Application of Physiology in Wheat Breeding (2001)
M,P, Reynolds, J,I, Ortiz-Monasterio, and A, McNab, Editors
Chapter 14 pp 160-170
Manipulating Wheat Development to Improve Adaptation
G.A. Slafer and E.M. Whitechurch1
1 Departamento de Producción Vegetal, Facultad de Agronomía,
Universidad de Buenos Aires, Av. San Martín 4453, 1417, Buenos Aires, Argentina.

Crop development, though a continuous process, may be divided into three major components: the vegetative, reproductive, and grainfilling stages. The duration of each stage and of the whole life cycle, as well as the number of primordia initiated, are determined by interactions between genetic and environmental factors. These responses largely determine a crop’s adaptability to a range of environmental conditions.

This chapter describes wheat’s main developmental responses to environmental factors and highlights opportunities for improving wheat’s adaptation to particular conditions. It also discusses how to use developmental responses to further increase yield potential. Breeders can capitalize on this knowledge by using different factors to make crops fit the targeted growing season.

Also included in this chapter is a summarized, simplified view of wheat phasic development, but a more comprehensive description can be found in two recent reviews of the interactions among environmental factors (Slafer and Rawson, 1994a) and inter-relationships between phasic and morphological development (Slafer and Miralles, 1998).

Wheat Adaptation

Wheat is cultivated throughout the world (from South America and southern Oceania to North America and the northern parts of Europe and Asia, from sea level to about 3000 m), and its wide adaptability is based on complex developmental responses to environmental factors. As wheat has adapted to different regions, its development patterns have been modified to suit particular environmental conditions, the key issue being that anthesis must occur when the risk of frost is small. Thus an important feature of wheat adaptability lies in its ability to sense the seasons so that development is accelerated or delayed depending on the environment. Different types of wheat—spring, winter, and Mediterranean—are adapted to the cold, harsh-winter temperate, mildwinter temperate, and tropical regions where they are grown (Figure 1).

Spring types

In cold regions, many wheat plants could not survive the winter to produce a reasonable yield; wheat is thus normally sown in spring. Spring type wheats sense how advanced spring is and accelerate their developmental rate accordingly. The length of day (or night) is the environmental factor they sense best, as it invariably increases from the beginning of winter to the beginning of summer. Photoperiod sensitivity may therefore help to delay anthesis in early sowings and accelerate development in late sowings.

Figure 1.

Winter types

Winters are quite severe in harsh-winter temperate regions, but not cold enough to keep crops from surviving. Autumn sowing means the wheat crop has a long growing season and relatively early anthesis, having produced much biomass by then. In these regions plants must sense the season independently of daylength. Winter wheat plants must be exposed to low temperatures before their reproductive phase can begin, and autumn-sown plants do not initiate their reproductive period until winter has ended. Winter wheats sense a period of low temperatures and accelerate development thanks to vernalization sensitivity. The adaptive role of this sensitivity is highlighted in these regions, since it prevents inflorescence initiation in autumn, when photoperiod and temperatures are similar to spring.

Mediterranean types

In mild-winter temperate regions (such as Australia and Argentina, and Mediterranean areas), where wheat may be sown in winter, strong photoperiod sensitivity or slight vernalization sensitivity guarantees that the crop will flower shortly after the onset of a period with low or no frost risk. Genotypes with slight vernalization requirements are frequently referred to “intermediate, semi-winter, or Mediterranean types.” High temperatures and few rains are major constraints in tropical regions, so the wheat growing season must fit within the humid season, with plants flowering towards the end of it. In the tropics wheats do not need vernalization and are normally photoperiod insensitive, since the humid season does not always coincide with appropriate photoperiods.

How Wheat Adapts to Different Environments

Environmental signals

Wheat adapts its growing cycle to the best environmental conditions by sensing the right season (through vernalization and photoperiod sensitivities) and regulating flowering time based on temperature per se, to drive growth associated with development (Slafer and Rawson, 1994a). Thus the main environmental signals are temperature and daylength.

