View metadata, citation and similar papers at core.ac.uk brought to you by CORE provided by PDXScholar Portland State University PDXScholar Biology Faculty Publications and Presentations Biology 11-1-2005 Contributions of Heterosis and Epistasis to Hybrid Fitness Mitchell B. Cruzan Portland State University Let us know how access to this document benefits ouy . Follow this and additional works at: https://pdxscholar.library.pdx.edu/bio_fac Part of the Biology Commons, and the Plant Breeding and Genetics Commons Citation Details Rhode, J. M., and Cruzan, M. B. (2005). Contributions of Heterosis and Epistasis to Hybrid Fitness. American Naturalist, 166(5), E124-E139. This Article is brought to you for free and open access. It has been accepted for inclusion in Biology Faculty Publications and Presentations by an authorized administrator of PDXScholar. For more information, please contact [email protected]. vol. 166, no. 5 the american naturalist november 2005 E-Article Contributions of Heterosis and Epistasis to Hybrid Fitness Jennifer M. Rhode* and Mitchell B. Cruzan† Department of Biology, Portland State University, Portland, Stebbins 1959; Grant 1963; Rieseberg et al. 1999). How- Oregon 97207 ever, genomic integration among taxa via hybridization is a complex process because offspring from crosses between Submitted October 24, 2004; Accepted June 7, 2005; Electronically published August 30, 2005 divergent lineages often express diminished viability and fertility (Muller 1942; Templeton 1981; Orr 1996; Arnold 1997; Dowling and Secor 1997). Reductions in hybrid fit- ness (hereafter “hybrid breakdown”) may be a conse- quence of chromosomal rearrangements (Grant 1981; abstract: Early-generation hybrid fitness is difficult to interpret Fishman and Willis 2001), maladaptive trait combinations because heterosis can obscure the effects of hybrid breakdown. We used controlled reciprocal crosses and common garden experiments (Schluter 1998; Hatfield and Schluter 1999; Schemske to distinguish between effects of heterosis and nuclear and cytonu- 2000; Via 2002), or epistasis among divergent genomic clear epistasis among morphotypes and advanced-generation hybrid elements (Dobzhansky-Muller incompatibilities; Dob- derivative populations in the Piriqueta caroliniana (Turneraceae) zhansky 1936; Muller 1942; Orr 1995; Turelli and Orr plant complex. Seed germination, growth, and sexual reproduction 2000; Orr and Turelli 2001). Alternatively, the fitness of of first-generation hybrids, inbred parental lines, and outbred pa- early-generation hybrids may be increased because of het- rental lines were compared under field conditions. Average vegetative performance was greater for hybrids than for inbred lines, and first- erosis (i.e., because of dominance and overdominance at season growth was similar for hybrids and outbred parental lines. loci fixed for inferior alleles in the parental populations) Hybrid survival surpassed that of inbred lines and was equal to or after crosses among inbred populations (Lynch 1991; greater than outbred lines’ survival, and more F1 than parental plants Hedgecock et al. 1995; Burke and Arnold 2001). While reproduced. Reductions in hybrid fitness due to Dobzhansky-Muller heterosis has the potential to counteract fitness losses due incompatibilities (epistasis among divergent genetic elements) were to epistasis in early-generation hybrids, the effects of out- expressed as differences in vegetative growth, survival, and repro- crossing are not typically considered in studies of early- duction between plants from reciprocal crosses for both F1 and back- cross hybrid generations. Comparing performance of hybrids against generation hybrid fitness. parental genotypes from intra- and interpopulation crosses allowed Our understanding of hybrid breakdown as a mecha- a more robust prediction of F1 hybrids’ success and more accurate nism for reproductive isolation is based primarily on the interpretations of the genetic architecture of F1 hybrid vigor. model proposed by Dobzhansky (1936, 1937) and Muller (1940, 1942) and then further developed by Orr and col- Keywords: Cytonuclear epistasis, field transplants, heterosis, inbreed- leagues (Orr 1995; Coyne and Orr 1998; Turelli and Orr ing depression, hybrid breakdown, introgression. 