Proc. Nati. Acad. Sci. USA Vol. 88, pp. 11413-11415, December 1991 Evolution A test of Fisher's theory of dominance (recessivity/Chlamydomonas) H. ALLEN ORR Center for Population Biology, University of California, Davis, CA 95616 Communicated by James F. Crow, September 12, 1991 (receivedfor review June 15, 1991)

ABSTRACT One ofthe first patterns noticed by geneticists (7) have suggested, this physiological theory of dominance was that mutations are almost always recessive to their wild- differs profoundly from Fisher's: it does not invoke modifi- type alleles. Several explnations of this sriking pattern have cation of heterozygotes by natural selection. been offered. The two most influential are Fisher's theory- Few tests of these theories have been performed. Indeed, which argues that dominance results from natural selection the strongest evidence against Fisher's theory is a statistical against deleterious mutations-and Wrights theo- pattern noted by Charlesworth (8): although Fisher's theory ry-which argues that dominance results from the physiology predicts no correlation between the homozygous effect of a of gene action. The debate over which of these theories is mutation on fitness (s) and its dominance coefficient (h), s correct represents one of the most pr ed controversies in and h show a strong inverse correlation. Here I perform a the history of evolutionary biology. Here I test Fisher's theory direct test of Fisher's theory of dominance by examining the by asseing the domince ofmutations in an organism that is dominance of mutations in a normally haploid eukaryote, the typically haploid, the alga Chiamydomonas hardtii. The unicellular alga Chlamydomonas reinhardtii. The results results show that mutations are recessive just as often among falsify the notion that dominance results from modification of haploid as among diploid species. This result falsifies Fisher's heterozygotes by natural selection. theory of dominance and provides strong support for the alternative physiological theory. THE TEST Almost all mutations are recessive. Over the last 70 years, Perhaps the most critical test of Fisher's theory involves many explanations ofthis striking pattern have been offered. assessing the dominance of mutations in "artificial diploids" The debate over which of these theories, if any, is correct from a normally haploid species. Because heterozygotes do became one of the longest and fiercest controversies in the not exist in haploid species, it is impossible to select for history of evolutionary biology. reduced expression of mutations in heterozygotes; thus, Perhaps the most famous-and contested-of these theo- Fisher's theory could not explain the recessivity of mutations ries was offered by Fisher (1). Fisher argued that the domi- in haploid species. Unfortunately, the well-known recessiv- nance of wild-type alleles results from natural selection ity of most mutations in bacteria and fungi, although sugges- against recurrent mutations: although most mutations were tive, does not cleanly settle the issue: bacterial cells are originally semidominant, selection against their deleterious "polykaryotic"-that is, they normally carry several copies heterozygous effects gradually reduced their expression ofa chromosome (9, 10). Thus, each time a mutation recurs- among heterozygotes until these mutations became nearly which must be very often given the enormous size ofbacterial completely recessive. Fisher claimed that this decrease in populations-it spends several generations as a heterozygote dominance resulted from the accumulation ofmodifier alleles and is exposed to Fisherian selection for recessivity. Many at other loci. Although Wright (2, 3) severely criticized this fungi are also polykaryotic: the compartments of haploid theory, showing that the selection pressure on dominance hyphae are frequently multinucleate (11). Furthermore, fungi modification would only be of the order of the mutation rate, do not possess true cells; instead, the gene products of Fisher (4) maintained that extremely small selection coeffi- adjacent