Temperature affects wheat development in two markedly distinct ways. First, the development rate is accelerated (and the time a developmental phase lasts is shortened) due to increased temperatures, in a wide range of thermal conditions. This general biological effect of temperature is probably caused by the activation of enzymatic processes. Second, wheat development may be accelerated by exposure to a period of relatively low (vernalizing) temperatures (vernalization response is thought to occur in the shoot apex). In sub-optimal temperatures, the relationship between development rate and temperature is linear, and progress towards flowering may be quantified in thermal time units. Vernalizing temperatures are defined by their effects rather than as particular thermal values. In the literature there is variation in the temperature ranges at which vernalization is most effective. Although this variation may reflect methodological differences, it most likely reflects genetic variation in thermal thresholds at which vernalization takes place. The vernalizing stimulus may be perceived by seeds imbibed in the soil (immediately after sowing and before seedling emergence), by young green plants (during the vegetative stage), and even by grains in the spike of the mother plant, if exposed during grainfilling to low temperatures. Most characterization has been done in seedlings. A generalized pattern of the most effective vernalization temperatures is shown in Figure 2.

Depending on the cultivar (and vernalizing conditions other than temperature), the maximum effectiveness has a lower threshold of between 1 and 4°C and an upper threshold of between 6 and 10°C (Figure 2). Temperatures higher than the latter—and as high as 18°C—are still vernalizing, but with reduced effectiveness.

Daylength is the most reliable environmental signal because it invariably changes with the season. The actual daylength for any particular site and date can be calculated easily if the latitude is known. In calculating actual daylength for plant responses, the length of the day includes periods of twilight. This is why the annual average for a particular site is always greater than 12 h (and daylength on 21 March and 21 September is also greater than 12 h). This factor is markedly affected by latitude: the further north or south from the equator, the greater the daylength variation during the year.

Response to environmental factors

General response of time to heading.

Heading is the first unequivocal external sign that a plant has reached the reproductive stage. Since heading occurs quite close to anthesis, the effects of environmental factors on time to heading are key determinants of wheat adaptability. For these reasons, and because it is easy to assess, time to heading is the most common variable for determining the effect of genetic and environmental factors on wheat development.

Sensitivity to temperature.

Time from sowing to heading is universally (i.e., in all cultivars, all vegetative and reproductive phases during time to heading are sensitive) affected by temperature (Angus et al., 1981; Del Pozzo et al., 1987; Porter et al., 1987; Slafer and Savin, 1991; Slafer and Rawson, 1995a). It is widely recognized that time to heading shortens, in curvilinear fashion, as temperature increases (Figure 3a). However, the reduction in the time elapsed to reach heading is the result of an accelerated rate of development in response to increased temperatures. The relationship between this rate and temperature is almost invariably linear (Figure 3b; see also Slafer and Rawson, 1995a). Figure 3 is just a schematic example, but a substantial amount of published data confirms its general shape (e.g., Gallagher, 1979; Angus et al., 1981; Monteith, 1981; Rickman et al., 1983; Morrison et al., 1989; Slafer and Savin, 1991; Slafer and Rawson, 1995a). Thus, rates of progress towards heading increase linearly with temperature from a theoretical threshold at which the rate is zero (it would take infinite time to reach anthesis at that temperature) to an optimum value at which the rate is maximized (and time to heading is minimized), and beyond which higher temperatures frequently reduce the rate of progress towards flowering (once again lengthening the period to heading). The thermal thresholds within which the rate of development increases linearly with temperature are the base and optimum temperatures (Figure 3).

Since the relationship is linear, there is only one slope for the entire interval between base and optimum temperatures. The reciprocal of the slope is thermal time (degree days) needed to reach heading at the designated base temperature; there is thus only one value of thermal time regardless of the temperature, provided the plants are exposed to thermal regimes between the base and optimum temperatures.

In practice, thermal time is simply calculated as the sum of daily effective temperatures (mean minus base temperature, the latter being the abscissa intercept of the relationship). By means of thermal time, development events can be expressed fairly independently of fluctuations in temperature.

Sensitivity to vernalization and photoperiod.

