Crop Science Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 1 March 2007
Published in Crop Sci 47:629-635 (2007)
© 2007 Crop Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Lee, E. A.
Right arrow Articles by Good, B.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Lee, E. A.
Right arrow Articles by Good, B.
Agricola
Right arrow Articles by Lee, E. A.
Right arrow Articles by Good, B.
Related Collections
Right arrow Maize
Right arrow Crop Genetics

CROP BREEDING & GENETICS

Re-examining the Relationship between Degree of Relatedness, Genetic Effects, and Heterosis in Maize

E. A. Lee*, M. J. Ash and B. Good

Dep. of Plant Agriculture, Crop Science Bldg., Univ. of Guelph, Guelph, ON, Canada, N1G 2W1. Financial support, in part, from the Ontario Ministry of Agriculture and Food, Natural Science and Engineering Research Council, Canadian Foundation for Innovation, Ontario Innovative Trust, and Ontario Corn Producers' Association.

* Corresponding author (lizlee{at}uoguelph.ca).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The dominance hypothesis is one of two major genetic hypotheses that have been proposed regarding heterosis in maize (Zea mays L.). This study examines two underlying tenets of the dominance hypothesis: (i) Dominant gene action must occur at many loci in order for heterosis to be expressed; and (ii) genetic diversity is a good predictor of heterosis (i.e., differences in gene frequency are required for the expression of heterosis). To examine these tenets, we used a unique set of genetic materials, sister-line inbred lines. Sister-line inbred lines are highly related inbred lines that are derived from a common parental cross. Three sets of six sister lines were used in this study, ranging between 47 and 77% identical-by-descent (IBD), creating a series of lines that potentially vary in gene frequency. The sister lines were mated using a partial diallel to form sister-line hybrids. The sister-line hybrids and the parental inbred lines were evaluated in replicated yield trials for grain yield, grain moisture, broken stalks, and test weight in five environments. The genotypic variance was partitioned using Gardner and Eberhart's Analysis III to examine additive and nonadditive genetic effects. Three relevant findings regarding heterosis for grain yield can be drawn from our results: Substantial genome-wide heterozygosity is not a requirement for the expression of heterosis, there is not a consistent relationship between degree of relatedness and the magnitude of heterosis, and the presence of nonadditive genetic effects is not a requirement for the manifestation of heterosis.

Abbreviations: GCA, general combining ability • IBD, identical-by-descent • MPH, midparent heterosis • OCHU, Ontario crop heat units • RCBD, randomized complete block design • SCA, specific combining ability • SL, sister line • SSR, simple sequence repeat.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
HYBRID MAIZE (Zea mays L.) traces its roots back to experiments on outcrossing and inbreeding conducted by Shull (1908, 1909) at Cold Spring Harbor Laboratories in New York and East (1908) at Connecticut State College. They observed that when maize plants were self-pollinated (i.e., inbred) in successive generations, vigor and grain yield rapidly deteriorated (Shull, 1908; East, 1908). However, when two inbred lines from unrelated populations were crossed, both vigor and grain yield of the F1 hybrid often exceeded that observed for the original source populations (Shull, 1908). It was these observations, made nearly 100 yr ago, and methodology outlined by Shull (1909) that gave rise to the modern hybrid maize industry (Crow, 1998). This phenomenon, termed heterosis by Shull (1952), refers to the superior performance of the F1 hybrid over either one of its parents.

Two major genetic hypotheses have been proposed regarding the genetics underlying heterosis: (i) dominance hypothesis and (ii) overdominance hypothesis. The dominance hypothesis attributes heterosis to the accumulation of favorable dominant genes or masking of deleterious recessive alleles in the hybrid (Davenport, 1908; Bruce, 1910; Keeble and Pellew, 1910). In quantitative genetic terms, heterosis results when there is some degree of directional dominance and the parents differ in gene frequency (Bruce, 1910; Falconer, 1981). The dominance hypothesis can be expressed in terms of a single-locus (B) with no epistasis as

Formula
where a and –a are the genotypic values of the parental genotypes (B1B1 and B2B2) and d is the genotypic value of the nonparental genotype (B1B2). The dominance theory is consistent with recent genomic evidence of differences in genic content between maize inbred lines (Fu and Dooner, 2002; Song and Messing, 2003; Brunner et al., 2005) and has been demonstrated as the underlying cause of a heterotic response for grain yield in a quantitative trait locus mapping study (Graham et al., 1997). The other hypothesis, overdominance, argues that the heterozygous combination of the alleles at a single locus is superior to either of the homozygous combinations (Shull, 1908; East, 1908). The overdominance hypothesis, unlike the dominance hypothesis, does not require the presence of either linkage or the involvement of multiple loci for heterosis to be expressed, nor is it necessarily based on classic Mendelian genetics. However, like the dominance hypothesis, it requires that the parents differ in gene frequency. While there is no direct evidence in support of this hypothesis in the literature, it has not been completely rejected as an underlying genetic cause.

