Sorghum (Sorghum bicolor (L.) Moench) is a cereal of remarkable genetic variability, which makes it difficult to classify. A few names of sorghum are milo, jowar, kafir corn, guinea corn, and cholam The plant originated in equatorial Africa and is distributed throughout the tropical, semi-tropical, and arid regions of the world. The seed or caryopsis of sorghum provides a major source of calories and protein for millions of people in Africa and Asia. Grain characteristics of sorghum have been documented by Rooney and Miller (1982). Appearance and quality of sorghum are affected significantly by genetically controlled characters.

Grain is marketed, according to the US grain standards (USDA 1987), as sorghum, tannin sorghum, white sorghum, and mixed sorghum. The sorghum grade does not contain more than 3% of sorghum with a pigmented testa or undercoat. The white grade contains sorghum with a white pericarp without a pigmented testa. It cannot contain more than 2% of sorghum with colored pericarp or testa. Mixed sorghum contains mixtures of sorghum with and without pigmented testa. Tannin sorghums contain proanthocyanidins as part of their phenolic compounds but do not contain tannic acid nor hydrolyzable tannins. Tannin sorghums have a pigmented testa on the innermost layer of the pericarp. The pigmented testa is seen as a dark layer between the light endosperm and the pericarp when the caryopsis is scraped to remove the pericarp. Bleaching using the bleach test causes the constituents in the pericarp and testa to oxidize and yields black pigments on the surface of the caryopsis (Waniska et al. 1992). Sorghums with a pigmented testa and tannins remain black longer during bleaching than do non-tannin sorghums.

Several interacting factors affect the color and overall appearance of sorghum caryopses. Appearance is mainly affected by pericarp color and thickness, presence of pigmented testa, and endosperm color (Rooney and Miller 1982). Secondary plant and glume colors, as well as, deterioration due to the environment, insects and molds, complicate the evaluation of pericarp color and other properties. Immature pericarp tissues, when damaged, respond with antimicrobial phenolic compounds, which form pigments that stain the pericarp and endosperm. Insect and mold damage of the pericarp commonly occurs together when caryopses mature in hot, humid environments.

Pericarp color is genetically controlled by the R and Y genes. The combination of these genes can produce white or colorless (R_yy or rryy), lemon yellow (rrY_), or red (R_Y_) color. The intensifier (I) gene increases the brightness of the pericarp color; this is especially prominent in red pericarp sorghums. The Z gene controls pericarp thickness. Sorghums with homozygous recessive (zz) genes possess a thick mesocarp. A thick pericarp contains small starch granules, which causes a chalky appearance that masks the color of the testa and endosperm (Earp and Rooney 1982; Rooney and Miller 1982). A pigmented seed coat or testa is present when both B₁_ and B₂_ genes are dominant. The tptp genes control testa color, which can be brown or purple. Caryopses with a pigmented testa (B₁_B₂_) and a recessive spreader gene (ss; type II) or dominant spreader gene (S_; type III) contain condensed tannins and are called brown or tannin sorghums (Hahn and Rooney 1986). Type III sorghums have more tannins than type II sorghums, whereas type I sorghums do not contain tannins or proanthocyanidins.

A black sorghum from Sudan, Shawaya, has very high levels of flavonoids that oxidize in the pericarp during caryopsis development and mask pericarp color. The black pigmentation is a recessive trait. Some sorghums have a dominant trait for intense pigmentation or blotching in the pericarp and peripheral endosperm when damaged. Blotching detracts from appearance and food utilization. Darker secondary plant and glume colors are dominant and increase the pigmentation of milling and food products. Each secondary plant color (tan, red, and purple) has a range of glume colors (from very light to very dark). The lighter glume colors yield products that are less pigmented and are preferred in most foods.

Yellow endosperm cultivars contain high levels of carotenoid pigments. Endosperm color affects appearance, especially in caryopses with a thin pericarp and without a pigmented testa. Caryopses appear yellow when the pericarp is thin and colorless, the testa is absent, and the endosperm is yellow. A thick mesocarp and colorless pericarp cause a white or chalky appearance. Heteroyellow endosperm sorghum results when sorghums with yellow and non-yellow endosperm colors are hybridized (Rooney and Miller 1982). Bronze sorghums contain a thin, red pericarp with yellow endosperm color, while cream sorghums contain a thin, white pericarp with yellow endosperm.

Waxy endosperm cultivars contain three genes (wx) in the recessive form. Heterowaxy genotypes contain one or two of these genes in the dominant form whereas normal or non-waxy endosperm sorghums contain all three genes in the dominant form. Waxy cultivars contain nearly 100% amylopectin and the endosperm looks like candle wax (Rooney and Miller 1982).

The environmental conditions during caryopsis development and maturation greatly affect the appearance of sorghum because the panicle or head is readily exposed to biotic and abiotic agents (Castor and Frederiksen 1980; Glueck and Rooney 1980). An environment that is hot and humid during maturation negatively affects quality. Deterioration results from insects, molds, moisture, and sunshine. Physical damage resulting from insect attack activates a phenolic defense mechanism that produces discolored spots in the pericarp and the endosperm. Mold colonization discolors the pericarp surface initially, breaks down the endosperm composition and structure, and adversely affects processing quality (Waniska et al. 1987). Sorghums that have an open panicle structure and caryopses with thin pericarp, condensed tannins, corneous endosperm, and large, tight glumes are generally considered to be more resistant to grain molding and weathering (Waniska et al. 1989).