compartments freely mix (11). Thus, mutations cients were adequate if selection were exerted over a very essentially arise as heterozygotes, just as in diploid orga- long time. nisms. Natural selection-even weak selection acting for Wright (2, 3), on the other hand, argued that dominance only several generations each time a mutation arises-can follows from the physiology ofgene action. Wright suggested then render these mutations increasingly recessive over long that there were simple metabolic reasons to expect a curve of periods -of time. Last, fungi have a diploid sexual stage in diminishing returns relating phenotype to genotype. If most which Fisherian selection might occur. wild-type alleles have very high enzyme activities, then Chlamydomonas escapes most of these difficulties and so having one allele will increase flux through some metabolic provides nearly ideal material for testing Fisher's theory. pathway from zero to a substantial level, while adding a Chlamydomonas, which can reproduce vegetatively for an second allele will cause a negligible increase in flux. Obvi- apparently endless number of generations, spends the over- ously, then, organisms will enjoy a margin of safety against whelming proportion of its time as a haploid (12, 13). More- loss-of-activity mutations: flux through mutant heterozy- over, its haploid cells are truly monokaryotic-i.e., they goteswill roughly equal that through wild-type homozygotes. carry only one copy of each chromosome (12). Thus, muta- As a result, most mutations will appear recessive. Regardless tions arise as hemizygotes, not heterozygotes, and so are not of whether the high activity of wild-type enzymes results subjected to Fisherian selection. Although Chlamydomonas from selection to withstand environmental fluctuations, as can enter into a diploid zygotic state, this stage is brief and Wright, Plunkett (5), and Muller (6) suggested, or is simply an dormant (12). Most important, as discussed below, many inevitable consequence of metabolism, as Kacser and Burns characters are expressed only in the haploid stage. The dominance of Chlamydomonas mutations has been The publication costs ofthis article were defrayed in part by page charge tested in two ways. First, one can form temporary dikaryons: payment. This article must therefore be hereby marked "advertisement" when two haploid gametes fuse to form a prozygote, one can in accordance with 18 U.S.C. §1734 solely to indicate this fact. observe the phenotype of the cell immediately prior to 11413 Downloaded by guest on September 30, 2021 11414 Evolution: Orr Proc. Natl. Acad. Sci. USA 88 (1991) nuclear fusion (13, 14). Second, one can screen for rare lethal mutations from the Chlamydomonas data (see below), diploid vegetative cells. In the laboratory, zygotes occasion- as Fisher did for the Drosophila data, the distributions ally (0.2-3.0%o) divide mitotically instead of meiotically, become even more similar (G121 = 0.65; P = 0.75). Thus, yielding temporary vegetative diploids (15, 16). These dip- mutations are recessive just as often in a typically haploid loids represent evolutionary dead-ends: they are mitotically eukaryote as in a typically diploid species. unstable and difficult to maintain and apparently cannot It is particularly interesting to note that in Chlamydomo- undergo imeiosis (12, 17, 18). nas, just as in Drosophila (4, 8), lethal mutations are usually I surveyed the literature for information on the dominance recessive. Among auxotrophic or severe photosynthetic- of Chlamydomonas mutations using Harris' (12, 19) mutant defective mutations, which would almost surely be lethal in lists as guides. To ensure that mutations having similar nature (i.e., ac, arg, cr, nit, pab, and thi mutations), 13 of 16 phenotypes represent alleles at separate loci, only mutations are recessive. (Note that this does not reflect a selective bias that were mapped to a known linkage group were included. toward recovering recessives: under permissive conditions, Because its evolutionary history is unclear, mutations map- dominant auxotrophs