Wheat genotypes may vary widely in their sensitivity to photoperiod and vernalization. Some genotypes are virtually insensitive, others have quantitative responsiveness (which may vary considerably), and still others show qualitative responses (Figure 4a). Although all these responses are possible, most cultivars exhibit a quantitative type of response. The slope of the curve (Figure 4b) indicates sensitivity to either vernalization or photoperiod, i.e., the degree of increase in the rate of development (reduction in time) per unit increase in vernalizing or photoperiod stimulus. This parameter varies widely among genotypes and is probably the reason wheat adapts so well to so many different climates.

Although, for the sake of clarity, the three examples in Figure 4a have the same optimum values of length of vernalization exposure or photoperiod, there is genetic variation for them (see examples in Slafer and Rawson, 1994a). Similarly, the example does not provide evidence for genetic variation in intrinsic earliness, but cultivars do differ in this trait.

In passing, it should be noted that the term “optimum” is used in developmental studies in reference to the values of environmental factors that maximize the rate of development (optimum temperature, optimum vernalization and optimum photoperiod), which does not at all mean that these conditions optimize yield (Slafer, 1996). In fact plants growing under optimum thermal and photoperiodic conditions would likely yield very poorly as they would experience an excessively short season.

Which phases are sensitive to each factor?

Although development is a continuous succession of changes progressing towards maturity, to facilitate understanding of the processes involved, it is frequently defined as a sequence of discrete phenological events controlled by external factors, each event causing important changes in the morphology and/or function of some organs (Landsberg, 1977). Thus, key developmental events marking the limits of phenophases must be identified. The most accepted markers of developmental progress from sowing to maturity are seedling emergence, floral initiation, terminal spikelet initiation, and anthesis (Figure 5a, b, c). These developmental stages limit the following phases.

Pre-emergence development. The crop is established during this vegetative phase, and the shoot apex initiates leaf primordia after seed imbibition. When soil moisture does not limit germination, development rate depends only on temperature per se. The length of this phase thus depends on soil thermal conditions at sowing depth and on sowing depth itself (the deeper the sowing the longer seedlings take to emerge at a given temperature). There is no evidence that vernalization affects the rate of development until seedling emergence and since daylength is perceived by the leaves (and the signal transmitted to the apex; Evans, 1987), it does not affect the length of this initial phase.

Vegetative development. All the leaves (and potential tillers) on the main shoot are initiated. This phase continues until the apex becomes reproductive (marking the end of the phase). Leaves start to appear at a regular thermal interval (known as phyllochron) and tillering begins; the appearance of the first tiller coincides with the appearance of the fourth leaf, and the subsequent primary tillers appear at regular one-phyllochron intervals (e.g., Masle, 1985). The theoretical relationship between leaf and tiller appearance most frequently holds for this phase (since plants are small and not very demanding, available resources usually match demand).

The rate of development during this phase is sensitive to all three major environmental factors. Although not strictly true physiologically (see examples in Slafer and Rawson, 1994a), it may be assumed that vernalization requirements must be satisfied before a cultivar becomes responsive to photoperiod. While temperature similarly affects the rate of development towards floral initiation and the rate of leaf initiation, final leaf number is hardly affected by temperature per se (Slafer and Rawson, 1994b). On the other hand, vernalization and photoperiod affect the rate of development much more markedly than the rate of leaf initiation. Therefore, the longer the phase (due to lack of satisfaction of vernalization or photoperiod requirements), the higher the final leaf number (Halloran, 1977; Kirby, 1992; Rawson, 1993; Rawson and Richards, 1993; Evans and Blundell, 1994; Slafer and Rawson, 1995b,c).

Early reproductive development.

All the spikelets and many florets are initiated. Leaves continue to appear, and tillering usually reaches its maximum rate. However, this rate is hardly the one theoretically expected from the relationship with leaf appearance, because intra- and/or inter-plant competition for resources begins, reducing the assimilates available for growth of all tillers that could appear. Depending on the length of the phase and on agronomic conditions, the maximum number of tillers may be reached by the end of this phase.