The high-lysine (hl) sorghum from Ethiopia has a soft, floury endosperm texture, a shriveled kernel structure, and is susceptible to deterioration in environments where it rains during caryopsis development (Rooney and Miller 1982). The improved protein digestibility sorghum (Weaver et al. 1998) and the chemically induced, high-lysine sorghum (Ejeta and Axtell 1987) have intermediate-soft to soft endosperm textures; they have reduced grain yield and increased deterioration due to grain molds and weathering.

Food-type sorghums have tan secondary-plant-color, straw-colored glumes, white or clear pericarp, no pigmented testa, intermediate to hard endosperm texture, and increased resistance to grain weathering that mill into products with bland flavor, white color, and no off-colors.

The caryopsis consists of three distinct anatomical components (Fig. 1): pericarp (outer layer), endosperm (storage tissue), and germ (embryo). The outer layer or pericarp originates from the ovary wall (Saunders 1955; Glennie et al. 1984) and is divided into three histological tissues: epicarp, mesocarp, and endocarp (Earp and Rooney 1982). The outermost layer, or epicarp, is generally covered with a thin layer of wax. The epicarp is two or three cell layers thick and consists of rectangular cells that often contain pigmented material. Unlike most cereals, the sorghum mesocarp contains starch granules. A thick pericarp usually contains three or four mesocarp cell layers filled with small starch granules (Fig. 1). The inner pericarp tissue, endocarp, is composed of cross cells and tube cells. Pericarp thickness (the Z gene) ranges from 8 ΅m to 160 ΅m (Blakely et al. 1979; Earp and Rooney 1982) and varies within an individual mature caryopsis. The areas below the style and near the hilum are the thickest with the sides of the kernel being thinnest. The inner tube cells conduct water during grain germination whereas the outer cross cells form a layer that impedes moisture loss. A stylar area is located on the tip of the caryopsis, opposite the germ. The black layer, or hilum, is the colored placenta scar tissue that develops at the germ tip when the caryopsis reaches physiological maturity. Cell walls of sorghum pericarp, aleurone, and endosperm exhibit a blue autofluorescence, which is mainly due to esters of ferulic acid.

The seed coat or testa is derived from the ovule integuments. The thickness of the testa ranges from 8 ΅m to 40 ΅m and varies within individual caryopses (Blakely et al. 1979; Earp and Rooney 1982). The thickest area usually is observed below the style and the thinnest on the side. In sorghums with dominant B₁ and B₂ genes, the testa is pigmented and contains tannins (proanthocyanidins). Tannins are present in the testa and the pericarp in sorghum with dominant B₁_ B₂_ and spreader (S_) genes.

Figure 1. Diagram of sorghum caryopsis showing the pericarp [cutin, epicarp, mesocarp, tube cells, cross cells, testa, pedicel, and stylar area (SA)], endosperm (E) (aleurone layer, corneous, and floury), and germ [scutellum (5) and embryonic axis (EA)] (Source: Rooney and Miller 1982, with pennission).

The endosperm tissue is triploid, resulting from the fusion of a male gamete with two female polar cells. It is composed of the aleurone layer, peripheral, and corneous and floury areas (Earp and Rooney 1982) (Fig. 1). The aleurone is the outer cover and consists of a single layer of rectangular cells adjacent to the testa or tube cells. The cells possess a thick cell wall, large amounts of proteins (protein bodies, enzymes), ash (phytin bodies), and oil (spherosomes). The peripheral area is composed of several layers of dense cells containing more protein and smaller starch granules than the corneous area. Both the peripheral and corneous areas appear translucent, or vitreous, and they affect processing and nutrient digestibility. Waxy sorghums contain larger starch granules and less protein in the peripheral endosperm than regular sorghums (Sullins and Rooney 1974, 1975).

The corneous and floury endosperm cells are composed of starch granules, protein matrix, protein bodies, and cell walls rich in cellulose, ί-glucans, and hemicellulose. Starch granules and protein bodies are embedded in the continuous, protein matrix in the peripheral and corneous areas (Seckinger and Wolf 1973; Hoseney et al. 1974). The protein bodies are largely circular and 0.4–2.0 ΅m in diameter (Taylor et al. 1984). High-lysine cultivars contain fewer and smaller protein bodies than do regular sorghums, and thus contain significantly less alcohol soluble kafirins (Seckinger and Wolf 1973; Paulis and Wall 1979). The starch granules are polygonal and often contain dents from the protein bodies. Their size varies from 4 ΅m to 25 ΅m, the average being 15 ΅m. Granules present in the corneous endosperm are smaller and angular whereas those in the floury endosperm are larger and spherical.

The opaque, floury endosperm is located near the center of the caryopsis. It has a discontinuous protein phase, air voids, and loosely packaged, round, starch granules (Hoseney et al. 1974). The presence of air voids diffracts incoming light, which gives an opaque or chalky appearance.

The germ is diploid due to the sexual union of one male and one female gamete. It consists of two major parts: the embryonic axis and scutellum (Fig. 1). The embryonic axis contains the new plant and is divided into a radicle and plumule. Upon germination and development, the radicle forms primary roots whereas the plumule forms leaves and stems. The scutellum is the single cotyledon and contains reserve nutrients, i.e., moderate amounts of oil, protein, enzymes, and minerals, and serves as the bridge or connection between the endosperm and germ.

