The effect of waterlogging in wheat (T. aestivum L.)

Abstract

Waterlogging is an important factor influencing yield and yield components in wheat. The objective of this study was to evaluate the effect of waterlogging on yield, yield components, protein and proline content, and chlorophyll a and b in wheat. In the study, seven levels of waterlogging treatment, 0, 5, 10, 15, 20, 25 and 60 days of flooding were applied. Increasing waterlogging stress decreased yield, spike number per m², seed weight and number per spike, protein content, and chlorophyll a and b; and caused increase in proline content. Results indicated significant linear responses for yield, spike number per m², seed weight and number per spike, protein content, chlorophyll a and b.

Keywords:

• Protein
• proline
• wheat
• waterlogging
• yield
• seed number and weight per spike
• spike number per m2
• protein content
• chlorophyll a
• chlorophyll b

Introduction

Wheat (Triticum aestivum L.) is one of the most important crops in Turkey, and increases in wheat production have come from irrigated land: irrigation is very important in Turkey, where it is applied to almost 1/8 of cultivated land (Anon., 2001). Wheat in many regions of Turkey is frequently subjected to waterlogging because of heavy irrigation, level topography, and/or inadequate soil drainage, and furthermore, unconscious irrigation causes important yield losses in such areas (Ceylan, 1994). In waterlogged soil, diffusion of gases through soil pores is so strongly inhibited by their water content that it fails to match the needs of growing roots. A slowing of oxygen influx is the principal cause of injury to roots, and the shoots they support (Vartapetian & Jackson, 1997). Oxygen deficiency caused by waterlogging reduces shoot and root growth of plants, as well as yield (Gardner & Flood, 1993). Waterlogging affects physiological processes of plants, such as absorption of water (Drew, 1991), root and shoot hormone relations (Huang et al., 1994), and decreases the uptake and transport of ions through roots, causing nutrient deficiencies (Greenway et al., 1994; Jackson & Ram, 2003). Tolerance to long-term waterlogging requires plants not only to ‘survive’ but also to grow during the waterlogging events. Wheat is very sensitive to waterlogging at sowing time, and during seedling, flowering, and grain-filling periods; waterlogging for 30 days during these periods reduced grain yield by 50–70% due to poor seed set and fewer spikes per unit area (Luxmoore et al., 1973; Misra et al., 1992). Heightened vulnerability at or just before flowering stage has been noted for wheat (Cannell et al., 1984).

The key strategy used for long-term waterlogging is the development of aerenchyma in roots to facilitate gas diffusion (Jackson et al., 1982; Fried & Smith, 1992). Other important traits in long-term adaptation include suberisation of nodal roots, which contributes to “effective” aerenchyma development. Several studies revealed significant reduction due to waterlogging occurred for yield, some yield components, proline and protein contents, as well as chlorophyll a and chlorophyll b in wheat (Luxmoore et al., 1973; Van Ginkel et al., 1991; Gardner and Flood, 1993; Musgrave and Ding, 1998). The aim of this trial is to assess the effect of waterlogging on yield, seed number per spike, seed weight per spike, spike number per m², proline and protein contents, and chlorophyll a and chlorophyll b in wheat.

Materials and methods

This study was carried out in the Ilica location of the Eastern Anatolia Research Institute, Erzurum, Turkey (29° 55′ N, 41° 16′ E, at an altitude of 1850 m) in the 2001–2002 and 2002–2003 cropping seasons. Precipitation was 375.4 mm in 2001–2002 and 348.5 mm in 2002–2003. Karasu-90, hard red winter wheat released for irrigated conditions, red and 45 g 1000-seed weight, tall (105 cm), resistant to lodging, cold, and stripe rust, was used. Seeds were sown in PVC containers (1 m width, 1 m length, and 0.75 m height) containing 80 kg of loamy textured soil (29.6% sand, 33.2% silt, and 37.2% clay). Soil also had 0.37% CaCO₃, 311.4 mmol/kg P₂O₅, 407.4 mmol/kg K₂O, and 2.06% organic matter, 7.44 pH, and 2.81 dS/m electrical conductivity. Our study was conducted under field conditions. Containers in the experiment were protected from bird damage by netting. Normal quality water (ECe = 1.0–2.5 dS m⁻¹) was selected in the study. Wheat was sown during the first two weeks of September at a seed rate of 475 seed/m². Sixty kg N ha⁻¹ (½ at sowing stage and ½ at tillering stage) and 60 kg ha⁻¹ P₂O₅ (at sowing) were applied. Ammonium sulfate (21% N) and triple superphosphate (46% P₂O₅) were used as fertilisers in the study.

Experimental design was a randomised complete block design (RCBD) with three replications. Normal irrigation as a control (C) at sowing, at stem elongation (Feekes 6.0), and at flowering (Feekes 10.51) was applied, and after this stage waterlogging was applied. Wheat was allowed to grow until flowering stage and, starting from the beginning of the flowering stage, waterlogging treatments consisted of six treatments: 5 days waterlogging (W ₅ ), 10 days waterlogging (W ₁₀ ), 15 days waterlogging (W ₁₅ ), 20 day waterlogging (W ₂₀ ), 25 days waterlogging (W ₂₅ ), and 50 days waterlogging (W ₅₀ ). Waterlogging was accomplished by using water from a nearby water service, flooding the containers assigned to the waterlogging treatment. Soil was kept saturated with water above field capacity by continuous flooding, usually every day to create an oxygen-deficiency environment. Grain yield and spike number per m², seed weight per spike, protein content, 1000-seed weight, harvest index, days to heading (Ceylan, 1994; Olgun et al., 2000), grain-filling period (Darwinkel et al., 1977), proline (Bates et al., 1973), and chlorophyll a and chlorophyll b (Arnon, 1949) were evaluated.

Seed protein content was determined by using near-infrared (NIR) analysis apparatus (Olgun et al., 2005). Proline was extracted from leaf samples of 100 mg fresh weight with 2 ml of 40% methanol. A 1 ml extract was mixed with 1 ml of a mixture of glacial acetic acid and orthophosphoric acid (6M) (3.2 v/v) and 25 ml of ninhydrin. After 1 h incubation at 100 °C, the tubes were cooled and 5 ml of toluene was added. The absorbance of the upper phase was spectrophotometrically determined at 528 nm. The proline concentration was determined using a standard curve (Bates et al., 1973). The amount of chlorophyll was determined according to Arnon (1949). Leaf material (0.5 g) was homogenised in acetone and centrifuged in a table centrifuge for 15 min. Then, the supernatant was treated with acetone to 15 ml. The absorbance value of the sample was read at 645–663 nm spectrophotometrically. Data were assessed in formulae (1) and (2) below, and amounts of chlorophyll were calculated as mg chlorophyll/g fresh leaf.

1

2

Results and discussion

Seed number per spike in the first year was significantly higher than in the second year (p<0.05) (Table I). Seed number per spike significantly decreased with increasing time of waterlogging (p<0.01) and reached its lowest level (6.7) at 50 days waterlogging (W ₅₀ ), and the highest seed number per spike (31.7) was obtained from control (C). If control was accepted as the reference, reductions in seed number per spike in W ₅ , W ₁₀ , W ₁₅ , W ₂₀ , W _25, and W ₅₀ were 2.0, 11.6, 20.8, 34.4, 52.9, and 78.8%, respectively. After the 20th day of waterlogging a rather important decrease occurs. These decreases in seed number per spike support the findings of Collaku and Harrison (2002) and they showed that increasing time of waterlogging caused decreases in seed number per spike (45%). Collaku and Harrison also pointed out that waterlogging tolerant genotypes showed a much higher seed number per spike (Collaku & Harrison, 2005). A significant linear response to waterlogging was observed (y= − 0.5433x+32.33, R ²=95.5%) for seed number per spike (Table II). Huang et al. (1994) reported that a response to waterlogging in seed number per spike was both significant and linear. The results of our study revealed that waterlogging stress caused a seed decrease of 0.54 day.

Table I. Yield, seed number per spike, seed weight per spike, spike number per m² of wheat genotypes subjected to different waterlogging applications.

Variables	Seed number per spike	Seed weight per spike (g)	Yield (kg/ha)	Spike number per m^2
2001–2002	23.4a	0.96a	3910.8a	419.5a
2002–2003	21.8b	0.78b	3521.1b	334.7b
Control (C)	31.7a	1.42a	6105.3a	587.3a
5 days Waterlogging (W 5 )	31.2a	1.30b	5866.0a	545.5b
10 days Waterlogging (W 10 )	28.0b	1.19c	4828.7b	495.1c
15 days Waterlogging (W 15 )	25.1c	0.8 d	3761.8c	524.8ab
20 days Waterlogging (W 20 )	20.8d	0.62e	3098.0d	456.7b
25 days Waterlogging (W 25 )	14.9e	0.54e	1589.5e	456.7b
50 days Waterlogging (W 50 )	6.7f	0.15f	651.4f	71.2
Mean	22.6	0.87	3700.1	527.8
LSD (%)	Year: 1.3 Waterlogging: 2.1	Year: 0.16 Waterlogging: 0.09 Waterlogging-year interaction: 0.12	Year: 253.4 Waterlogging: 266.0	Year: 34.1 Waterlogging: 29.4 Waterlogging-year interaction: 41.5
Year	23.61*	111.58**	43.79*	607.67**
Waterlogging	309.81**	403.71**	956.81**	621.19**
Waterlogging-year interaction	2.08ns	7.24**	2.43ns	16.29**
Coefficient of variation (%)	16.10	13.42	12.89	10.54
LSD = least significant difference. * Significant at 5%. ** Significant at 1%. ns: Not significant.

Table II. Orthogonal comparison table of waterlogging applications.

		F Values
Source of variation	Degree of freedom	Seed number per spike	Seed weight per spike (g)	Yield (t/ha)	Spike number per m^2
Linear effect	1	1722.6**	2353.8**	5624.4**	3586.4**
Quadratic effect	1	132.1**	22.6**	67.7**	108.2**
Cubic effect	1	1.4ns	1.1ns	16.4**	4.2ns
Quartic effect	1	1.8ns	18.4**	2.2ns	14.2**
Error	24
		F Values
Source of variation	Degree of freedom	Proline (µmol g^−1)	Protein content (%)	Chlorophyll a (mg chlorophyll/g fresh leaf)	Chlorophyll b (mg chlorophyll/g fresh leaf)
Linear effect	1	2777.7**	63.8**	1867.4**	1197.7**
Quadratic effect	1	133.1**	19.4**	151.1**	47.9**
Cubic effect	1	11.0**	0.0ns	2.8ns	5.9*
Quartic effect	1	49.6**	1.43ns	4.6*	0.5ns
Error	24
* Significant at 5%. ** Significant at 1%. ns: Not significant.

Seed weight per spike is one of the most essential yield components in wheat (Ceylan, 1994). Differences between years and times of waterlogging were found to be significant (p<0.01) in seed weight per spike (Table I). More seed weight per spike was obtained from the first year (0.96 g). Extending the time of waterlogging decreased seed weight. A significant waterlogging-year indicated that seed weight changed in time depending on the seven levels of waterlogging. The lowest seed weight belonged to W ₅₀ (0.15 g). Compared with control, reductions in seed weight per spike in W ₅ , W ₁₀ , W ₁₅ , W ₂₀ , W ₂₅ , and W ₅₀ were 8.4, 16.2, 39.4, 56.3, 61.9, and 89.4%, respectively, and after W ₅ this reduction was very severe. Results in this paper were similar to the findings of Saqib et al. (2004), who found that seed weight of waterlogged wheat was reduced by 50–85% relative to non-waterlogged applications. A significant linear response to time of waterlogging was observed for seed weight (y= − 0.0267x+1.3454, R ²=91.5%) and the decrease of seed weight per spike was determined as 0.026 g per day.

Yield was significantly influenced from years and times of waterlogging (p<0.05 and p<0.01). A higher yield in the first year (3910.8 kg/ha) was found, compared with the second year (3521.1 kg/ha) (Table I). A decrease in yield as a result of increasing times of waterlogging was observed. The highest yield was obtained from C (6105.3 kg/ha) and W ₅ (5866.0 kg/ha), while W ₅₀ gave the lowest yield (651.4 kg/ha). Accepting normal irrigation as a control (C), reductions in yield in W ₅ , W ₁₀ , W ₁₅ , W ₂₀ , W ₂₅ , and W ₅₀ were 3.9, 19.7, 38.3, 49.2, 73.9, and 88.6%, respectively (Table I). No important decrease occurred until W ₅ ; after that time of waterlogging the yield was severely depleted. Decreases in yield support the findings of Setter et al. (2001) and Barret-Lennard (2003) showing that reductions in yield with extending times of waterlogging occurred from 10 to 85%. A linear response to times of waterlogging occurred (y= − 117.76x+5818.8, R ²=88.3%) and the decrease in yield was 117.7 kg/ha per day.

Spike number per m² is a classical and important criterion in stress physiology of wheat (Musgrave & Ding, 1998). Significant differences were observed in both years and times of waterlogging (p<0.01). In adition, interaction of years and times of waterlogging was important at the 1% level (Table I). Spike number per m² in the first year was significantly higher (419.5) than in the second year (334.7). In our study, extending times of waterlogging caused a tremendous decrease in spike number per m² and this decrease was very apparent in W ₅₀ . With 587.3, the highest spike number per m² was obtained from C whereas W ₅₀ gave the lowest one (71.2). As compared with control (C), reductions in spike number per m² in W ₅ , W ₁₀ , W ₁₅ , W ₂₀ , W ₂₅ , and W ₅₀ were 7.1, 15.6, 35.3, 43.5, 60.9, and 87.8, respectively (Table I) and a considerable decrease was recorded after W ₁₀ . It has been reported that waterlogging reduces spike number per m² in winter wheat by about 20–50% and that the decrease in spike number per m² becomes important after W₁₀ (Musgrave & Ding, 1998; Collaku & Harrison, 2002). Table II shows that a significant linear decrease occurred in spike number per m². The best response to times of waterlogging in spike number per m² (Table II) was linear (y = −2.4166x+546.41, and R ²=60.4%), and the decrease in spike number per m² per day was calculated as 2.4.

Photosynthesis is an important process, which is altered by waterlogging. A decrease in CO₂ assimilation is induced by waterlogging (Mamdouh et al., 2001; Barret Lennard, 2003). Proline accumulation in waterlogged plants is due to stimulated synthesis due to loss of feedback inhibition, inhibited oxidation, probably due to effects on mitochondria, and impaired protein synthesis. It is likely determined by a combination of all changes mentioned, and by the rate of proline export via the phloem (Huang et al., 1994; Leul & Zhou, 1999). No significant differences occurred between years. Differences between waterlogging were found to be significant (p<0.01) in proline content (Table III). Increasing times of waterlogging caused significant increase in proline content until W _15. and after that time the increase was insignificant. Proline content reached the highest level at W ₂₀ and kept this level till W ₅₀ (8.55 µmol g⁻¹). Comparing with control, the increase in proline content in W ₅ , W ₁₀ , W ₁₅ , W ₂₀ , W ₂₅ , and W ₅₀ was 8.4, 16.2, 39.4, 56.3, 61.9, and 89.4%, respectively. Similar results were found by Mamdouh et al. (2001) and they found that proline content in waterlogged plants was at least four times higher than that in non-waterlogged plants. In Table III, orthogonal comparisons revealed that the best explanation of the effect of waterlogging in proline content was linear (y=0.1918x+1,3157, R ²=67.4%) and that the increase in proline content was 0.19 µmol g⁻¹ per day.

Table III. Protein and proline contents, chlorophylls a and b of wheat genotypes subjected to different waterlogging applications.

Variables	Proline (µmol g^−1)	Protein content (%)	Chlorophyll a (mg chlorophyll/g fresh leaf)	Chlorophyll b (mg chlorophyll/g fresh leaf)
2001–2002	2.99	12.4	0.799	0.248
2002–2003	3.44	12.1	0.793	2.243
Control (C)	0.15d	13.0a	1.163a	0.392a
5 days Waterlogging (W 5 )	0.24c	13.0a	1.112a	0.350b
10 days Waterlogging (W 10 )	2.57b	13.0a	0.998b	0.310c
15 days Waterlogging (W 15 )	4.58a	12.7a	0.897c	0.272d
20 days Waterlogging (W 20 )	8.55a	12.0b	0.798d	0.225e
25 days Waterlogging (W 25 )	8.55a	11.9b	0.417e	0.132f
50 days Waterlogging (W 50 )	8.55a	11.7b	0.187f	0.037g
Mean	3.22	12.3	0.796	0.245
LSD (%)	Waterlogging: 0.51	Waterlogging: 0.71	Waterlogging: 0.078	Waterlogging: 0.034 Year-waterlogging interaction: 0.049
Year	7.27ns	10.68ns	0.37ns	0.43ns
Waterlogging	495.40**	15.18**	341.59**	208.96**
Waterlogging-year interaction	1.83ns	0.75ns	1.30ns	10.27**
Coefficient of variation (%)	12.45	4.54	15.41	16.35
LSD = least significant difference. * Significant at 5%. ** Significant at 1%. ns: Not significant.

Protein content is susceptible to environmental conditions and varies with changing environment (Kattimani et al., 1996). The effect of times of waterlogging in the study was significantly important at the 1% level. Reducing protein content in seeds did not significantly vary until W ₂₀ (13.0–12.7%), but a sharp decrease occurred with W ₂₀ and thereafter this decrease was rather gentle (Table III). The highest protein content belonged to C, W ₅ , and W ₁₀ (13.0%) and W ₅₀ gave the lowest protein content (11.7%). Compared with control, reductions in protein content for W ₅ , W ₁₀ , W ₁₅ , W ₂₀ , W ₂₅ , and W ₅₀ were 0.0, 0.0, 2.3, 7.6, 8.4, and 10.0%, respectively (Table III). Plants respond to anaerobic conditions through changes in the expression of several genes (Ricard et al., 1994). Polyribosome dissociation occurs under anoxia together with a rapid decline in the synthesis of proteins (Kennedy et al., 1992). Oxygen deficiency restricts protein synthesis in roots (Sachs et al., 1980) and accelerates anoxic metabolism (Jackson et al., 1982). Therefore, prolonged waterlogging causes a significant decrease in seed protein content (Reggiani et al., 1988). There was a linear response to waterlogging (y= − 0.031x+13.025, R ²=77.6%) in protein content (Table III) and the decrease in protein content per day was 0.031%.

Waterlogging of tolerant genotypes with high chlorophyll content could be the most appropriate method for obtaining higher yield in high rainfall zones where waterlogging often occurs (Cannell et al., 1984; Gardner et al., 1993). With 1% significance, considerable variation in times of waterlogging occurred in chlorophyll a and b, besides which waterlogging-year interaction in chlorophyll b was significantly important at the 1% level. Prolonging times of waterlogging caused a significant decrease in chlorophylls a and b (Table III). Both chlorophyll a and chlorophyll b in C were found to be highest (1.163 and 0.392 mg chlorophyll/g fresh leave, respectively) and the lowest values were belonged to W ₅₀ (0.187 and 0.037 mg chlorophyll/g fresh leaf, respectively). If waterlogging applications are compared with control, decreases in W ₅ , W ₁₀ , W ₁₅ , W ₂₀ , W ₂₅ , and W ₅₀ were 4.3, 14.1, 22.8, 31.4, 64.1, and 83.9% in chlorophyll a; 10.7, 20.9, 30.6, 42.6, 66.3, and 90.6% in chlorophyll b, respectively; being exposed to waterlogging for longer than W ₁₀ created a tremendous decrease in chlorophylls a and b. Pang et al. (2004) observed that increasing waterlogging significantly decreased chlorophyll content and CO₂ assimilation rate. Smethurst and Shabala (2003) reported that chlorophyll a and b content gradually decreased over the time of the experiment in the waterlogged cultivars; chlorophyll content of leaves decreased by 19–45% after 16 days of waterlogging. A linear response to times of waterlogging occurred for both chlorophyll a (y= − 0.0212x+1.1739, R ²=91.9%) and chlorophyll b (y= − 0.0074x+0.3769, R ²=94.7%) and the decrease in chlorophylls a and b per day was 0.021 and 0.007 mg chlorophyll/g fresh leaf. Similar to our findings was the report that linear waterlogging reduces photosynthesis in wheat with a significant effect on yield (Drew, 1991).

Conclusion

Waterlogging damage occurs when heavy-handed irrigation is applied and ponding on the soil surface remains more than five days at the beginning of the flowering stage in wheat. This study will play an important role for wheat growers in the future as they will be able to make better crop management decisions when they encounter incidence of long waterlogging.