Puccinia graminis (stem rust of cereals)

P. graminis spreads by urediniospores with a generation time of 14+ days. More than two generations can lead to crop loss. The pathogen has spread worldwide from western Asia with the movement of crops. Local and regional spread is by wind but intra-continental spreads are unclear although rare wind patterns could be responsible. Initially, most stem rust incidents were caused by the movement of people and air travel has been involved in the recent movement of other rusts. Within each formaspecialis there are a number of different races (virulence combinations) that are due to mutation and asexual reproduction and more rarely due to sexual recombination on Berberis and Mahonia spp. Races moving from one continent to another are a major threat to wheat, an important food crop worldwide. P. graminis has forms that attack oat, barley, rye, orchard grass, timothy, fescue, ryegrass, bluegrass, wild barley and wheat relatives.

Several subspecies and either varieties (var.) or formae speciales (f.sp.) (Eriksson, 1882) are recognized. According to the nomenclature of Savile (1984), based on the descriptions of Urban (1967), there are three distinct morphological forms: P. graminis subsp. graminis var. graminis; P. graminis subsp. graminis var. stakmanii; and P. graminis subsp. graminicola.The distinction of subsp. graminis and subsp. graminicola is justified based on morphological features, host ranges and distinct evolutionary histories (Urban and Markova, 1984); it also simplifies the taxonomy. There is, however, less consensus regarding the morphologically based var. designations beyond the subspecies level. Many workers prefer an f.sp. nomenclature based on host specialization (Anikster, 1984). The appropriate taxonomy then becomes P. graminis subsp. graminis or subsp. graminicola, followed by an f.sp. designation.However, in most current literature, the causal fungi are directly named and the subsp. designation is not included; for example, P. graminis f.sp. tritici, f.sp. avenae, etc. A range of species, varieties, or formae speciales are reported in the literature, most of which have now been included under either P. graminis subsp. graminis or subsp. graminicola. An f.sp. form may be further subdivided into numerous races (or pathotypes).The taxonomy of the rusts was inaugurated in a 'quarta on new genera', by Senor Micheli in Florence, Italy, in 1729. An illustration of the first recognized genus of the rusts, Puccinia, named in honour of the lecturer and physician, Thomas Puccini (Arthur, 1928) is included in this publication. Scientific classification of the rusts began with Persoon (1801), who provided the first binomial for P. graminis and remains the authority for this and many other taxa.The taxonomy of the cereal rusts is complicated by morphological variation and specialization on numerous different hosts. From the morphological taxonomy of Savile (1984) and Urban (1967), a cladistic analysis by Baum and Savile (1985) separated var. stakmanii and var. graminis, based on urediniospore size and numbers of urediniospore germ pores.The oat and rye stem rusts have been placed in var. stakmanii, and the wheat stem rust in var. graminis (Savile, 1984). However, not all data support the above separations. In crossing studies, the wheat and rye stem rust fungi show greater genetic compatibility than the oat and wheat stem rust fungi, or the oat and rye stem rust fungi (Green, 1971a; Johnson and Newton, 1932; Stakman et al., 1930). Two-dimensional polypeptide mapping (Kim et al., 1984, 1987) supported the genetic data where the differences in numbers and distribution of polypeptides between the oat and either wheat or rye stem rust fungi were greater than those between the wheat and rye stem rust fungi. The inclusion of the oat and rye stem rusts into a single variety, separate from wheat stem rust, is questionable and indicates that purely morphological criteria are not adequate to clearly delineate intraspecific relationships in the rust fungi. Modern techniques, Anikster et al. (2005) may make it possible to make morphological separations that distinguish between the populations with unique host ranges.Host range may also be problematic. Tajimi (1976) noted that P. graminis f.sp. avenae, although classified as subsp. graminis, had a host range more like that of subsp. graminicola. In 1925, the Joint Committee on Nomenclature of the American Phytopathological Society recommended that the term forma 'should be applied to a subdivision of a species or variety which is characterized and distinguished primarily by physiological instead of morphological characters'. Johnson (1968) and Anikster (1984) put forward a strong case for host specialization as a taxonomic criterion. However, host specialization is complicated by numerous overlapping host compatibilities. For example, the tritici and secali forms are generally recognized by their adaptation to species of Triticum and Secale, respectively, but they also have common hosts including species of Agropyron, Elymus, Hordeum and Bromus.The forma speciales designations have been based either on the host species of the first observation or on the principal host affected. Despite the complexities of host specialization, most workers involved in cereal rust pathology find the f.sp. nomenclature more workable; the var. designations are mostly used with cultivated forage grasses, while f.sp. is commonly used worldwide. References to variations in P. graminis in the remainder of the text are based on host specialization (f.sp.).As indicated in the previous paragraphs, classification of P. graminis is confused by the use of two naming systems. Herbarium specimens currently require a morphological basis for classification, while the control of the disease is based on a host range classification (f.sp.). Classical genetic evidence for differences in P. graminis has been limited to due difficulty in making crosses between individuals in the various defined groups. Molecular genetics has solved this problem and current data indicates two or three subdivisons or even species within the current P. graminis Pers. Current data is similar but not identical to the f.sp. divisions (Leonard and Szabo, 2005). The problem is that the samples are mainly from initially introduced pathogen populations which probably lack the range of variation that exists. Furthermore, genetic data will probably always be lacking for most herbarium specimens.Races, or pathotypes, constitute a taxon below the forma speciales level. These are characterized by differences in physiological (pathogenic) reactions, resulting in different virulence/avirulence patterns on a selected set of differential resistance genotypes within a host genus. Pathogenic specialization has been most thoroughly studied for P. graminis f.sp. tritici and f.sp. avenae.The concept of pathogenic 'race' is often poorly understood. The differential set of Stakman et al. (1962) for P. graminis f.sp. tritici has been used as a worldwide standard and still provides the basis for race classification in some countries. One advantage of this system is that it provides a taxonomic base and a means of comparison and communication.However, the standard differentials are limited and include lines with complexes of several resistance genes. It is, therefore, not possible to define the virulences of fungal populations on a gene-for-gene basis. Although approximately 400 races have been identified using this system, the pathogenic variability within P. graminis f.sp. tritici is much higher. Rust fungal populations are highly variable within an ecological zone or in different regions of the world. It is unlikely that an internationally standardized set of differentials will be able to differentiate local virulences satisfactorily; this may have important implications in plant breeding for resistance.Virulence surveys relate the virulences within a rust fungal population to sources of resistance in breeding programmes and allow international communication. The standard set of differentials can be supplemented and entirely new sets of differentials can be devised; these are usually adapted to local requirements.Differentiation and nomenclature have been standardized for races of P. graminis f. sp. tritici (Roelfs and Martens, 1988) and P. graminis f.sp. avenae (Martens et al., 1979) for North America, but there are currently no worldwide standards for the identification of races of either of these fungi. Roelfs (1984, 1985b) summarized major race nomenclature systems for f.sp. tritici and f.sp. avenae. None of these systems has proved totally satisfactory, but the accumulation of information on the distribution of virulences may help in the design of an internationally usable system in the future.The traditionally used differentials for f. sp. avenae have been single-resistance gene lines; these are included in currently used differential sets; worldwide comparisons of virulence are, therefore, much easier for this fungus. International virulence surveys of f.sp. tritici (Luig, 1983) and f.sp. avenae (Martens et al., 1976), using enlarged common sets of differentials, provide a useful means of comparing international virulences.

