Responses of holocyclic and anholocyclic Rhopalosiphum padi populations to low‐temperature and short‐photoperiod induction

Abstract The different life cycles of aphid species make these organisms good models for studying the short‐term consequences of sex. The bird cherry‐oat aphid Rhopalosiphum padi has a wide geographic distribution and correspondingly different life cycles. In this study, the life cycles of R. padi collected from six different regions in China were characterized experimentally by comparing the responses of holocyclic and anholocyclic populations to low‐temperature and short‐photoperiod induction. Clones collected from Chuzhou, Taian, and Taigu consistently reproduced via obligate parthenogenesis, whereas clones from Hami and Baicheng were holocyclic in their response, and those from Lanzhou were both holocyclic and anholocyclic. Prolonged exposure to low temperature and a short photoperiod (LS) had negative effects on the offspring of anholocyclic aphids with regard to adult lifespan, total longevity, and fecundity compared with aphids maintained at a normal temperature and a long photoperiod (NL). Holocyclic LS R. padi had longer developmental times at all nymph stages, a shorter adult lifespan, shorter total longevity, and a lower fecundity than NL counterparts. The adult prereproduction period of gynoparae was significantly longer than that of virginoparae, and the total longevity of gynoparae was significantly shorter than that of virginoparae. Moreover, the net reproductive and gross reproduction rates, as well as the total fecundity, were roughly fivefold higher in virginoparae than in gynoparae, indicating that there is the short‐term cost of sex. When maintained on their secondary host (Triticum aestivum), gynoparae, males, and oviparae produced by holocyclic populations could survive, and gynoparae produced oviparae. However, under NL conditions, oviparae could not produce overwintering eggs on the secondary host, whereas a few overwintering eggs were generated by oviparae under LS conditions. Taken together, these results illuminate the complexity of insect responses and contribute to a complete understanding of the aphid life cycle and its evolution.


