Effects of dwar ﬁ ng genes on water use ef ﬁ ciency of bread wheat

Climate change has increased the risk of drought, which signi ﬁ cantly limits plant productivity. Various ways of increasing water availability and sustaining growth of crop plants in drought-prone environments are available. Genetic advances in grain yields under rainfed conditions have been achieved with the introduction of dwar ﬁ ng genes. A thorough understanding of the effects of different dwar ﬁ ng genes on root growth, coleoptile length, grain yields and water using ef ﬁ ciency (WUE) will provide opportunities to select appropriate Rht genes for breeding high WUE and grain yield cultivars. This review focuses on the mechanism involved in Rht genes that reduce plant height and affect root and coleoptile length, their consequent effects on grain yields and WUE, and suggests that for rainfed and irrigation-limited environments, combining GAR and GAI dwar ﬁ ng genes in breeding may help boost WUE and yields, and more materials from different parental sources should be collected to assess opportunities for potential comprehensive application of speci ﬁ c Rht genes.


Introduction
The Green Revolution, which was outlined by Dr. Norman Borlaug at the third international wheat genetics symposium in 1968 and which occurred from the 1940s to the 1960s, led to the development of high-yielding, disease resistant wheat cultivars with dwarfing genes (reduced height, Rht genes) that are lodging resistant and highly responsive to inputs such as heavy fertilizer and pesticide application. The success of these cultivars likely constitutes the most important event in the history of modern agricultural research [1] . Spectacular increases in wheat yields achieved during the 'Green Revolution' were enabled with the introduction of dwarfing traits into wheat plants [2] . Such increases in crop yields have also proven instrumental in providing for growing global populations over the past 40 years [3,4] . However, climate change has increased the instability of crop grain production by increasing the unpredictability of the timing and amounts of rainfall and by causing extreme heat and cold spells that have affected grain yields worldwide and threatened food security. Sources of specific adaptation related to drought and heat and the breeding of genetic traits will help maintain grain yields in dry and warm years. The effects of drought are expected to increase with climate change and increasing water shortages [5] . The adverse effects of climate change-induced rainfall patterns on agroecosystems exceed the effects of remedial agricultural measures such as fertilization or the physiological and morphological responses of wheat plants [6] . Thus, securing wheat productivity under conditions of climate change is essential for human nutrition and for meeting increasing demands for wheat from a growing world population [7] . The use of limited water resources to produce more grain is an important objective that can be achieved by breeding high water use efficiency (WUE) crop cultivars. According to previous study with eight varieties released from the 1940s to the 2010s under two water conditions, with an increase in plant height, WUE values decrease (Fig. 1) [8] . It is thus possible to enhance wheat WUE through the introduction of dwarfing genes.
The most important effect of dwarfing genes in wheat production was that it greatly strengthened the lodging resistance and improved the harvest index [9] . In recent decades, a number of studies conducted by breeders and agronomists have focused on dwarfing genes. More than 20 dwarfing genes have been identified (Table 1). According to their responses to exogenous gibberellic acid (GA), dwarfing genes are divided into two classes: GA-insensitive (GAI) and GA-responsive (GAR) genes. Up to 2000, more than 70% of commercial wheat cultivars world-wide contained the GAI dwarfing genes (Rht-B1b and Rht-D1b) [2,18] . In China, these two genes were also widely used in North China winter wheat region and Yellow and Huai Valley facultative wheat region, which had relatively high rainfall or irrigation and good climate conditions; while in the North-western spring wheat region and South-western China autumn-sown spring wheat region, in which agriculture was rainfed and rainfall was low, a large fraction of wheat cultivars contained the GAR dwarfing gene Rht8 or were without dwarfing genes [19,20] . This showed that different dwarfing genes might have different effects on the drought tolerance and WUE of bread wheat. Nowadays, the problem we all face is how to use limited water resources to produce more grain to feed increasing population. So measuring the effect of Rht genes on wheat grain WUE in detail and selecting genes with higher WUE is urgently needed. Thus, in this review, first, we summarize the mechanism of plant height reduction induced by Rht genes, then the effects of these dwarfing genes on the root system, coleoptile length, the physiological performance and on wheat biomass and grain yield under drought stress conditions, to obtain a thorough understanding for proper utilization of those Fig. 1 The relationship between plant height and water use efficiency (WUE) under two water conditions [8]

Rht19
Vic M1 -R S EMS [16] Rht20 Burt M860 -R R g rays [14] Rht21 XN0004 2AS I P - [16,17] Note: I and R in the sensitivity to GA 3 column denote insensitivity or responsiveness. P, S, D and R in the inherited character column denote partial dominance, semidominance, dominance and recessiveness, respectively. dwarfing genes in wheat improvement for drought tolerance and water use efficiency.