Numerous studies have examined the relationship between genetic distance and heterosis in maize (e.g., Melchinger et al., 1990; Smith et al., 1991; Melchinger, 1999) and the utility of genetic markers for classifying inbred lines into heterotic groups (e.g., Lee et al., 1989; Dudley et al., 1991; Melchinger et al., 1991). These works are based on the assumption that genetic diversity (i.e., differences in gene frequency) is a requirement for the expression of heterosis, consistent with either of the current genetic theories. The genetic materials used in these studies was quite diverse, including highly selected elite inbred lines from the central corn belt (Smith et al., 1990, 1991), germplasm pools, and populations representing tropical, subtropical, and temperate germplasm (Reif et al., 2003), and selected S3 lines from two breeding populations (Lanza et al., 1997).

This study re-examines the two underlying tenets of the dominance theory: first, that dominant gene action must occur at many loci for heterosis to be expressed, and second, that genetic diversity is a good predictor of heterosis (i.e., differences in gene frequency are required for the expression of heterosis). Unlike previous heterosis studies (e.g., Melchinger et al., 1990; Smith et al., 1990, 1991; Melchinger, 1999; Reif et al., 2003), the present study used a unique set of genetic materials to modulate gene frequency. Sister-line (SL) inbred lines are highly related inbred lines derived from a common parental cross. In theory, any two inbred lines derived from a common F1 should be 50% identical-by-descent (IBD). In other words, SL inbred lines should by chance alone share 50% of their alleles. The three sets of SLs used in this study represent lines that are between 47 and 77% IBD, creating a series of lines that potentially vary in gene frequency. The SLs have two distinct features: (i) A maximum of two possible alleles are present for a given locus, and (ii) very precise estimates of genetic relatedness can be determined. With this unique set of genetic materials, our specific objectives were to examine the requirement for genome-wide heterozygosity, the relationship between genetic distance and heterosis, and the role of additive and nonadditive genetics effects in heterosis.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Genetic Materials and Molecular Genotyping
This study used three families of six SL inbred lines and the resulting 45 SL hybrids (Table 1). Sister lines are highly related inbred lines (e.g., A and A*) that have been developed from the same parental cross, generally a F1 single-cross. The three families of inbred lines used in the present study were derived over a 15-yr period through inbreeding three single-cross hybrids: Pioneer 3902, Pioneer 3929, and Pioneer 3790. Each line was developed using traditional plant breeding methodology, starting with a F1 single-cross (only two possible alleles at a locus); rapid inbreeding via self-pollination; early generation (S3) testing using two unrelated inbred testers of defined heterotic patterns and a minimum of three environments to identify superior genotypes; and additional testing in advanced generations (S5 to S7) using additional inbred testers and three to six environments to identify superior genotypes. Within a set of lines derived from a common F1, any pair of inbred lines derived from different F2 plants should by chance share 50% of their alleles IBD (fij = 1/2). Because of the nature of the starting material (i.e., two distinct heterotic patterns represented in the F1) and the methodology used to develop the lines (i.e., self-pollination and using testers that potentially represent one of the heterotic patterns in the original F1), we anticipated that for any pair of related inbred lines, more than 50% of the genome has the potential to be IBD. Inbred lines from family 3902 and family 3790 belong to the Iodent heterotic pattern (Lee et al., 2000a, 2000b, 2001), while family 3929 represents an unrelated heterotic pattern (Lee, unpublished data). Sister-line hybrids were generated by mating two SLs (i.e., AxA*). Simple sequence repeat (SSR) primer pairs were used to establish IBD estimates between inbred lines within each family (Table 2) (Lee et al., 2006).


View this table:
[in this window]
[in a new window]

  		Table 1. Germplasm description of the 18 inbred lines used as female lines in the genetic distance experiment (Lee et al., 2006).