P. graminis is a macrocyclic, heteroecious rust, with five distinct spore stages. The characteristics and ontogeny of each of the spore stages have been described at the ultrastructural level by Harder (1984).The urediniospores of P. graminis are dikaryotic (n+n), dehiscent, thick-walled and covered with spines. They are elliptical and about 20 x 30 µm.The teliospores are two-celled, thick-walled (with up to five wall layers) and are thickened at the apical end. The formation and characteristics of teliospores are described by Mendgen (1984).For further information, see Biology and Ecology.

The table of geographic distribution does not distinguish between the various formae speciales as they may affect different crops in different regions. Where appropriate, these distinctions are made in the Epidemiology section of the text on Biology and Ecology.

P. graminis occurs in most areas of the world, either on wild grasses or cultivated cereals. Most of the geographic information concerns distribution on cultivated cereals, mainly on wheat. Information on the occurrence on wild grasses is more limited. Detailed surveys of grasses would probably reveal the presence of P. graminis in countries where occurrence is not indicated here.

Although stem rust on wheat has largely been controlled worldwide by the use of resistant cultivars (Roelfs et al., 1992), the disease is potentially a continuous problem in some areas. Areas where there has been a consistent disease problem, or where there is greater potential for disease, are indicated here as 'widespread'. Most regions of the world have conditions that are normally more marginal for rust development.

A new and virulent race of wheat rust, Ug99, has spread from East Africa to Yemen on the Arabian Peninsula. Ug99 has been recorded in Uganda, Kenya and Ethiopia and has now been confirmed in Yemen (FAO, 2007; HULIQ News, 2007), Iran (Nazari et al., 2009) and South Africa (Visser et al., 2011). For further information on the status, likely migration and strategies to mitigate the threat to wheat production from Ugg99, see Singh et al. (2006, 2007) and Joshi et al. (2008).

Consideration of the geographic distribution of the pathogen is complicated by prevailing climatic conditions, the movement of global air masses, geographical features, the availability of alternative grassy hosts or the alternate sexual host, and cropping practices. Geographic distribution must, therefore, be discussed in terms of disease epidemiology (see section on Biology and Ecology).

Teliospores of P.graminis have been found on wheat glumes dating to 3300 years ago in Israel (Kislev, 1982). The early history of wheat rusts was reviewed by Chester (1946). Reports before 1600 do not distinguish between the rusts. P. graminis on wheat was widely spread across Europe by the early 1700s. In 1660, in Rouen, France, a law was introduced to eradicate barberry, the alternate host of P. graminis, as farmers correctly believed it was responsible for wheat rust. In North America a law to eradicate barberry was enacted in Connecticut colony in 1726. The Jesuit colony in Rio Grande do Sul, Brazil, also reported crop failures due to stem rust in the 1700s. It is difficult to trace the initial introduction of P. graminis from western Asia to the rest of the world, but it was early. In North America an epidemic of stem rust in spring wheat in 1878 was described by Hamilton (1939). Additional epidemics occurred in 1904, 1916, 1919, 1923, 1925, 1927, 1929, 1935, 1953 and 1954 (Stakman and Harrar, 1957). The eradication of barberry by the late 1920s and the development of resistant cultivars has reduced the number of serious epidemics of the disease. The 1935 epidemic was due to race 56 and the susceptibility of the cultivars to that race. The epidemics of 1953 and 1954 resulted from the appearance of race 15B and its ability to overcome the Sr11 resistance in Lee. Since 1955, resistance breeding has kept ahead of the pathogen in stem rust prone areas of the northern plains. Overwintering in the southern plains has been reduced due to both cultural practices and resistant cultivars.

Although rust fungal urediniospores can travel across continents, most P. graminis populations remain within their larger epidemiological zones (Saari and Prescott, 1985). In this respect, the strategy of most cereal-breeding programmes is to incorporate resistance that is effective against the virulences in the local P. graminis populations. Researchers may wish to import exogenous isolates to use as testers because of their unique virulence or other properties, but there is a risk associated with importing such isolates because they may contain virulences that pose a threat to locally used resistance sources. Most countries do not list rust fungal species as quarantine pests. Some countries, such as Australia, have strict regulations regarding the introduction of any foreign biological material. Even with such strict regulations, it is not always possible to avoid new disease risks. There are two well-known instances where new cereal rust diseases have been inadvertently introduced into new areas aided by human travel (probably as urediniospores attached to items of clothing): the introduction of barley stripe rust into Columbia in 1975 (Stubbs, 1985); and wheat stripe rust into Australia in 1979 (Saari and Prescott, 1985; Wellings and McIntosh, 1981). Both diseases have since become widespread in the new areas.A new and virulent strain of P. graminis, Ug99, has spread from East Africa to Yemen on the Arabian Peninsula. Ug99 has been recorded in Uganda, Kenya and Ethiopia and has now been confirmed in Yemen. According to FAO (2007), the strain in Yemen may be more virulent than the one in East Africa and there is a high risk of the disease spreading to Sudan.

