Published online 6 May 2005
Published in Crop Sci 45:1017-1022 (2005)
© 2005 Crop Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
CROP BREEDING, GENETICS & CYTOLOGY
Responses to Divergent Phenotypic Selection for Fiber Traits in Timothy
Annie Claessensa,
Réal Michaudb,*,
Gilles Bélangerb and
Diane E. Mathera
a Dep. of Plant Science, McGill Univ., Sainte-Anne-de-Bellevue, QC, Canada H9X 3V9
b Agriculture and Agri-Food Canada, Soils and Crops Research and Development Centre, 2560 Hochelaga Blvd., Sainte-Foy, QC, Canada G1V 2J3
* Corresponding author (michaudr{at}agr.gc.ca)
 |
ABSTRACT
|
|---|
Selection based on fiber traits may make it possible to improve forage digestibility while maintaining plant biomass. Our objectives were to evaluate the effects of divergent phenotypic selection for fiber traits on timothy (Phleum pratense L.) digestibility, plant biomass, fiber component concentrations and their ratios, and to identify selection criteria that have effective and stable effects on timothy digestibility without affecting plant biomass. Fourteen populations derived by intercrossing plants selected for high or low values of neutral detergent fiber (NDF), acid detergent fiber (ADF), acid detergent lignin (ADL), and cellulose (CEL) concentrations, and for ADL/HEM, ADL/CEL, and ADL/(HEM+CEL) ratios were evaluated in a field experiment. Direct responses for the selected traits were significant for the NDF, CEL, ADL/HEM, ADL/CEL, and ADL/(HEM+CEL) populations. Indirect responses for in vitro true digestibility (IVTD) and in vitro NDF digestibility (IVNDFD) were greatest for the ratios involving ADL. The ADL/CEL selection resulted in the most stable responses across years for IVTD and IVNDFD. Averaged over 2 yr, the IVTD and IVNDFD of the low ADL/CEL population were 27 g kg–1 DM and 33 g kg–1 NDF greater than those of the high ADL/CEL population. Furthermore, the low ADL/CEL population maintained its HEM and CEL concentrations and its plant biomass. Phenotypic selection based on ADL/CEL could be used to improve timothy DM digestibility without reducing plant biomass.
Abbreviations: ADF, acid detergent fiber • ADL, acid detergent lignin • CEL, cellulose • DM, dry matter • HEM, hemicellulose • IVTD, in vitro true digestibility • IVNDFD, in vitro neutral detergent fiber digestibility • NDF, neutral detergent fiber
|
INTRODUCTION
|
|---|
IMPROVING the nutritive value of timothy depends on successful selection for traits that influence digestibility and intake. Forage digestibility is influenced by fiber concentration and composition. Forage digestibility can be improved by selecting for reduced fiber concentration (Shenk and Elliot, 1970; Carpenter and Casler, 1990; Casler, 1999), but this approach can reduce forage yield (Surprenant et al., 1988; Casler, 1999) since fibers represent more than 50% of forage DM in timothy. An alternative would be to increase the digestibility of the fiber. In forage grasses, the fiber is composed primarily of cell-wall cellulose, hemicellulose, and lignin. Lignin is by far the least digestible fiber component (Casler, 2001), whereas cellulose is considered to be less (Buxton and Brasche, 1991) or as digestible as hemicellulose (Casler, 1987). Since these three fiber components have different digestibilities, a decrease in less digestible fiber components or an increase in more digestible fiber components should increase forage digestibility. We therefore hypothesized that selecting for low values of ADL/HEM, ADL/CEL, or ADL/(HEM+CEL) or for high values of HEM/CEL would improve timothy DM digestibility through fiber digestibility.
We characterized timothy genotypes divergently selected for NDF, ADF, ADL, HEM, and CEL concentrations and for ADL/HEM, ADL/CEL, ADL/HEM+CEL, and HEM/CEL ratios in a previous study (Claessens et al., 2004). We demonstrated that divergent phenotypic selection for NDF, ADF, ADL, CEL, ADL/HEM, ADL/CEL, and ADL/(HEM+CEL) consistently resulted in divergent groups of genotypes, but selection for HEM and HEM/CEL did not. We also observed that selection for criteria involving ADL concentrations [ADL, ADL/HEM, ADL/CEL, and ADL/(HEM+CEL)] resulted in the most stable differences across years and had the most important effects on timothy IVTD. Selection for the ADL, ADL/CEL, and ADL/(HEM+CEL) selection criteria had no negative effects on plant biomass. Furthermore ADL/(HEM+CEL) and ADL/CEL were the most promising selection criteria to increase digestibility while increasing plant biomass.
