Socio-reproductive Conflicts and the Father’s Curse Dilemma*

Evolutionary conflicts between males and females can manifest over sexually antagonistic interactions at loci or over sexually antagonistic interests within a locus. The latter form of conflict, intralocus sexual conflict, arises from sexually antagonistic selection and constrains the fitness of individuals through a phenotypic compromise. These conflicts, and socio-reproductive interactions in general, are commonly mediated by hormones, and thus predictive insights can be gained from studying their mediating effects. Here, we integrate several lines of evidence to describe a novel, hormonally mediated reproductive dilemma that we call the father’s curse, which results from an intralocus conflict between mating and parental efforts. Essentially, a genetic locus exerts pleiotropic and antagonistic effects on the mating effort of one individual and the parental effort of a related individual who is the primary provider of parental care. We outline the criteria for operation of the father’s curse dilemma, provide evidence of the phenomenon, and discuss the predictions and outcomes arising from its dynamics. By integrating the effects of hormones into socio-reproductive conflicts and socio-reproductive effort, clearer links between genotypes, phenotypes, and fitness can be established.


Evolutionary Conflicts and the Father's Curse Dilemma
Evolutionary conflicts in reproduction arise due to divergences in evolutionary interests of loci depending on whether they are in males, females, parents, or offspring (Arnqvist and Rowe 2005;Bonduriansky and Chenoweth 2009;Mokkonen et al. 2016;Rowe et al. 2018;table 1). Such conflicts result in fitness trade-offs that can constrain adaptation to phenotypic optima during reproduction (Trivers 1974;Chippindale et al. 2001;Chapman et al. 2003;Arnqvist and Rowe 2005;Bonduriansky and Chenoweth 2009;Pennell and Morrow 2013). Prezygotic reproductive investment differs considerably between the sexes due to anisogamy and the resulting difference in reproductive opportunities for females and males. Consequently, males are typically the sex that is under stronger sexual selection: a biased operational sex ratio toward males gives rise to greater variance in reproductive success and greater opportunity for selection compared to females (Clutton-Brock 2007). Evolutionary conflict can arise between males and females when reproductive interests diverge to such an extent that sexually antagonistic selection acts on phenotypes with sex-specific fitness optima in females compared to males. Such sexual conflict can take the form of interlocus sexual conflict, whereby the conflict is between coevolving loci in females and males (Chapman et al. 2003;Arnqvist and Rowe 2005), or intralocus sexual conflict, whereby the conflict occurs within a locus that exerts sexually antagonistic effects on fitness (Pischedda and Chippindale 2006;Bonduriansky and Chenoweth 2009;Pennell and Morrow 2013). In addition to these prezygotic differences in reproductive investment between the sexes, postzygotic reproductive investment also differs between females and males due to factors such as greater paternity uncertainty experienced by fathers, a greater benefit or lower cost to the sex providing care, and the action of sexual selection (Trivers 1972;Clutton-Brock 1991;Andersson 1994;Arnqvist and Rowe 2005;McNamara and Wolf 2015). These asymmetries in pre-and postzygotic investment define the varying sex roles of females and males and indicate that there are many opportunities for conflict to arise during reproduction.
Mating and caring for offspring are reproductive components that trade off with regard to an individual's finite resources (Dawson 1996;Stiver and Alonzo 2009). Owing to divergent sex roles, males typically employ greater mating effort, while females expend more parental effort in reproduction (Queller 1997;Kokko and Jennions 2008;Stiver and Alonzo 2009). In both sexes, but more so in males, an individual that maximizes mating effort is expected to increase their fitness through this component of reproductive success: selection will primarily favor those traits that yield higher mating success and more offspring. Conversely, an individual that maximizes their parental effort is also expected to increase their fitness, especially through the component of offspring survival: selection will primarily favor those traits that produce higher-quality care of offspring during the postnatal period. Investments in mating effort and parental effort frequently result in parent-parent conflicts and parent-offspring conflicts, respectively, though clear demarcation between these various forms of conflict has become increasingly more challenging (Godfray and Parker 1992;Parker et al. 2002;Royle et al. 2004;Meunier and Kolliker 2012;Patten et al. 2014). For example, interlocus sexual conflict results when female reproductive success is constrained by assuming greater parental care, while male reproductive success is enhanced by obtaining more time for mating activities, thus extending the phenotypic realm of sexual conflict beyond premating social interactions (Arnqvist and Rowe 2005;Wedell et al. 2006;Aloise King et al. 2013;McNamara and Wolf 2015). When the same allele influences traits associated with both mating and parental efforts between individuals, an intralocus conflict that we call the father's curse dilemma may arise.