2000; Orr and Turelli 2001). In the Dobzhansky-Muller model, hybrid inferiority results from interactions among new mutations that have become fixed in diverging lin- The exchange of genetic material among divergent lineages eages (Dobzhansky 1936; Muller 1942; Orr 1995), so re- has been a potent evolutionary force in a variety of plant productive isolation can evolve without fitness loss. Ac- and animal groups (Stebbins 1959; Knobloch 1972; Har- cording to this model, if the ancestral genotype were aabb, rison 1990; Arnold 1997; Dowling and Secor 1997) and divergent lineages might become fixed for mutations at may be a mechanism for the acquisition of adaptive genetic different loci (a a bb and aab b ). First-generation hybrids elements (Anderson 1949; Anderson and Stebbins 1954; (aa bb ) would be expected to have reduced viability due to epistatic interactions between the a and b alleles (Orr * Present address: Department of Biological and Environmental Sciences, Georgia College and State University, CBX 081, Milledgeville, Georgia 31061; 1995; Coyne and Orr 1998). Similarly, reductions in hybrid e-mail: [email protected]. viability could be a consequence of epistasis between di- † E-mail: [email protected]. vergent nuclear and cytoplasmic genetic elements (i.e., Am. Nat. 2005. Vol. 166, pp. E124–E139. ᭧ 2005 by The University of aa -c or bb -c ). Because mutations contributing to hybrid Chicago. 0003-0147/2005/16605-40699$15.00. All rights reserved. breakdown are operationally neutral in their native genetic Heterosis, Epistasis, and Hybrid Fitness E125 backgrounds and because each allele has the potential for Table 1: Survey of primary literature published from 1993 to multiple interactions, the accumulation of Dobzhansky- 2003 that compared fitness of animal or plant F1 hybrids to Muller incompatibilities is expected to be exponential (Orr parentals under field or laboratory conditions 1995). Parental genotypes Many theoretical and empirical studies have focused Result Inbred, field Inbred, lab Outbred on loss of fitness in hybrids, but if populations of the parental taxa are inbred, early-generation hybrids may Equal or superior 6 7 0 instead display heterosis and increased fitness. Limited Intermediate 3 2 1 dispersal and small population size can lead to local in- Inferior 1 1 0 Depends on environment 0 3 0 breeding (Wright 1932; Fisher 1965; Hedrick and Miller 1992), which often results in fixation of mildly delete- rious mutations (Hartl and Clark 1997). Because in- breeding depression is likely to be the consequence of inbred parental genotypes (i.e., those derived from within- many alleles of small effect (Charlesworth and Charles- population crosses) for comparison may be misleading worth 1987), we would expect populations of divergent with respect to the expected persistence of hybrid geno- taxa to become fixed for different sets of deleterious re- types and hybrid populations. cessive alleles. Crossing between taxa would produce hy- The parental group to which F1 hybrid fitness should brids with higher fitness because a larger proportion of be compared depends on the study’s goals. For example, the loci segregating for deleterious alleles would be het- when analyzing the genetic architecture of hybrid break- erozygous. Examples of increases in the vigor of offspring down, it is best to use outbred (between-population in the generation following an intraspecific interpopula- crosses) parental lines for comparison. Contrasting hy- tion cross are common (Darwin 1876; Dobzhansky 1937; brid and inbred (within-population crosses) parental ge- Stebbins 1950; Mayr 1963; Charlesworth and Charlesworth notypes can underestimate the severity of hybrid break- 1987; Galloway and Fenster 1999; Fenster and Galloway down because parentals, with their fixed deleterious 2000). Similarly, interbreeding among closely related taxa recessive alleles, may seem relatively unfit (e.g., Emms is likely to result in heterosis for hybrid offspring (Lynch and Arnold 1997; Dunham and Argue 2000; Parris et al. 1991). However, the positive effects of isolated incidences 2001; Hauser et al. 2003). To predict the performance of of outcrossing will be transient, as the level of heterozy- newly formed hybrid genotypes within parental popu- gosity in offspring will decrease with each hybrid gener- lations, on the other hand, it may be more appropriate ation after the first (Burke and Arnold 2001). The expected to compare hybrids with inbred (within-population) pa- boost in vigor of early-generation hybrids due to increased rental crosses, the genotypes with which they would com- heterozygosity could be critical for the establishment of pete under field conditions. Knowing the degree to which hybrid populations and the initiation of introgression (Ar- populations are inbred will allow a more accurate as- nold 1997), but analyses isolating the specific contributions sessment of their invasibility by migrant genomes (e.g., of heterosis to early-generation hybrid vigor have not been Richards 2000; Ebert et al. 2002; Haag et al. 2002; Marr conducted. The high frequency of biparental inbreeding et al. 2002). Furthermore, comparing hybrid genotypes as a consequence of small effective population size in many from reciprocal crosses to both