In our previous study, the characterization of timothy genotypes divergently selected resulted in the identification of seven selections with divergent groups of genotypes. In the present study, we intercrossed the genotypes from those seven selections and evaluated their progenies, to (i) evaluate the progress achieved by one cycle of divergent phenotypic selection for NDF, ADF, ADL, CEL, ADL/HEM, ADL/CEL, and ADL/(CEL+HEM) on timothy digestibility, (ii) evaluate correlated responses to selection, and (iii) identify the most effective and stable selection criteria that can improve timothy digestibility through fiber digestibility without adversely affecting plant biomass.
|
MATERIALS AND METHODS
|
|---|
Plant Material
The genetic material used in this study originated from one cycle of divergent phenotypic selection for NDF, ADF, ADL, and CEL concentrations and ADL/HEM, ADL/CEL, and ADL/(HEM+CEL) ratios among 78 high yielding timothy plants (Claessens et al., 2004). For each of the seven selection criteria, eight plants selected for high values and eight plants selected for low values were vegetatively propagated and isolated in crossing cages inside a greenhouse during the winter of 1997. For each crossing cage, equal weights of seeds from each selected plant were bulked. This resulted in 14 populations. Some plants were selected according to more than one criterion and therefore contributed to more than one population.
The 14 populations were arranged in a randomized complete block design with eight blocks. Each plot consisted of five plants. All plants were started in the greenhouse in April 1998 and transplanted to the field at Lévis, QC, Canada (46°9'15'' N, 71°12'00'' W; altitude 
45 m) in May 1998, with plants spaced at 90 cm between and within rows. The herbicide 4-chloro-2-methylphenoxyacetic acid was applied at 2 L ha–1 about 1 mo after transplantation for control of annual weeds. In late August, plants were clipped with a flail-type harvester to a 10-cm stubble height and fertilized with 40 kg N ha–1. In 1999, plants were fertilized with 55 kg N ha–1 in early May and harvested in mid June. In late August, plants were clipped as in 1998 and fertilized with 40 kg N ha–1, and 1.3 L ha–1 of a mixture of 295 g L–1 2,4-D (2, 4-dichlorophenoxyacetic acid), 110 g L–1 mecoprop [2 (4-chloro-2-methylphenoxypropanoic acid)] and 80 g L–1 dicamba (3, 6-dichloro-2-methoxybenzoic acid) was applied for weed control. In 2000, plants were fertilized with 55 kg N ha–1 in early May and harvested in late June.
Sampling and Analyses
Average monthly temperatures in 1999 were 4.2, 14.7, and 18.2°C for April, May, and June respectively, and 3.7, 10.9, and 15.0°C for the same months in 2000. In 1999, plots were harvested between 14 and 16 June when most plants had reached growth stage 56 (3/4 of the inflorescence emerged) according to the Simon and Park (1981) maturity scale. There was some variability among plants within plots, with plants ranging from growth stage 54 (1/2 of the inflorescence emerged) to growth stage 58 (base of the inflorescence just visible). In 2000, plots were harvested between 21 and 23 June when most plants had reached growth stage 54. Individual plants ranged from growth stage 52 (1/4 of inflorescence emerged) to growth stage 56. Each plant was cut to a 10-cm stubble height with a hand sickle, the average biomass harvested per plant (plant biomass) was measured on a DM basis for each plot, and a composite sample was made by collecting plant material from each plant in each plot. Samples were air-dried at 55°C for 2 d and ground to pass through a 1-mm screen in a Wiley mill (Thomas-Wiley Laboratory Mills, Philadelphia, PA).
All 112 samples from the first harvest in each year were scanned on a NIR spectrophotometer (Pacific Scientific Model 6150, Silver Spring, MD). Each year, 60 forage samples were used for calibration purposes by cluster analysis (Infrasoft International, 1993). The calibration samples were analyzed in duplicate for NDF, ADF, and ADL concentrations by the detergent fiber system of Goering and Van Soest (1970), except that neither sodium sulfite nor 
-amylase were used and the initial sample dry weight was 0.25 g instead of 0.50 g. Separate samples were used for NDF and ADF analyses. Calibration samples were also analyzed for IVTD and IVNDFD. The NDF determination of the postdigestion samples was done with an ANKOM Fiber Analyzer (Ankom Technology, Fairport, NY), using sodium sulfite and -amylase. The rumen fluid digestion used to measured digestibility was done with an ANKOM Daisy PII incubator (Ankom Technology, Fairport, NY). Rumen fluid was obtained from a lactating ruminally fistulated dairy cow. Determinations from the 60 calibration samples were used each year to develop NIR spectroscopy calibration equations for NDF, ADF, and ADL concentrations, and IVTD and IVNDFD values (Table 1). Predicted values of NDF, ADF, and ADL were used to estimate HEM concentration by subtracting ADF from NDF and to estimate CEL concentration by subtracting ADL from ADF. No correction for ash was made.
View this table:
[in this window]
[in a new window]
|
Table 1. Statistics for the NIRS calibration equations used to predict trait values in populations derived from one cycle of divergent phenotypic selection for high and low concentrations or ratios of fiber components in 1999 and 2000.
|
|
Statistical Analysis
Analysis of variance (GLM Procedure, SAS Institute, 1996, Cary, NC) was conducted on the data of the seven selection criteria from 1999 and 2000, with populations and years as fixed effects and replications as random effects. Data from each year were also analyzed separately. The data from the populations were then grouped by selection criterion and evaluated by performing contrasts between the divergent groups. Statistical significance was defined as P < 0.05. Correlation analysis was conducted to study the relationships among IVTD, IVNDFD, fiber components and their ratios, and plant biomass, using population mean values within years and the correlation matrix method (SAS Institute, 1996, Cary, NC).
|
RESULTS AND DISCUSSION
|
|---|
Year and population effects and population x year interactions were significant for all variables measured (Table 2). Because of the interaction with years and the fact that we wanted to identify a selection criterion that would be stable across years, the data are presented for each year.
Direct Response to Selection
Direct divergent selection responses (High vs. Low) were significant for NDF, CEL, ADL/HEM, ADL/CEL, and ADL/(HEM+CEL) in 1999 and 2000 (Table 3), with responses in the expected directions, that is, populations derived from parents selected for low values of the selected trait having lower values than populations derived from parents selected for high values. Averaged over 2 yr, the divergent NDF populations differed by 30 g kg–1 DM in NDF concentration. This difference is greater than what has been reported for reed canarygrass (Phalaris arundinacea L.) (Surprenant et al., 1988) and smooth bromegrass (Bromus inermis Leyss.) (Carpenter and Casler, 1990; Han et al., 2001), in which divergent NDF populations differed by 10 to 17 g kg–1 DM in NDF concentration. The divergent CEL population differed by 18 g kg–1 DM in CEL concentration when averaged over 2 yr. In contrast, Carpenter and Casler (1990) reported no significant response to divergent selection for CEL concentration in smooth bromegrass.
View this table:
[in this window]
[in a new window]
|
Table 3. Concentrations of fiber components or their ratios in timothy from a field experiment conducted on spring growth in 1999 and 2000 involving populations derived from one cycle of divergent phenotypic selection for high and low concentrations or ratios of fiber components.
|
|
The low and high-ADF populations were not consistently divergent across years (Table 3). The low-ADF population had a lower ADF concentration than the high-ADF population in 2000 but not in 1999. In general, we observed higher ADF concentrations and lower differences in ADF concentration between populations in 1999 than in 2000. These differences between years could be due to higher average monthly temperatures in spring of 1999 than in 2000. Temperature has been reported to accelerate the rate of plant development (Buxton and Casler, 1993), which in turn may reduce the differences in fiber concentration between varieties (Mika, 1983). The ADF selection might have been more sensitive to accelerated plant development than the other selection criteria since a large proportion of the fibers deposited at more advanced development stages are cellulose and lignin. Successful divergent selection for ADF has been reported in alfalfa (Medicago sativa L.) (Coors et al., 1986) and smooth bromegrass (Carpenter and Casler, 1990).
Although parental groups selected for high and low ADL concentration were divergent for ADL concentration in both years (Claessens et al., 2004), the populations derived from these parents did not differ significantly in ADL concentration in either year (Table 3). These results suggest that lignin has a low heritability in this specific parental population. There are no previous reports on ADL heritability in timothy, but in smooth bromegrass, lignin is considered to have a low heritability (Jung and Casler, 1990).
Our results on direct response to selection suggest that (i) one cycle of divergent phenotypic selection for NDF, CEL, ADL/HEM, ADL/CEL, or ADL/(HEM+CEL) can be effective in producing genetically different populations for each of these selection criteria, (ii) the response to selection for ADF is not stable across years, and (iii) one cycle of divergent phenotypic selection for ADL is not effective in producing divergent ADL populations in timothy. Consequently, the indirect response to selection for ADF and ADL is not presented.
Indirect Responses to Selection
Digestibility
Selection Based on Fiber Ratios.
In the analysis of variance combined over years, the selections involving ADL ratios [ADL/HEM, ADL/CEL, and ADL/(HEM+CEL)] were the only ones significantly affecting forage IVTD and IVNDFD. These selection criteria were also among the most effective for IVTD and IVNDFD in the study characterizing the parents (Claessens et al., 2004). Analyses of variance conducted within years indicated that ADL/CEL was the only selection criterion significantly increasing IVTD and IVNDFD values in both years (Table 4). The low-ADL/CEL population had higher IVTD and IVNDFD values than the high-ADL/CEL population in both years (Table 4). Differences in IVTD averaged over 2 yr for the ADL/CEL (27 g kg–1 DM) selection were greater than for the ADL/HEM (19 g kg–1 DM) or ADL/(HEM+CEL) (16 g kg–1 DM) selection. This differs from what we found in the study characterizing the parents, in which selection for the ADL/HEM and ADL/(HEM+CEL) had greater effects on IVTD than selection for ADL/CEL (Claessens et al., 2004).
View this table:
[in this window]
[in a new window]
|
Table 4. In vitro true digestibility (IVTD), in vitro NDF digestibility (IVNDFD), and plant biomass of timothy harvested in 1999 and 2000 from a field experiment involving populations derived from one cycle of divergent phenotypic selection for high or low values of individual fiber traits.
|
|
Compared with the high-ADL/CEL population, the low-ADL/CEL population had low ADF and ADL concentrations, but these two populations had similar HEM and CEL concentrations (Table 3). Hence, the higher IVTD values of the low ADL/CEL population are likely due to a reduction in lignin, the least digestible fiber component. This is similar to what we found for the divergent ADL/CEL parental groups (Claessens et al., 2004).
Selection Based on NDF or CEL.
In the analysis of variance combined over years, the divergent NDF populations were not significantly different for IVTD and IVNDFD. Generally, NDF concentration is negatively correlated with DM digestibility and positively correlated with ADF and ADL (Coors et al., 1986; Casler, 1987; Fonseca et al., 1999), and it has been used successfully as a selection criterion to increase DM digestibility in smooth bromegrass (Carpenter and Casler, 1990; Casler, 1999). In general, then, low-NDF populations are expected to be more digestible and to have lower ADF and ADL concentrations than high-NDF populations. Similar associations were observed in this study, but only in 2000 (Table 5). In that year, the low-NDF population had a higher IVTD value and lower NDF, ADF, ADL, and CEL concentrations than the high-NDF population (Tables 3 and 4). In 1999, however, the low-NDF population had a lower IVTD value, lower NDF, ADF, and CEL concentrations, and a higher ADL concentration than the high-NDF population (Tables 3 and 4). The lower fiber concentration of the low-NDF population was therefore counterbalanced by a higher ADL concentration resulting in lower IVTD and IVNDFD.
View this table:
[in this window]
[in a new window]
|
Table 5. Correlation coefficients among fiber components and their ratios, in vitro true digestibility (IVTD), and in vitro NDF digestibility (IVNDFD) in timothy samples harvested in 1999 (in roman font) and in 2000 (in italic font).
|
|
Low- and high-CEL populations had similar IVTD, IVNDFD, and ADL values. In both years, the low-CEL population had lower NDF, ADF, and CEL concentrations than the high-CEL population (Table 3), yet these two populations were similar for IVTD and IVNDFD (Table 4). This is similar to what has been found in smooth bromegrass, where selection for CEL failed to change digestibility in smooth bromegrass (Carpenter and Casler, 1990).
Relationships between Digestibility and Fiber Components and Their Ratios.
The ADL concentration was the fiber component most consistently and negatively correlated with IVTD and IVNDFD (Table 5). Many other studies have also reported negative correlations between lignin and forage DM digestibility (Mowat et al., 1969; Coors et al., 1986; Casler, 1987; Buxton and Russell, 1988; Fonseca et al., 1999). Therefore, the best way to affect forage DM digestibility would be to identify a selection criterion, which would have a large and consistent effect on lignin concentration. Generally, NDF, ADF, and CEL concentrations are positively correlated with the ADL concentration (Coors et al., 1986; Casler, 1987). In our study, however, we observed these correlations in only one of the two years (Table 5) and divergent selection for NDF, ADF, or CEL did not always translate into significant and stable differences in ADL concentrations between divergent populations.
The ADL/HEM, ADL/CEL, and ADL/(HEM+CEL) ratios were always highly positively correlated with the ADL concentration (Table 5). These correlations did translate into significant and stable differences in ADL concentrations but only between the divergent ADL/CEL and ADL/(HEM+CEL) populations (Table 3). After one cycle of divergent selection, the ADL/CEL populations differed by 4.8 g kg–1 DM in ADL and the ADL/(HEM+CEL) populations differed by 3.8 g kg–1 DM in ADL. These selection criteria were also the ones that had the largest and most consistent effect on IVNDFD, and the ADL/CEL selection criterion also had the most consistent effect on IVTD (Table 4).
Plant Biomass
In the analysis of variance combined over years and for each year (Table 4), the divergent NDF, ADF, ADL, ADL/HEM, ADL/CEL, and ADL/(HEM+CEL) populations had similar plant biomass. Furthermore, there was no significant correlation between plant biomass and any of the traits measured (data not shown). This lack of response of plant biomass to selection could be attributed to the independent culling selection for plant biomass and fiber components. However, the lack of response for plant biomass between divergent ADL/CEL and ADL/(HEM+CEL) populations is in contrast to what was observed for the parents of these populations (Claessens et al., 2004); low ADL/CEL and ADL/(HEM+CEL) parental groups had higher plant biomass than the high ADL/CEL and ADL/(HEM+CEL) parental groups. Failure to maintain this yield advantage through one cycle of intercrossing suggests that the relationships of ADL/CEL and ADL/(HEM+CEL) with plant biomass observed in the parental group were not due to pleiotropy or close genetic linkage. The divergent CEL populations differed only slightly in plant biomass, with the difference being significant in the analysis of variance combined over years (data not shown) but not in the analyses of variance for each year (Table 4).
|
CONCLUSIONS
|
|---|
These results suggest that (i) one cycle of divergent phenotypic selection for NDF, CEL, ADL/HEM, ADL/CEL, or ADL/(HEM+CEL) can be effective in producing genetically different populations for each of these selection criteria, (ii) among those, the ADL/CEL ratio is the most effective and stable selection criterion for improving timothy IVTD and IVNDFD, (iii) lignin is the most important fiber component affecting timothy IVTD and IVNDFD, and (iv) improvement in timothy DM digestibility can be achieved with no reduction in plant biomass. Phenotypic selection based on ADL/CEL could be used to improve timothy DM digestibility (IVTD) without reducing the concentrations of the more digestible cell wall components and plant biomass.
|
NOTES
|
|---|
Contribution no 771 Agriculture and Agri-Food Canada.
Received for publication March 11, 2004.
|
REFERENCES
|
|---|
- Buxton, D.R., and M.R. Brasche. 1991. Digestibility of structural carbohydrates in cool-season grass and legume forages. Crop Sci. 31:1338–1345.[Abstract/Free Full Text]
- Buxton, D.R., and M.D. Casler. 1993. Environmental and genetic effects on cell wall composition and digestibility. p. 685–714. In H.G. Jung et al. (ed.) Forage cell wall structure and digestibility. ASA, CSSA, and SSSA, Madison, WI.
- Buxton, D.R., and J.R. Russell. 1988. Lignin constituents and cell-wall digestibility of grass and legume stems. Crop Sci. 28:553–558.[Abstract/Free Full Text]
- Carpenter, J.A., and M.D. Casler. 1990. Divergent phenotypic selection response in smooth bromegrass for forage yield and nutritive value. Crop Sci. 30:17–22.[Abstract/Free Full Text]
- Casler, M.D. 1987. In vitro digestibility of dry matter and cell wall constituents of smooth bromegrass forage. Crop Sci. 27:931–934.[Abstract/Free Full Text]
- Casler, M.D. 1999. Correlated responses in forage yield and nutritional value from phenotypic recurrent selection for reduced fiber concentration in smooth bromegrass. Theor. Appl. Genet. 99:1245–1254.[CrossRef]
- Casler, M.D. 2001. Breeding forage crops for increased nutritional value. Adv. Agron. 71:51–107.
- Claessens, A., R. Michaud, G. Bélanger, and D.E. Mather. 2004. Characteristics of timothy genotypes divergently selected for fiber traits. Crop Sci. 44:81–88.[Abstract/Free Full Text]
- Coors, J.G., C.C. Lowe, and R.P. Murphy. 1986. Selection for improved nutritional quality of alfalfa forage. Crop Sci. 26:843–848.[Abstract/Free Full Text]
- Fonseca, C.E.L., D.R. Viands, J.L. Hansen, and A.N. Pell. 1999. Associations among forage quality traits, vigor, and disease resistance in alfalfa. Crop Sci. 39:1271–1276.[Abstract/Free Full Text]
- Goering, H.K., and P.J. Van Soest. 1970. Forage fiber analysis (apparatus, reagents, procedures, and some applications). USDA Agric. Handb. 379. U.S. Gov. Print. Office, Washington, DC.
- Han, L.X., M.D. Casler, and C.R. Grau. 2001. Responses to divergent selection for fiber concentration at two disease potentials in smooth bromegrass. Crop Sci. 41:30–39.[Abstract/Free Full Text]
- Infrasoft International. 1993. NIRS2 version 3.00. Routine operation and calibration software development for near infrared instruments. Infrasoft International, Silver Spring, MD.
- Jung, H.G., and M.D. Casler. 1990. Lignin concentration and composition of divergent smooth bromegrass genotypes. Crop Sci. 30:980–985.[Abstract/Free Full Text]
- Mika, V. 1983. A comparison of the nutritive values of early and late varieties of timothy. Grass Forage Sci. 38:67–71.
- Mowat, D.N., M.L. Kwain, and J.E. Winch. 1969. Lignification and in vitro cell wall digestibility of plants parts. Can. J. Plant Sci. 49:499–504.
- SAS Institute. 1996. The SAS system for Windows. Release version 6.12. SAS Institute, Cary, NC.
- Shenk, J.S., and F.C. Elliot. 1970. Two cycles of directional selection for improved nutritive value of alfalfa. Crop Sci. 10:710–712.[Abstract/Free Full Text]
- Simon, U., and B.H. Park. 1981. A descriptive scheme for stages of development in perennial forage grasses. p. 416–418. In J.A. Smith and V.W. Hays (ed.) Proc. Int. Grassl. Congr.
- Surprenant, J., D.K. Barnes, R.H. Busch, and G.C. Marten. 1988. Bidirectional selection for neutral detergent fiber and yield in reed canarygrass. Can. J. Plant Sci. 68:705–712.
This article has been cited by other articles:
|
|
|
|
 
A. Claessens, R. Michaud, G. Belanger, and D. E. Mather
Leaf and Stem Characteristics of Timothy Plants Divergently Selected for the Ratio of Lignin to Cellulose
Crop Sci.,
October 27, 2005;
45(6):
2425 - 2429.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