The father's curse dilemma is an evolutionary conflict between parental effort in females and mating effort in males that arises due to pleiotropy and sexually antagonistic selection on a shared locus ( fig. 1). We refer to this conflict as the father's curse dilemma because it results in a trade-off between mating effort of males that experience greater variance (skew) in reproductive success and parental effort of females that are the primary caregivers during the weaning period (Clutton-Brock 2007;Kokko and Jennions 2008;Aloise King et al. 2013; fig. 1). We emphasize here that although such a conflict between mating and parental efforts has deleterious fitness consequences for both sexes, males have the potential to suffer larger fitness costs due to their greater variance in reproductive success compared to females; given such asymmetry in reproductive success between the sexes, the number of reproducing males is by definition less than the number of reproducing females in a given population. The genomic conflict occurs within a locus pleiotropically linked to both mating and parental efforts, which can impact fitness at different life stages (table 1). Early-life survival of males (during postnatal care) is predicted to be influenced by indirect genetic effects of the mother whereby the mother exerts nongenetic maternal effects associated with parental care on the survival of the offspring. Consequently, juvenile males that have higher-quality parental care and survive the juvenile period are predicted to fare worse in reproductive competition as adults due to the action of the sexually antagonistic locus. Similarly, daughters of reproductively successful males are predicted to provide a lower level of parental care ( fig. 1). The father's curse dilemma thus relies on the action of a gene with antagonistic fitness effects to give rise to the evolution- ary conflict between sexes and ontogenetic stages, with the additional consideration of indirect effects of parental investment Rice and Chippindale 2001). Such conflicts are expected to be mediated, in particular, by the presence of genetically based hormone signaling systems that exert strong pleiotropic effects on mating and parental efforts due to the central role of hormones in the pre-and postzygotic phases of reproduction (Mokkonen and Crespi 2015;Mokkonen et al. 2016). This article details how fitness trade-offs between mating effort and parental effort that are characteristic of the father's curse dilemma can lead to evolutionary conflict ( fig. 1). We outline theoretical criteria for the father's curse dilemma; provide evidence for this dilemma primarily from mammalian taxa in the context of the testosterone, arginine vasopressin, and oxytocin hormone systems; and describe predictions as well as evolutionary outcomes that derive from this phenomenon.

A Sexually Antagonistic Locus with Hormonally Pleiotropic Effects on Mating and Parental Efforts
Through pleiotropy, a genetic locus can exert effects on multiple fitness-relevant traits (Williams 1957;McGuigan et al. 2011). While it is still unclear whether genetic variation associated with mating success is more or less pleiotropic compared to the rest of the genome, we can focus on biological systems with known pleiotropic effects on mating effort: hormone signaling systems (Fitzpatrick 2004;Mank et al. 2008;). Recent evidence indicates that testosterone (T), oxytocin (OXT), and arginine vasopressin The closely related neuropeptide AVP is primarily responsible for mediating prosocial behaviors and aggression, including social aggression, dominance behavior, male bonding, and anxiety (Young et al. 1999;Donaldson and Young 2008;McCall and Singer 2012). These hormones actively shape fitness through their effects on both reproductive traits and social interactions: when phenotypic optima differ between individuals over expression of, or internal receptor-driven responses to, a given hormone, conflict is expected in the hormonally mediated trait and potentially within the genome.  Figure 1: Antagonistic fitness effect of an allele on parental and maternal efforts between generations. A, An allele that benefits parental effort (PE-beneficial allele) has a corresponding negative effect on mating effort, while an allele that benefits mating effort (ME-beneficial allele) has a corresponding negative fitness effect on parental effort. B, The predicted negative fitness correlation between male mating effort and female parental effort. C, The sexually antagonistic allele through generations, exerting alternating fitness costs on male mating effort and female parental effort.