| INTRODUCTION
Two much-debated issues in evolutionary biology concern the longterm persistence and the origin of sex and asexuality in organisms.
In general, asexual lineages have many evolutionary disadvantages because their offspring are of identical genotype, and favorable mutations, by-products of meiosis, are not possible, whereas deleterious mutations can easily accumulate (Dixon, 1985;Gilabert et al., 2009).
Thus, at least theoretically, asexual lineages should ultimately be eliminated (Hadany & Beker, 2003;Muller, 1964). Sexual reproduction, by contrast, can reduce the accumulation of deleterious mutations, a process referred to as Muller's ratchet (Muller, 1964), and play a role in the progressive changes in telomere length .
However, in addition to the ability of asexual populations to very rapidly outcompete sexual ones under some conditions, asexuality may be advantageous because of the twofold cost of sex (the cost of males and of passing on only half of the individual's genes to offspring). The many ecologically and genetically based theories and hypotheses addressing the paradox of sex and its long-and short-term advantages (Butlin, 2002;Kondrashov, 1994;Muller, 1964) offer a conceptual framework for studying geographic parthenogenesis (Law & Crespi, 2002).
Among insects, aphids provide an excellent model for studying the evolution of sex, empirically examining the relative costs and benefits of sexual versus asexual reproduction, and identifying the ecological cost of switching from one to the other. A holocyclic life cycle with cyclical parthenogenesis, in which many parthenogenetic generations alternate with a single annual sexual generation, is one of the most remarkable polyphenisms in aphids. The sexual phase of cyclical parthenogenesis is triggered by environmental changes, with photoperiod, which acts as the main seasonal cue, and temperature the most crucial factors (Lees, 1966). Aphids can assess the time still available to produce appropriate aphid forms based on the progress of the photoperiod (Nylin & Gotthard, 1998). Temperature often interacts with photoperiod and modifies its effect. The sexual morphs of aphids can be induced by the appropriate combination of a short photoperiod together and a low temperature (Simon, Blackman, & Le Gallic, 1991).
The mating of sexual females and male aphids results in overwintering eggs, which are more resistant than adults to harsh conditions, such as freezing, which typically kill asexual forms (Nespolo, Halkett, Figueroa, Plantegenest, & Simon, 2009;Rispe, Simon, & Pierre, 1996). The link between sexuality and cold resistance can be seen as a contingent short-term advantage for sex. Conversely, the aphid sexual phase can also be facultative or lost by some genotypes. Anholocyclic aphid lineages with obligate parthenogenesis reproduce parthenogenetically during the whole year and, despite the appropriate inducing conditions, do not produce sexual morphs.
Besides reproductive polyphenisms, organisms have other lifehistory traits, including development time, fecundity, survival rate, and longevity, that are influenced by natural selection and may be constrained by genetic factors that are differentially expressed depending on the specific conditions (Homeny & Juliano, 2007). In nature, adaptation to the spatial heterogeneity of the environment results in intraspecific geographic variations in relevant ecophysiological traits (Dmitriew, 2011;Rajpurohit, Nedvěd, & Gibbs, 2013). Lineage specialization governed by genetic factors is also a trait of aphids, including the relative investment in sexual reproduction, which may be influenced by energy constraints in a given environment. The relative investment in all forms of reproduction will determine the life-history strategy of an aphid lineage and, therefore, also its fitness (Stearns, 1989 In this study, we examined the influence of a prolonged exposure to low temperature and a short photoperiod on the reproduction of R. padi from different regions. Additionally, we investigated the life history of gynoparae, males, and oviparae on their secondary host. Our aim was to analyze the reproductive modes of R. padi from different regions and thus to elucidate the effects of low temperature and a short photoperiod on R. padi populations with anholocyclic and holocyclic life cycles. The survival and reproduction of sexual forms maintained on the secondary host were also investigated. Our results contribute to a better understanding of life-cycle evolution in aphids.  (Pitchers et al., 2013;Zhang, Qiao, & Peng, 2016).
To initiate the experiment, 15 second-instar nymphs from each area raised under normal conditions were selected, transferred to the inducing conditions, and monitored until they reached adulthood (G 0 ). They were then allowed to reproduce; 3 days later, the adults were removed from the plants and again reared under normal conditions. The aphid forms were identified using an anatomical microscope when the G 1 individuals became adults. At the same time, winged parthenogenetic females and gynoparae were reared separately to observe the offspring morphs, because they could not be discriminated clearly based on their morphological characteristics alone. G 1 individuals were allowed to reproduce for only 1 day because colonies reared at high density preferably give rise to winged aphids. The same process was followed to ascertain aphid morphs of the G 2 -G 5 generations.
The reproductive mode of a clone was determined according to its ability to produce different aphid forms during six generations of sex induction.

| Comparison of the life-history traits of anholocyclic and holocyclic Rhopalosiphum padi
After the determination of R. padi populations from different regions, at least 100 wingless adults per population were randomly chosen and the aphids from each population were reared separately. Following larviposition by these wingless adults, the newly born nymphs of each population were randomly divided into two groups. The NL group was maintained under normal temperature and a long photoperiod, whereas the LS group was maintained under a low temperature and a short photoperiod. Newly born nymphs of the NL group were maintained in environmental growth chambers at 24°C and a photoperiod of 16-hr:8-hr L:D cycle (normal conditions), and those of the LS group were maintained under the inducing conditions of a constant short photoperiod of 8-hr:16-hr L:D cycle at 12°C. After being fed for two generations, the apterous one-day-old adult aphids of the second generations (G 2 ) of groups NL and LS were transferred to caged host plants, where they were allowed to reproduce for 2 hr, after which they were removed from the plant to obtain cohorts of same-aged first-instar nymphs of R. padi. The cohort of each population was reared in a program-controlled incubator at 24°C, a relative humidity of 70%, and a photoperiod of 16-hr:8-hr L:D cycle to obtain a life table for each population. Only one newborn aphid nymph was placed in each of the prepared seedlings, which were used at the three-leaf stage. To minimize the effect of the nutritional conditions provided by the host plants, the wheat seedlings were replenished every 7 days.
For trait measurements, the test aphid was monitored twice daily, and molting and mortality were recorded at about the same time each day. During the reproductive period of adults, newborn nymphs were counted and removed twice daily. This process was continued until all the adult aphids died.