The spores of wheat rust are mostly carried by wind over long distances and across continents. P. graminis has spread worldwide wherever its host, a cultivated grass, is grown. The pathogen survives the non-wheat growing season on volunteer plants (self-sown) locally or elsewhere in the epidemic region. After crop planting the rust spreads from the over seasoning plants by wind. A study of the annual spread of the pathogen in North America (Roelfs, 1985, 1989) showed that P. graminis moves from overwintering infection foci usually south of the 30th parallel. As the season warms the rusts move northward with the winds into winter-sown wheat in the southern plains and into the central plains. Spring crops are planted in the northern plains and the rust moves into them in the summer. Most spores are deposited on non-host plants and on the ground where the spores are idle. Local movement within the canopy is by turbulence and deposition is by gravity. Most of the spores escaping the canopy travel less than 100 m. These spores are impacted on a new host plant or are lost. The pathogen survives by producing large numbers of urediniospores, 5000 spores per day per uredinium, with up to 1000 uredinia per stem. Long distance dispersal occurs when hot air currents lift the spores to 3000 m. At this altitude spores can be transported hundreds of kilometres. With a terminal velocity of about 1 cm/s in still air, few spores fall to the ground. However, the spores can be efficiently scrubbed from the air by rain of 25 mm or more. Spores have been known to move long distances (Watson and de Sousa, 1983) but spores moving this far rarely result in disease establishment.

About 410 graminaceous species in 79 genera are known hosts for the P. graminis complex (Cummins, 1971; Gäumann, 1959; Gerechter-Amitai, 1973). In addition, Gäumann (1959) lists over 70 species of Berberis, and there are several species of Mahonia and Mahoberberis (Berberis x Mahonia) (Roelfs, 1985b) that are alternate herbaceous hosts for this fungus. The trend in recent years is to place both Berberis and Mahonia in the genus Berberis.Host specialization in P. graminis is highly complex and is best understood in terms of co-evolution of the host(s) and fungus. Diversity is greatest near the centres of origin, where co-evolution has occurred in the fungus, grassy hosts and alternate sexual hosts. The evolutionary trend is to specialization, but there may be considerable overlap between formae speciales and host genera or species.The modern bread wheat and oat cultivars are hexaploids composed of three diverse genomes, and the contributing genomes have provided additional opportunities or restrictions for infection by certain formae speciales of the fungus. Durum wheat is tetraploid. Somatic and sexual hybrids between formae speciales that infect common hosts are also known to occur. Defining host ranges is, therefore, relatively open-ended and it is difficult to delineate a complete list of all of the formae speciales of the fungus and a precise listing of their hosts.It has been suggested that a genetically defined, international set of grassy hosts is devised to determine the range and identity of formae speciales, and to assist in the delineation of host ranges (Anikster, 1984). Some of the more important hosts affected are mentioned below.Forms of P. graminis subsp. graminicola mainly attack pooid grasses in several tribes. The most frequently affected hosts are Dactylis glomerata, Lolium sp., Poa sp. (Urban and Markova, 1984) and Elytrigia sp. (Baum and Savile, 1985). In a study using a limited rust fungal population, Cagas (1975) observed frequent infection of Agropyron pectiniforme and species of Festuca and Phleum; these grasses are frequently grown for forage. Cultivated cereals are rarely attacked.The various formae speciales of P. graminis subsp. graminis have a considerably wider host range than those of subsp. graminicola, having the ability to attack a wide range of grass species in addition to the cultivated cereals. The host ranges tend to be broader in regions such as the Transcaucasus, Middle-Eastern and western Mediterranean countries, where the host and pathogen have co-evolved, than in areas of introduction such as North America. There are distinct patterns of specialization on grassy hosts, but there are a number of exceptions. The main hosts for f.sp. tritici are species of Triticum, Aegilops, Elymus and Hordeum; for f.sp. avenae, species of Avena; and for f.sp. secalis, Secale and Hordeum sp. A host range study by Gerechter-Amitai (1973) illustrated both patterns of specialization and the potential host ranges for subsp. graminis.Most of the genera in the tribes Aveneae, Phalarideae, Agrostideae and Festuceae are hosts for f.sp. avenae but rarely for f.sp. tritici and f.sp. secalis. Exceptions include a number of species of Bromus (Festuceae) that are common hosts for f.sp. avenae and f.sp. tritici. Genera in the tribe Hordeae tend to favour f.sp. tritici, with the exception of four species of Lolium that are natural hosts for f.sp. avenae but not for f.sp. tritici. The f.sp. secalis has a more restricted host range.The main alternate herbaceous host for all forms of P. graminis is Berberis vulgaris.For detailed reviews on host ranges, see Anikster and Wahl (1979), Baum and Savile (1985), Cagas (1975), Cummins (1971), Gäumann (1959), Gerechter-Amitai (1973), Leppik (1961), Savile (1984), Urban (1969) and Urban and Markova (1984).

Uredinial StageThe uredinia may occur on leaves, stems, leaf sheaths, spikes, glumes, awns and occasionally on grains of their grassy hosts; stems and leaf sheaths are the main tissues affected. On stems, the uredinia are elongated and reddish-brown; loose epidermal tissue is conspicuous at the margins of the uredinia, giving a roughened feel to the stem surface. The uredinia coalesce to cover large areas of the host tissue in heavy infection. Since the urediniospores are dehiscent, they are released as powdery masses from the uredinia.Telial StageThe telial stage occurs in the same tissue as the uredinial stage, but becomes shiny-black. The teliospores are sessile, and the telial tissue is, therefore, firmer than the uredinial tissue; no spores are released.Pycnial StageThe pycnial stage occurs on the young leaves of the alternate host, mainly Berberis vulgaris. Pycnial infections initially appear as light, chlorotic areas on the adaxial leaf surface, then become light orange-brown lesions, consisting of individual small cone-shaped eruptions (the pycnia), often occurring in clusters.Aecial StageThe aecia develop on the abaxial surfaces of the leaves of the alternate host. When mature, they appear as bright-orange, closely-packed, raised clusters of individual aecia. The aecia are cylindrical in shape and flare out at their apices, appearing as a grouping of rings within the aecial cluster.