Research in avian and mammalian taxa has shown that T has opposing effects on mating and parental efforts within individuals (Clark and Galef 1999;McGlothlin et al. 2007;Mascaro et al. 2013;Roney and Gettler 2015). The pleiotropic effect of T manifests in primary and secondary sexual traits, including reproductive behaviors (Ketterson and Nolan 1999;John-Alder et al. 2009;Malo et al. 2009;Mills et al. 2009;Mokkonen et al. 2012;Rosvall et al. 2012). The resulting antagonistic effect of T on mating behavior in females compared to males underlines the direct importance and sex specificity of this hormone on mating effort (Mokkonen et al. 2012). However, given the social-context specificity of T, potential fitness benefits that it confers in mating effort may trade off with other fitness domains, primarily parental effort (Gray et al. 2002;Gray 2003;McGlothlin et al. 2007;Alvergne et al. 2009;Kuzawa et al. 2009;Muller et al. 2009; tables 2, 3). For example, an endocrine response of higher T in parents caring for offspring can result in lower parental attentiveness, greater aggression, and higher probability of deserting the offspring (Cavigelli and Pereira 2000;Mills et al. 2012;Rosvall 2013;Saltzman and Ziegler 2014). To potentially mitigate this cost of T, mammalian fathers in some species downregulate T during the critical period of postnatal care of offspring (Gray et al. 2002;Alvergne et al. 2009;Mascaro et al. 2013;Saltzman and Ziegler 2014). Thus, accounting for trade-offs between mating effort and parental effort will provide important insights into the fitness benefits and costs of sexually antagonistic traits mediated by T.

Uniparental Care of Offspring
Another requirement for the operation of the father's curse is that an individual belongs to a species in which uniparental care (e.g., maternal care) is the primary form of resource provisioning during the postnatal period. This asymmetry in parental effort is required for the evolutionary conflict between mating and parental efforts to be realized: lower parental effort by males allows them to invest more in mating effort relative to females. Among most taxonomic groups, biparental care is the exception, and males do not experience the fitness constraints and resource demands of parental care (Reynolds et al. 2002;Lukas and Clutton-Brock 2012). A phylogenetic comparison of vertebrate taxa has demonstrated that uniparental care is probably the ancestral mode (Reynolds et al. 2002). Most mammals (as well as other vertebrate taxa) adhere to this form of parental care, and as a result offspring survival is mainly determined by traits that optimize maternal care behaviors, while fathers influence offspring fitness only genetically ( fig. 2). Consequently, parental care likely evolves independently in females and males due to sex-specific expression and selection on associated traits (Walling et al. 2008;Bendesky et al. 2017). This presumed uncoupling of parental care behavior in females and males makes it less likely that selection on parental effort can reduce the fitness costs of the father's curse dilemma through correlated selection on the opposite parental sex; selecting for good-quality paternal care is unlikely to eliminate the fitness costs associated with the father's curse.

Context Specificity of the Selected Locus Determines Fitness Outcomes
As previously outlined, the father's curse dilemma is likely to manifest in a fitness trade-off between the parental effort of females and the mating success of males ( fig. 1). In practical terms, a female that provides high-quality parental care will produce sons that are relatively less competitive in reproduction. Similarly, a male with high reproductive success will produce daughters that provide a relatively lower quality of parental care. The conflict therefore exists between fathers and daughters, or mothers and sons. However, this observation does not necessarily exclude genomic conflicts between same-sex parent-offspring combinations if a locus can be shown to antagonistically impact parenting effort in one individual and mating effort in a related individual. While sexual conflicts are usually characterized as evolutionary conflicts between females and males, conflicts between parents and offspring are typically characterized as intergenerational conflicts between immediate relatives. However, in both cases, the locus or loci in conflict exert context-specific effects on fitness that depend on the genome or life stage in which the locus is expressed (e.g., female vs. male genome, juvenile vs. adult life stage, type of social interaction). Sexually antagonistic loci confer fitness benefits on one sex when found in its genome and fitness costs when located in the genome of the opposite sex Chapman et al. 2003;Arnqvist and Rowe 2005;Bonduriansky and Chenoweth 2009;. Similarly, imprinted genes exert conflicting effects on fitness in offspring due to the parent-of-origin manner of gene expression (Patten et al. 2013;Haig 2014). A maternally expressed (paternally silenced) locus such as IGF2R within the offspring may thus introduce similar conflictual dynamics whereby the maternal optimum (in this case, suppressed offspring growth in utero) is mismatched with the offspring optimum (greater offspring growth ;Haig 2004;Mokkonen and Crespi 2015;Saldivar Lemus et al. 2017). Thus, while we focus on sexually antagonistic loci in this article, it is conceivable that any locus that exerts such antagonistic effects could potentially result in father's curse-like dynamics.