| Comparison of the life-history traits of the sexual generation under normal and inducing conditions
To examine the survivorship and reproduction of males and oviparae on the secondary host (T. aestivum), 10 aphids were randomly collected from each of the three holocyclic colonies and then used to establish a clone. These 30 clones were raised separately under a constant short photoperiod of 8-hr:16-hr L:D cycle and a temperature of 12°C (inducing conditions) for three generations to obtain a sexual generation. Adult gynoparae and virginoparae of the third generation (G 3 ) from each clone were then randomly divided into two groups, one of which was maintained under a constant long photoperiod of 16-hr:8-hr L:D cycle at 24°C, and the other of which was maintained under low temperature (12°C) and a short photoperiod (8-hr:16-hr L:D cycle). The adult gynoparae and virginoparae of each group were allowed to reproduce for 12 hr on T. aestivum, after which they were removed from the plants. The offspring of these gynoparae and virginoparae were identified as males, oviparae, or gynoparae according to their morphological characteristics after they had molted into their final adult form (Duan et al., 2016;Simon, Rispe, & Sunnucks, 2002). The identified males and oviparae were then randomly divided into two groups; one group (~20 aphids per rearing condition) was still reared individually to observe the life history of each aphid using the same methods described for the life tables; the other group (~30 aphids per rearing condition) was used to observe reproduction on the secondary host (T. aestivum). Five males and five oviparae of the latter group were placed together for mating and egg laying in a breeding cage containing T. aestivum seedlings. The breeding cage was a transparent cubic Plexiglas container (10 × 10 × 10 cm) with the top replaced by a fine mesh gauze cover. A piece of white filter paper was placed above the nutritive medium of the seedlings to catch any eggs that fell. The number of eggs was counted 10 days later.

| Life tables and statistical analysis
The following traits were measured in both the NL and the LS groups according to the method of Birch (1948): age-specific survival rate (l x ), longevity of each aphid, age-stage-specific survival rate (s xj ), and agespecific fecundity (m x ). From these data, the mean generation time (T), reproductive value (V x ), intrinsic rate of increase (r m ), net reproductive rates (R 0 ), mean fecundity, age-stage life expectancy (e xj ), gross reproductive rate (GRR), and finite rate of increase (λ) were calculated using the TWO SEX-MSChart program (Chi, 2005).
The significance of the differences in the life-table parameters of group NL versus group LS was analyzed using a one-way ANOVA test.
For parameters determined to be significantly different (p < .05), the mean values of the two groups were then compared using the least significant difference (LSD) test (SAS Institute, 2003). Data of the survival rates and the percentages of the various aphid forms were log-transformed to meet the assumptions of normality and homoscedasticity required for these analyses. Then, the means were compared using the least significant difference (LSD) test (p < .05) when the overall variation in ANOVA was significant. All statistical analyses were performed using the SAS software (SAS Institute 2003).

| The reproductive mode of Rhopalosiphum padi from six populations
The percentages of aphid forms of clones collected from different regions and maintained under a low temperature and short photoperiod are shown in Table 1. Among the five generations, only alate and apterous parthenogenetic females, not sexuale (oviparae and males), were produced by clones collected in Taigu (Shanxi Province), Taian (Shandong Province), and Chuzhou (Anhui Province) (Table 1), indicating that the life cycle of R. padi from these regions is anholocyclic.
A holocyclic response consists of distinct reproductive periods.
Gynoparae are produced for the first time in the second generation, and their proportion of the total population reaches a peak in the third generation, after which it steadily declines. Males are produced by very few clones in the second generation, but the proportion increases thereafter and reaches a peak in the fifth generation. Oviparae are first produced in the third generation, after which production is similar to that of males. Alate and apterous parthenogenetic females are no longer produced in the fifth generation, respectively (Table 1). The clones from Baicheng of Jilin Province and Hami of Xinjiang Province were able to produce gynoparae, males, and ovipare and thus exhibited holocyclic life cycles.
Among the clones sampled from Lanzhou, eight produced only alate and apterous parthenogenetic females, not sexuale. Accordingly, these eight clones had an anholocyclic life cycle. Another seven clones produced only gynoparae, males, and oviparae (Table 1) (Table 1).
Among the three holocyclic populations, the developmental time of group LS nymphs was significantly longer than that of group NL nymphs (

