Reproductive adaptation in alate adult morphs of the English grain aphid Sitobion avenae under starvation stress

Adapting their reproductive physiology is a tactic that insects use in responding to conditions of food unavailability. The present study examined the potential effects of starvation periods on the ovarian development and reproduction of alate adult morphs of Sitobion avenae (Fabricius). Morphs both continuously fed and starved aphids contained two telotrophic ovaries, each comprising five ovarioles. As time increase after emergence, the number of offspring produced by the fed aphids increased gradually, whereas the number of embryos in their ovaries decreased gradually. Both the number of mature embryos and the volume of embryos rapidly increased at 24 h after emergence, and then remained at an approximately constant level between 24 and 144 h. Compared to the fed aphids, starved aphids only produced a small number of nymphs, and there was no significant change in the total number of embryos between 24 and 144 h, whereas both the number of mature embryos and volume of embryos increased significantly. Irrespective of starvation period, highly significant relationships between life span and fecundity were found. Adult aphids starved for longer periods presented lower longevity and fecundity, but dead females contained more mature embryos than those starved for shorter periods. These results suggested that, under starvation stress, S. avenae tends to invest in the development of larger embryos at the expense of reducing lifespan and future fecundity. This adaptive reproductive strategy under starvation stress could be one of the factors contributing to the successful establishment of new colonies of alate migratory aphids.

Aphidius smithi (Hymenoptera: Aphidiidae) also resorbed small embryos 18 . These data indicate that viviparous aphids may adopt an alternative tactic when they are confronted to adverse conditions. Egg production in female insects is related to the nutrients available. Under starvation stress, resorption of egg chambers at stages 8 and 9 in female flies caused a reduction in the number of their eggs 19 . Starved Sitobion avenae (Hemiptera: Aphididae) adults exhibited a significantly lower fecundity than fed adults 20 . Further studies are required in order to confirm this reduction in fecundity is due to resorption of small embryos in aphids.
Sitobion avenae is a major pest infecting cereals worldwide; it feeds on phloem sap and transmits viruses 21 . This species prefers temperate climates and cannot overwinter in the north of China. Thus, the establishment of new colonies in this area mainly depends on the migration of alate aphids from south to north; this is supported by suction trap monitoring results 22 . During long-distance dispersal, alates most likely undergo starvation because they lack a suitable host. Often teneral alate adults are starved for different days and fed subsequently, as a result of settling down in their new host plants in the field. Gonadal status is a useful indicator in understanding the reproductive strategy of aphids responding to adverse conditions 23 . To our knowledge, ovarian development has not been examined in alate aphids under starvation stress; nor has its ecological significance in terms of reproductive success. The current study (i) examined ovarian development in the alate aphid S. avenae in relation to the presence or absence of food, and (ii) estimated the effect of different starvation periods and subsequent feeding on life span and total fecundity. Based on these results, the reproductive adaptive strategies possibly adopted by migratory aphids during delayed location of the host plants are discussed. The results obtained here will contribute to an understanding of the ecological significance of the successful establishment and continuous reproduction of migratory aphids in a new habitat, particularly regarding their high pest status.

Results
Effects of starvation on ovarian morphology. Ovarian morphology in the adult stage subjected to the starvation and feeding treatments is shown in Fig. 1. A few embryos with pigmented eyes were observed in the basal end of ovarioles in newly molted aphids (Fig. 1a). Fed aphids within 144 h of emergence contained two or three embryos with red eyes in the ovary ( Fig. 1b-g). Compared to fed aphids, basal embryos within each ovariole of starved aphids continued to develop gradually with increasing starvation time, meaning that there was an increase in the volume of basal embryos and the proportion of embryos with pigmented eyes (Fig. 1h-m). After starvation for 72 h, embryos at the base of each ovariole had almost developed red eyes ( Fig. 1j-m). In live adult aphids starved for 144 h, the largest embryos at the base of each ovariole presented with red eyes and appendages, and a few penultimate embryos within the ovarioles occasionally presented red eyespots (Fig. 1m). These results indicated that, as the starvation level increased, larger embryos continued to develop and smaller embryos were still presented.
Effects of starvation on ovarian development. All aphids dissected contained two telotrophic ovaries, each consisting of five ovarioles, as expected from ovarian developmental symmetry. With increasing time after emergence, there was a gradual increase in the number of offspring produced by the fed aphids (Kruskal-Wallis: H = 151.01, P < 0.001, Fig. 2c), whereas the number of embryos retained in their ovaries gradually decreased (Kruskal-Wallis: H = 34.62, P < 0.001, Fig. 2a). At 24 h after emergence, both the number of mature embryos and the average volume of the largest embryo in each ovariole rapidly increased. This contrasted with the results for the periods between 24 and 144 h, when their ovaries contained between four and five mature embryos and the largest embryos ranged in volume from 0.015328 ± 0.000870 to 0.018025 ± 0.000793 mm 3 (ANOVA:  Fig. 2d, respectively). Between 48 and 144 h after emergence, the total number of embryos and the number of mature embryos were significantly higher, but number of offspring produced was significantly lower in starved than in fed individuals at the same stage, respectively. Within 72 h of emergence, the largest embryos in the starved aphids presented a greater volume in ovarioles than in the fed aphids at the same stage after emergence. These results further indicated that there was continuous development of the largest embryos with increasing starvation time.

