Biol Invasions DOI 10.1007/s10530-016-1179-1
ORIGINAL PAPER
Reduced population size can induce quick evolution of inbreeding depression in the invasive ladybird Harmonia axyridis Guillaume J. M. Laugier . Gilles Le Mogue´dec . Wang Su . Ashraf Tayeh . Laurent Soldati . Bruno Serrate . Arnaud Estoup . Benoıˆt Facon
Received: 31 March 2015 / Accepted: 25 May 2016 Ó Springer International Publishing Switzerland 2016
Abstract Understanding biological invasion is currently one of the main scientific challenges for ecologists. The introduction process is crucial for the success of an invasion, especially when it involves a demographic bottleneck. A small introduced population is expected to face a higher risk of extinction before the first stage of invasion is complete if inbreeding depression, caused by the expression of deleterious alleles, is important. Changes in mating regimes or in population size can induce the evolution of deleterious allele frequencies, either by selection or by drift, possibly resulting in the purging or the fixation of such alleles within the population. The harlequin ladybird
Electronic supplementary material The online version of this article (doi:10.1007/s10530-016-1179-1) contains supplementary material, which is available to authorized users. G. J. M. Laugier A. Tayeh L. Soldati B. Serrate A. Estoup B. Facon Inra, CBGP, UMR 1062, Montpellier, France
Harmonia axyridis became invasive on several continents following a scenario including at least one event of demographic bottleneck. Although native populations suffered from severe inbreeding depression, it was greatly reduced in invasive ones suggesting that deleterious alleles were purged during the invasion process. In this study, we performed an experiment designed to manipulate the effective population size of H. axyridis across successive generations to mimic contrasting introduction events. We used the measurement of two fitness-related phenotypic traits in order to test (1) if inbreeding depression can evolve at the timescale of an invasion; and (2) if the changes in inbreeding depression following a bottleneck in laboratory conditions are compatible with the purging of deleterious alleles observed in this species. We found that two generations of very low population size are enough to induce a substantial change in inbreeding depression. Although the genetic changes mostly consisted in fixation of deleterious alleles, purging did also occur, sometimes simultaneously with fixation.
G. J. M. Laugier (&) ´ ridas, ULS, Centro de Estudios Avanzados en Zonas A Rau´l Bitra´n, 1305 La Serena, Chile e-mail:
[email protected]
Keywords Inbreeding depression Bottleneck Purging Fixation Harmonia axyridis
G. Le Mogue´dec Inra, AMAP, UMR 0931, Montpellier, France
Introduction
W. Su Institute of Plant and Environment Protection, Beijing Academy of Agriculture and Forestry Sciences, Beijing, China
Biological invasions are one of the major threats to biodiversity and can have deleterious impacts on human health and economy (Pimentel et al. 2000).
123
G. J. M. Laugier et al.
Most biological invasions start with the introduction of one or several alien populations of small size. This step is critical for the success of the invasion because it implies a high extinction risk. Indeed, the introduction process usually involves a demographic bottleneck and such events imply a huge stochasticity in the growth rate of the population, but often also genetic drift, an increase of the inbreeding or selfing rate and a higher inbreeding depression (Barrett 2007; Keller and Taylor 2008). Inbreeding depression is defined as the decrease in fitness of the offspring of related individuals compared to that of the offspring of random individuals in the population. It is mostly caused by the expression at homozygous state of deleterious and (partially) recessive alleles (e.g. Roff 2002). Inbreeding depression is found in Nature in most species and can strongly affect the survival of concerned populations (Frankham and Ralls 1998; Hedrick and Kalinowski 2000). Inbreeding depression may evolve across generations (Charlesworth and Charlesworth 1987). Such variations may be due to a change in effective population size or mating regime, which instantaneously changes the probability of homozygosity in the offspring (Saccheri et al. 1998; Gle´min 2003). At longer term, both changes in mating regimes and in population size may induce the evolution of deleterious allele frequencies, either by selection or by drift, which can in turn impact both the level of inbreeding depression and the risk of extinction of the population (e.g. Bataillon and Kirkpatrick 2000; Gle´min 2003). The most extreme cases correspond to the purging or the fixation of such alleles within the population. In both of these cases, the inbreeding depression is no different from zero. Although both purging and fixation of deleterious alleles lead to a decrease in inbreeding depression, they have opposite effects on the fitness of the individuals and opposite consequences for the population (see Fig. 1 and Bataillon and Kirkpatrick 2000). Indeed, the mean fitness is decreased in a population when a deleterious allele has been fixed, threatening its survival, while it is increased if the allele has been purged. The evolution of inbreeding depression has been extensively studied in the framework of evolution of mating systems (e.g. Porcher and Lande 2005), and in a context of conservation biology (e.g. Hedrick and Kalinowski 2000; Madsen et al. 1996). More recently, research on inbreeding depression has been conducted
