made. Using a coefficient of parentage (r), a comparison of efficiency of identification of genetic origin on the basis of between two similar wheat lines (Larned and Scout 66, r = 0.98) and two unrelated pedigrees (Cimarron and Scout 66, r = 0.20) also was made. Again, MDA allowed an overall 73.3 % efficiency of identification of the genetic origin of the three wheat lines selected, with the greater error in misclassification going to the two most closely related wheat lines.

Heat coagulation, structure and temperature fractionation of wheat flour albumins and globulins. Albumins and globulins, coagulated at 70, 100, and 120 C, were studied by using microscopic and electrophoretic methods. The coagulated albumins appeared as aggregates with filamentous patterns; the albumins of a U.S. commercial HRW wheat blend sample included larger aggregates than those of an Israeli spring wheat (ISW). Dialyzed globulins appeared as finger-like patterns with embedded oily droplets, and heat-coagulated globulins appeared as aggregates with fibril-like patterns. Size of coagulated aggregates increased as temperatures increased. Differences in CE and SDS-PAGE patterns between coagulates from samples of the ISW and the U.S. HRW wheat blend were seen at all coagulation temperatures. SDS-PAGE patterns of coagulates obtained at lower temperatures exhibited relatively more intense bands of higher molecular weight (MW) components, and those obtained at higher temperatures exhibited relatively more intense bands of lower MW.

Protein distributions among hard red winter wheat varieties as related to environment and baking quality. Previous size-exclusion (SE)-HPLC studies of HRS wheat proteins correlated amounts of certain gliadin and glutenin fractions with `general score,' a quality descriptor based on loaf volume, mixing time, and other parameters. We extended these studies to 12 HRW wheat varieties, each grown at six midwestern locations. Gliadin and reduced glutenin fractions were isolated and fractionated by SE-HPLC. Amounts of fractions corresponding to specific MW ranges, determined by integrating SE-HPLC data, were correlated statistically with various baking quality parameters. Two fractions, one rich in gamma-gliadins and the other in low MW glutenin subunits, gave good correlations with loaf volume; correlations were even better than found previously for HRS wheats. Correlation coefficients tended to be equal to or greater than 0.6 when estimated among varieties, among locations, among varieties at each location, and among locations for each variety. Evidently both genetics and environment affect the relationship of quantitative protein distribution to loaf volume.

Environmental effects on protein components, chemical and physical properties, and milling and bread-making data of Karl wheat grown at six locations in Kansas and harvested in 1993. Environmental effects on protein components and various chemical and physical parameters were studied in a single cultivar grown in various locations. Six samples of a U.S. HRW wheat cultivar, `Karl`, were chosen that represented varying protein contents and hardness values. The wheats were grown in Kansas in 1993; three were grown in irrigated plots and the others on dryland. The grain properties tested were: test weight, thousand kernel weight, hardness scores by NIR and Single Kernel Wheat Characterization System (SKWCS), and wheat ash, moisture, and protein contents. Milling, baking, and protein fractionation characterization also were performed. The total peak area of the albumin and globulin proteins in the CE patterns of ethanol soluble proteins from the six wheats was related to baking quality.

Development of an automated digital image analysis system for the study of starch size distributions in wheat. Digital image analysis of isolated starch granules has provided a large amount of accurate information on the size distributions of the starch populations in wheat. A major problem associated with the use of image analysis has been the labor required for each sample analyzed. A new image analysis system coupled directly to a light microscope equipped with a computer-controlled step stage and automatic focus allows for automation of data acquisition and data processing. This new system has been used to standardize starch isolation methods and compare sample variations as well as determine operational variations. One major advantage of the new system has been the increased resolution. The use of higher magnifications and stage automation has allowed the analysis of starch granules as small as 0.9 micro m in diameter, while continuing to analyze a large number of granules. The new digital system has been compared to two other image analysis instruments by analyzing several isolated starch samples. All three image analysis systems gave similar starch size distribution patterns for the samples, even though the comparisons were conducted at different times and on different slides of the samples.

Microstructural and chemical examination of soft, hard, and durum wheats using fourier transform infrared microspectroscopy. Fourier transform infrared microspectroscopy was used to study in situ chemical differences within the endosperm of soft (Caldwell) and hard (Karl) red winter wheats and durum wheats (Vic and West Bred 881). Procedures were developed to cryosection the wheats into 2 micro m sections, which were mounted on barium fluoride disks and scanned using the Spectra-Tech IRuS Scanning Infrared Microprobe. Linear spectral maps were made using a programmable stage and scanning the aleurone area through central endosperm for a distance of 663 micro m. Functions for the carbohydrate peak 1,025 cm^-1 and protein peak 1,542 cm^-1 were developed to measure peak areas and plotted against the scanned area, giving a quantifiable concentration curve for these two components. The protein and carbohydrate levels in Karl and Caldwell were high in the subaleurone in comparison to the aleurone. Analysis of the central endosperm revealed that the carbohydrate content further increased, while the protein content decreased. The durum curves showed a less rapid increase in carbohydrate and protein in the subaleurone. The central endosperm analysis showed a decrease of both carbohydrate and protein. This technique may provide some new insights into chemical hardness differences of some varieties of wheat.