Although it is frequently assumed that vernalization affects only the length of the vegetative period (Halse and Weir, 1970; Flood and Halloran, 1986; Roberts et al., 1988; Ritchie, 1991), several authors (e.g., Halloran and Pennel, 1982; Fischer, 1984; Stapper, 1984; Masle et al., 1989; Manupeerapan et al., 1992; Slafer and Rawson, 1994a) have recognized that vernalization may also affect the duration of early reproductive development, though usually less than the vegetative phase. Sensitivity to photoperiod and temperature per se are widely acknowledged in this phase (see examples in Slafer and Rawson, 1994a). In keeping with the discussion of final leaf number, when the rate of development of this phase is accelerated by photoperiod and vernalization, the shorter period results in fewer spikelets being initiated. This is not necessarily true when the length of the phase is affected by temperature per se, since it also substantially affects the rate of spikelet initiation (Slafer and Rawson, 1994b).

Late reproductive development. The number of fertile florets is determined simultaneously with active growth of stems and spikes. Leaves continue to emerge until the flag leaf (initiated last) appears. Stems first and spikes later grow actively, dramatically increasing the demand for assimilates, which in turn markedly increases competition for resources. Growth becomes most sensitive to changes in resource availability and yield most reduced if the crop is exposed to stress. The period from terminal spikelet initiation to anthesis is thus considered crucial in determining yield potential.

Due to increased competition, resource availability becomes insufficient to maintain all young tillers, and some of them die, normally in reverse order to when they appeared, and the number of tillers per m2 is reduced from its peak to final number of spike-bearing tillers per m2, normally set by the time of heading. At the end of the early reproductive phase, when the terminal spikelet is initiated, the final number of spikelets per spike is determined and several floret primordia initiated, particularly in spikelets in the middle third of the spike. From then on, floret initiation increases rapidly to its peak value, more or less coinciding with full flag-leaf expansion, when apparently no further florets are initiated (Kirby, 1988; Miralles, 1997). From booting to heading/anthesis, many florets degenerate during stem and spike growth, and only a few floret primordia become fertile and are fertilized by anthesis. Most wheats requiring vernalization, if grown in the right season, have by this time satisfied their requirements. Thus, even if vernalization sensitivity of this phase can be proven experimentally (Masle et al., 1989; Slafer and Rawson, 1994a), it may be assumed insensitive to vernalization under reasonable agronomic conditions (e.g., a wheat cultivar with strong winter habit should not be sown in spring). However, photoperiod may keep influencing the length of the late reproductive phase (Allison and Daynard, 1976; Rahman and Wilson, 1977; Masle et al., 1989; Connor et al., 1992; Manupeerapan et al., 1992; Slafer and Rawson, 1996); this effect may be direct (Slafer and Rawson, 1997), rather than simply mediated by a higher final leaf number due to exposure to short photoperiods during the vegetative phase. As discussed above, temperature per se affects the rate of development of all phases, and the fact that the time from terminal spikelet initiation to anthesis is reduced by increasing temperatures has been well documented (see examples in Slafer and Rawson, 1994a).

Post-anthesis development, or grainfilling. Grains develop their potential size and grow to their maximum dry weight by maturity. Most endosperm cells develop during early grainfilling; they are the actual sinks for the accumulation of assimilates during the next phase of active grainfilling, when grains grow and gain weight linearly with thermal time. At the end of the phase, grain growth declines until grains reach their maximum dry weight. During this phase the embryo is formed and the shoot apex initiates the first (normally, of four) leaf primordia.

The rate of post-anthesis development towards maturity in wheat is insensitive to photoperiod and vernalization and only appears to respond positively to temperature per se. Thus, the length of the grainfilling phase is quite conservative in terms of thermal time, unless severe water stress occurs. This will virtually end grainfilling, regardless of how many degree days have elapsed since anthesis. Since grain growth is most frequently limited by sink size (Slafer and Savin, 1994), accelerated development due to higher temperatures will reduce final grain weight (Sofield et al., 1977; Chowdhury and Wardlaw, 1978; Slafer and Miralles, 1992) much more than total protein content (since nitrogen is mostly source limited). Thus, higher temperatures during grainfilling will reduce yield-increasing protein percentage.