The kernel or grain is considered a naked caryopsis although some African types retain their glumes after threshing (Saunders 1955). Caryopses differ widely in weight (3–80 mg), test weight (708–785 g L^-1) and density (1.15–1.38 g cm-³). Commercial US sorghums are generally 4 mm long, 2 mm wide, and 2.5 mm thick with a kernel weight of 25–35 mg, test weight of 747–772 g L^-1 and density of 1.28–1.36 g cm^-3 (Rooney and Miller 1982).

McDonough and Rooney (1990) monitored grain development of six sorghum cultivars using scanning electron microscopy. The pericarp begins to differentiate into distinctive layers at 3–6 days after anthesis (DAA). It is fully developed at 6–9 DAA and, thereafter begins to compress. Cells in the testa layer are apparent at 3–6 DAA, whereas the aleurone layer takes longer to develop (6–12 DAA). Ovary walls contain simple and compound starch granules at anthesis (Glennie et al. 1984; McDonough and Rooney 1990). Simple starch granules and protein bodies begin to develop in endosperm cells at 3–6 DAA. Protein bodies at this stage are covered by a thin filamentous webbing, which later develops into a distinctive protein matrix. During the milk stage (8–14 DAA), the endosperm cell walls contain a large number of holes to facilitate translocation of storage and biochemical components. The holes in the cell walls gradually disappear by the hard dough stage (20–25 DAA). The endosperm gains most of its storage material by the hard dough stage of caryopsis development, while moisture content decreases continuously from soft dough stage (15–19 DAA) (Glennie et al. 1984). Physiological maturity (or black layer) marks the end of nutrient delivery and the beginning of senescence and desiccation of the caryopsis. The vitreous endosperm has a continuous protein matrix, which is attached to the starch granules, protein bodies, and cell walls (McDonough and Rooney 1990). The floury endosperm has a discontinuous protein matrix with many small voids between the starch granules. Rate of endosperm development is faster in sorghum with hard endosperm than that with softer endosperm.

In many regions of the world, grain mold is a major disease of sorghum. Grain weathering is the added deterioration caused by the environment, sunshine, precipitation, and insects, during caryopsis maturation and exposure to ambient conditions in the field. This reduces sorghum grain yield and quality, which affect physical properties, processing, and nutritional and market value (Glueck et al. 1978).

Mycotoxins are less of a problem in sorghum than in maize (Zea mays L.). Ergosterol levels correspond to mold colonization of weathered sorghum (Seitz et al. 1979) but have not been linked to animal health problems. Fumonisin has not been found at significant levels in sorghum probably because a different species that does not produce fumonisin, Fusarium thapsinum Klittich, Leslie, Nelson et Marasas sp. nov. colonizes sorghum (Klittich et al. 1997). Aflatoxin is not found at significant levels in the field but does appear on improperly stored, high-moisture sorghum (G.N. Odvody, Texas A&M University, Corpus Christi, USA, 1999, personal communi-cation). These differences cause aflatoxin levels to be excessive in maize while sorghum is aflatoxin free in similar drought-prone environments.

Ergot, caused by Claviceps africana Fredrickson, Mantle, & de Milliano, has recently infested sorghum in the Western Hemisphere (Isakeit et al. 1998). The sclerotia contain alkaloids that may cause problems when fed to animals (Berde and Schild 1978). However, significant livestock or human health problems have not occurred in Africa and Asia where ergot has existed for centuries. This may change as more sorghum production occurs in cooler, moist environments. The hybrid seed industry must limit ergot infestation since the organism invades sterile florets and prevents seed production.

An array of fungal species, the most prominent being Curvularia lunata (Wakker) Boedijn, Fusarium spp, Alternaria spp, Phoma sorghina (Sacc.) Boerema, Dorenbosch, & van Kesteren, and Dreschlera spp, cause grain molding in the field (Castor and Frederiksen 1980). Wet, humid weather cause molding of caryopses after anthesis and post-physiological maturity. Limiting grain mold in sorghum has been very difficult. Chemical control is cost -prohibitive, and biological control mechanisms have not been feasible. The most cost-effective and realistic method to control grain mold is by genetic resistance. However, selection for grain mold resistance is difficult because numerous genetic factors are reported to influence grain mold resistance.

Plant traits such as panicle shape, plant height, and glume structure have been associated with grain mold resistance (Castor and Frederiksen 1980; Rao and Rana 1989). Caryopsis traits such as endosperm hardness, a pigmented testa layer, and red pericarp color are related with grain mold resistance (Glueck and Rooney 1980; Jambunathan et al. 1992; Esele et al. 1993; Audilakshmi et al. 1999). Although many sorghums with tannins and higher levels of phenol-based pigments are resistant to molding, these compounds cause dark colors, astringency, and/or decreased nutritional value in foods or feeds (Earp et al. 1983; Hahn et al. 1984). Accordingly, sorghums for food use are either grown in environments that are not conducive to deterioration, with a correspondingly lower yield, or cultivars resistant to grain mold are selected which yield less desirable food. Hence, grain mold resistance is necessary in food-type sorghums. Many food-type sorghums have been developed that have improved grain mold resistance; but none, currently, are resistant to grain mold in hot, humid environments. Many red pericarp sorghums have more resistance to grain mold than the white pericarp sorghums. None of these traits solely explains the variation in grain mold resistance found in sorghum. Additional factors necessarily underlie the visible factors in the expression of grain mold resistance. Continued efforts to overcome the limitations imposed by grain molds are required to facilitate increased production and utilization of sorghum in warm, humid environments. The following two sections discuss tannin and antifungal protein-based approaches to grain mold resistance in sorghum.