123
in the context of biological invasions (Daehler 1999; Keller and Fournier 2002; Facon et al. 2011; Verhoeven et al. 2011). The introduction step of an invasion is often linked to low population densities (Davis et al. 2004). In such conditions, the probability of mating with a related individual is high, even in panmictic populations, and inbreeding depression is expected to be stronger in recently introduced populations than in source populations of bigger size before the possible evolution of deleterious allele frequency. A small introduced population in which inbreeding depression is important is hence expected to face a higher risk of extinction due to the expression of deleterious alleles before the first stage of the invasion is complete (Lande and Schemske 1985; Kirkpatrik and Jarne 2000). Conversely, drift and selection against a deleterious allele during a bottleneck can theoretically purge it from the population, reducing the level of inbreeding depression within the population, which in return could boost the invasion success. For instance, the harlequin ladybird Harmonia axyridis, originating from eastern Asia, was introduced on several continents following a scenario including at least one event of demographic bottleneck (Lombaert et al. 2010; Facon et al. 2011). To investigate inbreeding depression in this species, Facon et al. (2011) measured several fitness-related traits (including the hatching rate and the survival of larvae until adulthood) in the offspring of full-sibs and unrelated couples from several native and invasive populations of H. axyridis. Their results showed that although native populations suffered from severe inbreeding depression, it was greatly reduced in invasive populations, suggesting that the deleterious alleles involved were purged during the invasion process. The present study is designed to deepen some of Facon et al. (2011)’s results and aims to test (i) if inbreeding depression can evolve at the time-scale of an invasion (i.e. a few generations), and (ii) if the changes we might observe in controlled laboratory conditions are compatible with the purging of deleterious alleles inferred by Facon et al. (2011). We used an experimental design of laboratory controlled crosses designed to manipulate the effective population size of a native population of H. axyridis. We followed the evolution of two groups of populations of different effective sizes and mating systems (reduced panmixia and full-sib mating) for several generations,
Reduced population size can induce quick evolution Depression
Purging
Fixation
Fitness
No depression
N
F
N
F
N
F
N
F
Fig. 1 Theoretical illustration of measurement of inbreeding depression, purging and fixation of deleterious alleles. Measurement of a fitness-related trait for non-sibling (N) or full-sib
(F) crosses. Black and white dots respectively represent the control and focal groups, while gray dots are generic
along with a control population of larger effective size. Our results are based on the measurement of two early survival traits: hatching rate and larval survival.
create G2 and on forth generations included only fullsibs. The second generation (G2) consisted in 105 families, each of them sired from randomly-created G1 couples. Each G2 family contributed equally to three groups of populations (namely, Serially Inbred, Restricted Panmixia and control groups), by producing two adults each (see Fig. 2a). This experimental design allowed us to minimise the differences between genetic pools of each group at the start of the experiment. The Restricted Panmixia group (RP) was split into independent sub-populations of three couples each (triplets, N = 6). The Serially Inbred group (SI) was also split in independent sub-populations of one single couple each (sibling-families, N = 2). The control group was kept as a single population, although each couple was reared individually for a better control of the population effective size (N [ 30). The 60 first families of G2 to produce three adult males and three adult females each—i.e. enough founding individuals to contribute equally to each group—were assigned to experimental block 1. The 45 remaining families took longer to produce enough founding individuals, and were assigned to experimental block 2. It is worth noting that because of this, the two blocks were not statistical replicates but rather two independent cohorts. During the following generations, the crosses were arranged within each population units (i.e. population, triplet and sibling-family respectively for control, RP and SI groups). Full-sib crosses were avoided, except within the SI group (Fig. 2b, c1, c3 and c6). The fifth generation (G5) was a generation of outcrossing that allowed to shuffle the alleles within each experimental