Distinguishing selected hard red and soft red winter wheats by image analysis of starch granules. Starch was isolated from 24 wheat samples representing 14 HRW and 10 SRW wheats grown in various areas of Kansas. Samples with a wide range of NIR hardness values were selected from the Kansas Winter Wheat Performance Test. Isolated starch images were video recorded using darkfield light microscopy, and the digital images were analyzed by extracting various morphometric parameters. Equivalent diameters and aspect ratio were useful in discriminating HRW from SRW wheats. SAS procedures were used to transform the raw data of equivalent diameter and aspect ratio into frequency percentage within defined ranges. This transformation resulted in a new set of distributional data called Counts. A plot of Count 4, equivalent diameter, 5.5-7.0 micro m, versus Count 3, aspect ratio, 1.65-1.95, resulted in separation of HRW from SRW wheats, even though the NIR hardness values overlapped. Distributional analyses of the starch granule size and shape descriptors proved useful in classifying hard and soft wheats.

Nonstarch lipids in wheat flours suitable for noodle making. Total nonstarch lipids (TNSL) were extracted at 25 C from 11 wheat flours with a mixture of ethanol-diethyl ether-water (2:2:1, v/v/v). The 11 samples consisted of four sets: set (A) for 3 Australian Standard White wheat flours, the most favored for Japanese salt noodles (Udon); set (B) for 4 flours milled at Flour Mill (I) in Japan for Udon; and sets (C) and (D) for two flours each, milled at Flour Mill (II). The two samples each in sets (C) and (D) were the top-grade flours for Udon and Chinese noodles, respectively. Thus, 9 samples were Udon flours (sets A and C being the best) and 2 samples of (D) were Chinese noodle flours. The extracted lipids were further fractioned into nonpolar lipids (NL), glycolipids (GL), and phospholipids (PL). TNSL amounts ranged from 1.27 to 1.60 %; NL from 0.53 to 0.78 %; GL from 0.43 to 0.51 %; and PL from 0.20 to 0.32 % of the flour weights (db). The average quantities of TNSL were 1.32 % for set (D), 1.38 % for (B), 1.46 % for (A), and 1.56 % for set (C). The lipid compositions were similar for sets (A), (B), and (C) and substantially different for set (D). The average lipid compositions of 11 Udon flours vs. 2 Chinese noodle flours were 48.9 vs. 43.2 % NL; 33.4 vs. 34.3 % GL; 17.7 vs. 22.6 % PL; and 51.1 v.s 56.8 % for polar lipids (PoL = GL+PL). The ratios of lipid fractions were good indices differentiating Udon noodle flours from Chinese noodle flours: 0.96 vs 0.76 for NL/PoL and 2.76 vs. 1.92 for NL/PL. Flour lipids higher in NL/PoL or NL/PL ratio seem more desirable for Udon noodles.

Free lipids in straight grade flours and their air-classified high-protein fractions. Free lipids (FL) were extracted with petroleum ether on a Soxhlet from straight grade flours (SGF) milled from 2163, Tam 200, Newton, and Karl harvested at Manhattan, KS, in 1993 and their air-classified high-protein fractions (ACHPF). Protein contents increased from 9.1 % in the SGF to 19.5 % in the ACHPF for 2163, from 1.09 to 20.0 % for Tam 200, from 11.6 to 24.1 % for Newton, and from 12.7 to 24.2 % (on a 14 % mb) for Karl, respectively. FL contents also increased from 82.5 mg in 10 g (14 % mb) SGF to 180.9 mg in 10 g (14 % mb) ACHPF for 2163, from 84.2 to 173.7 mg for Newton, from 74.0 to 148.5 mg for Karl, and from 84.1 to 147.6 mg/10 g (14 % mb) for Tam 200. The extracted FL were fractionated further into NL, GL, and PL. The amounts of NL, GL, and PL were higher in the ACHPF than those in the SGF. The average increases were 97 % (77-114 %) for protein contents, 101 % (76-119 %) for FL, 108 % (87-128 %) for NL, 59 % (28-91 %) for GL, 84 % (44-146 %) for PL, and 68 % (36-82 %) for PoL (GL + PL). The smallest increases in all parameters were shown with Tam 200. The lipid compositions differed among the SGF and the ACHPF: the average NL were 81.7 % (80.2-83.1 %) for the SGF and 84.8 % (82.8-86.5 %) for the ACHPF; GL were 10.9 % (9.2-11.9 %) and 8.6 % (7.1-10.6 %); PL were 7.4 % (5.7-8.6 %) and 6.5 % (6.4-7.1 %); and PoL were 18.3 % (16.9-19.8 %) and 15.2 % (13.5-17.2 %). Thus, ACHPF-FL were richer in NL than SGF-FL.

