1. Introduction

Inbreeding depression, the loss of fitness resulting from homozygosity across the genome, can occur within otherwise outbred populations as a result of matings between relatives. It is caused by the accumulation and expression of deleterious alleles or the loss of advantageous heterozygosity, although the relative contributions of these factors are not well understood [1]. Inbreeding depression in captive animals is well documented [2], and there are a growing number of examples from natural populations [3,4]. Many vertebrates actively disperse from their natal sites, and manipulative studies suggest that this behaviour may be driven by benefits of reduced inbreeding [3]. Dispersal can relax selection for active inbreeding avoidance behaviour [5]). In species with dispersal strategies, the risk of inbreeding may be too low to drive the evolution of sibling recognition mechanisms [6]. In some species, individuals not only distinguish between kin and non-kin when choosing mates, but may even discriminate among close kin and distant kin [7]. Grosberg and Hart [8] reported that an allorecognition polymorphic gene may regulate the reproductive interaction between mates or between gametes to avoid inbreeding. Overall, the available evidence suggests that costs of inbreeding represent a significant selection pressure across a broad range of taxa [9–11].

Inbreeding avoidance has been proposed as a potential explanation for the prevalence of polyandry in groups where multiple matings do not provide direct benefits to females [12,13] and inbreeding costs are high [14]. Although no evidence for such a process was found in shrews, for which the hypothesis was originally proposed [15], evidence from 3 species of shorebirds [16] and from blue tits (Parus caeruleus) [17] indicates that females utilise extra-pair copulations to reduce inbreeding. Furthermore, studies of guppies and 2 species of field crickets indicate that multiply mating females can avoid costs of inbreeding through increased fertilisation success of unrelated males [18–21], suggesting it would be worth looking for similar effects in other species [22–25].

Studies in Drosophila melanogaster failed to find any evidence for potential pre-or post-copulatory inbreeding avoidance [26,27]. Evidence for inbreeding avoidance in invertebrates in general is limited, although this may reflect less attention being paid to the phenomena in comparison to studies of vertebrates [3,4]. de Boer et al. [23] found inbreeding avoidance in a handful of laboratory species (e.g., [28–32]) after identifying 139 experimental studies. It is most likely to be found in species whose ecology leads to high risks of encountering closely related individuals. Such species will include those with strong population structure and philopatry, and also those that are frequently exposed to founder events, e.g., [6]. Stored product pests represent such a group since they exist on resources that are often ephemeral and they rely on their ability to locate and colonise new food patches, e.g., [33]. Such colonisation events may frequently involve a small number of individuals arriving at a new patch, with concomitant risks of mating amongst relatives.

The red flour beetle, Tribolium castaneum is a stored product pest found in flour stores throughout the temperate and tropical zone. It has been used as an experimental model system across many fields of ecology and evolution [34,35]. Populations have been shown to experience inbreeding depression, suffering reduced egg laying rate and hatchability [36], sperm length, testes volume, reproductive output and lifespan [37], with variation between populations in the extent of inbreeding depression [37]. In T. castaneum, it has been shown that inbreeding increased levels of female promiscuity compared with noninbred cohorts [38]. T. castaneum is known to exercise mate choice using scent cues to distinguish between mates on the basis of female maturity and previous mating history [39,40] so there is potential for olfactory kin recognition. More attractive males are known to have higher P₂ (P₂ being the proportion of eggs fertilised by the 2^nd male to mate with a female [41]). The fertilisation success of male T. castaneum was not depressed under strong inbreeding, but inbreeding affected the outcome of sperm competitiveness [42]. Whether copulations result in complete failure of fertilization was not influenced by the relatedness of mating pairs [43].

In this study we examine costs of inbreeding in T. castaneum and investigate whether adults can distinguish between related and unrelated potential mates, whether they discriminate against relatives when choosing to mate and whether females can bias paternity towards unrelated mates when mated to both a full sibling and an unrelated male.