Possible mechanisms involved in Rht gene-induced reduction in plant height
The mechanism of GAI gene-induced reduced plant height is associated with DELLA proteins [21] . Wild-type DELLA proteins repress plant growth and are degraded in the presence of GA, thereby removing this growth restriction.
Rht-B1b and Rht-D1b each contain a single nucleotide polymorphism that encodes a premature stop codon, producing a truncated DELLA protein with reduced GA sensitivity and reduced plant height [10] . Rht-B1c, the allele of DELLA protein-encoding gene Rht-B1a, differs from Rht-B1a by one 2 kb Veju retrotransposon insertion, three coding region single nucleotide polymorphisms (SNPs), one 197 bp insertion, and four SNPs in the 1 kb upstream sequence, resulting in DELLA motif primary structure disruption [11] . The causes of GAR Rht gene-induced reduced plant height are less well understood. Gasperini et al. [22] examined the endogenous GA content of 6-week-old wild-type and semidwarf Rht8 lines and found there was no significant difference between these test lines. Traditionally, Rht8 has been classified as a GA-sensitive gene because plants respond to the exogenous application of GAs [12] . Thus, Rht8 is not involved in gibberellin biosynthesis or signaling [22] . Finally, for exogenous brassinosteroid assays, it was found that the leaves of Rht8 lines are less sensitive to brassinosteroids [22] . In studies on Rht12, results show that Rht12 mutants may be deficient in GA biosynthesis rather than GA signal transduction [23] and that Rht12 lines also exhibit a normal brassinosteroid response [16] .
Relatively few studies have been conducted on the mechanism involved in Rht gene-induced reduction in wheat plant height. However, according to studies on rice and Arabidopsis thaliana, the dwarfing was associated with gibberellin biosynthesis or signaling/brassinosteroid signaling [24][25][26] . Moreover, deficiencies of brassinosteroid biosynthesis-induced plant height reduction have been found in tomato and Arabidopsis [27,28] . Therefore, we hypothesize that these two phytohormones have key roles in facilitating Rht gene-induced wheat plant height reduction.
A central role of the GA class of growth hormones in responses to abiotic stress is becoming increasingly evident. A reduction of GA levels and signaling has been shown to contribute to plant growth restriction with exposure to several stresses, including cold, salt and osmotic stress. Conversely, increased GA biosynthesis and signaling promotes growth in plant escape responses to shading and submergence. In several cases, GA signaling has also been linked to stress tolerance [29] . In the GA signaling process, the DELLA protein is a major repressor of GA-dependent processes, and GA promotes the proteasome-dependent degradation of DELLAs to promote growth [21] . Studies on the DELLA function have shown that DELLAs restrain growth and promote survival through adverse conditions by reducing levels of reactive oxygen species (ROS) and that DELLAs cause ROS levels to remain low following exposure to biotic or abiotic stress, thus delaying cell death and promoting tolerance [30] . Other studies have confirmed that under abiotic stress, GA 3 levels decline rapidly in maize leaves subjected to water stress [31] , and for Arabidopsis thaliana seedlings, exposure to salinity triggers a reduction in endogenous bioactive GAs that coincides with DELLA accumulation [32] . Again, all of these results show that GA biosynthesis and DELLA functions contribute to stress tolerance.
Brassinosteroid (BR) has been found to provoke response to many abiotic stresses like cold, drought, salt and so on [33][34][35] . In the process of root water uptake, the symplastic and transcellular pathways are collectively referred to as the 'cell-to-cell' pathway and studies have confirmed that this pathway plays important roles in root water uptake under drought stress [36,37] . It had been reported that this pathway could be largely controlled by the activity of aquaporins, which respond relatively rapidly and reversibly, causing changes in root hydraulic conductance [38,39] . In Arabidopsis thaliana, BR was shown to be involved in modification of transmembrane transport of water by controlling aquaporin activities [40] .
Taking all those into consideration, we can conclude that the Rht genes reduce plant height by controlling the signal or biosynthesis of GA or BR, which affects the stress response, so the introduction of Rht genes might affect the stress response and WUE of wheat.