Stabilities of several forms of vitamin C during making and storage of pup-loaves of white pan bread. White-pan bread was fortified with three commercial samples of fat-coated L-ascorbic acid (AsA) and uncoated AsA at a level of 64 mg per 100 g flour. Sample 1 contained large crystals of AsA coated with 30 % fat (m.p. 58-71 C) and resisted leaching in 6 % metaphosphoric acid at 25 C. Sample 3 contained mostly small crystals coated with 10 % fat (m.p. 49-69 C) and lost 70 % of its AsA during leaching. Proofed bread dough mixed with Sample 1 retained 80 % of fat-coated AsA in proofed dough, whereas that with Sample 3 retained only 7 % of fat-coated AsA. Bread (0-3 days old) fortified with Sample 1 showed approximately 20 % higher retention of vitamin C than bread fortified with Sample 3 or with uncoated AsA. Bread fortified with Sample 2 showed intermediate retention of vitamin C. Bread also was fortified with AsA in the forms of L-ascorbic 2-polyphosphate (AsPP) and L-ascorbate 2-monophosphate (AsMP). After mixing, fermenting, and proofing, about 85 % of AsPP and AsMP was hydrolyzed to AsA. The 2-phosphorylated AsA remaining in proofed dough largely survived the baking step, whereas the free AsA underwent approximately one third destruction. Bread after 3 days of storage retained 10-15 % higher vitamin C when fortified with AsMP or AsPP than with AsA.

Fortification of bread with beta-carotene. A water-dispersion form of beta-carotene (5 mg active ingredient) was incorporated into 100 g-straight dough formulation, and the retention of beta-carotene was determined in dough and bread. The lyophilized and ground dough or bread (1 g) was extracted vigorously with a mixture of methanol, acetonitrile, chloroform, and water (200:250:90:11 by volume) (25 ml) at 25 C on a magnetic stirrer plate for 15 min. After centrifugation, supernatant (20 micro l) was injected into a chromatograph. Reverse-phase HPLC with detection at 445 nm was used, and base-line resolution of beta-carotene was obtained in the extract. Freeze-dried bread spiked with beta-carotene showed 96 % recovery. Retentions of beta-carotene in proofed dough and freshly baked bread were 96 % and 73 %, respectively. The stability of beta-carotene in stored pan bread and in other breads is under examination.

Volatile compounds in wheat cultivars from several locations in Kansas. The need for objective off-odor detection in grain grading has stimulated research on volatile components in grains. Our recent research has shown that specific compounds are associated with off-odors, but more data were needed on variation in volatiles among samples with mostly normal odors and whether conditions during grain maturation or inherent properties of the cultivars themselves influenced volatile composition. Five cultivars of wheat grain harvested at six locations across Kansas in 1992 and 1993 were analyzed for volatiles. Even though odors of the samples were mostly normal, >100 compounds were detected. Alcohols were most abundant, followed in order by aldehydes, alkanes, alkylbenzenes, ketones, methyl esters, naphthalenes, terpenes, and other compounds. Amounts of some volatiles differed among locations and cultivars, but none of the differences appeared to be associated with intrinsic properties of the cultivars themselves. Slightly elevated amounts of three compounds in a few samples provided evidence for dampness and slight insect infestation, but these were the only compounds detected that our previous research had shown to be associated with off-odors.

Detection of specific compounds that indicate off-odors in grains. A gas chromatograph-mass spectrometer (GC-MS) system attached to a purge and trap instrument was set up with `target compound analysis' software for quantitative detection of specific volatile compounds that our previous research had shown to be indicators of off-odors in grains. The GC-MS system was used to analyze three sets of samples from the FGIS, including corn, sorghum, soybeans, and wheat with a variety of normal and off-odors. The odor of each sample was classified by FGIS inspectors and by personnel at our laboratory. For each sample, the odor(s) determined by human inspectors was compared to the odor(s) assigned by the GC-MS system based on what odor-indicating compounds were detected and their concentrations. After each sample set was analyzed, relationships between compounds and odors were reevaluated, then adjustments were made in the list of odor-indicating compounds and the concentrations of those compounds that had to be exceeded to indicate an off-odor. Neural networks also were evaluated for potential use in classifying odors from data generated by the GC-MS system.