Sorghum is unique among major cereals because some cultivars produce polymeric polyphenols known as tannins (Butler 1990). All sorghums contain phenols and most contain flavonoids; however, only cultivars with a pigmented testa, B₁_B₂ _ genes, produce condensed tannins or pro-anthocyanidins. Most cultivated sorghums do not contain condensed tannins even though non-tannin, phenolic compounds are occasionally reported as tannins. Sorghums are classified as type I (without tannins), type II (tannins present in pigmented testa), or type III (tannins present in pigmented testa and pericarp). Tannins have antioxidant properties and are currently being considered as nutriceuticals (Hagerman et al. 1998).

Phenolic compounds have been divided into three major categories: phenolic acids, flavonoids, and tannins (Chung et al. 1998). Phenolic acids are derivatives of benzoic or cinnamic acid (Fig. 2). Flavonoids consist of two units: a C6-C3 fragment from cinnamic and a C6 fragment from malonyl-CoA. The major groups of flavonoids in sorghum are the flavans— flavan-3-en-3-ols with double bond between C3 and C4 and hydroxylated at C3 are anthocyanidins. Tannins are polymers of 5–7 flavan-3-ol units (catechin) linked through acid labile carbon-carbon bonds (Hahn et al. 1984; Mehansho et al. 1987b; Butler 1990).

Sorghum containing tannin is called tannin or brown sorghum even though the pericarp color may be white, yellow, or red. Grain appearance is not necessarily related to tannin presence or content. Phenolic compounds and tannins have been reported in much higher levels (10- to 100-fold) in pericarp, glumes, and leaf sheaths than in the endosperm. Phenols isolated from the leaf sheaths (Sereme et al. 1993) and stems (Rey et al. 1993) can be used as pigments for clothes, pots, etc. Tannins (proanthocyanidins) apparently occur only in the pericarp, pigmented testa layers (Hahn and Rooney 1986), and glumes (Doherty et al. 1987) of some type II or type III sorghums but not in glumes of type I sorghums.

Figure 2. Basic structures of phenolic acids, flavonoids, and tannins (A = cinnamic acid; B = benzoic acid; C= the major flavonoid groups are flavanones (carbonyl at 4), flavonols (carbonyl at 4, hydroxyl at 3), flavones (carbonyl at 4, double bond between 2 and 3), and flavans (no carbonyl at 4); the major flavans are leucoanthocyanidins (h~ls at 3 and 4) and catechin (hydroxyl at 3, double bond between 3 and 4); and D = proanthocyanidin (tannin) polymer (n= 2-4) (Source: Rooney and Serna-Saldivar 2000, with Permission).

Hahn et al. (1983) separated, by reverse phase HPLC (high performance liquid chromatography), free and bound phenolic acids of sorghum. Eight main phenolic acids with different polarity were identified in extracts (Table 1). A typical brown sorghum contained the highest amount of free phenolic acids. Sorghum cultivars resistant to fungal attack contained both a greater variety and larger amounts of phenolic acids in the free form. Waniska et al. (1989) also partitioned sorghum phenolic acids and concluded that white cultivars without pigmented testa contained the lowest amount of phenolic acids. Brown cultivars contained higher levels of free phenolic acids and compounds and were more resistant to grain weathering. Phenolic acids and compounds increase during caryopsis development with a maximum at physiological maturity (and a decrease afterwards). This decrease may be due to decomposition, covalent linkage to structural polymers, or to entrapment with the solid endosperm matrix. Higher levels of phenolic compounds relate to improved grain mold resistance. This observation, however, may be the simple result of the strong correlation of phenolic compounds with tannins, since most tannin containing sorghums are more resistant to grain mold.

Tannins protect the grain against insects, birds, fungi, and weathering (Waniska et al. 1989). Rates of pre-harvest germination or early sprouting are lower in brown sorghums. These beneficial effects ensure that brown sorghums will continue to be produced in certain pest-ridden areas of the world (Butler 1990). These agronomic advantages are accompanied by nutritional disadvantages and reduced food quality. The antinutritional effects of tannin sorghums have been reviewed (Butler 1990; Butler and Rogler 1992; Chung et al. 1998; Hagerman et al. 1998).

The colors of sorghum grain and flour play an important role in its acceptance. In general, white sorghums produce the most acceptable colored food products. Food color is the result of factors such as grain color, pericarp color, pigmented testa, endosperm color, presence of tannins, degree of milling, and pH of the food system. Pigmentation in the pericarp and testa is primarily due to phenolic compounds. The color intensity greatly depends on pH. Anthocyanins are very unstable in acid medium and are readily converted to the corresponding anthocyanidin under slight acidic conditions (Hahn and Rooney 1986).