Materials and methods Over 200 beetles were sampled in apple orchards around Beijing (China) during the summer of 2011. These individuals constituted the wild generation called G0. They were kept in plastic boxes and fed with live aphids for 2 weeks at Beijing Academy of Agriculture and Forestry Science before being transferred at the Institute of Biology for Population Management in Montferrier-sur-Lez, France. From this point on, the beetles were reared in Petri dishes, with one couple (two adults) or sibling family (c. a. 10 larvae) per dish; and fed with ionised eggs of Ephestia kuehniella (Lepidoptera, Pyralidae). The rearing conditions where 14:10 (L:D) light cycle and a constant temperature of 23 °C. The experimental design is presented in detail in Fig. 2. The first generation in the laboratory conditions (G1) was obtained through crosses designed to (i) maximise the effective size of the population (and thus limiting the loss of genetic diversity) by mating each individual with as many other as possible, and (ii) get rid of females infected by male-killing organisms (such as Spiroplasma sp.) that sometimes parasite Asian populations of this species (Majerus et al. 1999) by discarding the females bearing no male offspring. All the individuals used in the subsequent stages were thus clear of male-killing bacteria. In any subsequent generations (from G1 to G6), each adult was only mated once to make sure that the families used to
123
G. J. M. Laugier et al. Fig. 2 Experimental design. a Plans of the different generations during the experiment. Rectangles in light gray and dark gray backgrounds respectively represent Restricted Panmixia groups (RP) and Serially Inbred groups (SI). Rectangles in white backgrounds represent control groups except for G0, G1 and G2 that represent the preliminary steps of the experiment. See main text for details. b Detailed description of the different controlled crosses processed during the experiment. Each circle represent a full-sib family, and each coloured rectangle is a (sub)population unit. Crosses between full-sibs were systematically avoided, except for crosses C2, C4 and C6. For group RP, all crosses were arranged within a single triplet (except for crosses C5 where they were arranged between triplets). All crosses were arranged between individuals from the same generation. This protocol was inspired by the one used in Fox et al. (2008)
A
B
group (i.e. control, RP and SI) and restore heterozygosity in all the groups (Charlesworth and Charlesworth 1999; Fox et al. 2008). In order to compensate the loss of lines and to avoid the extinction of (sub)populations, replicates have been made for generation G3 of block 2 and generation G4 of block 1. The last generation (G6) is a measurement generation, allowing the estimation of the inbreeding depression in each group at the end of the experiment. In this generation, full-sib matings and outcrosses
123
were performed in each group. The data obtained were measures for the two phenotypic traits that displayed the strongest evidences for purging in the study of Facon et al. (2011): the hatching rate and the rate of survival from hatching to adulthood. Two to three clutches per females were collected and each was placed in a Petri dish of 7 cm in diameter in order to limit the acts of larva-on-egg cannibalism. The counting of the eggs before hatching and of larvae just after hatching allowed us to estimate the hatching rate. Up to eight larvae per clutch were kept until their
Reduced population size can induce quick evolution
emergence as imagos. The low density of larva in each Petri dish allowed to limit the cannibalistic behaviour, which can be important in this species (Osawa 1993). The ratio of the number of emerging imagos over the number of larvae kept was used to estimate the rate of survival from hatching to adulthood (larval survival). Overall, 105 couples were established in G3, divided in 1, 20 and 60 populations in experimental block 1 and 1, 15 and 45 populations in block 2, for control, RP and SI groups respectively. Despite the loss of family lines during the experiment, data could be gathered for a total of 42 families in the first experimental block and 33 in the second one. Details can be found in Tables S1 and S2 in the supplementary materials. Statistical treatment of data The data was merged by family within each group and type of crossing, then analysed with the R statistical software (R Core Team 2012). Because survival in a set of independent individuals is a binomial process, we analysed the data using quasi-binomial Generalised Linear Models (GLMs) (Crawley 2012, Chapter 16). The use of such models allows to take into account over-dispersion. The effects of crossing, population size and experimental block on hatching rate or survival to adulthood, along with all their interactions, were used as explicative variables to build a full model. Then the non-significant effects were removed one by one using the backwards method. The estimators of the GLMs have an asymptotic normal (Gaussian) distribution from which we constructed contrast tests (Sokal and Rohlf 1995; Crawley 2012). In other words, the statistical distributions of GLM-estimated effects were used to test the significance of the differences represented in Fig. 1, i.e. the presence of inbreeding depression and of purging and/ or fixation of deleterious alleles.
