Introduction

Wheat flour products fall in a wide range, from yeast-leavened dough products requiring high mixing strength (as in pan bread), to chemically leavened products where gluten development is not desired (such as in biscuits or cakes). Within the latter category, factors other than gluten are considered more important in end-use quality. One factor that is critical in batter-based applications is viscosity. In batter coatings, cake or pancake production, viscosity is a crucial quality parameter. Batters for coating must be thick enough to cling to the product, but without clumping (as in tempura). Pancake or donut batter must be thick enough to retain leavening gasses and prevent settling and to retain leavening gas, but without being so thick as to inhibit flow and spread.

There exists a substantial amount of unexplained variation in batter viscosity. Better understanding and control of the biochemical components contributing to viscosity would translate into cost savings in processing as well as increased consumer acceptance of a product. The ability to produce a uniform mix is crucial. Further, the elucidation of the biochemical underpinnings of viscosity variation has the potential to guide breeding programs to produce wheat that will perform in end-use applications without additional additives, leading to “shorter” labels and concomitant consumer acceptance and decreasing the need for expenditures on remedial, modifying ingredients.

Variation in batter viscosity has traditionally been attributed to two sources: HMW glutenins and arabinoxylans (also known as pentosans, hemicellulose or non-starch carbohydrate). In batters themselves, bubbles formed from leavening gasses can also contribute to viscosity. Current tests for viscosity-contributing components include the lactic acid Solvent Retention Capacity (SRC) test and SDS-sedimentation test that identify the contribution of HMW glutenins. Sucrose SRC identifies the contribution of arabinoxylans, and, to an extent, gliadins.

Arabinoxylans consist of a ß-1,4 xylose backbone, variously substituted at the 2- and 3-carbon position with arabinose. The degree and frequency of the substitutions determine the water solubility of the polysaccharide. As such, arabinoxylans are conveniently separated into water-soluble and water-insoluble fractions. The water insoluble fraction can adsorb about ten times its weight in water and hold it to the extent that water activity is affected. Water soluble arabinoxylans have varying amounts of ferulic acid moieties attached to the arabinose units. Under the correct circumstances, usually the presence of free radicals, the ferulic acid moieties crosslink, forming a large network of arabinoxylans. The resulting matrix entraps, or sequesters water, leading to a gel. Neukom and Markwalder (1978) indicated that similar free radical-induced cross-linking can also occur between tyrosine residues contained in proteins or between protein, via tyrosine, and arabinoxylans, via ferulic acid.

The potential for oxidative gelation has long been known. Durham published a manuscript in 1925 indicating that an unidentified water soluble fraction caused gelation when exposed to hydrogen peroxide. Further work by Ciacco and D’Appolonia (1982) identified the water-soluble fraction as arabinoxylans and the mechanism as ferulic acid cross-linking. Oxidative gels are mechanically fragile. They break apart fairly easily under shear. This indicates that their impact would be seen less in mixed doughs (e.g. pan bread dough or biscuit dough) and more in batter or cake formulations where shear is minimized.

Thus, the primary effect of oxidative gelation is on batter viscosity. Obtaining the correct range of viscosity in batter is crucial in preventing “settling” of the batter during storage (e.g. pancake batter in a restaurant sitting between orders) or loss of leavening gas. Direct effects on end-product quality are apt to be seen as sticky textured products due to sequestered water.

In this research, the focus was on the potential contribution of not only the quantity of arabinoxylans and proteins, but also on their potential to form oxidative gels that contribute to viscosity, as estimated with a new test method.

Materials and methods

An oxidative gelation test was developed that measures the combined impact of free radicals (endogenous or induced), the arabinoxylan-ferulic acid complex and protein on viscosity. The method is as follows: 10 g flour (14%mb) weighed into a 50 mL conical-bottom, capped, polypropylene tube. Hydrogen peroxide (75ppm, per Hoseney and Faubion, 1981; 25 mL) is added and the flour suspended through vortexing. The tube is rocked to allow complete hydration for 20 min on a Labquake Shaker (Fisher Scientific). Alternatively, the tube can be shaken by hand each 5 min to assure hydration of the flour. Horseradish peroxidase (Type II, Sigma; 60 µL at 1 PU/µL) is added, the tube capped and inverted three times and the slurry dispensed into the reservoir of a Bostwick Consistometer (VWR Intl). The gate allowing the reservoir to empty is tripped after one min and the distance the slurry runs down the trough is measured at 40 sec.

This preliminary research used 89 soft white winter (U.S. market classification) wheat flours from an advanced Oregon State nursery that varied in end-use quality and gelation. The oxidative gelation test was run as indicated above, but with water, xylanase (flour hydrated with Bioxylanase 10L, Kerry Bio-Science, 2 U without hydrogen peroxide for 20 min), and xylanase followed by peroxide+peroxidase (xylanase without hydrogen peroxide for 20 min, then 3% hydrogen peroxide added in sufficient amount to yield 75 ppm, the slurry inverted three times, then 60 µL peroxidase added, the slurry inverted three more times and then proceeding as usual). In future the peroxide+peroxidase treatment will be referred to only as “peroxidase” for simplicity.