can be recovered and maintained just ping to the unusual circular UNI chromosome (12) were as readily as recessives.) Fisherian selection to modify the excluded; inclusion of these loci would not, however, qual- dominance oflethals, while possible in bacteria and fingi, is, itatively change the present results. of course, impossible in a monokaryotic haploid: lethal mutations are immediately lost each time they arise regard- less of their dominance in artificial diploids. It is difficult to RESULTS AND DISCUSSION see, therefore, how the recessivity oflethals in Chiamydomo- Table 1 lists all loci for which information on dominance was nas could result from natural selection. found. Table I also indicates whether the mutations were One could object, however, that recessivity of Chlamy- dominant, semidominant, or recessive. The results are un- domonas mutations results from Fisherian selection in its ambiguous: most mutations in Chlamydomonas are recessive brief diploid zygotic stage. Although unlikely, one can elim- (Tables 1 and 2). Indeed, no fully dominant mutations were inate this possibility almost entirely. Many characters are found. -Most important, the distribution ofdominance effects simply not expressed in the zygote; indeed almost al of the found in Chlamydomonas does not differ from that Fisher (1) mutations listed in Table 1 have no discernible phenotypic found in Drosophila (Table 2; log-likelihood ratio test G[21 = effect among zygotes, explaining why dominance tests must 2.25; P = 0.33). These Drosophila data, of course, provided be performed in artificial, not meiotic, diploids. Perhaps the the original motivation for Fisher's theory. If we exclude clearest and best studied of these "haploid-limited" charac- ters is the flagellum. The mature zygote does not possess Table 1. Dominance of mapped nuclear mutations flagella; moreover, flagellar phenotype is determined solely in Chlamydomonas by the haploid genotype (14). One can therefore compare the dominance of flagellar mutations in Chlamydomonas (some Locus Dominance Ref. Locus Domdnance Ref. unknown fraction of which block some metabolic pathway) ac-9 r 19.29 nic-7 r 15 with all mutations in Drosophila. Although the sample size ac-i4 r 19 nic-13 r 19.20 here is obviously smaller, the result is clear: these mutations ac-20 r 19.68 nit-i s 25 are recessive just as frequently as those from a diploid ac-29 r 16 nit-2 s 25 organism despite the fact that they have neverbeen subjected ac-11S r 19.56 pab-2 r 12 to selection in diploids (Table 2; G121 = 0.18; P = 0.91). Ibis ac-141 r 19.56 Pf-I* r 19.50 result seems fatal to Fisher's theory of dominance. act-2 r 19.78 pf-2* r 19.20 Interestingly, these data also falsify two other theories of agg-i r 19.78 r 19.16 dominance, both offered by Haldane (28, 29). In his first arg-i r 15 pf-i6* r 19.16 theory, which is closely related to Fisher's, Haldane (28) arg-2 r 15 pf-li7* r 27 argued that natural selection provides a safety net against the arg-7 r 19.60 pf-12* r 19.1 heterozygous effects of mutations by replacing wild-type arg-9 r 21 Pf-I9* r 19.1 alleles that produce "just enough" enzyme with those pro- apm-2 r 22 pf-20* r 19.1 ducing "too much" enzyme. In short, the evolution of cr-i 19.36 sad-i r 19.49 dominance may involve substitutions at the locus suffering cr-7 r 19.68 shf-i* r 19.51 the heterozygous effects of recurrent mutations, not substi- ery-) 20 shf-2* r 19.51 tutions at other modifier loci as Fisher had maintained. In his ery-3 19.20 smr-i s 24 second theory, Haldane (29) suggested that, during a favor- fla-i* r 19.2 sr-i r 15 able mutation's sweep to fixation, modifiers might accumu- fla-2* r 19.2 sup-l* s 19.47 late that render the new mutation dominant to the allele it is fla-3* r 19.2 sup-2* r 19.47 replacing. Both of these theories, like Fisher's, require fla-4* r 19.2 sup-3* r 19.47 selection on heterozygotes and so cannot explain dominance fla-S* r 19.2 thi-3 r 19.78 in haploid organisms. fla-6* r 19.2 vfl-i* r 19.3 fla-8* r 19.2 y-i r 26 Table 2. Dominance of mutations in Chlamydomonas and Drosophila