Table 1 Effects of cross types, population size and their interaction on the hatching rate Tested effect
LR v2
df
p value
Cross type
5.1425
1
0.02335
Population size
2.3476
2
0.30918
Cross type:population size
8.8134
2
0.01220
Bold value indicates statistical significance (with a threshold of 0.05) Deviance analysis of the GLMs of the effects of crossing type (full-sibs or non-siblings), population size (defined by groups RP, SI or control), and their interaction on the hatching rate. Type III tests
population size (RP, SI or control). The results are not significantly different between the two experimental blocks (p [ 0.5). The observed effects correspond to a difference in the level of inbreeding depression across the three population sizes (Table 2; Fig. 3a). Hatching rates are significantly lower from full-sib crosses than the ones from non-sibling crosses for groups RP and control (RP: 0.40 and 0.56; control: 0.38 and 0.49, for full-sibs and non-sibling crosses respectively, see Fig. 3a). This pattern corresponds to a significant inbreeding depression (Table 2). In contrast, there is no significant inbreeding depression in SI group (hatching rates of 0.46 and 0.40 for full-sibs and non-sibling crosses respectively, Table 2). For this group, hatching rate in non-sibling crosses is significantly lower than the one observed in the control group which corresponds to the fixation of deleterious alleles (Table 2). For SI group, although Table 2 Tests of inbreeding depression, purging and fixation for hatching rates Groups
Tested effect
Control
Inbreeding depr.
RP
SI
Results Hatching rate Table 1 shows that hatching rate is affected by crossing type (full-sibs or non-sibling), and by the interaction between type of crossing and the type of
Inbreeding depr.
t 2.262
df
p value
176
0.012 0.002
2.963
176
Purging
-0.406
176
0.657
Fixation
-1.128
176
0.870
Inbreeding depr.
-1.007
176
0.842
Purging
1.225
176
0.111
Fixation
1.868
176
0.032
Bold value indicates statistical significance (with a threshold of 0.05) The tests are unilateral for purging, fixation or inbreeding depression and bilateral for differences between groups SI (N = 2) and RP (N = 6)
123
0.6 0.5 0.4 0.3 0.2
GLM estimation of hatching rate
A
0.7
G. J. M. Laugier et al.
N
F
Cross type
C 1.0 0.9 N
F
Cross type
0.5
0.6
0.7
0.8
0.9 0.8 0.7 0.6 0.5
GLM estimation of larval survival
1.0
B
N
F
Cross type
Fig. 3 Survival rates for non-sibling (N) and full-sib (F) crosses as estimated by the GLM, for the three population sizes: N = 2 (SI; dark gray, solid line), N = 6 (RP: light gray, dashed line) and N [ 30 (control; empty circles, dotted line).
Error bars represent the 95 % confidence interval around estimations. For hatching rates (panel A), the experimental blocks are merged. For larval survival, experimental block 1 (panel B) and 2 (panel C) are represented separately
there is a trend for purging (hatching rate from full-sib crosses in SI is higher than in the control group, Fig. 3a), this trend is not significant (Table 2).
(Table 3). The mean survival rate of larvae from full-sib crosses is lower than those from non-sibling crosses for all groups (control: 0.59 and 0.82; RP: 0.72 and 0.75; SI: 0.76 and 0.86 for full-sibs and non-sibling crosses, respectively; see Fig. 3a). This pattern corresponds to inbreeding depression. This decrease is significant for groups SI and control (although the depression is less important in group SI than in the control group), but not for RP group (Table 4). For RP group, the survival rate of larvae from nonsibling crosses is significantly lower than in control group, which correspond to fixation of deleterious alleles (Table 4; Fig. 3). For the same group, the survival rate of larvae from full-sib crosses is significantly higher than in control group, which correspond
Larval survival For this trait, the deviance analysis of the GLM shows a significant block effect, whether alone or in interaction with the effect of population size (p \ 0.5). The results are thus accordingly shown separately for each block. Experimental block 1 The effects of cross type and population size and their interactions on larval survival are significant
123
Reduced population size can induce quick evolution Table 3 Effects of cross types, population size and their interaction on the larval survival LR v2
df
p value
Cross type
27.6226
1
1.474 3 1027
Population size
12.9605
2
0.001533
8.3293
2
0.015535
15.8050
1
7.022 3 1025
Population size
5.9471
2
0.05112
Cross type:population size
0.7035
2
0.70345
Tested effect Experimental block 1
Cross type:population size
In SI group, the survival rate of larvae from nonsibling crosses is not lower than in control group (Fig. 3a), which corresponds to the lack of fixation of deleterious alleles. For this same group, the survival rate of larvae from full-sib crosses is significantly higher than in control group, which corresponds to the purging of deleterious alleles (Table 4; Fig. 3b).
Experimental block 2 Cross type
Bold value indicates statistical significance (with a threshold of 0.1) GLM deviance analysis of the effects of crossing (full-sibs or non-siblings), population size (defined by groups RP, SI or control) and their interaction on larval survival rate in the two experimental blocks. Type III tests
to the purging of deleterious alleles (Table 4; Fig. 3b). In this group, the data thus display the signature of both the purging and the fixation of some deleterious alleles. Table 4 Tests of inbreeding depression, purging and fixation for larval survival in the two experimental blocks Groups
Tested effect
t
df
p value
80
8 3 1027
Block 1 Control RP
SI
Inbreeding depr.
5.182
Inbreeding depr.
0.647
80
0.260
Purging Fixation
2.400 1.664
80 80
0.009 0.050
Inbreeding depr.
2.137
80
0.018
Purging
3.423
80
5 3 1024
Fixation
-1.053
80
0.852
3.935
74
9 3 1025
Block 2 Control RP
SI
Inbreeding depr.
2.557
74
0.006
Purging
Inbreeding depr.
-2.275
74
0.987
Fixation
2.715
74
0.004
Inbreeding depr.
3.329
74
0.001
Purging
0.110
74
0.457
Fixation
-0.729
74
0.766
Bold value indicates statistical significance (with a threshold of 0.05) The tests are unilateral for purging, fixation or inbreeding depression and bilateral for differences between groups SI (N = 2) and RP (N = 6)
Experimental block 2 The main effects of Cross type and Population size are significant, those of the interaction between these effects are not (at the threshold of 0.1, see Table 3). In all groups, the mean survival rates of larvae from full-sibs are significantly lower than those from nonsibling crosses (control: 0.72 and 0.82; RP: 0.64 and 0.74; SI: 0.72 and 0.83 for full-sibs and non-sibling crosses, respectively, see Table 4; Fig. 3). This pattern corresponds to inbreeding depression in all groups for this experimental block. In RP group, the survival rate of larvae from nonsibling crosses is significantly lower than those from control group (Table 4; Fig. 3c). This pattern corresponds to the fixation of deleterious alleles. In addition, the survival rate of larvae from full-sib crosses was not higher than those in control group (Table 4; Fig. 3c). This pattern corresponds to the lack of purging of deleterious alleles. In group SI, the survival rate of larvae from nonsibling crosses was not lower than those from control group (Fig. 3c). This pattern corresponds to the lack of fixation of deleterious alleles. In addition, the survival rate of larvae from full-sib crosses was slightly higher than those from control group, which would correspond to the purging of deleterious alleles; but this difference is not significant (Table 4; Fig. 3c).
Discussion Variability in evolution of inbreeding depression Overall, our results show that experimental populations of H. axyridis undergoing a short demographic bottleneck event display variable changes in inbreeding depression and fitness in general. This variability seems mostly due to stochastic processes and may represent a significant feature in the event of
123
G. J. M. Laugier et al.
demographic bottlenecks associated with the introduction of an alien species. More explicitly, our experimental results underline the great variability of responses to a decrease of the effective size of the population. This variability exists at several levels: between both measured traits, between the two groups (i.e. RP and SI), and also for a single trait between the experimental blocks. This variability is in accordance with the work of Pray and Goodnight (1995), who found a high variability in several phenotypic traits following inbreeding between full-sib lines in the red flour beetle Tribolium castaneum. Byers and Waller (1999) also found an overall strong variability in the efficiency of purging in plants in their meta-analysis. The genetic drift, which is particularly strong in our experiment, can be accounted for a large part of the variation in fitness observed in our study. The great variability of responses to experimental conditions can also be partly explained by the genetic variation caused by sampling at the start of the experiment. Lacy et al. (1996) also found a great variability between populations of mice, which they say is due to differences in genetic composition between the initial lines. In our experiment, we minimised such variation by insuring that each source family (in generation G2) produced an equal number of founders for each groups (SI, RP and control) within each experimental block. Admitting that inter-block differences could be due to a genetic heterogeneity, it is possible that the initial sampling of wild populations represented a substantial genetic bottleneck despite our efforts to atone it. The purging of deleterious alleles might already have begun in preliminary generations (G1–G2) and that some deleterious alleles would then be expressed in some individuals, causing (among other) to slow their development down. The corresponding families would then be more present in experimental block 2, which could explain the lower inbreeding depression observed in the control group of this block (compare Fig. 3b, c). This suggests that the purging of deleterious alleles could occur very quickly during realistic, less severe bottlenecks. The variability in fitness (proxied by hatching and survival rates) observed can also come from experimental variations (Swindell and Bouzat 2006). Our study indeed spanned on over 8 months, during which occasional incidents (e.g. food shortage, climatic chamber break-down) might be responsible for a part
123
of the variation observed in the results, particularly between experimental blocks, which were staggered in time, and might thus be affecting a given incident at different development stages. In the context of biological invasions, this huge stochasticity implies that different introductions of alien populations may have different establishment success, assuming inbreeding depression affects establishment success, and that more introductions will increase the probability of one being successfully established. Despite the high variability of response to a demographic bottleneck, changes in fitness were almost systematic, suggesting that every single introduced population represents a chance of successful introduction. Rapid evolution of inbreeding depression Overall, a demographic bottleneck of short duration was sufficient to induce significant changes in inbreeding depression in some experimental populations of H. axyridis, including partial purging of deleterious alleles. These results are compatible with Facon et al. (2011) hypothesis that the lack of inbreeding depression displayed by invasive populations of this species is due to the purging of deleterious alleles during the short bottleneck experienced by introduced populations. However, conversely to the case displayed by invasive populations of this species, the genetic changes mostly consist in partial fixation of deleterious alleles, sometimes along a partial purging. Two generations of full-sib crosses were sufficient to considerably affect inbreeding depression of hatching rate (Table 5). This result is congruent with those from Fox et al. (2008) in the seed-feeding beetle Stator limbatus, in which inbreeding depression is reduced Table 5 Main results summarised from Tables 2 and 4 Purging
Fixation
No
Yes
No
No
Hatching rate Serially inbred (N = 2) Restricted panmixia (N = 6) Larval survival Serially inbred (N = 2)
Yes
No
Restricted panmixia (N = 6)
Yes
Yes
For larval survival, the results are considered significant if the corresponding tests are significant at a 5 % threshold for at least one experimental block
Reduced population size can induce quick evolution
by purging in three generations of full-sib mating. This kind of rapid variation of inbreeding depression is not limited to insects, but also shown in various plant species (Barrett and Charlesworth 1991). Our results for larval survival rate in RP group suggest that a brief decrease in effective size, even without a change in mating regime, could be sufficient to reduce the level of inbreeding depression in H. axyridis, although the purging of deleterious alleles is not systematic. Simultaneous purging and fixation of deleterious alleles Our study shows that purging and fixation can occur simultaneously. Although little discussed in the literature, the co-occurring of these two mechanisms has already been suggested (Crnokrak and Barrett 2002). There is two means by which inbreeding depression can be decreased: (i) through an increase in the fitness of full-sib offspring, and (ii) through the decrease in the fitness of non-sibling offspring (cf. Fig. 1). If these two mechanisms occur simultaneously, the decrease in inbreeding depression is great, but the decrease of the genetic load is small (if any). This genetic load is indeed the result of the expression of deleterious recessive alleles at several loci. It is possible that some of them were fixed by drift during the experiment, causing the mean fitness of the population to decrease (for any type of crossing). Simultaneously, deleterious alleles at other loci could have been eliminated by selection and/or by drift. When segregating in the population, these alleles mainly affect the fitness of individuals from full-sib crosses, which are often homozygous. Alleles that are less sensitive to drift (being highly deleterious or mildly recessive) are expected to be more frequently purged than others of lesser effect (Kirkpatrik and Jarne 2000; Gle´min 2003). Alternatively, (quasi)simultaneous purging and fixation of deleterious alleles can occur through selective sweep in case of linkage disequilibrium (Charlesworth and Charlesworth 1999; Wang et al. 1999).
crosses (G5, Fig. 2), although unavoidable to restore heterozygosity, allow the crossing between first cousins in the next generation (C5 et C7), notably in the case of group SI, where these cousins come from full-sib lines and have a high level of kinship. These crosses may have produced low fitness individuals from ‘‘non-siblings crosses’’ in G6, possibly leading to over-estimating the fixation and under-estimating the inbreeding depression in groups RP and SI. On the other hand, an opposite bias may have been caused by the limited effective size of control group used in this study, in which the genetic drift was most likely non-negligible. To better explain the implication of this feature, let us imagine that a single rare deleterious mutation is only present in one family at generation G2. Suppose as well that one single copy of this allele was transmitted in each experimental group. As, in SI lines, this mutation cannot be fixed in more than a single line, it may not be present in more than a few individuals in this group at generation G6. Conversely, as control group is panmictic, this mutation can be fixed by drift and thus be present in all individuals of control group at the end of the experiment. The case of group RP is intermediary between those of groups SI and control. In these conditions, the mean fitness of individuals in the control group may be under-estimated in both full-sib and nonsibling matings. It is thus possible that the results presented here over-estimate, to some extent, the purging of deleterious alleles compared to the control group. Finally, it is possible that the purging process has been dampened by over-stressing laboratory conditions that can, as would do too mild conditions, decrease the efficiency of the selection against deleterious alleles (Kishony and Leibler 2003). In too adverse conditions, even individuals that are homozygous for a wild-type (non-deleterious) allele have a low fitness. During the six generations constituting this experiment, the population was probably exposed to selection pressures due to rearing conditions. In particular, the food offered to the beetles was not the main feeding source in natural conditions (Ferran et al. 1997). If the population we used was preliminarily adapted to experimental condition, the results might have been a stronger or more frequent purging.
Limits of this study Conclusions and perspectives The analysis of the results brought by this study might be slightly biased by two factors of the experimental design. On the one hand, the generation of non-sibling
It is not possible to conclude, based on the results of this study that reduction of the genetic load
123
G. J. M. Laugier et al.
observed by Facon et al. (2011) in invasive populations of the harlequin ladybird in fact occurred due to a demographic bottleneck. However, the high variability of outcome observed here highlight the importance of stochastic processes that can occur during the bottleneck period often associated with biological invasions. It is thus possible that the outcome observed in wild harlequin ladybird invasive populations correspond to outliers in the distribution of the possible outcomes of an introduction. This is in accordance with Reed and Bryant (2001)’s study, in which the best trade-off between purging and fixation is an exceptional value, reached by a population from a single founder couple (see Fig. 1 of their publication). If the introduction events are repeated, then the chances increase that one of these iterations produces such an extreme case. In the peculiar case of a biocontrol agent species such as H. axyridis, releases are often repeated. In particular, in Northern America, a number of releases coming directly from the native area were conducted during nearly 60 years, before an invasion occurred in the East then in the West of the USA (Koch 2003). It is thus possible that this invasion would be the consequence of an introduction associated with a demographic bottleneck characterised by an exceptionally strong purging and an exceptionally low fixation of deleterious alleles. The laboratory conditions (in terms of effective size, number of generations, and environmental stress) used in this experiment were probably far from the optimal conditions for purging the deleterious alleles responsible from the genetic load in H. axyridis. This is particularly true for SI group in which most of deleterious allele fixation occurred. Although useful to induce rapid changes in inbreeding depression, the design of full-sib mating remains the most drastic case of decrease in population size. In the case of H. axyridis, Facon et al. (2011) showed that the bottleneck endured by invasive populations was probably longer (&20 generations) and less intense (i.e. with greater effective sizes; &150 individuals) than the ones used in this study. It is unfortunately difficult to reproduce experimentally bottlenecks of similar intensities, and/or to use a greater number of replicates to fully describe the variation of responses. Only a modelling approach would allow to further expand the result of our study.
123
References Barrett SCH, Charlesworth D (1991) Effects of a change in the level of inbreeding on the genetic load. Nature 352(6335):522–524 Barrett SC, Colauti RI, Eckert CG (2007) Plant reproductive systems and evolution during biological invasion. Molecular Ecology 17:373–383 Bataillon T, Kirkpatrick M (2000) Inbreeding depression due to mildly deleterious mutations in finite populations: size does matter. Genet Res 75(01):75–81 Byers DL, Waller DM (1999) Do plant populations purge their genetic load? effects of population size and mating history on inbreeding depression. Annu Rev Ecol Syst pp 479–513 Charlesworth D, Charlesworth B (1987) Inbreeding depression and its evolutionary consequences. Annu Rev Ecol Syst 18:237–268 Charlesworth B, Charlesworth D (1999) The genetic basis of inbreeding depression. Genet Res 74(3):329–340 Crawley MJ (2012) The R book. John Wiley, Hoboken Crnokrak P, Barrett SCH (2002) Perspective: purging the genetic load: a review of the experimental evidence. Evolution 56(12):2347–2358 Daehler CC (1999) Inbreeding depression in smooth cordgrass (spartina alterniflora, poaceae) invading san francisco bay. Am J Bot 86(1):131–139 Davis HG, Taylor CM, Civille JC, Strong DR (2004) An allee effect at the front of a plant invasion: Spartina in a pacific estuary. J Ecol 92(2):321–327 Facon B, Hufbauer RA, Tayeh A, Loiseau A, Lombaert E, Vitalis R, Guillemaud T, Lundgren JG, Estoup A (2011) Inbreeding depression is purged in the invasive insect Harmonia axyridis. Curr Biol 21(5):424–427 Ferran A, Gambier J, Parent S, Legendre K, Tourniere R, Giuge L (1997) The effect of rearing the ladybird harmonia axyridis on ephestia kuehniella eggs on the response of its larvae to aphid tracks. J insect Behav 10(1):129–144 Fox CW, Scheibly KL, Reed DH (2008) Experimental evolution of the genetic load and its implications for the genetic basis of inbreeding depression. Evolution 62(9):2236–2249 Frankham R, Ralls K (1998) Conservation biology: inbreeding leads to extinction. Nature 392(6675):441–442 Gle´min S (2003) How are deleterious mutations purged? drift versus nonrandom mating. Evolution 57(12):2678–2687 Hedrick PW, Kalinowski ST (2000) Inbreeding depression in conservation biology. Annu Rev Ecol Syst pp 139–162 Keller L, Fournier D (2002) Lack of inbreeding avoidance in the argentine ant Linepithema humile. Behav Ecol 13(1):28–319 Keller SR, Taylor DR (2008) History, chance and adaptation during biological invasion: separating stochastic phenotypic evolution from response to selection. Ecol Lett 11(8):852–866 Kirkpatrik M, Jarne Ph (2000) The effects of a bottleneck on inbreeding depression and the genetic load. Am Nat 155(2):154–167 Kishony R, Leibler S (2003) Environmental stresses can alleviate the average deleterious effect of mutations. J Biol 2(2):14
Reduced population size can induce quick evolution Koch RL (2003) The multicolored Asian lady beetle, Harmonia axyridis: a review of its biology, uses in biological control, and non-target impacts. J Insect Sci 3(1):p32 Lacy RC, Alaks G, Walsh A (1996) Hierarchical analysis of inbreeding depression in peromyscus polionotus. Evolution pp 2187–2200 Lande R, Schemske DW (1985) The evolution of self-fertilization and inbreeding depression in plants. i. genetic models. Evolution pp 24–40 Lombaert E, Guillemaud T, Cornuet JM, Malausa T, Facon B, Estoup A (2010) Bridgehead effect in the worldwide invasion of the biocontrol harlequin ladybird. PLoS One 5(3):e9743 Madsen T, Stille B, Shine R (1996) Inbreeding depression in an isolated population of adders vipera berus. Biol Conserv 75(2):113–118 Majerus TM, Graf von der Schulenburg JH, Majerus ME, Hurst G (1999) Molecular identification of a male-killing agent in the ladybird Harmonia axyridis (pallas)(coleoptera: Coccinellidae). Insect Mol Biol 8(4):551 Osawa N (1993) Population field studies of the aphidophagous ladybird beetle Harmonia axyridis (coleoptera: Coccinellidae): Lifetables and key factor analysis. Res Popul Ecol 35(2):335–348 Pimentel D, Lach L, Zuniga R, Morrison D (2000) Environmental and economic costs of nonindigenous species in the United States. Bioscience 50(1):53–65 Porcher E, Lande R (2005) The evolution of self-fertilization and inbreeding depression under pollen discounting and pollen limitation. J Evol Biol 18(3):497–508
Pray LA, Goodnight CJ (1995) Genetic variation in inbreeding depression in the red flour beetle tribolium castaneum. Evolution 49(1):176–188 R Core Team (2012) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, http://www.R-project.org/, ISBN 3-90005107-0 Reed DH, Bryant EH (2001) Fitness, genetic load and purging inexperimental populations of the housefly. Conserv Genet 2(1):57–61 Roff D (2002) Inbreeding depression: tests of the overdominance and partial dominance hypotheses. Evolution 56(4):768–775 Saccheri I, Kuussaari M, Kankare M, Vikman P, Fortelius W, Hanski I (1998) Inbreeding and extinction in a butterfly metapopulation. Nature 392(6675):491–494 Sokal RR, Rohlf FJ (1995) Biometry, 3rd edn. WH Freman and company, New York Swindell W, Bouzat J (2006) Reduced inbreeding depression due to historical inbreeding in drosophila melanogaster: evidence for purging. J Evol Biol 19(4):1257–1264 Verhoeven KJ, Macel M, Wolfe LM, Biere A (2011) Population admixture, biological invasions and the balance between local adaptation and inbreeding depression. Proc Roy Soc B Biol Sci 278(1702):2–8 Wang J, Hill W, Charlesworth D, Charlesworth B et al (1999) Dynamics of inbreeding depression due to deleterious mutations in small populations: mutation parameters and inbreeding rate. Genet Res 74(2):165–178
123
