L567: From last time:

Transcription

1 L567: From last time: Fisher: in Fisher s original presentation, he considered female preference for male traits that were initially favored by natural selection. This could lead to a covariance between the mail trait and the female preference. If the covariance is strong enough, this covariance could lead to a runaway sexual selection, such that the male trait continues to be favored (due to female choice) even after it is no longer favored, or even disfavored, by Natural Selection. But, as we will see, Runaway Sexual Selection can lead to an interesting paradox. Kirkpatrick: Kirkpatrick showed (using a haploid model) that the male trait need not be initially favored by NS to generate linkage disequilibrium between female preference and the male trait. In fact, the male trait was initially selected against by NS. The model also showed that female preference would not increase, unless (1) D > 0, and (2) the male trait was increasing. Kirkpatrick s model also showed how drift and selection could interact to drive allopatric populations in different directions, with respect to female choice and the male trait. 1

2 New topic. How to choose a mate. Darwin considered two kinds of sexual selection (i.e., variation among individuals in mating success). 1. Intrasexual selection due to male-male combat (We won t cover this). And 2. Intersexual selection due to female choice But, does that mean that 1. Plants cannot have sexual selection? 2. Hermaphrodites cannot have sexual selection? See Delph and Ashman 2006 (posted on class web site) for a contrast of interactiondependent and interaction-independent sexual selection and the application to plants. 2

3 Why, in general is there female choice (and sometimes male choice)? 1. whims sensory exploitation This is assumed in Kirkpatrick s model 2. Males provide resources (nuptial gifts) or paternal care. 3. Males are selected for their good genes. a. Incest avoidance. b. genetic advantages to offspring. c. favorable combinations of paternal genes with maternal genes. Problems for the good-genes model. a. Natural selection and sexual selection for good genes erodes the additive genetic variance for the male trait. Hence, at some point in the future, there could be no effect of female choice. Why then be choosy? Possible solution to the problem (Hamilton and Zuk 1982). Conditiondependent choice for high-fitness males. a. The favored combination of genes (epistasis) changes over time. b. Good combinations of genes results in good condition c. Thus, condition may be a reliable indicator of good genes in the male. 3

4 But what causes the favored combination of genes to change over time? à Hamilton and Zuk (1982, Science) proposed that host-parasite coevolution could cause the favored combinations to change over time (The Red Queen again). As such, the variation underlying the male trait would be maintained by frequency-dependent selection. Question: Given that host-parasite coevolution is taking place, would female choice for good condition in males (Condition-Dependent (CD) choice) increase when rare, and go to fixation, in a population of randomly mating (RM) females? à An alternative: opposites attract (OA). What if instead of conditiondependent choice, the females could directly select males that had different genotypes from themselves, so as to increase the genetic diversity among their offspring? Would an OA female preference gene increase when rare in a population of random maters (RM)? Would an OA gene go to fixation? Would an OA female gene increase when rare in a population of females practicing condition-dependent (CD) choice of males? Would an OA gene go to fixation? 4

5 Introduction of Condition-Dependent (CD) or Opposites Attract (OA) alleles into Randomly Mating (RM) populations. Frequency Linkage Disequilibrium a c Condition-dependent choice Host genotype 11 choice allele b d Opposites attract Host genotype 11 choice allele Both forms of mate choice increase when rare in a RM population.!! CD choice goes to fixation, and increases the amplitude of oscillations.!! OA choice reduces the amplitude of oscillations. " CD choice increases linkage disequilibrium." " OA choice reduces linkage disequilibrium. " Generations Remember: linkage disequilbrium refers to a statistical association between alleles at two different loci. Here the loci are disease-resistance loci. Q: why does OA not go to fixation? Source: J. Evol. Biol. 16:

6 Top Left. OA increases against RM, but does not go to fixation. Two kinds of choice could coexist. Top Right. CD increases when rare against RM, and goes to fixation. Left. OA increases when rare against CD, but does not go to fixation. Two kinds of choice could coexist. 6 Source: BMC Evol. Biol. 2004, 4:48

7 Evidence for Opposites Attract 1. Mice and MHC. 2. Fish and MHC. 3. Humans and MHC: the T-shirt study (Wederkind et al Posted on web site.) But see this update from Fabio (class of 2013): for humans: and, for salmon see: Methods for the T-shirt study 1. Male (49) and female (44) Swiss college students typed for MHC genotype. 2. Men wore standard T-shirts for two nights (no showers, no wine, etc.) 3. Women were given T-shirts to smell in a small, unmarked box. 4. Women were asked to rate the odors for three men with similar MHC, as well as three men with dissimilar MHC. A score was given for pleasantness and intensity. Results: Odors of men having dissimilar MHC genotypes were preferred by women; and these odors were more likely to remind them of a boyfriend, at least by women not taking oral contraceptives. 7