In general, the optimal "social phenotype" for attaining a mate differs from the phenotype for optimal parental care. For example, intramale aggression and dominance behavior (promoted with higher T) may yield a higher probability of mating success, while evidence indicates that females (and to a lesser extent males in biparental taxa such as humans) experience greater OXT levels after parturition that facilitate parent-offspring bonding, parental attentiveness, emotional empathy, and physiological effects such as the promotion of lactation (McCall and Singer 2012; Ham-mock 2015; Crespi 2016; tables 2, 3). The affiliative effects of OXT are beneficial for both sexes during this period of parental care; however, such an affiliative effect on male mating behavior-particularly during the competition for mates-is predicted to be costly in terms of lower mating success since greater OXT levels in males are predicted to reduce competi- tive motivation and aggression (Dhakar et al. 2012;Calcagnoli et al. 2015). Thus, the context specificity of hormones can lead to trade-offs between mating effort and parental effort.
Most mammalian species meet the criteria for the father's curse dilemma, given that most of these systems feature overlapping generations, practice uniparental care by the mother (or mainly maternal care), share anciently conserved hormone signaling systems that are associated with sexually antagonistic traits, and involve sophisticated social interactions (Reynolds et al. 2002;Donaldson and Young 2008;Mokkonen and Crespi 2015). Next, we focus on one such species, the bank vole (Myodes glareolus), to illustrate how fitness-related hormones demonstrate the father's curse dilemma in this species.

Bank Vole System
Bank voles are broadly distributed throughout Europe, primarily inhabiting forests and fields (Hansson 1979;Bujalska and Hansson 2000). The densities of field populations have  Figure 2: Father's curse dilemma and associated life-history consequences. A, A male mating with low-T (or high-OXT) females will have lower (relative)-fitness sons that are provided more parental care, while a male mating with high-T (or low-OXT) females will have higher-fitness sons that are provided with less optimal parental care. In females (B), the mating dilemma is minimized since males do not generally provide parental care to offspring and provide only genetic benefits to offspring fitness; males will maximize mating effort and minimize parental effort. More competitive males (e.g., high T or low OXT) are predicted to produce daughters that provide less parental care. OXT p oxytocin; T p testosterone.
implications for fitness, as females are territorial and must contend with infanticidal threats during the breeding season Ylönen et al. 1997;Poikonen et al. 2008).
The mating system is polygynandrous, whereby males and females mate with multiple individuals (Shuster and Wade 2003;Mills et al. 2007Mills et al. , 2014Mokkonen et al. 2012). Furthermore, males establish dominance hierarchies in reproductive competition that are mediated by testosterone (Mills et al. 2009;Mokkonen et al. 2011;. While females, and not males, provide parental care to offspring, both sexes can increase their reproductive success by acquiring additional mates in a reproductive bout; females are mechanically induced to ovulate more with each additional mate (Clarke et al. 1970;Mokkonen et al. 2012;Mills et al. 2014).
Given the propensity of this species to mate with multiple partners, previous work has shown that up to 50% of field litters are sired by multiple males (Mills et al. 2007). However, the prevalence of multiple mating may be even higher, as another recent estimate of lab-based mating trials indicated that approximately 65% of females mated with multiple males, which supports other work that has found an appreciable level of postcopulatory competition in the form of sperm competition in this species (Lemaître et al. 2011Mokkonen et al. 2016). These studies indicate that the mating system of bank voles possesses great potential for evolutionary conflicts between mates, parents, and offspring, as well as siblings.