 

View this table:
[in this window]
[in a new window]

  		Table 2. Identity-by-descent (IBD) estimates for the sister-line (SL) combinations. For family 3902, the estimates are based on 178 simple sequence repeat (SSR) primer pairs; for family 3929, the estimates are based on 199 SSR primer pairs; and for family 3790, the estimates are based on 192 SSR primer pairs. Significant deviations from the expected IBD of 50% were determined using Chi-square goodness-of-fit test. (Lee et al., 2006).

 
Experimental Design
The 18 inbred lines and 45 SL hybrids were grown in five environments in a replicated yield trial at four southwestern Ontario locations: Alma (2550 Ontario Crop Heat Units [OCHUs] [Brown and Bootsma, 1993]) and Waterloo (2750 OCHUs) in 2002 and 2003, and Woodstock (2850 OCHUs) in 2002, using a randomized complete block design (RCBD) with two replications at each location (Lee et al., 2006). The RCBD involved the randomization option for nearest neighbor analysis (Agrobase IV software, Agronomix Software, Portage la Prairie, MB; Mulitze, 1992) to minimize the number of times that the same treatments are adjacent in each replicate. Experimental units were two-row plots, 11.56 m long, with a spacing of 0.76 m between rows. To meet the minimum plot weights required for the HM2200 HarvestMaster dual GrainGage (Juniper Systems, Inc., Logan, UT), plot lengths in the inbred trial were twice the length of those used in the hybrid trial. Competition between plots of SL hybrids and inbred lines was viewed to be minimal, and therefore 2-row, rather than 4-row plots were used. This decision was based on 2 yr of observations of the SL hybrids and the inbred lines in the breeding nursery (Cambridge, Ontario, 2000–2001). The major competition effect between plots in a yield trial is based on differences in plant height and leaf area. Visual assessment indicated that the SL hybrids and inbred lines used in this study did not differ substantially in plant height or leaf area.

The trials were machine planted (Wintersteiger precision planter, Wintersteiger, Inc., Salt Lake City, UT) in May and machine harvested using a New Holland split-plot combine (ALMACO, Nevada, IA) equipped with a HM2200 HarvestMaster high-capacity dual GrainGage in October or November. Trials were overplanted and thinned at the six-leaf tip stage to uniform stands of 68 000 plants ha–1 (2002) and 64200 plants ha–1 (2003). The soil type at all Ontario locations is Guelph loam (Typic Hapludalf). Fertilizer was applied based on soil tests at the rate of: 102, 54, and 43 kg ha–1 (N, P2O5, and K2O, respectively), supplemented with 50 000 L ha–1 liquid swine manure in Alma (2002); 100, 39, and 39 kg ha–1 in Alma (2003); 157, 39, and 39 kg ha–1 in Waterloo (2002); 140, 50, and 50 kg ha–1 in Waterloo (2003); and 140, 20, and 50 kg ha–1 in Woodstock (2002). Weeds were controlled using conventional herbicides. Four traits were measured: machine harvested grain yield (Mg ha–1) adjusted to 155 g kg–1 grain moisture, grain moisture at harvest (g kg–1), percentage of broken stalks (plants broken below the ear or inclined more than 45° from the vertical), and test weight (kg hl–1).

Statistical Analysis
Individual plot means were adjusted using nearest neighbor analysis (Agrobase IV software; Mulitze, 1992) according to the BEST option, which compares the adjustments made on a longitudinal and latitudinal basis and chooses the option with the greatest precision. The adjusted plot means were then combined across locations and analyzed as an RCBD using PROC GLM (SAS, 1999) according to the linear model

Formula
where Yrge is the measured trait of genotype g in replicate r at environment e, µ is the grand mean, {propto}g and ße are the genotype and environment main effects, {rho}r(ße) is the replicate effect nested within an environment, {propto}gße is the interaction between main effects, and {varepsilon}rge is the random experimental error. Genotypes were considered fixed, while environments and blocks were considered random effects.