The diagnosis of stem rust infections is principally by means of their symptoms. If there is doubt as to the cause of infection, a scraping of the lesion(s) may be taken with a scalpel, transferred to a glass microscope slide and placed in a droplet of water. The slide can then be examined under a low magnification (100 x) for urediniospores. The urediniospores of P. graminis generally appear as oblong, spiny spores, about 20 x 30 µm, of a light-orange to tan.The host species normally indicates the identity of formae speciales of the pathogen, particularly if a cultivated crop; barley is an exception. However, collections are frequently made from wild grassy hosts or as aecia from the alternate host. These collections could be one or more of several different formae or even rust species. Inoculum from these sources must therefore be transferred to a range of differential hosts for identification.In cereal cropping, the rusts are likely to be either P. graminis f.sp. tritici, f.sp. avenae or f.sp. secalis. For collections from barley or wild Hordeum, inoculation of wheat and rye will determine the identity of the pathogen as either f.sp. tritici or f.sp. secalis. For identification of forma specialis of collections from alternate hosts, wheat, rye and oat must be inoculated with the pathogen. A series of hosts may need to be selected for inoculation for the identification of rusts on grasses. Routine identification of a range of formae speciales would be greatly aided by an internationally accepted, clearly defined set of host differentials but such a set does not yet exist. Other rusts of wheat are Puccinia triticina, P. tritici-duri and P. striiformis; of oat P. coronata; of barley P. hordei, P. striiformis and P. coronata; and of rye P. recondita. Wild grasses related to these cereals also are hosts for these and other rust species (Cummins, 1971).Efforts are currently underway to identify specific rust fungi down to the pathotype level by molecular methods.

Life Cycle

P. graminis is a macrocyclic, heteroecious rust, with five distinct spore stages. The characteristics and ontogeny of each of the spore stages have been described at the ultrastructural level by Harder (1984). Advances during the past century in studies of physiology and biology of Puccinia have been reviewed by Staples (2000).

Uredinial Stage

The uredinial, or red summer, stage is initiated by germination of a urediniospore on its grassy host, penetration, development of an intracellular mycelium with intracellular haustoria, and subsequent sporulation of uredinia to form new urediniospores. The recycling of the uredinial stage is the major means whereby the fungus initiates and perpetuates an epidemic. The urediniospores of P. graminis are dikaryotic (n+n), dehiscent, thick-walled and covered with spines. They are elliptical and about 20 x 30 µm.

Telial Stage

As infected plants mature, urediniospore formation ceases and teliospore formation commences, either in the same, or in new (telia), fruiting structures. At this stage, the infections become black, hence the name black rust. The ontogeny of teliospores is the same as urediniospores, but the teliospores remain attached. The teliospores are two-celled, thick-walled (with up to five wall layers) and are thickened at the apical end. Teliospores are important because they are constitutionally dormant, enabling the fungus to survive severe cold or drought. The mature teliospore represents the only true diploid state of the fungus. The formation and characteristics of teliospores are described by Mendgen (1984).

Basidiospore Stage

The germination of teliospores and subsequent meiosis in the basidium results in the formation of haploid basidiospores. Four basidiospores, two of each opposite mating types, are produced from each basidium. If basidiospores are deposited on the surface of the alternate host (mainly Berberis vulgaris) they germinate, penetrate directly through the host epidermis and form a haploid mycelium. The fungus is most capable of infecting Berberis only when the leaves are young and tender. The fruiting structure, formed as a result of basidiospore infection, is called a pycnium.

Spermatial Stage

The pycnia are normally formed on the adaxial leaf surface, often in clusters. The important features of the pycnia are the formation of flexuous (receptive) hyphae and haploid spermatia. The spermatia, produced successively from the terminal ends of sporophores, are exuded in a nectar. The nectar attracts insects, which in addition to splashing rain drops, serve to transport the spermatia to flexuous hyphae of the pycnia of opposite mating types, where fusion occurs.

Aecial Stage

Following union of the opposite mating types, dikaryotization occurs. The spermatial nuclei migrate to the protoaecium, where mitosis occurs, the nuclei reassort into dikaryons and the aecial structure forms. The aecia of P. graminis are elongated, cylindrical structures. The ornamented, dikaryotic aeciospores are produced successively in chains from the aeciosporophores. The aeciospores infect the grassy host, completing the fungal life cycle.

Epidemiology

It is only in the most extreme climatic regions, such as very hot and dry, or tropical humid, that P. graminis does not occur, although irrigation can result in the occurrence of stem rust in some of these regions (Saari and Prescott, 1985). Although stem rust on wheat has largely been controlled worldwide by the use of resistant cultivars (Roelfs et al., 1992), the disease is potentially a continuous problem in some areas. Most regions of the world have conditions that are normally more marginal for rust development.

Appropriate conditions such as source and timing of inoculum, susceptible hosts, temperature and humidity are required for rust development (Roelfs et al., 1992). When these conditions occur, disease losses may become severe. The distribution of the pathogen is affected by prevailing climatic conditions, the movement of global air masses, geographical features, the availability of alternative grassy hosts or the alternate sexual host, and cropping practices. Urediniospores may be transported by wind over long distances, thus the occurrence of stem rust is only limited by the deposition pattern of the spores, suitable weather conditions and the availability of susceptible hosts. Geographic distribution must, therefore, be considered in terms of disease epidemiology.

Saari and Prescott (1985) suggested 10 major epidemiological zones for the cereal rusts; these major zones normally have one or more subzones because of geography or distance. The major zones, their principal cereal-growing countries, and the movement of rusts within these zones are detailed below.

1. South Asia

India, Pakistan, Nepal, Bangladesh, south-east Afghanistan. Stem rust is a major disease, generally distributed in India (Joshi et al., 1978). However, it is most damaging in parts of central and southern India and in the eastern regions, comprising parts of eastern Uttar Pradesh, Bihar and Bengal. P. graminis does not normally survive the hot, dry summers of the central Indian plains. However, the valleys of the Himalayas and the hills of southern India provide suitable conditions for the survival of rusts throughout the year, and are the foci of infection for crops grown in central India during the winter; the major source is the south Indian hills. The fungus survives on several grass species, volunteer crop plants or on Berberis (Nagarajan and Joshi, 1985). Stem rust is not normally a serious problem in Bangladesh. In Pakistan, P. graminis oversummers in the Sulaiman ranges, Hindukush Mountains and the Baluchistan area, then spreads to the wheat-growing areas of the Indus plains. The most severe infections occur in the coastal regions.