2. Methods

The lifecycle of T. castaneum consists of eggs laid in flour or similar substrates that hatch in 3–5 days into larvae that pass through multiple instars before pupating and emerging as long‑lived adults capable of reproducing for over a year [33]. Experimental beetles were provided by Dr R. Beeman, US Grain Marketing and Production Research Centre, and maintained in the laboratory under identical standard rearing conditions and sexed and isolated as pupae into single-sex groups [33]. Pupae were kept individually in a 2 cm x 2 cm cell of a 5x5 cell plastic box with a mixture of 95% organic flour and 5% dried yeast in a dark incubator under standard rearing conditions of 30^ºC and approximately 67% relative humidity. Pupal isolation was crucial as males can engage in same sex behaviours and become sperm limited [44]. Previous work on T. castaneum suggested that mate preference for strain was affected by whether or not adults were kept in isolation or with members of their own strain [35]. However, this difference has minimal statistical support (chi-square test of independence of lines (a) and (b) in Table VI in [35], χ² = 5.22 p = 0.16. Our experiment is motivated by the possibility that in the wild individuals may be able to identify they siblings either before or after mating and discriminate against them. In this context we would not expect adult beetles, to be co-located with a group of exclusively siblings (or they would not encounter both siblings and non-siblings and potentially exercise mate choice). For this reason, we kept adults in isolation rather than in groups of full siblings. The eclosion date of each beetle was recorded to minimise age effects during mating trials (see, online supplementary material). All beetles used in this experiment were between 9 and 21 days post-eclosion, and the maximum age difference between any beetles in a trial was 2 days. Two strains were used, the wild type GA1 strain (Georgia 1, with filiform, wild type antenna) and the reindeer strain (Rd strain, with swollen antenna, see [45,46], which carries a visible dominant marker that causes the antennae to have a pronounced terminal club which is easily distinguished phenotype from the wild type antenna [45]. This strain has been used in numerous previous studies of sexual selection in Tribolium [45,47] and the antennal deformity does not appear to have a measurable effect on behaviour or mate discrimination, e.g., [46]. For the pre-mating inbreeding avoidance assays, just the GA1 strain was employed, for subsequent assays where we needed to identify the offspring of non-sib males, we employed both GA1 and Rd strains.

2.2. Post-mating inbreeding avoidance

We used a similar design to that used to investigate post-mating inbreeding avoidance in field crickets [20]. As in Tregenza and Wedell’s study [20] we had four treatments: a female mating to either two full siblings, two non-siblings or a sibling and a non-sibling in either order (see Fig 2). However, to assign parentage to offspring, in our design all females and their full sibling males were of the GA1 strain, whereas unrelated (non-sib) males were from the Rd strain. We chose these two strains as previous studies have found no difference between them in their reproductive output [50], also see [45]. A preliminary study (data from [50]) of the number of offspring produced by ten GA1 females mated with either a sib or non-sib GA1 male, or an Rd strain male revealed that sib crosses produced significantly fewer offspring than crosses involving Rd males (estimate: −0.39, z = −3.4, P < 0.0001, Fig 1 panel D). Non-sib GA1 male and Rd male crosses with GA1 females produced similar numbers of offspring (estimate: −0.16, z = −1.25, P = 0.209, Fig 1 panel D). Our design has the limitation that it confounds non-sib with ‘non-strain’; i.e.,. the non-sib crosses are an Rd male and a GA1 female whereas the sib crosses are a GA1 male and a GA1 female. This means that if we find a difference between the post-mating reproductive success of sibs and non-sibs, we need to be aware that this difference could arise if there are interactions between strains that affect fertilisation probability or offspring viability (even though the available evidence suggests that such differences do not exist).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Pre-mating (“Male.Choice”) and female (“Fem.Choice”) behavioural traits when interacting with either a full-sibling or a non-sib mate.

Panels A − D represent show raw data (full-siblings [squares] and non-sibs [circles]) as jittered points; mean ± SE are shown as a square or circle plus error bars. Panel A: contacts initiated by males (“Male.Choice”) and females (“Fem.Choice”) towards full-sib or non-sib mates; Panel B: mating attempts (counts) towards either a full-sib or non-sib mates; Panel C: Mating duration when the mate is a full-sib versus non-sib across both sets of assays; male choice and female choice trials. Panel D quantifies offspring produced (mean ± SE) by full-sibling GA1 pairs versus when paired with an unrelated GA1 and Rd non-sib males with a filled large central point indicating treatment average (shapes indicate full-siblings [squares] and non-sibs [circles, diamonds]).

https://doi.org/10.1371/journal.pone.0335301.g001

Twenty-five full-sib families were reared and used in forty-two experimental blocks, with no family used in more than 2 blocks. Each block consisted of four GA1 females each mated to two males in one of the following combinations: either to two full-sib GA1 males (SS); two non-sib Rd males (NN); a sib GA1 male followed by a non-sib Rd male (SN); or to a non-sib Rd male followed by a sib GA1 male (NS). Within a block, all ‘GA1’ and ‘Rd’ males were full-sibs and of similar age groups.

In each mating trial, the male and female were placed together in the same chamber as described above for the pre-mating study and observed until a mating lasting more than 40 seconds took place. As soon as the pair separated, the male was removed and the female was left individually in the cell for 24 hours. Females are reluctant to lay eggs in the absence of a suitable substrate, and we did not observe any egg laying over this period. After 24 hours the second male was placed with the female, and the pair were observed until they mated for at least 40s. All females were virgin at the start of the experiment and mated with both males. All males were used only once and were mated once to a stock female on the day before they were used in the experiment to check that they displayed normal mating behaviour. When the female’s second mating finished, she was kept in a 100 ml pot containing 40 ml of flour in a dark incubator at 30°C and 67% RH. After one week, the female was removed and the pot returned to the incubator under the same climate regime as above. The number of adult offspring were counted 45 days after the last day of oviposition, which was before any eggs laid by the first emerging offspring could themselves reach adulthood. Additionally, in 17 randomly chosen blocks, females were kept for 5 further weeks (we only did this with a subset because of limitations on time). Their offspring were collected and counted each week by moving the female to a new pot every week for 5 weeks to calculate P₁ (offspring fathered by the first male) and P₂ (offspring fathered by second the male). This design allowed us to examine any changes over time in P₁ and P₂.

3. Results

3.2. Post-mating inbreeding avoidance

According to our earlier study [50], there fewer adult offspring were produced by GA1 females mated to GA1 full-sibling males compared to GA1 females mated to Rd non-sibling males (mean ± S.E.) number of offspring = 70.64 ± 4.6 and 104.5 ± 6.9 for GA1 sibling and Rd males respectively z = −3.4, P < 0.0001, Fig 1 panel D). This indicates that GA1 females do suffer costs of mating (lower offspring viability when they mate with related males as opposed to unrelated GA1 males or non-strain Rd males.

The difference in fecundity of females in the 4 treatments (sib then sib (SS), non-sib then non-sib (NN), non-sib then sib (NS) is shown in Fig 2, panel A. These data include females that did not receive any sperm despite their two copulations and produced no offspring (10, 4, 4 and 3 females for SS, NN, NS and SN treatments respectively (N = 42 females in each of the treatments), χ² = 5.9, P = 0.12), and females that received sperm from one male but not the other (which cannot be distinguished from double fertilised females in the NN and SS treatments). Failure to transfer sperm at mating is common and its frequency does not differ substantially between within and among strain crosses [49]. The difference in the offspring produced over one week was significant between SS and NN pairs (estimate: −0.50, z = −2.35, P < 0.01, Fig 2 panel A).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Panel (A) The mean number of adult offspring produced by females mating to differing combinations of sibling (S) and non-sibling (N) males with means and standard errors; n = 42 blocks of 4 sibling females, one per treatment.

Panel (B). P₂ ± S.E. of sibling and non-sibling males over 5 weeks showing the decline in P₂ (the proportion of offspring sired by the 2^nd male to mate) over time. P₂ values are adjusted to take into account the lower survival of offspring from sibling males (offspring of sibling males have lower survival to adulthood than offspring of non-sibling males, so sibling male offspring numbers are multiplied by 1.42 to calculate fertilisation success (see ‘calculating adjusted P₂ below’)).

https://doi.org/10.1371/journal.pone.0335301.g002

The total number of offspring sired by the first and second male to mate with a female was calculated for those females that produced offspring from both males and ceased offspring production during the 5-week period. The total number of offspring produced by the first male until the female ceased offspring production was 329 ± 10.9 (27%) and by the second male was 894 ± 22.7 (73%).

In the NS and SN groups, offspring paternity can be determined using the visible genetic markers carried by the swollen antenna of the Rd males. Overall, the pattern of P₂ between a sibling and non-sibling male when competing against a Rd male was similar (estimate: −0.21, z = −0.48, P = 0.62, Fig 2 panel B).

4. Discussion

Males placed with a sibling and a non-sibling female attempted to mate with their sisters and actually succeeded in mating with them slightly more often than unrelated females. Given that there is substantial inbreeding depression in our population, and in the species in general [36], these findings are unexpected. One possible explanation is that there are inclusive fitness benefits of mating with closely related individuals. These benefits stem from the fact that an allele that encourages females to mate with their brothers (who have a 50% chance of carrying that allele) will tend to increase in frequency [59,60]. These kin selected benefits mean that it is arguably quite surprising that we do not see more examples of individuals preferentially mating with close relatives in nature [61]. Perhaps Tribolium beetles with their relatively low costs of inbreeding are adaptively choosing to mate with relatives because the inclusive fitness benefits outweigh the costs in terms of lower survival of inbred offspring.

As mentioned in our methods, an earlier study [35] found some evidence that whether T. castaneum beetles are reared on their own or with conspecifics could affect their mate preferences as an adult. Mate copying, a form of social learning has been shown in Drosophila melanogaster [62,63], and acoustic experience during development is known to affect mating tactics in field crickets [64]. Rearing individuals in complete isolation may remove developmental and social cues that are themselves biologically important to mate recognition. Hence, it is important to note that there is a possibility that our experimental design in which individuals are reared alone might generate a different outcome to a study in which rearing took place in groups.

Our findings reveal that siblings can recognise one another (otherwise they could not preferentially mate with one another). We do not know what the mechanism is for this. Siblings were separated as pupae but had spent all their larval stages (about 3 weeks since hatching) together and hence had some experience of each other which may have provided them the opportunity to learn to identify their brothers and sisters. It is also possible that they could identify siblings using the products of highly polymorphic genes [8], enabling them to identify close relatives directly via inherited traits or through phenotype matching [65].

Our approach of determining reproductive success via adult offspring has the drawback that survival to adulthood must be taken into account. An ideal study would genotype eggs using molecular markers and assign parentage at this stage. In relation to potential post-mating inbreeding avoidance, females mated to 2 non-sibling males produced significantly more offspring than those mated to 2 sibling males or to a non-sib and then a sib male. This difference is as expected given the high P₂ in this species [66,67]; females have less viable offspring if they mate last to a sibling, but their offspring will not suffer very much if an earlier mate is their sibling. It appears that female T. castaneum do not bias paternity in favour of unrelated males, which is consistent with de Boer et al.’s [47] meta-analysis which found little support for inbreeding avoidance in animals in general. However, our study suggests that Tribolium could avoid costs of inbreeding if they could somehow ensure that the last male they mate with before laying eggs is an unrelated male. Similarly, a study on this species showed that polyandry rescues female reproductive fitness within inbred populations [38].

As well as confirming the previously reported high P₂ immediately following a double mating, our results show that P₂ decreases significantly over a 5-week period. This is consistent with the suggestion by Lewis and Jutkiewicz [67] that high P₂ is a result of sperm being stratified in the female spermatheca, but suggests that this stratification breaks down over time. Lewis and Jutkiewicz [67] used dissection of mated females to show that a single insemination only fills about two-thirds of the total capacity of the spermatheca, with the remaining one third being filled by the next copulation. Our results show that sperm displacement occurs during the second copulation otherwise we would expect that when a female uses all the sperm she has available, approximately two thirds of her offspring will be sired by the 1^st male to mate. We show that the majority of offspring produced by doubly mated females that ceased offspring production over the 5 weeks (presumably because they run out of sperm) were sired by the second male. This suggests that sperm displacement is a factor driving the high P₂ seen in this species, a proposal which is consistent with Arnaud and Haubruge’s [39] suggestion that the morphological characters of the male genitalia of T. castaneum may allow removal of previously deposited sperm from the female reproductive tract.

Although inbreeding avoidance mechanisms are taxonomically widespread and were thought to be the rule rather than the exception amongst vertebrates [3], more recent analyses have shown that inbreeding avoidance is not as common as is sometimes assumed [23]. There are clear cases where inbreeding avoidance is employed; for example, crickets Gryllus bimaculatus have been shown to prefer unrelated males as mates [68,69] and appear to be able to avoid using sperm from closely related males and appear to be able to avoid using sperm from closely related males [19,20], and there is evidence that another species of cricket may show similar abilities [18]. Females of the predaceous mite Phytoseiulus persimilis, [70], avoid mating with kin. However the Argentine ant, Linepithema humile shows similar sibling and non-sibling mating rates [71]. Our population of T. castaneum also failed to show pre- or post-mating inbreeding avoidance despite survival to adulthood being substantially reduced in inbred offspring and the potential for other adult fitness costs that we have not measured (e.g., [72]). For instance, Vasudeva et al. [72] found that males experiencing greater inbreeding load suffered from shorter sperm, smaller testes volumes, smaller body sizes and reduced lifespan compared to males that were under control non-inbred conditions. Reproductive output was affected as well, but that value was a composite of both male and female inbreeding status (as inbred females were paired with inbred males). The results presented here has three potential explanations; either that inbreeding risk is not sufficient to drive the evolution of inbreeding avoidance, that inclusive fitness benefits of inbreeding outweigh the costs [59,60] (as discussed above), or that there are constraints on the evolution of kin recognition. The risks of, and costs imposed by inbreeding depression in wild invertebrate populations remain important open questions.

Our study supports the observation that some insect species with an ecology that is likely to expose them to inbreeding risk have not evolved the capacity to avoid it. Whether this is because inbreeding is adaptive as a result of inclusive fitness benefits, or because of constraints on avoidance of inbreeding, remains to be determined. However, the fact that we observe not only a lack of inbreeding avoidance, but a specific preference for mating with kin suggests that this mating with kin may be an adaptation in this species and potentially in others.