Pericarp color of sorghum appears to be due to a combination of anthocyanin and anthocyanidin pigments as well as other flavonoid compounds (Hahn and Rooney 1986). Kambal and Bate-Smith (1976) found no flavonoids in the pericarp of a white sorghum whereas Nip and Burns (1971) found four major anthocyanins in the pericarp of six white sorghums. Lutoforol was identified as the major pigment in red pericarp sorghums (Nip and Burns 1969). Cooking in alkali, i.e., alkaline tτ and tortillas, usually promotes color formation especially in non-white sorghums. The glume or secondary plant color also affects the appearance of the caryopsis and food products. Lighter colored products are prepared using grain with straw-colored rather than darker glumes.

Several methods have been used to characterize phenolic compounds and tannins in sorghum (Maxson and Rooney 1972; Butler 1982; Hahn et al. 1984; Hagerman et al. 1997) (Table 2). Many tannin methods measure phenolic compounds, which may or may not be condensed tannins. The amount of tannin in the sorghum kernel is impossible to determine because a significant proportion cannot be extracted and assayed (Hahn and Rooney 1986; Butler 1990) and a suitable standard for sorghum tannins is unavailable. The highly polymerized or condensed tannin molecules are the most difficult to extract which may account for the decreased extractability of tannins as caryopses mature. Different solvents and methods yield different tannin values because they extract and measure different chemical parts of the tannin molecule (Hagerman and Butler 1989). Methods that measure the reducing power of phenolic compounds are not specific for tannins. Protein precipitation methods do not differentiate between the types of tannins; they are sensitive to reaction conditions and the protein used. The acidic butanol method is specific for proanthocyanidins but other anthocyanidins in sorghum interfere with the reaction. The vanillin-HCl method using catechin as the reference with blanks subtracted indicate that significant levels of tannins are only present in type II and type III sorghums (Table 3). Tannins in type II sorghums are not extracted with methanol but with acidic methanol. Low values of tannin in type I sorghums are due to interfering compounds; hence these values are erroneous. Tannic acid is not present in sorghum, even though this laboratory and others have used tannic acid as a reference compound.

The antinutritional effects of tannins include diminished growth rate, protein digestibility and feed efficiency in rats, hamsters, swine, poultry, and ruminants (Jambunathan and Mertz 1973; Maxson et al. 1973a, 1973b; Cousins et al. 1981; Muindi and Thomke 1981; Mehansho et al. 1987a, 1987b; Knabe 1990; Mole et al. 1990, 1993). Tannins in sorghum reduce digestibility and efficiency of utilization of absorbed nutrients from 3% to 15%. Most animals consume brown or tannin sorghums at the same or slightly higher rates but do not gain as much weight.

Extracted tannins bind to proteins and inhibit many enzymes in in vitro assays (Hagerman and Butler 1994; Hagerman et al. 1997). Tannins from sorghum stimulate production and secretion of proline-rich, salivary proteins which bind tannins during mastication in many mammals (Asquith et al. 1987; Mehansho et al. 1987a); and some tannin-protein complexes are recovered unchanged in the feces (Butler and Rogler 1992). Tannin labeled with C¹⁴ was not absorbed by chicks; however, C¹⁴ labeled unknown polyphenols were absorbed and recovered in serum and other tissues. Apparently, lower molecular weight, absorbable polyphenols cause the systemic effects attributed to tannins (Butler and Rogler 1992).

Phenolic compounds (monomers, polyphenols, and tannins) do not interfere with absorption of iron in the gastrointestinal lumen. Iron absorption, however, is strongly inhibited by tannic acid (not present in sorghum) and gallic acid followed by chlorogenic acid; whereas, no inhibition was observed when catechins or condensed tannins were added to the diet (Brune et al. 1989).

Several approaches for detoxification of brown sorghums have been proposed. The most common and practical ways are by decortication and malting. Decortication removes the pericarp and testa and, therefore, most tannins. However, decortication is inefficient because most brown sorghums have soft endosperm texture (Chibber et al. 1980; Nwasaru et al. 1988; Reichert et al. 1988). Parboiling of brown sorghum prior to decortication is a better alternative for tannin removal (Young et al. 1990). Chemical decortication with alkali solutions followed by neutralization with acid decreases the amount of assayable tannins (Kock et al. 1985; Butler 1990; Waichungo and Holt 1995). Steeping alone or steeping and malting effectively lowers measurable tannins up to 43% (Osuntogun et al. 1989). Sorghums treated with moisture and wood ashes lowers measurable tannins, up to 97%, improves protein digestibility and increases weight gains in rats (Mukuru et al. 1992). Similar effects are achieved when brown sorghums are treated with Magadi soda. Several experimental alkaline treatments to detoxify brown sorghums were proposed by Price et al. (1979), Butler (1990), and Waichungo and Holt (1995).

Proteins inhibitory to fungal growth have been identified in sorghum endosperm. Extracts of immature and mature hard and soft endosperm areas were inhibitory to growth of Fusarium moniliforme Sheld. (Kumari et al. 1992; Kumari and Chandrashekar 1994). Hard endosperm sorghums had more of three proteins that exhibited antifungal protein (AFP) activity. One of these proteins, a mycelia-hydrolyzing protein, probably is related to a cysteine protease inhibitor in pearl millet (Pennisetum glaucum (L.) R. Br.) (Joshi et al. 1998).