2. Western Asia

Iran, Iraq, Turkey, Syria, Israel, Jordan, Lebanon. This area includes the primary centres of evolution of grasses and the secondary centres for cultivated cereals, along with the co-evolution of their rust parasites (Anikster and Wahl, 1979; Baum and Savile, 1985; Leppik, 1961; Urban and Markova, 1984). The alternate hosts, species of Berberis or Mahonia, also have their origins in Eurasia. P. graminis is thought to have originated as a parasite on Mahonia/Berberis (Green, 1971a; Leppik, 1961), and continued to evolve in grassy plant communities where Berberis was also present (Urban and Markova, 1984). P. graminis subsp. graminicola appears to have been the first to evolve on its grassy hosts, followed by subsp. graminis with the advent of cereal cultivation within the past 10,000 years. Thus there are numerous related wild species growing in this region, and rust fungal diversity is high. In Israel, Gerechter-Amitai (1973) found a high degree of diversity in specialization of indigenous formae speciales of P. graminis on native and introduced grasses. The most diverse of the formae speciales, f.sp. avenae, was isolated from 71 species in 36 genera, and was successfully inoculated to 144 species in 42 genera. Both subspecies of P. graminis occur throughout this region. The most serious infections of cultivated crops occur in the coastal areas, where humidity is higher than inland.

The situation in Egypt is different, as the rusts do not survive between cropping seasons and must be re-introduced each year (Saari and Prescott, 1985). Wind directions are highly variable in Egypt, and urediniospores can be introduced from almost any direction. The race composition of P. graminis f. sp. tritici populations in Egypt and the Near East countries indicates that the primary inoculum arrives from the Near East (Abdel-Hak et al., 1966). Stem rust is most severe in the northern coastal region, and decreases southward as rainfall diminishes.

3. Africa south of the Sahara and the south-western Arabian Peninsula

Saudi Arabia, Yemen, Ethiopia, Kenya, Tanzania, Zimbabwe, Lesotho, Angola, South Africa, Nigeria. Most cereals in this zone are grown at high altitudes, which favours the development of stem rust. Berberis is not a factor in disease epidemiology in this region. The moderate climates and the ability to grow cereals at any time of the year creates local endemic disease cycles. The fungus may persist on volunteer cereal plants, or on successive crops, as they are planted at different times of the year depending mainly on altitude and rainfall patterns.

The microclimate has a strong effect in the highlands of Kenya (DE Harder, Cereal Research Centre, AAFC, Winnipeg, Manitoba, personal communication, 1973). Stem rust may become severe, particularly in areas at the mid altitude range in Kenya (Pinto and Hurd, 1970; Harder, 1974). There is evidence of inoculum exchange within the Rift Valley, between Kenya and southern Tanzania (Harder et al., 1972), and it appears to act as a subepidemiological zone. The role of grasses as reservoirs of inoculum for the main crop, wheat, may not be important. Green (1969) was able to isolate only P. graminis f.sp. avenae or f.sp. secalis, but not f.sp. tritici, from local grasses. Isolated pockets of cereal cultivation occur south of the Rift Valley and there are endemic rust fungal populations (Saari and Prescott, 1985).

The Ethiopian highlands are somewhat unique as this area is considered a centre of diversity for durum wheats and barley. There is also diversity in the rust fungal populations; this area provides much of the inoculum for cereals grown in the Sudan.

There is no information on the origins of stem rust infections in western Africa but, as with eastern Africa, conditions are favourable for the establishment of local endemic rust fungal populations.

4. North Africa

Morocco, Algeria, Tunisia, Libya. This region also represents a secondary centre of diversity for cereals. Arthaud et al. (1966) reported 40 grass species occurring in the Atlas Mountains of Morocco as endemic hosts for P. graminis. Stem rust occurs annually on cereals in this region, the primary inoculum originating mainly as urediniospores from self sown plants in irrigated coastal areas (Saari and Prescott, 1985).

5. The Far East

China, Japan, Korea, Mongolia, the Soviet Far East. The epidemiological patterns in this region appear to be related to the extensive cultivation of wheat in China (Saari and Prescott, 1995). Wheat is cultivated at various times of the year throughout China, and stem rust occurs to varying degrees wherever wheat is grown. There are, however, several subzones within China where stem rust is a serious problem. These are the spring wheat areas of north-eastern China, Mongolia and north-western China, the autumn-sown areas of the Yangtze River basin and the south-eastern maritime provinces (Hu and Roelfs, 1985). The pathogen overwinters in the south-east, and airborne urediniospores then move northward to infect crops in north-eastern China and north-western regions. The fungus oversummers in the north and north-west, occurring on volunteer wheat and early planted winter wheat in the autumn. The cycle is then completed by the rust reappearing in the south and south-east of China (Hu and Roelfs, 1985). China is probably a source of inoculum for Korea, Japan and the Soviet Far East; Berberis may be found in the latter region (Saari and Prescott, 1995).

6. South-East Asia

Philippines, Sri Lanka, Indonesia, Malaysia, Thailand, Myanmar. P. graminis has only been observed in Thailand and Myanmar in this region. Myanmar may have a population of P. graminis f.sp. tritici that has developed in isolation (Saari and Prescott, 1985).

7. North America

Northern Mexico, USA, Canada. Arthur (1934) reported that P. graminis is present over all of North America, either on cultivated cereals or on grassy hosts. Collections of P. graminis f.sp. agrostidis, avenae, phlei-pratense, poae, secalis and tritici, from the USA and Canada, were able to infect 188 species from 52 genera of native and introduced wild grassy hosts (Fischer and Levine, 1941). Stem rust remains a potential problem on cultivated cereals, mainly in the northern plains region of the USA (Minnesota, North Dakota and South Dakota) and the eastern prairie region of Canada (Manitoba and eastern Saskatchewan). Historically, the occurrence of stem rust may have been more widespread, since Berberis eradication laws were passed in Massachusetts as early as 1754, followed soon thereafter by Rhode Island and Connecticut (Roelfs, 1985b). Berberis vulgaris has now been eliminated from most of the USA and Canada small grain cereal growing area (Roelfs, 1982).