Genotype variance was partitioned into among inbred line families, family 3902, family 3790, and family 3929. Within each family partition, the family variance was partitioned using Analysis III of Gardner and Eberhart (1966) into parents (i.e., inbred lines), crosses (SL hybrids), and parents vs. crosses. The crosses variance was further partitioned into general combining ability (GCA) and specific combining ability (SCA). Genotype variance partitions were tested using their respective source x environment mean squares, while the pooled error term was used to test sources partitioned from the entry x environment source. The approximate F-test of the ratio of the GCA to SCA mean squares was used to test the significance of the amount of additive vs. nonadditive variation, assuming that SCA is randomly and normally distributed with constant variance (Lin and Binns, 1991). Midparent heterosis (MPH) was calculated for each trait as [(F1– MP)/MP] x 100, where F1 is the mean of the F1 hybrid performance and MP = (P1 + P2)/2, in which P1 and P2 are the means of the inbred parents, respectively. Simple correlations between parameters were computed using PROC CORR of SAS ver. 8.2 (SAS Institute, 1999).


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
In theory, any two inbred lines derived from the same F1 should be 50% IBD. Most of the IBD estimates in the 3902 and 3790 families were significantly greater than 50% (Table 2). Fewer IBD estimates in the 3929 family were significantly greater than 50%; however, in all cases, no IBD estimates significantly less than 50% were observed. The IBD estimates for all possible SL combinations ranged from 47 to 77%. In other words, inbred lines CG42 and CG83, from family 3929, have inherited only 47% of their alleles from the same parent, while inbred lines CG64 and CG85, from family 3790, have inherited 77% of their alleles from the same parent. The lines within family 3902 and family 3790 are more closely related to one another and are considered SLs. Some of the lines within family 3929 show the same levels of IBD observed in the other two families. However, about half of the inbred line combinations within family 3929, particularly those involving CG42, are not considered SLs, based on the lower levels of IBD (47–54%) that were not significantly different from the 50% IBD expectation. This is consistent with molecular fingerprinting data suggesting that CG42 is an Iodent line, while the other five lines belong to a distinct and unrelated heterotic group (Lee and Lukens, unpublished data).

The highest average grain yield, 6.3 Mg ha–1, was recorded at the Alma location in 2002 (7.5% broken stalks, 72.6 kg hl–1, 23.8% grain moisture). The lowest average grain yield, 5.2 Mg ha–1, was recorded at the Alma location in 2003 (13.3% broken stalks, 66.5 kg hl–1, 31.4% grain moisture). Genotype and the genotype x environment interaction were significant sources of variation for all four traits (Table 3). Environment was a significant source of variation for test weight, grain moisture, and broken stalks. However, it was not a significant source of variation for grain yield. Within genotypes, there were significant differences among and within the three families of inbred lines for all of the traits, except broken stalks within family 3902. The parents vs. crosses contrast, an indication of heterosis, was significant for grain yield in all three families. Even though the inbred lines are highly related (i.e., ≥ 50% IBD), significant heterosis for grain yield was observed (Table 4).


View this table:
[in this window]
[in a new window]

  		Table 3. Mean squares from the analysis of variance (Analysis III [Gardner and Eberhart, 1966]) of the inbred and sister-line yield trial combined across five locations for grain yield, test weight, grain moisture, and broken stalks.

 

View this table:
[in this window]
[in a new window]

  		Table 4. Family means and standard errors for the midparent (MP) grain yield and sister-line hybrid (SL) grain yield, midparent heterosis (MPH) for grain yield, and identity-by-descent (IBD).

 
Grain yield of the inbred lines was significantly lower than the grain yield of the SL hybrids for all three families of lines (Table 4). Families 3902 and 3790 exhibited similar average inbred and SL hybrid grain yields, while lower average grain yields were observed for family 3929 for both the inbred lines and the SL hybrids (Table 4). The average MPH estimates for the three families of lines are similar to MPH estimates observed for modern commercial maize hybrids. Modern commercial hybrids and their inbred parents grown under varying plant densities (30 000 plants ha–1 to 79 000 plants ha–1) exhibit MPH estimates between 92 and 111%, respectively (data from Duvick, 1999). Midparent heterosis for grain yield in the SL hybrids ranged from a low of 59.5% for CG40 x CG105 to a high or 214.1% for CG42 x CG83. Family 3929 exhibited the highest average MPH for grain yield (132%), while family 3902 exhibited the lowest average MPH for grain yield (81%) (Table 4). The larger MPH estimates in family 3929 are due to the relatively poor performance of the inbred parents, rather than to superior performance of the SL hybrids (Table 4).