Testosterone and Oxytocin in Bank Voles
The father's curse dilemma is exemplified by two hormones, T and OXT ( fig. 2). For the case of T, male bank voles benefit from higher T levels during male-male competition and courtship, while for females, selection for lower T levels results in higher mating rates, which increases their reproductive success (Mills et al. 2009Mokkonen et al. 2012). In bank voles, selection for greater male behavioral dominance (in male-male competition) results in (high-T) sisters with reduced litter sizes as well as reduced postnatal maternal care, characterized by lower growth of sons during the period between birth and weaning (Mokkonen et al. 2011; fig. 3). However, as adults, these sons have significantly greater T levels, which is correlated with greater mating and reproductive success and contrasts with reduced reproductive success of sisters (Mokkonen et al. 2011). Thus, for a male, the dilemma is that females that have high-T alleles will provide less effective parental care (characterized by lower postnatal growth; fig. 3), which may result in fewer offspring surviving to adulthood, whereas those high-T alleles being transmitted to offspring will also result in surviving sons having greater future reproductive success. For the case of OXT, the situation is reversed: females with greater OXT levels are predicted to produce more offspring and provide better care to offspring that results in better survival during the postnatal period of care (Lonn et al. 2017). Males selected for higher OXT are predicted to fare worse in male-male competition and suffer reduced reproductive success due to the action of OXT in promoting affiliative behaviors (Mokkonen and Crespi 2015;Crespi 2016;Lonn et al. 2017). Evidence supporting these predictions is found in recent empirical work on the bank vole that has characterized a microsatellite in the promoter region of the oxytocin receptor locus (Oxtr), whereby the number of repeats in the promoter microsatellite directly influences the level of gene expression. These data revealed balancing selection acting on the divergent fitness optima for microsatellite length and thus expression level of the gene (Lonn et al. 2017). In terms of the individuals that survive and reproduce in the field, selection favored females with longer (and males with shorter) Offspring growth (g ± 1 SEM) 6.50 Figure 3: In bank voles, sexually antagonistic selection on male dominance behavior produces families in which mothers provision less during postnatal care and sons have higher relative testosterone (T) values. Females in the "higher male dominance and T" line had lower reproductive success, while females in the "lower male dominance and T" line had higher reproductive success (Mokkonen et al. 2011). Females were artificially selected under principles of sexually antagonistic selection: the brothers were selected for male dominance behavior in reproductive competition, which also resulted in dominant males having significantly greater T levels compared to subordinate males (Mokkonen et al. 2011). Sons (filled circles and solid lines) and daughters (open circles and dashed lines) had greater growth during postnatal parental care from "good" females in the line selected for subordinate males with lower T (generalized linear mixed model: line: F 1, 70 p 5:72, P p :019; sex: F 1, 70 p 6:69, P p :012, litter size: F 1, 70 p 0:66, P p :42). Offspring growth (20-dayold body mass 2 birth body mass) was the dependent variable, "line" and "sex" were fixed factors, the "litter size" was the covariate, and the "mother identity" was treated as a random effect. Further details on animal husbandry procedures and selection are described elsewhere (Mokkonen et al. 2011). Oxtr allele lengths. Thus, males benefited from lower expression of Oxtr and presumably less activity of OXT, in terms of their survival and reproductive success. Given how OXT mediates parent-offspring bonding and provisioning of resources through lactation, greater OXT production is predicted to result in better-quality parental care (Lee et al. 2009;McCall and Singer 2012). Thus, for a male, the dilemma is that females that have high-OXT alleles will provide better care to offspring that result in greater offspring recruitment but will also result in sons having reduced future reproductive success. For both of these hormones, their positive effect on maternal care (i.e., greater OXT, lower T) is predicted to trade off with their negative effect on the reproductive success of males.

Predictions, Fitness Outcomes, and Conclusions
While this father's curse dilemma has been outlined for the general case of hormones in mammals and specifically for bank voles, given the widespread prevalence of hormonally mediated sexual conflicts in mammals (Mokkonen and Crespi 2015;Mokkonen et al. 2016), this phenomenon can be generalized to other taxa and traits that experience trade-offs between mating and parental efforts. Essentially, any trait or locus that results in a fitness trade-off between mating and parental effort between related individuals would be subject to the father's curse dilemma. Natural selection optimizes traits for greater survival, which can constrain adaptation in traits that improve reproductive success. Conversely, sexual selection optimizes traits for greater reproductive success, often at the expense of survival (Kokko and Brooks 2003). Thus, we predict trade-offs between social traits that improve survival and reproductive traits that increase reproductive success through mating or parental effort.