| The effect of inducing conditions on the life-table parameters of anholocyclic and holocyclic populations
The effects of the inducing conditions on the intrinsic rate of increase (r), finite rate of increase (λ), mean generation time (T), net reproductive rate (R 0 ), and gross reproduction rate (GRR) of the R. padi populations from six regions are summarized in Table 3. Overall, the values of all five parameters differed significantly between groups NL and LS. In holocyclic populations, four values (r, λ, R 0 , and GRR) were significantly higher in group NL than in group LS, whereas the mean generation time (T) of group LS was longer than that of group NL. Among  significantly higher in anholocyclic than in holocyclic R. padi, whereas the latter had a longer mean generation time.
The two populations (GL-OP and Gl-CP) from the same region but with different life cycles were used to directly compare life-table parameters, including age-specific survival rate (l x ), age-specific fecundity (m x ), reproductive value (V x ), and life expectancy (E x ). The results are shown in Figure 1. In the GL-CP population, but not the GL-OP population, the curves of survival rate, fecundity, reproductive value, life expectancy, and female fecundity differed significantly between group NL and group LS.
The major life-table parameters of gynoparae and virginoparae produced by the three holocyclic populations are summarized in Table 5. Overall, the intrinsic rate of increase, finite rate of increase, fecundity, net reproductive rate, and gross reproduction rate varied significantly among the three populations. Interestingly, the values of these same parameters were significantly higher in virginoparae than in gynoparae at each location. The difference in the net reproductive rate, gross reproduction rate, and total fecundity was roughly fivefold higher in the GL-CP than in the GL-OP population.

| Comparison of male and oviparae produced by three holocyclic populations under normal and inducing conditions
When maintained on the secondary host, males and oviparae sur-  L1-L4 represents the developmental period of the first-, the second-, the third-, and the fourth-instar nymph stages, respectively; NL, the population which was fed in normal temperature and long photoperiod conditions; LS, the population which were induced by low temperature and short photoperiod conditions; p value of LSD test is shown in the last column. suggest that R. padi is able to adapt to local climatic conditions by investing in two overwintering strategies, overwintering eggs and parthenogenetic females, which provide a biological advantage and maximum energy use in terms of survival and spread. The maintenance of reproductive polymorphism was discussed in an earlier study (Halkett et al., 2008). The relative frequency of different life cycles varies geographically, with the proportion of holocyclic clones declining and that of anholocyclic clones increasing along a gradient from higher to lower latitudes. Life-cycle polymorphisms were also discovered for R. padi populations in Sweden (Wiktelius, 1982), the UK Tatchell, Plumb, & Carter, 1988), and France (Halkett et al., 2008;Hulle et al., 1999;Simon et al., 1996). The distribution of clones differing in their life cycles is probably maintained by powerful selective forces that prevent homogenization of the reproductive modes of different populations. These forces are predominantly ecological, mainly regional differences in winter severity (Dedryver et al., 2001;Duan et al., 2016;Halkett et al., 2004;Rispe, Pierre, Simon, & Gouyon, 1998;Simon et al., 1999). According to the model of , "sexual lineages" predominate in regions with cold winters because of the ecological advantage conferred by overwintering eggs, whereas essentially asexual lineages predominate in regions with mild winters, which support higher fecundity and fit-

| The effect of long-term inducing conditions on offspring of anholocyclic populations
Previous reports showed that holocyclic aphid clones maintained over the long term at a low temperature and short photoperiod could produce sexual generation ( 1998; Simon et al., 1999). By contrast, in this study, anholocyclic clones exposed to long-term inducing conditions could not produce males or oviparae, and the aphid offspring had a shorter adult lifespan, shorter total longevity, and reduced fecundity. These results indicate the existence of environmentally based maternal effects, even in anholocyclic clones.
Maternal heat stress was also shown to negatively affect the developmental time and nymphal birthweight of the G1 progeny (Ismaeil et al., 2013;Jeffs & Leather, 2014). Our results similarly suggest that exposing aphids to a low temperature and short photoperiod negatively impacts certain life-history traits and parameters of their progeny, mimicking the effects of seasonality.