P. graminis does not normally overwinter in the northern prairie region, although a severe epidemic on susceptible winter wheat in Manitoba in 1986 may have been due, in part, to the pathogen overwintering in sites in North Dakota (Roelfs and Long, 1987). Isolated heavy infections of Hordeum jubatum plants by P. graminis f.sp. secalis early in the growing season in Manitoba indicate overwintering of the fungus in this region (DE Harder, Cereal Research Centre, AAFC, Winnipeg, Manitoba, personal communication, 1995). However, the major source of inoculum for the cereal-growing region of the northern Great Plains is infected material, mostly from autumn-sown crops growing along the Gulf Coast of the USA. Rajaram and Campos (1974) include northern Mexico in this southern epidemiological subregion.

The two main stem rusts in North America, those on wheat and oat, follow very similar epidemiological patterns. There are no geographic barriers in the south to the movement of inoculum northwards. Rust fungal inoculum usually begins to increase along the Gulf Coast in March, then spread northwards, reaching the eastern prairies of Canada in late June or early July. Large areas of autumn-sown crops in the southern to mid portions of the Great Plains of the USA have traditionally provided much of the inoculum for crops further north (Eversmeyer and Kramer, 2000). Recently a race of wheat stem rust appeared on barley in the northern plains (Fox and Harder, 1995; Roelfs et al., 1997, Dill-Macky and Roelfs, 1998, 2000); that caused concern among barley producers and users.

Wild grasses appear to play a minor role as alternative hosts for f.sp. tritici (Roelfs et al., 1992). Although wild oat is highly susceptible to f.sp. avenae and is widespread in northern areas, it is less common in the southern USA and does not play a major role in initiating epidemics of oat stem rust. Reduced cultivation of oat in the USA over the past decade has resulted in less inoculum being generated to infect oat crops further north.

In addition to the Great Plains epidemiological zone, there are also the eastern USA/Canada and the Pacific Northwest zones in North America (Rajaram and Campos, 1974). In Canada, virulence analysis of populations of P. graminis f.sp. tritici and f.sp. avenae over a number of years has shown the maintenance of distinct populations in eastern Canada, the prairie region and the Pacific region (from western Alberta across the Rocky Mountains to the west coast of British Columbia) (Harder, 1994).

Stem rust may be widespread in Mexico, but few serious epidemics have been recorded because of crop timing and the use of resistant cultivars (Roelfs, 1985a). There is currently little inoculum exchange of wheat stem rust between the more southerly Mexican wheat growing regions and the southern USA. Rajaram and Campos (1974) considered an area including southern Mexico and Guatemala to be a distinct epidemiological zone. However, the distribution of races of oat stem rust was identical in material collected from northern and southern Mexico, and from Manitoba, during 1994 (Harder et al., 1996), indicating possible common inoculum sources.

8. South America

Stem rust occurs throughout the South American cereal-producing areas, but losses are not generally severe, except for localized epiphytotics. South America has two distinct epidemiological regions:

The Andean region: (Columbia, Ecuador, Peru and western Bolivia). The Andean region is characterized by changes in altitude over short distances, and wild species of Hordeum, often quite severely rusted, are common (Saari and Prescott, 1985). The diverse climates and wide-ranging occurrence of susceptible hosts create conditions suitable for the survival of an endemic rust population, except in the extreme southern end of this region.

The Pampas region: (Argentina, Uruguay, Paraguay, southern Brazil and lowland Bolivia). Double-cropping is common, planting and harvesting occurs over several months, volunteer crop plants are common, and the climate is mild over most of this region; this creates suitable conditions for endemic rust fungal populations. In Brazil, most grain production occurs in the southern states of Rio Grande do Sul, Santa Catarina, Parana, Sao Paulo and Mato Grosso do Sul. Stem rust occurs over all of this area.

A large area, known as the Cerrados, including parts of the states of Minas Gerais, Goias, Mato Grosso and Bahia, is potentially a great wheat producing area (da Silva, 1991; McMahon, 1984). Stem rust occurs in this region, but the potential for an epidemic under conditions of large scale cropping is not known.

9. Australia (including Tasmania) and New Zealand

The geographical position of Australasia provides a unique, epidemiologically-isolated region for the cereal rusts. However, the isolation is not complete: there is evidence that at least three times in the past 75 years, inoculum of wheat stem rust has been introduced into Australia by the aerial movement of inoculum from East Africa, a distance of about 8000 km (Watson and de Sousa, 1983).

The Australasian epidemiological zone has been divided into four regions by Luig (1985), based on the occurrence of stem rust on wheat:

- The summer rainfall area of eastern Australia (Queensland, northern New South Wales, north-western plains), which is traditionally the most rust-prone part of the wheat growing areas. Inoculum of P. graminis f.sp. tritici can overseason on volunteer plants and on wild grasses such as Agropyron scabrum and Hordeum leporinum.

- Southern New South Wales, Victoria, South Australia and Tasmania. Rust development is more limited in this area. Berberis occurs in Tasmania, but the predominant rust is f.sp. secalis, and there appears to be little effect on the main cereal growing areas.

- Western Australia, separated from the southern region by about 1300 km of desert. Inoculum may be exchanged between these two regions, mainly west to east.

- New Zealand. Most wheat is grown on the South Island, where P. graminis survives readily, but temperatures are below the optimum for stem rust development. Later maturing crops in the southern region of Australia also appear to be a source of inoculum for the South Island. The North Island provides more favourable conditions, and epidemics may become severe. Inoculum may arrive onto the North Island from Queensland.

In Australia, barley is affected by f.sp. tritici, f.sp. secalis and hybrids between the two. Both of these formae speciales have common hosts in the grasses A. scabrum and H. leporinum, and readily form hybrids. These hybrids are often referred to as f.sp. hordei because of their virulence to commercial barley. Oat is grown over a wider range of environments in Australia than other cereals. Oat may be severely affected by P. graminis f.sp. avenae. The pathogen may overseason on volunteer plants and several wild grasses, among them A. fatua, A. ludoviciana, A. strigosa, Vulpia bromoides, Amphibromus neesii, Lamarckia aurea, Dactylis glomerata, Phalaris sp. and H. leporinum (Luig, 1985).