Seetharaman et al. (1996) identified three other AFPs, sormatin, chitinase, and glucanase, that increase during caryopsis development; they were high at physiological maturity and decrease at combine harvest maturity of the grain. Ribosome-inactivating protein (RIP) levels were high at 15 DAA, and then subsequently decreased. Effectiveness of sorghum AFPs in vitro was demonstrated when a mixture of sormatin, chitinase, glucanase, and RIPs (110 ΅g ΅L^-1) inhibited spore germination and ruptured hyphal tips in both C. lunata and F. thapsinum (Tables 4 and 5), but only inhibited hyphal elongation for <36 h (Seetharaman et al. 1997). These AFPs are positively charged and water-soluble and elute from an anionic column with a salt gradient. The amount of AFPs in physiologically-mature sorghum caryopses was estimated to be approximately 260 ppm (30–35 mg in each caryopsis).

Sormatin is a thaumatin-like protein, i.e., a small, basic protein, approximately 22 kDa, which acts by causing membrane permabilization (Vigers et al. 1991:). A ί-1,3-glucanase of 30 kDa was identified in sorghum (Darnetty et al. 1993; Krishnaveni et al. 1999a, 1999b). Chitinases hydrolyze the N--acetylglucosamine polymer, chitin, a compound of fungal cell walls (Boller 1985; Yun et al. 1997). Several endogenous and stress inducible chitinases ranging from 21 kDa to 33 kDa were reported in sorghum (Darnetty et al. 1993; Krishnaveni et al. 1999a, 1999b). The RIPs range from 28 kDa to 31 kDa and inhibit protein synthesis in target cells by specific ribonucleic acid (RNA) N-glucosidase modification of 28s RNA (Logemann et al. 1992). Maize RIP antibodies cross react with a 30 kDa sorghum protein (Seetharaman et al. 1996). Results from these investigations suggest that understanding the interplay between AFP levels in caryopses, i.e., stage of maturity and environment (humidity/imbibition), and time of pathogen infection is important to maximize the bioactivity of AFPs in vivo.

Mobility of AFP within caryopses is a prerequisite for their possible role as a fungal growth inhibitor. Sorghum AFP leached from immature caryopses, but was retained in the pericarp of mature caryopses during water imbibition (Seetharaman et al. 1996). A similar study by Swegle et al. (1992) on the mobility of barley (Hordeum vulgare L.) chitinases is consistent with how AFPs help protect caryopses and seedlings from pathogenic fungi before and during germination. They found that AFPs were released from the endosperm into surrounding environment 24 h after water imbibition. If AFPs are mobile and ‘bound’ to the pericarp during imbibition, then AFPs could be concentrated in <10% of the caryopsis. This would increase AFP concentration about 10-fold, thereby, increasing their antifungal potential.

The AFPs are potentially important in fungal inhibition in the field; however, field data has been unclear about the relationships of AFP and grain mold resistance. Seetharaman et al. (1996) were the first to observe a significant inverse correlation coefficient between sormatin content in caryopses at 30 DAA with sorghum grain mold rating at harvest time. Subsequently, this was observed in seven environments with high levels of grain mold incidence (Rodriguez-Herrera et al. 1999). Levels of four AFPs were significantly affected by the 16 F_2:5 lines in all environments with higher levels of sormatin and other AFPs observed in the resistant lines compared to the susceptible lines (Fig. 3). The data strongly suggest that sormatin is associated with grain mold resistance. Orthogonal contrasts showed that the ί-1,3-glucanase levels were significantly higher in the resistant group in only three of seven environments. This suggests that ί-1,3-glucanase may be associated with grain mold resistance in sorghum. However, its function on grain mold resistance may not be as direct as other AFPs or perhaps the ί-1,3-glucanase level in plants is also a result of other mechanisms of response, different from a disease-resistance response. Chitinase and RIP concentrations in the resistant compared to susceptible lines were 1.5- to 14-fold higher and were associated with grain mold resistance.

Sormatin and chitinase contents were not significantly different (P <0.05) between resistant and susceptible lines at Halfway, Texas, USA in 1996. This supports the constitutive expression of sormatin in caryopsis. Sormatin and chitinase contents in caryopses of resistant and susceptible lines grown under a grain mold-free environment were higher than when grown in environments with grain mold incidence. There are several possible causes of this result. Resistant lines may maintain higher levels of sormatin for longer periods than susceptible lines under grain mold pressure and/or resistant lines may induce more AFP production than susceptible lines upon infection. Resistant lines had higher levels of sormatin one or two weeks after being stressed with grain mold fungal species compared to susceptible lines (Bueso et al. 2000). Resistant tissues accumulated chitinases more rapidly and in some instances to higher final concentrations than susceptible tomato tissues (Punja and Zhang 1993). Since AFP levels were affected by environment/fungal pressure and by degree of resistance, proteins that exhibit antifungal properties appear to have an important role in sorghum grain mold resistance.

Correlation estimates among the AFPs were positive and significant at most of the environments, except when grown in a grain mold-free environment. A significant and positive correlation indicates that under grain mold incidence, as one AFP increases in the sorghum caryopsis, so do the others. Co--expression of all AFPs seems to be a strategic biochemical action of the plant’s resistance mechanism to grain mold. In vitro studies with Trichoderma reesei E. Simmons and Fusarium sporotrichioides Sherb demonstrated that a combination of barley RIP and chitinase inhibit fungal growth more efficiently than do either enzyme alone (Leah et al. 1991).

Figure 3. Mean antifungal protein levels (~g gl) in mature caryopses of eight grain mold resistant (GMR) and eight susceptible (GMS) lines grown in seven environments. Resistant lines have more AFP than susceptible lines except for glucanase. For glucanase means followed by the same letter are not significantly different (within environments).