Other attributes of wheat and flour were obtained. Single Kernel Characterization System (SKCS) hardness was measured with the Perten 4100 SKCS (Perten Instr., Springfield, IL, USA). Ash was measured by combustion by Leco TGA-601 and protein by Leco FP-528 (Leco Corp., St.Joseph, MI, USA). Cookie spread was measured after baking using AACC method 10-52 (sugar-snap cookie).

Results and discussion

Consistometer readings for water (measuring only the oxidative gelation potential of the arabinoxylans, protein and endogenous free radicals) ranged from 9.9 to 15.7 with a mean of 13.3. Consistometer readings for peroxide (maximum oxidative gel potential with free radicals from the peroxidase system) ranged from 3.8 to 14.0 with a mean of 11.1. Xylanase treatment (like water consistometer, but with xylanase added) produced a consistometer range of 12.4 to 17.0 (near the theoretical limit of measurement, for the volume of sample used) with a mean of 14.8. Xylanase treatment followed by peroxidase treatment ranged from 10.6 to 15.8 with a mean of 13.7.

The xylanase treatment had a correlation with the peroxidase treatment of r = 0.05. The oxidative gel normally observed from peroxidase treatment was prevented from adequately forming because the polymeric backbone that is interconnected through di-ferulic acid bridges was hydrolysed by the enzyme. Visual examination of the relationship showed that there are two groups of points (Fig 1A). One group correlated more strongly to protein content than the other. However, this correlation between flour protein and the peroxidase treatment was -0.31. When treated with xylanase before being treated with peroxidase, the correlation improved to -0.72 (Fig 1B). The group of samples that reacted strongly to gelation, but not so much to protein, in Fig 1A (the group in the lower left quadrant of the chart) were much better correlated after xylanase treatment removed the potential for arabinoxylan cross-linking to form a gel. The explanation for these observations was that not only was an oxidative gel being formed among arabinoxylans, but among proteins as well, through the mechanism described above.

Figure 1. Relationship between flour protein and Bostwick Consistometer viscosity resulting from peroxidase treatment (A, left) and xylanase treatment, followed by peroxidase treatment (B, right).

Water consistometer readings were correlated with ash content (r = 0.60), indicating that water insoluble arabinoxylans, likely in bran and cell wall material in the flour, were adsorbing water. Treatment with xylanase resulted in consistometer readings that were correlated with ash only at r = -0.12, providing further indications that insoluble arabinoxylans do adsorb a certain, fixed amount of solution.

It appears oxidative gelation also has an effect on end-product quality when sugar-snap cookie diameter is examined in light of the viscosity test. Although the oxidative gel was too fragile to withstand the mixing action involved in cookie dough production, the results indicated that perhaps some oxidative gel fragments remain sufficiently intact to sequester water and affect water relationships within the dough system.

A relationship between cookie diameter and peroxidase consistometer was reflective of the relationship with flour protein. Protein has long been associated with cookie quality, but the relationship has unexplained variation. Fig 2A indicated that arabinoxylans oxidative gelation may be a part of the unexplained variation. When treated with xylanase, no real relationship was seen that could be explained by either protein or by arabinoxylans. However, when treated with xylanase, followed by peroxidase, the effect of protein gelation potential could be seen. Again, as above with relationship between protein and the xylanase – peroxidase system, the relationship showed a more clear relation between the residual protein gelation and cookie diameter.

Figure 2. Relationship among cookie diameter and consistometer readings with peroxidase treatment(A; r = 0.60), xylanase treatment (B; r = 0.16) and xylanase followed by peroxidase treatment (C; r = 0.57).

To examine the combined effect of factors contributing to end-use quality as measured by cookie diameter, a statistical model was created, taking only the best combined independent variables. Cookie diameter was most influenced by peroxidase (F-value = 22), the xylanase followed by peroxidase (F-value = 16), ash (F-value = 8) and SKCS hardness (F-value = 8). The statistical model indicated that, in order, the combined oxidative gelation effect of arabinoxylans and proteins, protein gelation alone, insoluble arabinoxylans (estimated by ash) and starch damage (estimated by SKCS hardness) most contributed to cookie quality. Other variables had no significant effect on the model.

As this research is preliminary, further investigation is necessary. However, indications are that oxidative gels were formed and measured in this method; oxidative gelation occured among arabinoxylan polymers, likely through di-ferulic acid bridges; among proteins, likely via di-tyrosine bridges and possibly among arabinoxylans and proteins through ferulic acid – tyrosine bridges. Oxidative gels appeared to affect cookie diameter, but are more likely to have a much larger influence on the quality of high-moisture products that are prepared with little shearing (e.g. pancakes, cake donuts and other batter systems).

Further research into this oxidative gelation method is required to: establish the effect of pH on viscosity and gel formation; measure the impact of oxidative gelation on pancakes and Japanese sponge cake quality; measure the concentration of water-soluble and total arabinoxylans and determine their relative molecular weights; determine damaged starch content in flour, and to acquire this information on more location-year flour sample sets.

References

American Association of Cereal Chemists, Approved Methods, 10^th Edition (2000) Approved Method 10-52.

Ciacco, C.F. & D’Appolonia, B.L. (1982). Cereal Chem. 59:96-99.

Durham, R.K. (1925). Cereal Chem. 11:297-305.

Hoseney, R.C. & Faubion, J.M. (1981). Cereal Chem. 58:421-424.

Neukom, H. & Markwalder, H.U. (1978). Cereal Foods World 23:374-376.