8 Genetic correlation, indirect selection, sexual selection, and selfish genetic elements Featuring stalk-eyed flies Males compete for position on root hairs (nighttime roosts). Male with the longer eyestalk usually wins. Females mate with territory holders. Neither males nor females care for the offspring. 8

9 Genetic correlation. 1. Graphical version showing a genetic correlation (r) between the family means for daughters preference plotted against the mean for sons traits, for Stalk eyed flies. Correlation: r = cov(y,x) var(y) var(x) Note the difference between correlation and the slope of a regression line β y,x = cov(y,x) var(x) 9

10 2. Using what we know from Quantitative Genetics, selection on the father s trait should produce a response to selection in the Daughter preference, as well as in the male trait in sons. How to test? Truncation selection on father s eyestalk should produce a response to selection in daughter s preference. 10

11 Thirteen-generation artificial selection experiment with Stalk-Eyed flies (Wilkinson and Reillo 1994). Prediction: selection on male eye-span length will produce a correlated response in the preference by females. 1. Control line: 10 males and 25 females picked at random every generation. 2. Long line: 10 males with longest eye span (relative to body size) picked each generation, along with 25 randomly selected females. **male and female eyespans increased. 3. Short line: 10 males with shortest eye span (relative to body size) picked each generation, along with 25 randomly selected females. **male and female eyespans decreased. The experiment ran for 13 generations, and then female choice was determined. Specifically, at generation 13, females were given a choice between two males, matched for body size, where one had a long-eye stalk and one had a short eyestalk). 11

12 Results. 1. Females from the Short line preferred to mate with males with short eyestalks. Why? 2. Females from the Control line preferred to mate with males with long eyestalks. What does that mean? 3. Females from the Long line preferred to mate with males with long eyestalks. But they did not prefer males with long eyestalks more than females from the control lines. Why? (ps: no evidence that long eye span was associated with good genes for development) 12

13 Curious result Stalk-eyed flies are female-biased in nature. But, the female bias became less extreme in the lines where males were selected for long stalks (i.e., sex ratio became closer to 1:1). And the sex ratio became more female-biased in the lines were males were selected for short eyestalks. Why? A surprising twist. Suppose we had two kinds of males: X d Y males and XY males, where the d refers to driving. At meiosis, the driving X chromosome, X d, kills Y-bearing sperm. This kind of drive is called meiotic drive. Normal X chromosomes do not kill the Y-bearing sperm. What happens to X d? 13

14 And what does this have to do with the stalk-eyed flies? Remember that the offspring of the lines selected for short-eye stalks had female-biased sex ratios. Perhaps short eyes are genetically associated (in linkage disequilibrium) with the driving X chromosome, X d. Why do females prefer long-stalk males in Nature? It would now appear that either: 1. Males with long eye-stalks are more likely to have a normal X (non driving) chromosome. Or, 2. that males with long eyestalks are more likely to have Y chromosomes that prevent the action of the driving X (Wilkinson et al. 1998). (Best supported at present?) 14

15 More on selfish DNA? Cytoplasmic Male Sterility (CMS) in plants (simplified). CMS is a mitochondrial gene that inhibits pollen production in hermaphroditic plants. (see page 164 in Burt and Trivers: Genes in conflict.) This was a chalk talk with hand-drawn figures. The main points were 1. CMS is a mitochondrial gene that inhibits pollen production. Resources are shifted to making ovules, which increases the transmission probability for cytoplasmic genes in general. CMS spreads when rare. 2. The population becomes female-allocation biased, but the CMS does not go to fixation. (show graph of CMS spread when rare.) 3. The female bias creates selection that favors allocation to pollen. This could result in increased allocation to male function in male-fertile hermaphrodites, and could lead eventually to the evolution of separate sexes, where female is determined by the CMS mutation. 4. But there is also selection favoring nuclear genes that restore male function in CMS plants. Basically, CMS causes a departure from the ESS sex ratio. Nuclear genes that move non-cms plants closer to the ESS sex ration are favored by selection. 15