Effects of dwarfing genes on drought resistance traits of wheat
The most important function of dwarfing genes is to reduce plant height to improve lodging resistance. Compared with lines without dwarfing genes (rht), Rht lines reduced height from 10% to 64% ( Table 2). The Rht-B1c and Rht-D1c had the strongest effect on stem reduction (64%), while GAR Rht8 gene was the weakest dwarfing gene (10% reduction).
The other effect of dwarfing genes in wheat is to influence the seedling vigour. This result is similar to the result for coleoptile length i.e. the GAR dwarfing genes have no negative effect on seedling early vigour, while GAI genes reduce it (Table 2) [46][47][48] .
Water from precipitation or irrigation can be lost in the form of crop respiration, soil evaporation and percolation into deeper soil layers. Plants can re-access water that has entered deep percolation only when they sustain long and vigorous root growth at early stages [49] . It is important to assess the effects of Rht genes on wheat root growth at the seedling stage as results vary depending on experimental media used (Table 3). Studies have shown that Rht genes have a direct effect on root growth during seedling establishment rather than a secondary partitioning effect [50] . However, this study also showed that dwarfing alleles did not alter either the total root length or the root architecture (average root diameter and lateral root/total root ratio). According to research conducted in a greenhouse, the effects of dwarfing genes (Rht1, Rht2 and Rht3) on root dry matter are relatively minor [55] . However, root dry matter was sampled at maturity in this study. In another study, under three different experiment conditions (gel, soil-filled columns and field grown), Rht8 was not found to have a negative effect on total root length compared to the control line (rht lines) while Rht-B1c and Rht-D1c were found to produce significantly shorter total root lengths under the latter two conditions [50,56] . Arabidopsis semidwarfs, independent of loss-of-function mutations at GA locus 5 (GA 5 ), which is an ortholog of green revolution dwarf alleles in rice and barley, showed decreased root length in comparison to related wild-type [57,58] . Another study on the Rht8 gene also found reduced root proliferation [59] . Thus it appears that the effects of Rht genes on root growth are influenced by genetic features and experimental conditions. If we wish to achieve an ideal root system using different Rht genes, it is necessary to breed lines using as many parental combinations as possible under a specific environment. In the future, more data on the different growth stages of Rht wheat roots must be generated.

Effects of Rht on wheat WUE under drought conditions or osmotic stress
Given the objective of using fewer water resources to produce higher grain yields, plant cultivars exhibiting higher WUE are urgently needed. WUE is defined as the ratio of grain yields (GY) to evapotranspiration (ET). ET includes water from plant transpiration (PT) and evapora-
One study focusing on the adaptation strategies of replacement cultivars subjected to short-term osmotic stress found that following the introduction of dwarfing genes, leaf transpiration rates decrease (P < 0.05) [62] . Field experiments on different wheat cultivars/lines containing different dwarfing genes showed that water consumption over whole growth periods showed no significant difference between these lines while the grain yield decreased with an increase in plant height (from 70 to 150 cm, R 2 = 0.25, P < 0.05, Fig. 2a and Fig. 2b) and the WUE decreased with an increase in plant height (R 2 = 0.57, P < 0.01, Fig. 2). Another study also found the grain yield was the highest at a plant height of 70 cm [64] . This shows that under drought conditions, the introduction of dwarfing genes may increase WUE by increasing grain yield.
Regarding the relationship between plant height and grain yields, one experiment has shown that higher yields can be achieved by using lines of between 70 and 100 cm in height [9] . That experiment also found no specific advantages associated with major dwarfing genes (Rht1 and Rht2). It was also found that tall lines perform equally well or better (P < 0.05) than all other classes (Rht-B1b and Rht-D1b) for grain yields, test weight, and kernel weight in all environments under a range of soil moisture condi-tions [65] . While another study showed that Rht-B1b and Rht-D1b lines generate higher yields than wildtype lines (rht) and Rht8 lines in most environments and are superior to Rht8 as a source of height reduction for spring wheat [66,67] . Rht-D1b was found to yield more than Rht-B1b lines while their yields were still higher than rht lines in spring wheat [68] . Other studies also confirmed that there was no significant difference in yield between these two genes [69][70][71] . Butler et al. [65] considered that Rht-B1b lines yield more than Rht-D1b lines only in the fully irrigated environment. Thus, the same dwarf gene showed different effects in different environments.
More recently, researchers have focused more on GAR genes. Rht5 lines decrease grain yields more than wild-type lines in north-western China and Australia [13,72] . Rht4 and Rht12 lines increase grain yields (P < 0.05) [13] . The Rht13 gene increases grain quantities and wheat yields in Australia [73] while in north-western China, the Rht13 gene decreases grain yields [45] . As the only GAR gene that has been used in commercial breeding, Rht8 has been studied more than all of the genes listed above [22,74,75] . However, results achieved from different areas and stock plants have varied.
Rht-B1b, Rht-D1b and Rht-B1c have been found to depress WUE by 15% as a result of specific genetic backgrounds, with the reduction in total water used being less than that of grain yields [55] . Another study found that Rht1 and Rht2 lines reduce WUE under certain environmental conditions [76] . However, higher levels of soil water depletion by standard height sunflower hybrids in agronomy trials are attributed to deeper rooting depths compared to those of dwarf hybrids [77] . Grain yields, however, were not given, and so the effects of dwarf genes on sunflower WUE remain unknown. Dwarf Rht-B1c (Rht3) seedling plants are characterized by a reduced decreased in photosynthesis, delayed non-stomatal limitations on photosynthesis, higher instantaneous WUE levels and better functional states of the photosynthetic apparatus under drought conditions compared to tall Rht-B1a Fig. 2 The relationship between plant height and grain yields (a) and water consumption (b) (data from reference [63] and unpublished data) plants [78] . However, Rht-B1b plants have an intermediate character between wild type and Rht-B1c lines that often does not significantly differ from that of the Rht-B1a plants examined in the study. Rht-B1c seedling plants maintain a lower injury index and less malondialdehyde content under osmotic stress [79,80] . These results suggest that Rht-B1c seedlings can better modulate antioxidant defense systems to sustain membrane integrity and higher levels of photosynthesis. However, in field experiments, lines with tall alleles (Rht-B1a and Rht-D1a) have been found to perform equally well or better than (P < 0.05) all other classes for grain yields, test weight, and kernel weight under drought stress [71] . In another study, Rht-B1b and Rht-D1b genes were not found to improve yields under marginal environmental and management conditions [67] . In previous studies on winter wheat cultivars released from the 1940s to the 2010s, it was found that under wellwatered conditions, grain yields of cultivars released after 1980s were higher than those released before, but under drought stress, grain yield improvements vanished [8] . Taking these results into consideration, plants with GAI Rht genes may be able to adapt to short-term osmotic stress rather than to mitigate field water stress.
GAR Rht genes are expected to improve grain yields, reducing plant heights [73,81] . The semidwarfing Rht8 was also found to exhibit greater drought tolerance in one of Alghabari's experiments [82] . However, specific studies on the effects of GAR genes on drought stress have been quite rare. Under natural conditions, the Rht10 gene negatively affects the IWUE via higher levels of leaf transpiration [52] . Under two contrasting water conditions, Rht13 and Rht8 wheat lines were found to show increased WUE at grain yield levels. In this study, the authors found that Rht13 and Rht8 lines generate higher grain yields under two water conditions (P < 0.05), but water consumption over the growth period showed no significant difference (P > 0.05) [63] . All of these results show that higher grain yields contributes to higher WUE and that the introduction of Rht13 and Rht8 genes has no negative effects on root water uptake.
Although the water consumption over the growth period was not significantly different between different lines, there may be a change in the proportion of PT and ES. The reduced plant height reduces the distance between the leaves and the soil, which may increase the coverage rate and reduce the ES. And the increase in early seedling vigor can also reduce the evaporative loss from the soil surface [61] .

Avenues for future research
The effects of Rht genes on wheat grain appear to vary for different trial materials and under different environments, and it seems that specific Rht genes exhibit low levels of eurytopicity. Therefore, in future studies on specific Rht genes in specific areas, more materials from different parental sources should be collected to assess opportunities for potential comprehensive application. GAR and GAI dwarfing genes may be combined during breeding to increase WUE and yields. Studies show that the simultaneous selection of Rht8 + Rht-B1b or Rht-D1b sesquidwarfs during early generation may facilitate the rapid development of high-yielding cultivars targeting both favorable and unfavorable environments [81] . Breeders may also collaborate with crop cultivation experts to develop better cultivation methods, as ES from wellmanaged wheat crops represents 40% of evapotranspiration, and this value is often substantially higher for poorly managed crops. In addition, selecting cooler canopies of wheat may prove useful for PT decline, and results confirm that newer wheat cultivars maintain cooler canopies under drought conditions [83] .
Although researchers have done a lot of fruitful jobs, the effect of the most dwarfing genes on yield and other physiological and agronomy traits still remain unclear. In the future, we should give more focus to the effect of dwarfing genes on drought resistance and WUE of wheat to promote their application. In addition, exploring more dwarfing genes with higher drought resistance and WUE is needed.
Compliance with ethics guidelines Jiakun Yan and Suiqi Zhang declare that they have no conflict of interest or financial conflicts to disclose.
This article is a review and does not contain any studies with human or animal subjects performed by any of the authors. Fig. 3 The relationship between plant height and water use efficiency (WUE) under two water conditions (data from reference [63] and unpublished data)