TCK smut in wheat. Preliminary plans were made to obtain wheat contaminated with dwarf bunt spores and determine their fate during handling and milling of the wheat. All samples that were received from the Pacific Northwest had either extremely high numbers of spores or very low numbers, such that they were inappropriate for the planned milling study. Various mixtures were made in the laboratory using clean and contaminated wheat to see if we could make a batch with a predictable level of contamination. Some spores apparently transferred from the contaminated to the clean wheat, but many remained on the contaminated kernels, so that spore counts on the mixture were highly variable. Another problem with mixed wheat was that the spores that readily come off of the contaminated kernels also readily come off of the initially clean kernels. Small-scale milling tests showed that most spores were found in the bran fraction and about 95 % of them were broken during milling. Destruction of remaining viable spores in wheat mill-feed with a small pellet press will be tested on a large scale using a pellet mill and an extruder.

Volatile flavor components of breads made from hard red winter and hard white winter wheat. Volatile components of white pan bread and whole wheat bread made from HRW or HWW wheat were collected on a Tenax TA trap by dynamic headspace concentration. Bread crumb and crust were studied separately. A total of 74 compounds was separated and identified by gas chromatography-Fourier transform infrared spectrometry-mass spectrometry. Among the compounds that possessed some odor qualities, 11 of them had bread-like odors. Most of the carbonyls and pyrazines were found to be higher in relative quantities in the crust than in the crumb. Many differences were noticed between white pan and whole wheat breads. However, many fewer differences were found between breads made from HRW and HWW wheats.

Recent efforts to develop technologies needed to implement a `Total Quality Grain Marketing System'

. In 1989, The Office of Technology Assessment (OTA) of the U.S. Congress published a study that recognized the strengths and weaknesses of the U.S. grain trade system. One of the major problems identified was a lack of research, programs, and policies that can supply a consistent and stable source of physically sound, insect-free grain with a known end-use quality to domestic and international customers. The USGMRL and other USDA/ARS research programs have been developing and improving technologies that have some potential to be used in an integrated "end-use quality"-driven marketing system for the grain industry. Recent research accomplishments have resulted in technologies based on NIR, digital image analysis, and automation of characterization data for hardness, moisture, size, and weight of single kernels that need to be applied to the grain handling and marketing industry. The potential to implement a `Total Quality Grain Marketing System', based on potential objective end-use measurement technologies, has occurred only recently because of a strong realization by members of the U.S. grain industry that an objective quality prediction system is needed to keep the U.S. grain industry competitive in the future world grain trade.

Relationship between single-kernel characteristics and end-use quality. I. Hard Wheats. The SKWCS, developed by the USGMRL was used to study the relationships between the SKWCS parameters and some quality data. To eliminate environmental effects on end-use quality, the set of 12 HRW and 12 HRS wheats, grown at the same location (Sacramento Valley, CA) during 3 crop years (1988-1990) were used. The equations were derived by multiple stepwise or principal component regression analysis to study the relationships between single kernel (SK) parameters and quality data. Experimental micro-milling yields and micro-bread loaf volumes were significantly correlated with SK. The most significant kernel parameters were sizes for milling yields and weights for loaf volumes. The r values of mathematical relationships were higher for the HRW wheat classes or HRS wheat classes separately (n = 36 each) than for the combined two wheat classes (n = 72). In general, the higher r values were obtained by the principal component regression analysis than by the multiple stepwise regression analysis. II. Soft Wheats. The SKWCS was used to study the relationships between the SKWCS parameters and some quality data for soft wheat samples. Soft wheat samples (140 cleaned and 136 uncleaned) and their quality data were provided by the USDA/ARS Soft Wheat Quality Laboratory, Wooster, OH. The USGMRL-SKWCS data were used to predict various quality parameters of soft wheats by multiple stepwise regression. Softness Equivalent (S.E.) of both cleaned and uncleaned wheat would be estimated well by SKWCS (r = 0.94 for uncleaned vs. 0.92 for cleaned set). However, the Adjusted Flour Yields and Milling Scores were fitted significantly better with a prediction equation for the cleaned set than for the uncleaned set (r = 0.87 vs. 0.77). The reverse trend was shown with Baking Scores and Combined Quality Scores with an r of 0.91 and 0.85 for the uncleaned set and an r of 0.78 and 0.80 for the cleaned set. Some other parameters such as NIR hardness score, flour protein content, etc. were not affected greatly by cleaning grain prior to testing.