| Comparisons of the life-history traits of gynoparae, virginoparae, males, and oviparae
In the autumn, gynoparae, which fly from the secondary host back to the primary host, are produced as a response to long nights and low temperatures. They differ from virginoparae in their olfactory responses at the peripheral level (Park, Elias, Donato, & Hardie, 2000) and show significantly larger electroantennogram responses to nepetalactol and nepetalactone than do alate and apterous virginoparae (Park & Hardie, 2002). This may explain the significantly reduced fecundity and shorter longevity of gynoparae than virginoparae, because of associated energy restrictions. Over the first 20 days of reproductive life, S. avenae virginoparae maintained under long-day conditions were significantly more fecund than gynoparae (Newton & Dixon, 1988). In our study, the chosen measure of fitness was the intrinsic rate of increase (r), which is a commonly used metric in invertebrates (Carter, Simon, & Nespolo, 2012; Emery, Rice, & Stanton, 2010). We showed that the intrinsic rate of increase in virginoparae was higher than that of gynoparae, which suggested the better adaptation of the former to varying environmental conditions.
When maintained on wheat seedlings, males and oviparae did not produce overwintering eggs under normal environmental conditions (24°C and a L:D cycle of 16:8 hr), but they were able to mate with each other and produce overwintering eggs under inducing conditions (12°C and a L:D cycle of 8:16 hr). These results suggest that the long photoperiod and high temperature accounted for the absence of overwintering egg. Additionally, one of the most important extrinsic factors determining aphid fitness is the host plant species (Wool & Hales, 1997). In nature, on the primary host, wingless oviparae produced by gynoparae mate with males and produce overwintering eggs.
However, diapause eggs are also generated on the secondary host.
These results point to the ecological advantages of holocyclic clones.
Specifically, holocyclic lineages can overwinter on the secondary host when the primary host is scarce or cannot be found by males and gynoparae.

| The short-term cost of sex
The survival rate of virginoparae increased, the development time decreased, and aphid fecundity increased in mild versus cold winters.
The short-term cost of sex was evidenced by the following: (1) the switch from asexual to sexual reproduction of holocyclic lineages in autumn spanned several generations; (2) the net reproductive and gross reproduction rates, as well as the total fecundity, were roughly fivefold higher in virginoparae than in gynoparae; (3) the adult prereproduction period of gynoparae was significantly longer than that of virginoparae; and (4) the total longevity of gynoparae was significantly shorter than that of virginoparae. Moreover, the longevity of males and oviparae was shorter under control than under inducing conditions, which suggests fewer opportunities for the mating of males and oviparae. Consequently, asexual lineages would have a > twofold advantage in terms of increasing their population numbers (Artacho, Figueroa, Cortes, Simon, & Nespolo, 2011).
Although cyclical parthenogenesis is the ancestral mode of aphid reproduction, these insects exhibit four distinct reproductive modes. Delmotte et al. (2001) reported that R. padi has multiple routes to asexuality, including the complete, spontaneous loss of sex and repeated gene flows from essentially asexual to sexual lineages. The high costs of sex under mild winter conditions might explain the complete, spontaneous loss of sex. In this study, we examined the effects of altering temperature and photoperiod on the life-history traits of R. padi. Our findings will be enhanced by studies of the evolution of the life cycles of R. padi.

| CONCLUSION
This study identified two different life cycles (holocyclic and anholocyclic) in six geographically distant populations of R. padi. Holocyclic and anholocyclic populations differed significantly with respect to their F I G U R E 2 Comparisons of the total longevity (mean ± SE) of males and oviparae under different temperatures and photoperiods; **significant differences between populations (p < .01, ANOVA followed by LSD tests) development, reproduction, and responses to low temperature and a short photoperiod. Specifically, the net reproductive rate, gross reproduction rate, and total fecundity were approximately fivefold higher in virginoparae than in gynoparae, whereas the adult prereproduction period was significantly longer in the latter. Together with the significantly shorter total longevity of gynoparae than virginoparae, our results provide evidence for the short-term cost of sex. Interestingly, oviparae were able to survive and produce overwintering eggs on a secondary host, which demonstrated the adaptive plasticity of this life form to primary and secondary hosts. However, there is still only scant information on the genetic basis of the different life cycles and life-history traits. Revealing the intrinsic mechanisms underlying these processes will be the next challenge in understanding the responses of R. padi to seasonal constraints.