10. Europe and Central Asia

This area is very diverse in its range of host plants, cultural practices, climate and geography. Along the southern edge of this zone is the Mediterranean Sea and North Africa, and to the east are the Black and Caspian Seas, and the western Asia zone. A subzone exists in eastern Europe from the Ukraine and the Caucasus across the Russian Steppes (Saari and Prescott, 1985). The major geographical feature throughout Europe is the west/east orientation of the Alps and the Pyrenees. This feature affects both cropping practices and the movement of rust fungal inoculum. Uredinial infections of P. graminis do not generally overwinter in northern and central Europe (Hassebrauk, 1967).

Zadoks and Bouwman (1985) discussed the existence of two major north-south tracts for movement of wheat stem rust inoculum. The West European tract follows the Atlantic coast from Morocco to Great Britain and the Netherlands. The East European tract may have its origin in Greece, tracking both north-west along the Danube and north along the Carpathian range, then fanning out west and east over northern Europe and the Ukraine.

Stem rust has been a more serious problem historically in Europe than it is today. Much of the stem rust problem was due to the widespread occurrence of Berberis sp. over much of Europe (Zadoks and Bouwman, 1985). The association of Berberis and the occurrence of rust was recognized in Europe in the early 1800s, and voluntary eradication was widely practised. Berberis eradication laws were generally proclaimed in western Europe in the early 1900s. These eradication programmes had a pronounced effect on reducing stem rust infection in some areas (Hermansen, 1968; Hinke, 1964). However, there still remain some regions, particularly in eastern Europe and western Asia, where Berberis is common, and where cereals or grasses may show considerable stem rust infection. For further information on distribution of the pathogen in Europe, see Santiago (1961).

Stem rust, caused by P. graminis subsp. graminicola, can be a serious problem on forage grasses, principally Phleum pratense, in eastern Europe and western Asia (Cagas, 1975), and control has been a problem. A wide range of grass species, many growing wild, are susceptible to this pathogen and numerous sources of infection are available. Many of these sources also grow in proximity to stands of Berberis, creating possible endemic cycles. 

Stem Rust Ug99

The story of Ugg99 begins in the 1930s in Germany where Georg Riebesel transferred the 1RS chromosome arm of rye (Secaliscereale cv. Petkus) replacing the 1BS chromosome of wheat. One of the derivatives, Riebsel 47/51, was used as a yellow (stripe rust) differential host. Riebsel 47/51 was released in Germany as the cultivar Weique. Crosses between local cultivars and Riebsel were released in 1972 as Aurora and Kavkaz in Russia, Lovrin 10 and Lovrin 13 in Romania, and as Clement in the Netherlands. These wheat cultivars had a wide agronomic adaptability that made them useful over large areas. Additionally they carried Sr31, Lr26 and Yr17 that gave resistance to all three wheat rusts. The Lovrin cultivars were widely used in crosses in China by the 1970s and Aurora and Kavkaz had spread to the breeding programmes in Europe and Mexico. By 1980, the 1B/1R chromosome derivatives were used worldwide in breeding programmes. The rust resistance was desirable and its agronomic adaptability made it possible for them to be used in a wide range of environments (Zadoks and Bouwman, 1985). The International Center for Maize and Wheat Breeding (CIMMYT) released Bobwhite and Veery populations that consisted of excellent lines with high yield. The Bobwhite and Veery sibs are major cultivars in Mexico, South America, Africa and southern Asia. The Lovrin derivatives are widespread in China. However, in high output bakeries, the 1B/1R cultivar’s ‘tough dough’ limited their use in the bread wheat cultivars of Canada and the USA.   Virulence by yellow rust to Yr7 was known in Europe by 1975 (Stubbs, 1985) and virulence to leaf (brown) rust (Lr26) was known in Europe by 1970, and in Romania and Russia by 1973 (Bartos, 1975). The stem rust resistance was effective until 1998, when William Wagoire of Uganda found stem rust on CIMMYT lines carrying the 1B/1R in a wheat plot at Kalengyere, Uganda. A rust culture was sent to from Uganda to South Africa where virulence to Sr31 was confirmed (Ug99) (Pretorius et al., 2000). Although the Sr31 resistance lasted over 20 years, longer than most resistances, it was the combination of virulent genes in Ug99 that made the culture so threatening to wheat production. Added virulences were soon found (Jin et al., 2008, 2009). This, along with the continuous presence of wheat throughout the year, will enable this area to continue to produce new races and vast numbers of urediniospores, some of which will be transported out of the epidemiological zone.   The exchange of rust inoculum within the south Africa epidemiological zone was reported by Saari and Prescott (1984). Wheat is an important crop for Ethiopia, Kenya and Yemen, all of which have two crops per year, some of which is irrigated. The west Asia zone is adjacent and to the east and north of the south Africa zone. Studies of wheat stem rust in Kenya go back to at least 1933. Due to the severity of stem rust epidemics in Kenya it was used as a testing area for resistance from 1950 to 1980 as the disease survived there in the uredinial stage throughout the year. Wheat is grown at altitudes below 1000 m in the Rift Valley to over 3000 m on Mt. Kenya. Ethiopia and Yemen wheat is grown under similar environments and in isolated areas uredinial infections can survive between seasons on the alternate season crop and on volunteer wheat. When virulence to Sr31 appeared in Kenya, Ethiopia and Yemen it was obvious that this was a serious threat to wheat production. Additionally Ug99 was virulent on most of the known resistance genes. The father of the green Revolution, Dr Norman Borlaug, voiced concern for the people of south Asia indicating that the historical destroyer of wheat again threatens Pakistan and India. Saved from starvation by the Green Revolution in the 1970s, this area now faces the threat of a shortage of wheat along with much of the rest of the world. Dr Borlaug’s efforts resulted in the Borlaug Global Rust Initiative in 2005. This project focused worldwide attention on the problem and challenged funding and research institutions worldwide to pool efforts and resources against P.graministritici. Further movement by local spread of stem rust would be north along the eastern Mediterranean and eastward through Turkey or Syria into Iraq and Iran, then following the path of yellow rust virulence to 8156 cultivars to Afghanistan, Pakistan and India (Saari and Prescott, 1984). Pakistan and India depend on Sr31 protection for wheat production which is essential for feeding their people. However Ug99 race TTKSK was found in west central Iran in 2009 (Nazari et al., 2009) and perhaps in southern Iran a year earlier where virulence on Sr31 has been since been reported. This long distance spore movement is typical for stem rust and raises the threat to southern and central Asia. For establishment, weather conditions must be favourable for urediniospore production, release and wind transportation to the distant wheat fields. Urediniospores in long distance transport are scrubbed from the air by rain and deposited on susceptible wheat, and with conditions favourable for germination, infection and spore production, the disease is established.

