THE YEAR IN EVOLUTIONARY BIOLOGY 2009 What Can Asexual Lineage Age Tell Us about the Maintenance of Sex? Maurine Neiman,a Stephanie Meirmans,b and Patrick G. Meirmansc aDepartment of Biology and the Roy J. Carver Center for Comparative Genomics, University of Iowa, Iowa City, Iowa, USA bCentre for the Study of the Sciences and the Humanities, University of Bergen, Bergen, Norway cLaboratoire d’Ecologie Alpine, Universite´ Joseph Fourier, Grenoble, France Sexual reproduction is both extremely costly and extremely common relative to asex- uality, indicating that it must confer profound benefits. This in turn points to major disadvantages of asexual reproduction, which is usually given as an explanation for why almost all asexual lineages are apparently quite short-lived. However, a growing body of evidence suggests that some asexual lineages are actually quite old. Insight into why sex is so common may come from understanding why asexual lineages persist in some places or taxa but not others. Here, we review the distribution of asexual lineage ages estimated from a diverse array of taxa, and we discuss our results in light of the main mutational and environmental hypotheses for sex. Along with strengthening the case for wide variation in asexual lineage age and the existence of many old asexual taxa, we also found that the distribution of asexual lineage age estimates follows a surprisingly regular distribution, to the extent that asexual taxa viewed as “scandalously” ancient merely fall on the high end of this distribution. We interpret this result to mean that similar mechanisms may determine asexual lineage age across eukaryotic taxa. We also derive some qualitative predictions for asexual lineage age under different theories for sex and discuss empirical evidence for these predictions. Ultimately, we were limited in the extent to which we could use these data to make inferences about the mainte- nance of sex by the absence of both clear theoretical expectations and estimates of key parameters. Key words: sex; asexual; parthenogenetic; ancient asexual; lineage age; scandal The phylogenetic distribution of asexual line- not a successful long-term strategy relative to ages is traditionally considered to be “twiggy,” sex (Maynard Smith 1978; Lynch & Gabriel meaning that asexual lineages are short-lived 1983; Hurst et al. 1992; Normark & Lanteri relative to sexual lineages (Weismann 1889; 1998; Burt 2000; Rice 2002). Williams 1975, pp. 162–167; Maynard Smith The ubiquity of sexual reproduction runs 1978, pp. 51–54, 1986; Bell 1982; Avise 1994; counter to the notion that sex faces substantial Normark & Lanteri 1998; Burt 2000; Rice and immediate costs relative to asexual repro- 2002; Simon et al. 2002; Schurko & Logsdon duction (Maynard Smith 1971, 1978; Williams 2008). This pattern implies that asexuality is 1975). These costs are so profound that un- derstanding why sex is so common has been termed the “queen of questions” (Bell 1982) in evolutionary biology and has been the fo- Address for correspondence: Maurine Neiman, Department of Biology, cusofalargebodyoftheoreticalandem- University of Iowa, Iowa City, IA 52242. Voice: 319-384-1814; fax: 319- 335-1069. [email protected] pirical research. However, despite decades of The Year in Evolutionary Biology 2009: Ann. N.Y. Acad. Sci. 1168: 185–200 (2009). doi: 10.1111/j.1749-6632.2009.04572.x c 2009 New York Academy of Sciences. 185 186 Annals of the New York Academy of Sciences study and a great deal of attention, this ques- lineate patterns, and discuss their implications. tion remains largely unanswered (Butlin 2002; A main goal of our review is to present the Normark et al. 2003; de Visser & Elena 2007; case that the consideration of the biology, ecol- Hadany & Comeron 2008). ogy, and phylogeography of asexual lineages of all ages is an integral component of a com- prehensive evaluation of the advantages of sex. Insights from Studying Asexual We also consider whether merely “old” asexu- Lineage Age als can be distinguished from exceptionally old (i.e., ancient) asexuals and whether the distri- The first direct challenge of the assumption bution of asexual lineage ages that we charac- that asexual lineages almost never persist came terize changes assumptions about the phyloge- from two different studies presenting mitochon- netic distribution of asexuality. We discuss all of drial sequence–based evidence for the exis- this in light of the main classes of hypotheses ∼ tence of 5,000,000-year-old asexual salaman- for sex. der lineages (Hedges et al. 1992; Spolsky et al. 1992). These findings raised awareness that asexual lineages may not be the dead-ends they Review of Asexual Lineage were previously assumed to be, and motivated a Age Estimates flurry of similar studies. A few years later, grow- ing evidence for a diverse array of “ancient Methods asexual” lineages was described as a “power- ful challenge to all theories of sex” (Judson & To determine how asexual lineage ages are Normark 1996). Understanding why and how distributed within and across taxa, we reviewed such “evolutionary scandals” (Maynard Smith the scientific literature for studies reporting es- 1978) persist can help to explain why most or- timates of asexual lineage age in eukaryotic ganisms are sexual (Judson & Normark 1996). taxa containing obligately asexual forms. We This perspective has been a primary mo- included some asexual taxa where asexual line- tivation for the considerable recent efforts to ages have been documented but their age has understand the exceptional status of a few an- not been estimated (e.g., the weevil Aramigus cient asexual species, such as the bdelloid ro- and all included plant taxa), with the goal of tifers (e.g., Gladyshev & Meselson 2008; Mark emphasizing that asexual age estimates are still Welch et al. 2008) and darwinulid ostracods needed for many taxa. When more than one (e.g., Schon¨ et al. 1998, Van Doninck et al. estimate of asexual lineage age was published 2003). Considerably less attention has been de- for a given taxon, we used the most recent one. voted to characterizing and understanding the We also determined whether there were sin- body of asexual lineage age estimates that are gle versus multiple origins of asexual lineages now available from dozens of other taxa. This within each taxon, because the rate of asexual evidence, however, could help to solve the prob- lineage origin is likely to be a major determi- lem of sex. As Butlin (2002) pointed out, insights nant of the extent to which sex faces challenges into the mechanisms maintaining sex can come from asexual invaders (Lively & Howard 1994; from comparing empirical estimates of asexual Burt 2000). Finally, to generate a quantitative, lineage age distribution to the theoretical ex- visual depiction of asexual lineage age distri- pectations under different models for sex. A bution across taxa, we plotted the cumulative comprehensive review of these data has never number of taxa in which asexual lineage age been conducted with this in mind. has been estimated versus the logarithm of the Here, we review the body of existing data maximum asexual lineage age most recently on the distribution of asexual lineage ages, de- reported for a given taxon. Neiman et al.: Asexual Lineage Age and the Maintenance of Sex 187 There is an active debate about how to deter- Results and Interpretation mine whether a putatively asexual lineage is ac- tually asexual (Hurst et al. 1992; Judson & Nor- Our survey indicates that the common as- mark 1996; Lunt 2008; Schurko et al. 2009). sumption that asexual taxa are almost al- One issue that comes up repeatedly is that ways short-lived is frequently violated (Table 1, nearly all asexual lineage age estimates (and Fig. 1; see also Butlin 2002 and Normark et al. determination of asexuality itself) rely upon 2003). For example, more than half of the taxa negative evidence, such as failure to find males (56%) were represented by asexual lineages es- or to detect a recent sexual ancestor (Judson & timated to be >500,000 years old. Not surpris- Normark 1996; Little & Hebert 1996; Normark ingly, the famous “evolutionary scandals,” such et al. 2003; Schurko et al. 2009). This means as the bdelloids, are among the oldest asexual that age estimates are subject to downward re- lineages reported. vision as long as there is, for example, poten- Even so, asexual lineages that have been her- tial for an undiscovered close sexual relative alded as being of exceptional antiquity do not (Robertson et al. 2006), functional males (e.g., appear exceptional when considered against Smith et al. 2006), or cryptic sex (Mikheyev the background of asexual lineage ages esti- et al. 2006; Cooper et al. 2007; Thompson mated from a diverse array of animal taxa. In- et al. 2008). Given that this debate remains un- stead, we found that the relationship between resolved, we simply presented asexual lineage the cumulative number of taxa in which asex- ages as currently estimated. ual lineage age has been estimated and the loga- Also, there exist confounding factors that rithm of the maximum asexual lineage age most could influence our review. For one, there is al- recently reported for a given taxon is quite reg- most certainly a publication bias toward papers ular and nearly linear (Fig. 1). This result has at reporting evidence for ancient asexual lineages, least two interesting implications: (1) there is no because “young” asexual lineages are merely obvious point of demarcation between young behaving as predicted. Another certain source and “ancient” asexuals, with the consequence of bias is taxonomy: for example, no reliable that distinguishing potentially exceptional estimates of asexual lineage age are available (and potentially illuminating) ancient line- forplantseventhoughasexualityiscommonin ages from merely “old” asexual lineages is diffi- plant taxa. This is probably due at least in part cult, and (2) the most parsimonious explanation to the complexity of plant reproductive systems of this pattern of asexual lineage distribution and the perceived lack of suitable molecular is that similar types of mechanisms determine tools.