The wealth of research in sexual-conflict theory indicates that sexually antagonistic traits and associated genetic variation should be ubiquitous within the genome and across taxa, though empirical data are still lagging behind the theory (Chapman et al. 2003;Arnqvist and Rowe 2005;Bonduriansky and Chenoweth 2009;van Doorn 2009;Pennell and Morrow 2013;Rice 2013;Rowe et al. 2018). Emerging challenges include identifying the genetic loci associated with sexually antagonistic phenotypes and assessing the prevalence of genetic pleiotropy associated with such sexually antagonistic genetic variation. Bioinformatic tools such as gene ontology assays may allow pleiotropy to be more efficiently characterized (Ashburner et al. 2000). Even if the identified sexually antagonistic loci are pleiotropic but do not impact parental effort per se, this information would provide a more comprehensive understanding of the fitness benefits and costs of the evolutionary conflict.
Genetic evidence of the father's curse is limited; however, focusing on the gene regulatory regions for T, OXT, and AVP hormone receptors may yield further insights into this evolutionary phenomenon. In mammals, the lengths of microsatellite sequences in the 5 0 regulatory region upstream of the coding region for these receptor genes correspond to the level of gene expression in a growing number of mammalian taxa (Hammock and Young 2005;Donaldson and Young 2013;Keane et al. 2014;Lonn et al. 2017). Mounting evidence indicates that these receptor genes-AR, Avpr1a, and Oxtr-experience sexually antagonistic selection and mediate mating and parental efforts (Summers and Crespi 2008;Lonn et al. 2017). However, challenges still abound: for example, the prevalence of regulatory region-associated microsatellites for the oxytocin receptor gene remains unresolved, as the Oxtr microsatellite has only recently been characterized (Lonn et al. 2017) and appears to be lacking in other studied mammalian species despite the known relevance of this regulatory region for gene expression (Inoue et al. 1994;Young et al. 1997). Data are also needed that relate hormone receptor microsatellite polymorphisms with the quality of parental care provided by mothers and how the reproductive fitness of male offspring is associated with this relationship (taking into account potential confounding maternal effects). The mediating effects of the hormones T and OXT present an evolutionary dilemma between mating and parental efforts, and we have used these hormone systems here as examples of the father's curse. While we have focused on these examples of loci in hormonal pleiotropy, it is conceivable that other pleiotropic loci that exhibit sex-differential effects on mating and parental efforts would be candidate genes for participation in a father's-curse-dilemma form of conflict. Nonetheless, the evolutionary conflict characterized by the father's curse dilemma supports the perspective that relationships between kin need to be further incorporated to better understand the nature of evolutionary conflicts (Haig 2014;Mokkonen et al. 2016;Faria et al. 2017).
During reproduction, fitness constraints manifest in sexual conflicts between mates and conflicts between parent and offspring because of the differences in maternal/paternal and maternal/progeny interests, respectively (Parker et al. 2002;Mokkonen et al. 2016;Rowe et al. 2018). An implication of the close relatedness between parent and offspring is the fact that the individuals in conflict share a high proportion of genes (e.g., 50% of autosomes and the x chromosome shared between mothers and sons), which differentiates this type of evolutionary conflict from interlocus sexual conflicts. Any evolutionary modification of a parental trait that negatively impacts the survival of offspring will therefore indirectly harm the fitness of the parent as well. Thus, we predict that the fitness trade-off between male mating effort and female parental effort arising from the father's curse dilemma will be akin to fitness trade-offs arising in parent-offspring conflicts and (cross-generational) intralocus sexual conflicts, whereby the survival and recruitment of the offspring influ-ences the parent's fitness. Nonetheless, in the father's curse dilemma, while the father exerts a deleterious direct genetic effect on the future parental effort of daughters, the mother exerts two different effects on the fitness of sons: a beneficial indirect genetic effect mediated by maternal investment and a deleterious direct genetic effect on future reproductive success of sons. Thus, a never-ending fitness constraint in reproduction is borne by males.