IncidenceInfection of Festuca arundinacea seeds by P. graminis subsp. graminicola consisted of mycelium within the embryo and urediniospores carried on both surfaces of the glumes. It is not known whether caryopses that are internally infected can germinate to produce infected individuals, but it is thought that this could be an important quarantine consideration (Kulik and Dery, 1995).Effect on Seed QualityNone.Seed TransmissionNot reported.Seed TreatmentsNone.Seed Health TestsNot reported.

Historically, stem rust has caused major devastation to wheat crops in most of the wheat-growing areas of the world (Roelfs et al., 1992). In ancient Rome, and probably a wide area around the city, damage to crops, mainly caused by stem rust, was so severe that a number of ancient authors referred to the problem. Rites and processions were organized to appease the numen (spirit), Robigo (Zadoks, 1985).

Globally, stem rust was the most important disease of wheat until the late 1950s, when the use of resistant cultivars became more widespread (Saari and Prescott, 1985). Epidemics of stem rust can be spectacular, reducing an apparently healthy crop to a tangle of blackened stems and shrivelled grain within a few weeks. Widespread epidemics occur relatively infrequently, but disease within a region or in individual fields is frequently severe, often completely destroying the crop. Widespread epidemics have been documented for Australia (Luig, 1985), Europe (Zadoks and Bouwman, 1985) and North America (Roelfs, 1985a).

Epidemics also occur regularly in Africa, China and Asia (Saari and Prescott, 1985). Accurate assessment of losses is difficult and, as a result, losses are often poorly documented. Losses in North Dakota, during the severe epidemics of 1935 and 1954, were estimated at US $356 million and US $260 million, respectively, based on wheat prices in late 1995 (Roelfs, 1978).

The annual value of stem rust resistance for eastern Saskatchewan and Manitoba, Canada, was estimated at $217 million (in 1977 Canadian dollars), based on the annual acreage yield loss (25%) expected if susceptible cultivars were grown in this area (Green and Campbell, 1979). The value relates to about $307 million at late 1995 US $ prices for wheat.

Wheat losses to stem rust in Chile, monitored over a 30-year period (1960-90), averaged about 0.25% (Hacke, 1992). Losses in southern Europe, mainly in Portugal, Spain, France, Italy, southern Germany, Romania and Bulgaria can average 10%, but losses as high as 60-80% have been reported (Santiago, 1961). Stem rust on wheat is, at present, largely under control worldwide. Even with the widespread use of resistant cultivars, P. graminis remains ubiquitous and heavy localized losses are possible.

The major loss due to stem rust currently is the costs incurred to find, incorporate and evaluate resistance in new cultivars. The need is constantly demonstrated by the appearance of new pathogen virulences. In recent years Enkoy rusted in Ethiopia, Sr24 in South Africa (LeRoux, 1985), Sr31 in Uganda (Pretorius, 2000) and the Rpg-1 barley cultivars in North America (Roelfs et al., 1997).

The economic impact of stem rust has been reduced mainly through the breeding of cultivars with resistance to wheat stem rust. However, resistance in cultivars must continue to be improved to keep up with pathogen evolution (McIntosh and Brown, 1997). In most years, 10-20% of the cost of cultivar development is related to stem rust resistance. A novel stem race in Africa now is of great concern. Records of yield losses caused by cereal rust diseases in the USA have been maintained by the USDA since the early twentieth century (Roelfs, 1978) (see the USDA Cereal Disease website for information from recent years). Losses in dollars are difficult to estimate as rust reduces the quantity of grain, which in turn increases the price. Stem rust also reduces the quality of grain, resulting in a lower price. The USDA data only estimates the reduction of quantity.

Stem rust disease can cause a severe reduction in food availability and in some areas wheat is the base of the diet. In the 1950s, epidemics in the USA caused losses not only to the farmer but to the community. Grain handlers, transportation, millers and bakers found themselves without a crop to move and process. The indirect effect on the community of unpaid loans and the loss of jobs can cause severe hardship in areas where wheat is the main crop. In subsistence agriculture, where bread is the main component of some diets, the effect on human health is significant.

Grain fields or forages can be inspected for the disease at any stage of growth. From the seedling stage until about the five- or six-leaf stage, stem rust infections are most obvious on the leaves. The main time for inspection is from the stage when the crop begins to head (late boot stage) until near maturity. Stem rust pustules are easily detected by sweeping aside the canopy of growth and looking into the crop with strong sunlight coming from behind.The red-brown pustules are easily recognized against the normally blue-green colour of healthy stem tissue. However, there may be other causes for similar discoloration. Infections due to stem rust can usually be differentiated from these other causes by gently feeling the plant tissue between the thumb and forefinger for the roughened surface caused by rust infection. This test may also be used for detached plant tissue, where a rough feel of red-brown or black lesions may indicate the presence of rust fungal uredinia or telia, respectively.

There is a major lack of knowledge in the ability to identify P. graminis when only a single stage exists. The morphology of the teliospore is the basis for identification (Urban and Markova, 1984; Abbasi et al., 2002) and teliospores seldom exist, except at the end of the season when the damage is done. Historically, pathogen identification was based on the host and disease symptoms. The urediniospores are the main agent of spread but they are difficult to distinguish between forma specialis. Recent studies have shown that single spore and developing mycelium in tissue can be identified to the source population (Szabo, 2005; Szabo and Barnes, 2005; Szabo and Nguyen, 2005; Stoxen and Szabo, 2008). These techniques make it possible to distinguish between the original race and mutants of that race (Jin et al., 2009). The current variation in existing populations on each continent or isolated area needs to be determined. The path of movement and any evolutionary change can then be monitored as it occurs over time, thus allowing breeding programmes (McIntosh and Brown, 1997) more lead time (6-12 years is required) before the anticipated arrival of a given pathogen race.