Two of the families exhibited a significant ({alpha} = 0.05), negative relationship between IBD estimates and MPH (r = –0.77 for family 3902 and r = –0.62 for family 3790) (Table 5, Fig. 1 ). In other words, as the SLs share more alleles in common, less heterosis is observed. However, no significant relationship was detected between IBD estimates and MPH in family 3929 (r = –0.23) (Table 5, Fig. 1). Interestingly, it was in family 3929 with the lowest IBD estimates (i.e., the greatest genetic diversity) that the highest levels of MPH were observed (Table 4).


View this table:
[in this window]
[in a new window]

  		Table 5. Simple correlations between hybrid (H) grain yield, midparent (MP) inbred grain yield, midparent heterosis (MPH), and identity-by-descent (IBD) within each family of inbred lines.

 

Figure 1
View larger version (12K):
[in this window]
[in a new window]

  		Figure 1. Relationship between the identity-by-descent (IBD) estimates for sister-line inbred line pairs and heterosis (%) for grain yield observed in the sister-line hybrids, involving sister lines from three families: {diamond} Family 3920 lines; {blacksquare} Family 3790 lines; {blacktriangleup} Family 3929 lines.

 
General combining ability was a significant source of variation for grain yield in all three inbred line families; SCA was a significant source of variation only in family 3929 (Table 3). There was no significant relationship between hybrid grain yield and midparent grain yield for any of the families (Table 5, Fig. 2 ), meaning that average per se performance of the inbred parent is not a predictor of SL hybrid performance. The hybrid type used in this study is what Melchinger (1999) refers to as an intragroup hybrid. In previous studies examining both "intragroup" and "intergroup" hybrids, the relative importance of SCA effects to GCA effects is greater for intragroup hybrids (Melchinger, 1999). This was not the case for two of the intragroup hybrid families used in this study. In both family 3902 and family 3790, a significantly greater proportion of the genetic effects influencing grain yield is additive, as indicated by the approximate F-tests (Table 3). The contribution of nonadditive genetic effects to SL hybrid grain yield in family 3929 appears to be more important than in the other two families (i.e., nonsignificant approximate F-test), consistent with Melchinger's (1999) observations for intragroup hybrids (Table 3). This difference in relative importance of genetic effects may in part explain the lack of a significant relationship between degree of relatedness and heterosis in family 3929.


Figure 2
View larger version (10K):
[in this window]
[in a new window]

  		Figure 2. Relationship between midparent grain yield (Mg ha–1) for sister-line inbred line pairs and hybrid grain yield (Mg ha–1) observed in the sister-line hybrids, involving sister lines from three families: {diamond} Family 3920 lines; {blacksquare} Family 3790 lines; {blacktriangleup} Family 3929 lines. Grain yields were averaged across five environments.

 

   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
This study took a unique approach to examining heterosis by using three families of six SLs. We were consequently able to definitively establish the relationship between degree of relatedness and heterosis for grain yield in the SL hybrid by exploiting the phenomenon of IBD. Three relevant findings regarding heterosis can be drawn from our results. First, significant heterosis can be observed, even between pairs of closely related inbred lines. Therefore, it would appear that substantial genome-wide heterozygosity is not a requirement for the expression of heterosis. Second, the relationship between degree of relatedness and the magnitude of heterosis is not consistent. Two of the families showed a significant, negative relationship between the level of IBD and heterosis, meaning that the more alleles that two inbred lines shared in common, the less heterosis that was observed. However, the other family of inbred lines did not demonstrate a significant relationship between level of IBD and heterosis. This also supports our contention that substantial genome-wide heterozygosity is not a requirement for the expression of heterosis. Third, the nature of genetic variation cannot be inferred from the relationship (or lack there of) between per se performance and hybrid performance. In two of the families, mostly additive genetic effects controlled grain yield, even in the absence of a significant relationship between per se performance and hybrid performance. Hence, the presence of nonadditive genetic effects is not a requirement for the manifestation of heterosis.


   ACKNOWLEDGMENTS
 
Technical support by C. Grainger and A.K. Singh; original single-cross seed of Pioneer 3902, Pioneer 3929, and Pioneer 3790 supplied by Pioneer Hi-Bred International is gratefully acknowledged.

Received for publication May 12, 2006.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Lee, E. A.
Right arrow Articles by Good, B.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Lee, E. A.
Right arrow Articles by Good, B.
Agricola
Right arrow Articles by Lee, E. A.
Right arrow Articles by Good, B.
Related Collections
Right arrow Maize
Right arrow Crop Genetics

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Vadose Zone Journal
Journal of Natural Resources
and Life Sciences Education		 Soil Science Society of America Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome