|
|
||||||||
a Dep. of Plant and Soil Sciences, Oklahoma State Univ., Stillwater, OK 74078
b Dep. of Entomology and Plant Pathology, Oklahoma State Univ., Stillwater, OK 74078
c The Land Institute, 2440 E. Water Well Rd., Salina, KS 67401
* Corresponding author (bfc{at}okstate.edu).
| ABSTRACT |
|---|
|
|
|---|
| INTRODUCTION |
|---|
|
|
|---|
Because the awns in wheat are photosynthetically active and have direct vascular linkage with the developing spike, their suppression potentially reduces the total photosynthetic contribution to developing kernels (Weyhrich et al., 1995). Reduced photosynthate supply could translate into lower kernel weight and test weight (Olugbemi et al., 1976) and possibly reduced grain yield (McKenzie, 1972). More recent evidence from the southern Great Plains was furnished by a comprehensive analysis of the contribution of awns in three genetic backgrounds and 78 sister-line pairs of HRW wheat (Weyhrich et al., 1994). For the majority of sister-line comparisons, awn suppression had no significant effect on grain yield, but it did reduce test weight (mean difference of 31 kg m-3) and kernel weight (mean difference of 2.2 g per 1000 kernels). Wheat protein and hardness measurements were not consistently influenced by awn suppression in the parent bulk populations of those sister-line pairs.
We would expect foliar disease pressure to influence comparisons of awned and awnletted genotypes for grain yield and kernel weight. Grain yield and kernel weight may decline with disease development on the flag leaf in amounts proportional to the decline in green leaf area duration (Gooding et al., 2000). Leaf rust remains a significant foliar disease of winter wheat in the southern and central Great Plains. From 1976 to 2000, the estimated average grain yield loss caused by leaf rust was 3.5%, exceeding other diseases by more than two-fold (Bockus et al., 2001). The benefit of leaf rust resistance to grain production and quality can be substantial, even among a wide array of conditions, which benefits are measured (Drijepondt et al., 1990; Cox et al., 1997; Singh and Huerta-Espino, 1997). We hypothesized that the presence of effective leaf rust resistance genes might extend flag leaf area duration and thereby compensate for the absence of awns in awnletted genotypes. Conversely, awns could provide some photosynthetic compensation in the event of premature flag leaf senescence under rust development.
Field experiments were designed, featuring natural infection by Puccinia triticina, to more precisely estimate the single and joint effects of awns and leaf rust resistance. We chose to investigate two independent leaf rust resistance genes, Lr41 and Lr42, given their current role in providing effective resistance in the southern Great Plains. Our analysis focused on agronomic properties previously shown to be influenced by these genes, as well as several quality characteristics under continual selection pressure in hard wheat breeding programs. A secondary objective was to assess the relative merits of Lr41and Lr42 for trait improvement.
| MATERIALS AND METHODS |
|---|
|
|
|---|
KS93U62 and KS93U50 were crossed with OK92G205 (PI 561731), an awnletted NIL of Century with the pedigree Century*5/McNair 1003 (Carver et al., 1993). The soft red winter wheat, McNair 1003, provided the awnletted character. Each cross produced a F2 population cosegregating at one of the leaf rust resistance loci (Lr41 and Lr42, respectively) and a locus controlling awn formation. Phenotypes were scored among >200 F2 plants per cross grown in the greenhouse at Manhattan, KS, with artificial inoculation of seedlings with P. triticina. Four genotypic classes were identified: resistant, homozygous awned; susceptible, homozygous awned; resistant, homozygous awnletted; and susceptible, homozygous awnletted. Detection of homozygosity at the awn locus was aided by intermediate expression of awn formation by heterozygotes. The homozygous condition for leaf rust reaction was confirmed in the following generation at Stillwater, OK (May, 1997). We identified nonsegregating F2:3 families on the basis of adult-plant responses in the field under moderately heavy, natural development of leaf rust. Random families representing the four genotypic classes were grouped according to their original cross, and these were identified as series 41A (KS93U62/OK92G205 progenies) and series 42A (KS93U50/OK92G205 progenies) (Table 1).
|
Experiments were conducted at Stillwater, OK, in 1998 and at Stillwater and Lahoma, OK, in 1999. Each of the three series was arranged independently in the field in a randomized complete block design with four replicates. The four parents, KS93U62, KS93U50, OK92G205, and OK92G206, were included as checks depending on the parental origin of each series. The same cultural practices were applied to each series, including fertilizer applications (according to soil-test recommendations for a 2600 kg ha-1 yield goal), planting date (early to mid-October), planting rate (approx. 60 kg ha-1), and plot size (either 1.83 m2 in 1998 or 3.45 m2 in 1999).
Development of leaf rust depended entirely on natural infection. Approximately 2 wk after heading, leaf rust reactions were recorded at Stillwater on 13 May 1998 according to the modified Cobb scale (Peterson et al., 1948) and on 19 May 1998 and 12 May 1999 according to a stay-green scale of 1 to 9. For the latter, increasing values represented increasing levels of susceptibility, in which 1 (not applicable in this study), 2, and 3 = 4, 3, and 2 leaves, respectively, below the flag leaf mostly green; 4 = penultimate and flag leaf mostly green; 5 = flag leaf mostly green; 6 = flag leaf partially chlorotic; 7 = flag leaf mostly chlorotic; 8 = flag leaf severely chlorotic; and 9 = flag leaf necrotic.
As a supplement to these adult-plant field reactions, seedling reactions were determined in the greenhouse according to Browder (1971). A mixture of P. triticina urediniospores was collected in May 1999 from 10 HRW cultivars growing at three central Oklahoma locations. The avirulence-virulence formula of this mixture on a single-gene differential series (Thatcher background) plus Siouxland (Lr24 + Lr26) and Century (Lr24) was 9 17 19 26 SXL/1 2a 2c 3 3ka 11 16 24 30 CTY. We inoculated first leaves of 10 to 15 seedlings of each NIL and parent by brushing them with Danne seedlings on which P. triticina was profusely sporulating. After inoculation, seedlings were kept in a mist chamber at 20 to 22°C for 16 h and then moved to greenhouse benches. Leaf rust reaction was rated 10 to 12 d later according to Stakman et al. (1962).
Agronomic attributes were measured on all plots, totaling four replicates in each of three environments. Grain yield was determined as the weight of threshed, cleaned grain. Test weight was measured according to standard procedures with a 0.95-L container. Kernel weight per 1000 kernels was calculated from the weight of a 200-kernel sample per plot.
All quality attributes were determined at the Oklahoma State University Wheat Quality Laboratory. Kernel hardness and kernel diameter were determined with the Perten Single Kernel Characterization System (SKCS, Perten Instruments, Reno, NV) from a sample of 300 sounds kernels per plot. We measured grain protein content by near-infrared reflectance (NIR) spectroscopy of a 9-g ground-wheat sample (method 39-70, AACC, 1995) and reported it on a 140 g kg-1 basis. Three replicates from each environment were sampled for kernel hardness, kernel diameter, and wheat protein.
After cleaning and tempering 125-g grain samples to 155 g kg-1 moisture, they were milled on a Brabender Quadrumat senior mill (C.W. Brabender instruments, South Hackensack, NJ) to estimate flour yield by AACC method 12-10A (AACC, 1995). Flour yield was reported on a 140 g kg-1 basis. Mixing characteristics of the milled flour were evaluated with a computer-assisted mixograph and 10-g bowl (National Manufacturing Co., Lincoln, NE) according to method 54-40 (AACC, 1995). Mixing time was the number of minutes needed for optimal dough development and was adjusted for flour samples with <120 g kg-1 protein on the basis of NIR spectroscopy. Mixing tolerance was determined in two ways: (i) by a subjective scale of 1 to 10 based on visual comparison of the mixogram to 10 standard tracings with comparable flour protein content and (ii) by the actual width of the mixogram curve at 2 min past peak development. Scores from the subjective scale were interpreted as 1 to 2, poor mixing tolerance; 3 to 6, moderate tolerance; and 7 to 10, strong tolerance. We determined SDS-sedimentation values after adjusting for flour protein content, according to the method described by Carver (1994). Flour yield, mixograph properties, and SDS-sedimentation were reported from two replicates in two environments (Stillwater, 1998 and 1999).