A more comprehensive explanation of these (and other) markers and the main features of each phase can be found in Slafer and Rawson (1994a) and Slafer and Miralles (1998).

Genes affecting physiological Responses

Although the rate of development may respond markedly to vernalizing temperatures, daylength, and temperature per se, photoperiod and vernalization sensitivities apparently account for most genetic variation for this trait. In other words, most differences among wheat cultivars in time to heading (or in the length of any phase from seedling emergence to anthesis) can be ascribed to differences in their sensitivity to photoperiod and/or vernalization. Temperature per se has universal impact on the rate of development in all wheat cultivars, which would seem to imply there are no genetic differences in sensitivity to temperature per se. However, “residual” differences are frequently found among genotypes, once their vernalization and photoperiod requirements have been satisfied. While these differences are usually less marked than differences in photoperiod and vernalization sensitivities, they are nonetheless both statistically and agronomically significant.

Residual differences have long been thought to reflect the impact of a third group of genetic factors determining differences in “basic development rate” or “intrinsic earliness,” also termed earliness per se (e.g., Major, 1980; Flood and Halloran, 1984; Masle et al., 1989; Worland et al., 1994). Although reasons have been put forward for considering intrinsic earliness genes as responsive to temperature per se (in which case differences in intrinsic earliness would be better defined as differences in temperature sensitivity) (e.g., Slafer, 1996), for simplicity’s sake we will use the term intrinsic earliness in this chapter.

Genetic control of the rate of development in wheat is complex enough that almost any development pattern is possible in the duration of the seedling emergence to anthesis period (Slafer and Rawson, 1994a), which means that almost any length of time to heading can be achieved through genetic improvement. The three groups of genes (photoperiod-sensitive, vernalization-sensitive, and intrinsic earliness genes; Worland, 1996) combine to determine the precise time of anthesis in a specific environment.

Bread wheat is a hexaploid species, i.e., it has three sets of genetic material (the A, B, and D genomes) and seven homologous groups. While photoperiod-sensitive and vernalization-sensitive genes appear to be located in certain homologous groups, intrinsic earliness genes are apparently distributed among different groups. Evidence for the location of these genes:

Photoperiod sensitivity genes (Ppd1/ ppd1, Ppd2/ppd2 and Ppd3/ppd3) are located on the short arms of homologous group 2; dominant alleles (Ppd) confer insensitivity, and recessive alleles (ppd) sensitivity. Chromosomes 2D, 2B, and 2A carry the Ppd1/ppd1, Ppd2/ppd2, and Ppd3/ ppd3 genes, respectively (Welsh et al., 1973; Scarth and Law, 1983; Sharp and Soltes-Rak, 1988). Ppd1/ ppd1 genes are believed to have the strongest effects, and Ppd2/ppd2 and Ppd3/ppd3 progressively milder effects.

Vernalization sensitivity genes (Vrn1/ vrn1, Vrn2/vrn2, and Vrn3/vrn3) are located on the long arms of homologous group 5. As with photoperiod-sensitive genes, sensitivity is conferred by recessive alleles and insensitivity by dominant alleles. The Vrn1/vrn1 gene, considered responsible for the strongest responses, is located in chromosome 5A; Vrn2/vrn2 (also termed Vrn4/vrn4; Snape, 1996) and Vrn3/vrn3 were found in chromosomes 5B and 5D, respectively (Law et al., 1975; Maistrenko, 1980; Hoogendoorn, 1985). A Vrn5/vrn5 gene has been reported to be located in chromosome 7B (Snape, 1996). Wheats with strong winter habit reportedly possess the three recessive alleles (vrn1, vrn2, and vrn3), but spring wheats may have different combinations of recessive and dominant alleles (Pugsley, 1972), which means that some may respond to vernalization (Slafer and Rawson 1994a).