If-?I* r 19.65 y-5 r 19.22 r 23 y-6 r 19.22 Dominant Semidominant Recessive mbo-i* r 19.76 r 19.22 y-7 All r 19.76 r 19.23 Chiamydomonas mbo-2* y-8 mutations 0 7 52 msr-i r 19.78 y-9 r 19.23 Chlamydomonas y-iO r 19.23 flagellar Most loci harbor only one known mutant allele; in cases in which mutations 0 1 24 multiple alleles exist, dominance refers to the first mutation de- Drosophila scribed at the locus. r, Recessive; s, semidominant. References ofthe 0 13 208 form "19.x" refer to the numbered citations provided in Harris' mutations mutant list (19). Drosophila data are taken from Fisher (1). For statistical analysis, *Flagellar mutation. the zeros were set to 0.01. Downloaded by guest on September 30, 2021 Evolution: Orr Proc. Natl. Acad. Sci. USA 88 (1991) 11415

Thus, neither Fisher's theory nor Haldane's theories of 11. Barratt, R. W. (1974) in Handbook of Genetics, Vol. 1, ed. dominance can account for the recessivity of mutations in a King, R. C. (Plenum, New York), Vol. 1, pp. 511-530. haploid species. Indeed, the recessivity of mutations in 12. Harris, E. H. (1989) The Chlamydomonas Sourcebook (Aca- Chlamydomonas falsifies any theory of dominance that in- demic, San Diego). 13. Levine, R. P. & Ebersold, W. T. (1960) Annu. Rev. Microbiol. vokes modification of heterozygotes by natural selection. In 14, 197-216. short, these data imply that most mutations are, from the 14. Randall, J. & Starling, D. (1976) in The Genetics ofAlgae, ed. beginning, recessive. Although natural selection might alter Lewin, R. A. (Univ. of California Press, Berkeley), pp. 49-62. the precise activity of wild-type enzymes in particular cases, 15. Ebersold, W. T. (1967) Science 157, 447-449. the present result lends strong support to the notion that the 16. Nichols, T. (1980) in Polyploidy: Biological Relevance, ed. recessivity of mutations is a simple consequence of metab- Lewis, W. H. (Plenum, New York). olism, as Wright suggested. 17. Lee, R. W., Gillham, N. W., Van Winkle, K. P. & Boynton, J. E. (1973) Mol. Gen. Genet. 12, 109-116. I thank B. Charlesworth, J. Coyne, J. Crow, J. Gillespie, C. 18. Lee, R. W., Whiteway, M. S. & York, M. A. (1976) Genetics Langley, T. Prout, and M. Turelli for comments and criticism. This 83, s44. work was supported by a Sloan Postdoctoral Fellowship in Molecular 19. Harris, E. H. (1987) in Genetic Maps 1987, ed. by O'Brien, Evolution to H.A.O. and by National Institutes of Health Grant GM S. J. (Cold Spring Harbor Lab., Cold Spring Harbor, NY), pp. 38462 to J. Coyne. 257-265. 20. Hanson, M. R. & Bogorad, L. (1977) Mol. Gen. Genet. 153, 271-277. 1. Fisher, R. A. (1928) Am. Nat. 62, 115-126. 21. Loppes, R. & R. Heindricks (1986) Arch. Microbiol. 143, 2. Wright, S. (1934) Am. Nat. 68, 24-53. 348-352. 3. Wright, S. (1969) Evolution and the Genetics of Populations 22. James, S., Ranum, L., Silflow, C. & Lefebvre, P. (1988) (Univ. of Chicago Press, Chicago), Vol. 2. Genetics 118, 141-147. 4. Fisher, R. A. (1958) The Genetical Theory ofNatural Selection 23. Barsel, S.-E., Wexler, D. E. & Lefebvre, P. A. (1988) Genetics (Dover, New York), 2nd Ed. 118, 637-648. 5. Plunkett, C. R. (1933) Am. Nat. 67, 84-85. 24. Hartnett, M., Newcomb, J. & Hodson, R. (1987) Plant Physiol. 6. Muller, H. J. (1932) Proc. Sixth Int. Cong. Genet. 1, 213-255. 85, 898-901. 7. Kacser, H. & Burns, J. A. (1981) Genetics 97, 639-666. 25. Fernandez, E. & Matagne, R. F. (1986) Curr. Genet. 10, 8. Charlesworth, B. (1979) Nature (London) 278, 848-849. 397-403. 9. Watson, J. D., Hopkins, N. H., Roberts, J. W., Steitz, J. A. & 26. Wang, W.-Y., Boynton, J. & Gillham, N. (1977) Mol. Gen. Weiner, A. M. (1987) Molecular Biology of the Gene (Benja- Genet. 152, 7-12. min/Cummings, Menlo Park, CA), 4th Ed. 27. Starling, D. & Randall, J. (1971) Genet. Res. Camb. 18, 10. Hayes, W. (1957) in Drug Resistance in Microorganisms, eds. 107-113. Wolstenhlome, G. & O'Connor, C. (Ciba Found. Symp., Lon- 28. Haldane, J. B. S. (1930) Am. Nat. 64, 87-90. don), pp. 197-205. 29. Haldane, J. B. S. (1956) Proc. R. Soc. London B 145, 303-306. Downloaded by guest on September 30, 2021