Variations between resistant and susceptible lines in wet and dry field environments and when inoculated with F. thapsinum and C. lunata at anthesis were studied. Mold resistance was based on the history of each cultivar during previous growing seasons (Table 6). We intentionally chose those differing in mold resistance and AFP content. Hence, there were

resistant-high AFP, resistant-low AFP, susceptible-high AFP, and susceptible-low AFP cultivars. Mold rating of sorghums grown at Halfway was 1.35±0.15 indicating that the environment was not good for grain mold. Mold ratings between the resistant and susceptible cultivars exhibited no remarkable distinctions at 30 DAA even with fungal inoculation and sprinkling treatments. Accordingly, mold resistance did not correlate significantly with mold rating scores or chitinase and sormatin levels at 30 DAA. Nevertheless after 30 DAA, mold growth was favored by a warm-humid environment in the field and susceptible sorghums deteriorated. Hence, mold ratings increased considerably at 50 DAA and corresponded inversely with mold resistance.

Figure 4. Scatter plot of chitinase in caryopsis at 30 days after anthesis (DM) and grain mold rating of cultivars at 50 DM. Labels are in the following format: pericarp color-presence of tannins-endosperrn hardness (grain mold resistance); where W = white pericarp, R = red pericarp, tannins = tannins present, no = no tannins, H = hard endosperrn, M = medium endosperrn hardness, L = low endosperrn hardness, S = susceptible, and R = resistant.

The sorghums varied in maturity, pericarp color and composition, and endosperm hardness (Table 6), traits that have been associated with grain mold resistance (Bandyopadhyay et al. 1988; Esele et al. 1993; Menkir et al. 1996). At 50 DAA, two of three sorghums with pigmented testa, i.e., condensed tannins, had less deterioration but the third sorghum with pigmented testa molded (Fig. 4). Low mold ratings were observed in two sorghums with red pericarp and no tannins and in two sorghums with white pericarp and no tannins. Three other sorghums with white pericarp and no tannins exhibited much mold damage. So red pericarp color (vs white) was related to less molding, although the red pericarp sorghum, BTx638, exhibited more mold damage in previous years. Two sorghums with hard endosperm had the same low deterioration, as did sorghums with low and medium endosperm hardness; whereas the four sorghums with the most deterioration had low or medium endosperm hardness. Hard grain had low mold ratings when stressed with sprinkling or inoculation with fungal pathogens and loss was less during decortication. This trend was also observed in control and the combined treatment but with P<= 0.1.

Previous reports indicated that pigmented testa, red pericarp color, and hard endosperm confer mold resistance to sorghum (Castor and Frederiksen 1980; Hahn and Rooney 1986). Presence of proanthocyanidins but not flavan-4-ols and 3-deoxyanthocyanidins correlated with grain mold resistance (Melake-Berhan et al. 1996). Mold rating at 50 DAA was less when sorghum had red pericarp and/or hard endosperm traits. However, hidden factors, such as the presence of AFPs in the caryopsis might be involved. Seetharaman et al. (1996) noted differences in the mobility and extractability of chitinase and sormatin from sorghums of varying tannin contents. They found that Type I sorghums had variable levels of AFPs, Type II had increased sormatin and unchanged chitinase, while Type III had decreased AFPs. Therefore, they speculated that several mechanisms could interact to achieve grain mold resistance.

Amounts of AFPs in the caryopsis at 30 DAA were affected by field conditions of College Station and Halfway, Texas. Sormatin levels in caryopses were either higher (5) or unchanged (5) for sorghums grown in College Station compared to those in Halfway. Chitinase levels in resistant sorghums were either decreased (3) or unchanged (2) while chitinase levels in susceptible sorghums increased in three, decreased in one, and was unchanged in another cultivar when grown in College Station compared to those in Halfway. This suggests that less than ideal field conditions promote higher levels of sormatin and chitinase in susceptible cultivars and higher levels of sormatin but lower levels of chitinase in resistant cultivars.

The control caryopsis response between 30 and 50 DAA varied in AFP content and amount (Fig. 5). Sormatin increased in four resistant cultivars while chitinase increased in only one resistant cultivar. Sormatin and chitinase decreased in one resistant and all susceptible cultivars. It appears that elevated humidity induced AFP synthesis in some resistant cultivars but not in any susceptible cultivars. This partially contradicts the findings of Seetharaman et al. (1996) where AFPs decreased in caryopses after 30 DAA; however, the results could be evidence of an active defense response of AFPs in the caryopsis.

Figure 5. Changes in antifungal proteins as effected by sprinkling and/or inoculation in caryopses at 50 days after anthesis (DM). (All values were compared to the control at 30 DM in College Station, Texas, USA.)

The sprinkling treatment (intended to mimic brief showers) was conducted to determine mobility of AFPs in vivo. Sprinkling decreased sormatin in most cultivars at 30 and 50 DAA. Sprinkling increased sormatin in one susceptible cultivar, BTx638, at 30 DAA and one resistant cultivar, SC719-11E, at 50 DAA. Likewise, sprinkling decreased or did not change chitinase in most cultivars; however, sprinkling increased chitinase in two resistant cultivars, SC719-11E and Hegari*Dobbs, at 30 and 50 DAA respectively. Sprinkling caused mobility or a loss of AFPs in the caryopsis of most susceptible and several resistant cultivars.

Panicles were inoculated with fungal pathogens to determine whether AFPs were accumulated actively or passively. AFP levels decreased in most susceptible cultivars at 30 and 50 DAA (Fig. 5). Inoculation and fungal stress increased sormatin and chitinase at 30 DAA in two resistant cultivars (Malisor and R9025) and sormatin in two susceptible cultivars (IS 2319 and BTx638). Sormatin increased at 50 DAA in one susceptible and three resistant cultivars and chitinase increased in two resistant cultivars. Sormatin levels at 50 DAA increased more consistently than chitinase levels in resistant cultivars. Fungal stress induced AFP synthesis more in resistant than in susceptible cultivars. Mold resistance appears to be related to the degree of AFP induction and not on the relative amount of AFP in the unstressed sorghum caryopsis.

The caryopsis response to inoculation and/or sprinkling between 30 and 50 DAA varied in AFP content and amount (Fig. 5). Most cultivars had unchanged or decreased AFP levels at 30 and 50 DAA. The resistant cultivar, SC719-11E, had increased sormatin and chitinase at 30 and 50 DAA. Susceptible BTx638 had increased sormatin at 30 DAA while two resistant cultivars, Malisor and Hegari*Dobbs, had increased chitinase at 30 DAA. The resistant-like behavior of BTx638, a susceptible cultivar, may be an anomaly of the growing season, misclassification of resistance, or due to other factors. The combined treatment of fungal inoculation and sprinkling resulted in AFP changes more similar to the sprinkling than to the inoculation treatment. Sprinkling generally caused reduced AFP in all cultivars, while inoculation increased AFP levels in some of the resistant cultivars.

The resistant cultivar that initially had low levels of AFP (SC719-11E) was induced by the treatments to accumulate more sormatin and chitinase at 30 and 50 DAA (Fig. 5). The other resistant cultivars, except Sureno, responded to one or more of the stresses with higher levels of sormatin or chitinase during caryopsis development. The susceptible cultivars, regardless of AFP level, responded with lower levels of AFP with only one exception. This implies that the amount of AFP in caryopsis is less important than induction of AFP accumulation in the caryopsis by stresses leading to grain molding.

Mold rating of sorghum caryopses did not correlate to AFPs (discussed earlier), but correlated with change (%) in AFP (Table 7). Significant correlation coefficients were obtained at 50 DAA but not at 30 DAA. Grain molding increased during 30 to 50 DAA. Mold rating at 50 DAA was negatively correlated to change (%) in sormatin after inoculation and in chitinase after inoculation and sprinkling. This means that less fungal infection occurred when AFP levels increased. The imposed stresses either diminished AFP content in some cultivars and/or induced AFP synthesis in other cultivars. Hence, inherent grain mold resistance was more clearly expressed in less ideal field conditions. Conversely, grain mold resistance was associated with the amount that AFP synthesis that was induced by the stress.

We attempted to determine how anatomical tissues in the caryopsis respond to stresses (Table 8). We chose wounding the embryo by puncturing and by soaking as stresses to determine which tissues (embryo, endosperm, or pericarp) were active in the defense mechanism. Evidence of increased AFP induction in sorghum caryopsis in response to stress (wounding or soaking) was demonstrated using white sorghums without a pigmented testa [Dorado (resistant to grain mold) and RTx2536 (susceptible to grain mold)]. Soluble AFP in caryopsis (from caryopses 15, 30, and 50 DAA) were identified using antibodies for chitinase, glucanase, or sormatin. More chitinase was in the endosperm while more glucanase was in the pericarp of the controls. Most of sormatin was found in the endosperm of Dorado and in the pericarp of Tx2536. We observed differences in how each AFP accumulated after wounding and/or soaking of caryopses. Very little AFP leached from caryopses. AFPs were inducible in both the resistant and the susceptible cultivar. When caryopses were stressed, changes in AFP were greater in the endosperm where chitinase and sormatin increased, while AFP levels generally decreased in embryo and pericarp. Wounding caryopses (30 DAA) increased chitinase in embryo and endosperm and sormatin in endosperm. Hence, chitinase and sormatin levels were affected by these stresses. Much needs to be learned about what affects AFP levels, sequence of events, tissue location of accumulation, potency in vivo of AFP against fungi, etc.

Our long-term interest is to limit deterioration of sorghum to enable its use in food products. Our data suggests various phenolic compounds and AFPs can play roles in limiting fungal infection of caryopses. We know, from previous research, that many other chemical, structural, and biological mechanisms are involved in lowering grain molding and weathering. Caryopses resistant to grain molding respond to some stresses by accumulating chitinase and sormatin than did susceptible caryopses. Pyramiding the defense mechanisms (phenols, AFPs) with other attributes that contribute to grain mold resistance is a good strategy to increase grain mold resistance.

We thank God for all the hard work and dedication of legions of former graduate students, colleagues, and scientists from all over the world, especially those from the Texas A&M University Sorghum Improvement Program, Texas, USA who have provided so much inspiration and reliable information and help over the past 30 plus years. Long-term support of the Texas Agricultural Experiment Station and private industry is appreciated. This research has been partially funded since 1979 by the Title XII Sorghum and Millet Collaborative Research Support Program, INTSORMIL under Grant AID/DSAN/XII/G-0149 from the Agency for International Development (AID), Washington, DC, USA.