Intrinsic earliness genes, unlike photoperiod and vernalization sensitivity genes, are not well documented (likely because their impact on time to flowering is less than that of Ppd or Vrn genes; also, they behave more like “minor” genes), having received much less attention. However, there is some evidence they are located on several chromosomes, including the long arms of homologous group 2 (Scarth and Law, 1983; Hoogendoorn, 1985). For example, in wheat these genetic factors have been reported on chromosomes 2B (Scarth and Law, 1983), 3A, 4B, 4D, 6B, and 7B (Hoogendoorn, 1985), 2A and 5B (Major and Whelan, 1985) 7B (Flood and Halloran, 1983), 6D (Law, 1987) and 3A (Miura and Worland, 1994). In barley they are distributed throughout the genome (Laurie et al., 1995).

There are no apparent associations among these genes, so that a particular genotype may possess any combination of alleles for photoperiod sensitivity, vernalization sensitivity, and intrinsic earliness genes.

Improving adaptation Selecting for improvied adaptation appears simple, since it may be accomplished by including time to heading in the traits considered when selecting progeny. If a breeding program is local, targeting the region is simpler than when a program attempts to release cultivars to be grown over large areas (Figure 6).

In the first case, mechanisms controlling the rate of development towards heading may be disregarded, since the priority is to obtain cultivars that reach heading within a certain time and will not be distributed except in areas with characteristics similar to those of the breeding program (not only environmentally but also agronomically influencing time to heading, such as sowing time). There would also be little interest in choosing parents based on specific sensitivities.

If a second generation per year is obtained by growing the plots in an inappropriate season or a completely different environment, we recommend applying the least amount of selection pressure possible for time to heading, since the environment would be too different from the target environment and variation in time to heading might not be related to time in an appropriate growing season. For example, a vernalization-sensitive line may show optimum time to heading in a normal growing season, but if the second generation is grown in warmer conditions, the line may appear unsuitably long and be wrongly discarded.

When the program’s target environment is extensive, the lines should be adapted to most environments they will be exposed to. Empirically this may be done by running the program simultaneously in many sites representing the range of conditions under which the released cultivars will be grown. In this case, it may be sensible, when choosing parents, to consider not only their time to heading in particular circumstances, but also their genetic sensitivities to major environmental factors governing rate of development (and the range of environmental factors in the target region). For some areas the requirements of plasticity given by vernalizationor photoperiod-sensitive genes could be predicted; in those cases choosing parents carrying the required genetic information may help to increase the likelihood of obtaining a reasonable number of well adapted lines that could be selected for yield or other targeted characteristics to provide the best possible cultivar. In fact, knowing what limits yield in modern cultivars grown in the region may also help identify the best possible combination of genetic sensitivities so that not only adaptation may be improved but also yield potential. Figure 4 shows how different phases are sensitive to different factors; therefore it may be possible to customize cultivars that reach heading or anthesis at a specific time and have a certain combination of durations for component phases. As sources and sinks are predominantly formed in different phases, it is speculated that manipulation of genetic sensitivity to photoperiod and vernalization, together with appropriate combinations of intrinsic earliness genes, may present additional opportunities for further increasing yield potential.

Identification of key phenological phases

To visualize phenological changes occuring during plant development, an accurate identification of the different stages is necessary. Although external morphological observations can give a general idea of phenology, microscopic observation of apex morphology is much more accurate in determining the stages of development.

In the field, development wheat plots can be monitored periodically, and plant samples taken at random.The few stages (described above) that limit phenological phases are mostly seen externally, with the exception of floral initiation and terminal spikelet initiation. The former cannot be unequivocally determined by simple observation (i.e., it is not marked by a clear morphological change in the apex) and is many times replaced by the observation of the first double ridge, which is the first sign that the plant is undoubtedly floral, and occurs a bit later than floral initiation (see Slafer and Rawson, 1994a). Determination of both double ridge and terminal spikelet initiation requires dissecting the apex and the use of an optic magnifier. (See in Figure 7 how the apex looks at different stages, from vegetative stage to terminal spikelet initiation.)

A plant is taken and leaves are extracted by means of a sharp cutting instrument; the apex can be observed under the most rec-

Intrinsic Earliness Bibliography