16 Chase-away sexual selection (see page 516 in text, 4 th ed. Freeman&Heron). 1. A new male trait arises that increases the mating success of the males, but reduces the survivorship of females. 2. The male trait spreads in the population. 3. A female-expressed mutation arises that allows resistance to manipulation by the male trait. Resistant females become more common; males with the novel trait become less numerous. 4. A new mutation arises that is expressed in males, which again increases the mating success of males, with a cost to females. 5. The cycle continues. Genes expressed in males and genes expressed in females are engaged in a kind of antagonistic coevolution. Genomic conflict. Design an experiment But first, Bill Rice s experiment. 16

17 Rice s experiment. (see Rice s 1996 paper in Nature on web site.) 1. Experimental group. Males compete with each other for females. Females mate with multiple males. Sons were taken from this group to start the next generation. But the females for the next generation were taken from non responding stock population (meaning the females could not coevolve with the males). Males could evolve, females could not. 2. Control group. No competition for mates. 17

18 3. Results: males from experimental lines had higher fitness than control males. They were also more likely to re-mate with the same female, and they were more successful at defense (defense was measured as the relative number of offspring sired after the female was allowed to remate with a different male). 4. Results. Females mated to males from the experimental lines had significantly higher mortality rates, which apparently resulted from the toxic seminal fluid produced by males from the experimental lines. Design a new experiment.. 18

19 Remember parent-offspring conflict? Now for parent-parent conflict More on conflict between male and female parents: Zeh and Zeh Live birth (viviparity) offers the opportunity for conflict that is not possible in spawning species or egg laying species. Specifically, the womb (or similar structure) provides an arena for competing interests to play out. There may be a difference, for example, of the effects of natural selection on males and females. In the mother, selection should favor equal allocation to offspring within a brood, as they are all equally related to her. In addition, for interoparous species (species reproduce more than once), the mother may hold back resources for future broods. How does selection act on males in polyandrous species (where females mate with more than one male)? Males would be selected to manipulate the mother into investing more in their (the dad s) offspring, at the expense of the progeny sired by other males and at the expense of the mother s future reproductive success. How could they do this? 19

20 In mice embryos, the maternal allele for IGF2 is mostly turned off. IGF2 is insulin-like growth factor 2, which stimulates cell development, increases the growth rate of embryos, and effectively garners resources from the placenta. Why should a mother imprint her IGF2 alleles to reduce transcription of this gene in her offspring? Does she want to starve her offspring? In the same embryos, the paternal copy of the allele IGF2 is turned on. Why? In addition, the mother expresses another gene, IGF2-r, which codes for a receptor that mops up extra growth factor. Conversely, the male copy of the same gene is not expressed. Why? Who wins? 20

21 lgf2 on in sperm, off in eggs. lgf2r on in eggs, off in sperm. Top: transfer activity Bottom: suppression activity Figure 2. Left, Igf2 (insulin-like growth factor type 2) and Igf2r (insulin-like growth factor type 2 receptor) are reciprocally imprinted genes in eutherian mammals. In the developing embryo, gene expression occurs only from the paternally inherited (red) copy of the Igf2 locus and only from the maternally inherited (blue) copy of the Igf2r locus. The two loci function in opposition: insulin-like growth factor type 2 promotes resource transfer from the mother to the embryo but is degraded by the product of the Igf2r locus. With a number of loci contributing to the stimulation of resource transfer by the paternal genome and suppression by the maternal genome, maternal and paternal genomes are likely to vary in the extent to which they are genetically compatible (right). Since multiple paternity drives the evolution of more aggressive paternal genomes, such incompatibility is predicted to be greatest in crosses between populations which differ in level of polyandry. From Zeh&Zeh Bioessays paper (2000). 21

22 Figure 2. Functionally distinct roles played by maternally and paternally inherited genomes in mammalian development. Nuclear transplantation experiments have revealed that early development of the embryo proper is controlled largely by expression of the maternal genome, whereas paternally inherited genes control growth primarily of the trophoblast and placenta (McGrath & Solter 1984; Surani et al. 1984). Naturally formed conceptuses with two paternally inherited genomes and no maternal genome are known as complete hydatidiform moles (CHMs). CHMs closely resemble the androgenic embryos produced in the nuclear transplantation experiments. (From Zeh and Zeh 2008) 22

23 Additional points from the Zeh s paper (read for next time.) 1. Mammals develop post-zygotic isolating barriers much faster that frogs and birds. Why? 2. Viviparity creates a post-fertilization arena, allowing for the possibility of reallocation of resources within the arena, which indirectly aid the spread of cytoplasmic male killers. Why? 3. Polyandry drives conflict, which might drive greater degrees of polyandry, possibly resulting is a positive feedback situation. 23