After confirming error homogeneity across environments, the data were combined across environments for partitioning of the genetic variance in each series. Genetic effects associated with awn type and leaf rust resistance class (series 41A and 42A) or with resistance gene background (series 41/42) were considered fixed, while other effects were considered random. For series 41A and 42A, the genetic variation among 32 NILs was partitioned into single-df contrasts of resistant versus susceptible classes, awned versus awnletted classes, and their interaction. For the series 41/42, variation among the 32 NILs was partitioned into single-df contrasts of resistant versus susceptible classes, Lr41 versus Lr42 backgrounds, and their interaction. The error term used in F-tests and in constructing LSD values was derived from pooling the sums of squares within classes.
| RESULTS AND DISCUSSION |
|---|
|
|
|---|
Mean stay-green values for the four genotypic classes indicated distinct differences between the resistant (flag leaves mostly green) and susceptible lines (flag leaves mostly chlorotic to necrotic). Visual differences between resistant and susceptible classes were more apparent among NILs segregating for Lr41 than those segregating for Lr42. This was reflected in higher severity levels among susceptible lines derived from KS93U62 (Lr41 source) than susceptible lines from KS93U50 (Lr42 source) (Table 1). We found no visible infection among resistant lines with Lr41 (0R rating), whereas the resistant lines with Lr42 showed minimal visible infection (0-1R ratings). Seedling tests confirmed the resistance of lines with Lr41, but Lr42 appeared to confer resistance only in adult plants. As expected, no visual differences were noted in leaf rust reaction between awned and awnletted NILs within the same leaf rust response class.
Agronomic Attributes
From the analysis of variance for grain yield, test weight, and kernel weight, the genetic variation among NILs from series 41A and 42A consistently traced to a significant main effect of leaf rust reaction type (R vs. S, Table 2). The main effect of awn type (A+ vs. A-), with one exception, was also significant for these attributes. Resistance to leaf rust and the presence of awns generally improved agronomic performance. The interaction of leaf rust reaction type x awn type was not significant in five of six tests. Hence, the effect of adding awns was independent of the genotype for leaf rust reaction, or for the corollary, the effect of adding resistance was indifferent to the presence or absence of awns. Grain yield of the NILs segregating for Lr42 did show a significant interaction, as yield increased with the addition of awns on susceptible lines but not on resistant lines.
|
Examination of the leaf rust reaction type x awn type means in Table 2 indicated that leaf rust resistance was more effective in compensating for the loss of awns (compare RA+ vs. RA- classes) than awns were in compensating for the loss of leaf rust resistance (compare RA+ vs. SA+ means). Awned and awnletted classes had similar grain yields as long as they were resistant. In contrast, awned classes of resistant and susceptible genotypes differed substantially, with the susceptible lines having significantly lower yields. Making the same class comparisons for test weight and kernel weight, differences were always significant, but the patterns were similar to grain yield. Test weight and kernel weight were both reduced with the loss of awns, even with leaf rust resistance; the reduction was doubled when comparing resistant and susceptible NILs with awns.
We infer from the comparison of genotypic class means that awnletted and awned genotypes are indistinguishable for yield provided they are resistant to leaf rust, but the awnletted genotype may show some loss in test weight and kernel weight. Individual genotypes within classes did not show exception to those trends, though the within-class variation was significant (P < 0.01) for all attributes in each series (not shown). Variation for grain yield was nonoverlapping between susceptible and resistant classes (awned or awnletted). The resistant, awned NILs produced the greatest kernel weights and test weights, whereas the susceptible, awnletted NILs produced the lowest. Differences between those classes amounted to about 40 kg m-3 (3.1 lb/bu) in test weight and 4 g in 1000-kernel weight, constituting substantial losses in physical grain quality.
As noted above, the average effect of a resistance gene appeared to be larger for the Lr41 source of resistance than for the Lr42 source. Direct comparison of Lr41 and Lr42 was possible in series 41/42 because NILs for the two genes were developed independently from those in 41A and 42A and tested concurrently. The interaction of gene source x leaf rust reaction was indeed significant for grain yield and for kernel weight (Table 3). This interaction reflected a larger difference between resistant and susceptible NILs from the Lr41 source (for grain yield, 1030 kg ha-1) than the Lr42 source (550 kg ha-1). This yield compression from Lr42 corresponded with visual observations of leaf rust severity and stay-green ratings. Differences in green-leaf retention of R and S NILs from Lr42 were not as obvious as those from Lr41, which translated into significant (P < 0.01) gene source x leaf rust reaction interactions for stay-green readings (Table 1). Comparison of resistant Lr41 NILs with resistant Lr42 NILs (Table 3) revealed only a slight advantage (P = 0.05) of Lr41 for yield (150 kg ha-1) and 1000-kernel weight (1.1 g).
|
|
The genotypic classes performed similarly for wheat protein quantity and for protein quality as measured by the mixograph. One exception was the mixogram curve width at 2 min past peak dough development for resistant and susceptible lines (Table 4). The greater width of susceptible types implies greater mixing tolerance, which might be associated with greater flour protein content caused by stress-induced senescence of the flag leaf, much as drought stress elevates flour protein content in hard red spring wheat (Guttieri et al., 2000). However, the resistant isolines tended to have higher flour protein content, particularly in the 41A series where the R and S means were 116 and 112 k kg-1 (P < 0.01, data not shown).
Without experiments conducted in the absence of disease, we cannot rule out the alternative explanation that these resistance genes may affect mixing tolerance or gluten strength through linkage with other genes introgressed from T. tauschii. Those experiments were performed by Cox et al. (1997), who showed that bake-mixing time and water absorption were lower in the presence of Lr41. Mixing tolerance based on curve width was unaffected, and baking performance did not appear to be negatively affected in limited observations. The level of mixing tolerance reduction we observed did not significantly change the overall appearance or visual rating of the mixogram between classes (Table 4). In addition, mixing time varied among classes within the acceptable range of 4.5 to 6.4 min but was not influenced by either rust reaction or awn type (data not shown). Therefore, the effect of leaf rust reaction or awn type on dough-handling properties was either minor or nonexistent.
Only those quality attributes showing significant genotypic effects were reported for series 41/42 (Table 5). Resistance increased wheat protein content, more so for Lr41 than for Lr42 (interaction significant at P = 0.01). The parallel comparison in series 41A (RA+ vs. SA+, Table 4) did not produce the same result, so we cannot conclude in general terms that leaf rust resistance leads to a substantially favorable change in protein content. Consistency was found, however, for kernel diameter patterns. Kernel diameter increased by 0.2 mm with resistance, averaged for Lr41 and Lr42. Small differences were also found in kernel hardness but not of the expected magnitude to affect milling quality.
|
| CONCLUSIONS |
|---|
|
|
|---|
| NOTES |
|---|
|
|
|---|
Received for publication July 27, 2002.
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
B. F. Carver, R. M. Hunger, J. T. Edwards, D. R. Porter, T. F. Peeper, B. W. Seabourn, P. Rayas-Duarte, A. R. Klatt, and B. C. Martin Registration of 'Okfield' Wheat Journal of Plant Registrations, September 1, 2007; 1(2): 102 - 103. [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Vadose Zone Journal | |||
| Journal of Plant Registrations | Soil Science Society of America Journal | ||||
| Journal of Natural Resources and Life Sciences Education |
Journal of Environmental Quality |
||||