Relationships between adult mass, life span, and total fecundity. Adult mass was directly related
to life span of alate S. avenae adults continuously fed and starved for 72 h (Fig. 3a,d), but not directly related to life span of alate S. avenae adults starved for 24, 48, 96, 120, and 144 h (Fig. 3b,c,e-g). Adult mass was not directly related to fecundity ( Fig. 3h-n) of alate S. avenae adults either continuously fed, or starved for 24, 48, 72, 96, 120, and 144 h. However, relationships between life span and fecundity under the different treatments were determined to be highly significant (Fig. 3o-u).
As starvation time increased, adult life span decreased more rapidly, and dead females contained more mature embryos within their ovaries (Kruskal-Wallis test: H = 33.83, P < 0.001, Fig. 4). This result highlighted a marked tendency for females mortality to occur before the release of mature embryos.

Discussion
Starvation is a common challenge and is an important factor in inducing oosorption 1,24 . Many insect species have been reported to resorb large or specific stages of oocytes in their ovarioles during vitellogenesis under conditions of starvation stress; this leads to a reduction in the number of oocytes and egg load in the ovaries 3 ovariole increased significantly with increasing starvation time (Figs 1 and 2). Thus, larger embryos within ovarioles seem to develop continuously under starvation, in agreement with findings observed in M. viciae 16,17 . These data indicate that the resorption of large oocytes is not a generalized reaction in response to starvation stress in insects. Under starvation stress, the smallest embryos were resorbed in M. viciae 16 . This phenomenon was also observed in parasitized A. pisum under conditions of nutrient stress 18 . However, the present work revealed that there was no significant decrease in the total number of embryos with increasing starvation time in S. avenae (Fig. 2a), indicating that the smallest embryos within ovarioles might not be resorbed. The life span of S. avenae decreased with increasing starvation time, but more mature embryos were found in female bodies after death (Fig. 4). Thus, aphids might not benefit from the nutrients provided by the resorption of small embryos intended to ensure female survival. Starvation tolerance and embryo development mainly depend on the lipid content of the fat body 31 . The buffer ability of resorption and the reallocation of the smallest embryos are both extremely limited compared to that in the fat body 32 . It has long been suggested that apoptosis limits the reproductive potential of females in many animal groups 33 . In the present study, potential fecundity showed no significant reduction tendency based on the total number of embryos found in the gonads under starvation conditions (Fig. 2a). Thus, why did starvation stress lead to a significantly decreased fecundity? Indeed, under starvation stress, the larger embryos within each ovariole of S. avenae continued to develop (Fig. 2c). Reproductive investment increases to some extent at the expense of reduced lifespan 34 , which in turn leads to a reduction in reproductive potential. This hypothesis is compatible with the observation of the tendency for mortality to occur before release of the largest mature embryos in alate S. avenae adults experiencing starvation stress. The results obtained here indicate that the reduction in S. avenae fecundity was strongly correlated with reduced adult life span under starvation stress (Fig. 3o-u).
Ovarian response to starvation, which is vital for reproductive success, is highly variable among insect species under selection pressure 11 . Insects resorb large oocytes in order to store energy and reduce reproductive costs, therefore increasing their survival potential 15 . When resources improve, they resume the reproductive rate in order to enhance their future reproductive efforts 8,35 . Osawa (2005) 29 report that, under conditions of food availability, Harmonia axyridis (Coleoptera: Coccinellidae) presented both ovary development and resorption of oocytes. Moreover, the oosorption rate under starvation conditions was significantly higher than when food was available. This intermediate strategy might be advantageous when there are high fluctuations in food availability. The present study revealed that, under starvation stress, aphids might employ a further strategy, in which large embryos develop continuously in ovarioles, as shown for S. avenae (Fig. 2d).
Alate aphids, as host-selective morphs, exhibit a higher resistant to starvation and have a relatively greater number of olfactory sense organs than apterous individuals 20,36 . This phenomenon is beneficial to aphids in finding suitable host plants during dispersal. Reproductive physiology is a means by which insects can respond to resource-poor conditions 14 . Sitobion avenae is small and has a short life span. It shows a marked tendency to invest in offspring but not in survival. Under starvation stress, alate aphids display a high mortality rate whereas large embryos in each ovariole show continuous development until reaching a mature stage. However, they are rarely produced offspring. Live aphids survived a starvation treatment can deliver most offspring within a short

Methods
Insect rearing. One nymph of S. avenae was originally collected from the wheat fields at Yangling, Shaanxi, China (34°18′N, 108°5′E) in 2011 and successively maintained in the laboratory using the water-cultured wheat seedling method 40 . Aphids and wheat seedlings were cultured in a growth chamber at 21 ± 1 °C, 65 ± 10% relative humidity (RH), and a light:dark photoperiod of 16:8 h.
experimental setup. Alate aphids molted within the last 6 h were individually weighed using a microbalance (ME36S; Mettler Toledo, Schwerzenbach, Switzerland, 0.001 mg) and adult mass was measured as fresh mass. Subsequently, they were individually placed in Petri dishes containing only with water-saturated absorbent paper and starved for 24, 48, 72, 96, 120, or 144 h (30 individuals per treatment per experiment). Aphids at the same stage after emergence were fed on wheat seedlings as a control. Live aphids survived each treatment were used in the experiments.
Effects of starvation on ovarian morphology. In order to relate ovarian morphological changes to starvation levels, ovaries were dissected from live aphids that had not produced offspring, and put in a Ringer's solution on a concave slide, which was then examined under a binocular microscope (SMZ-140-FB; Motic, Shenzhen, China). Aphid starvation time points were for 24, 48, 72, 96, 120, or 144 h. We dissected fed aphids in order to collect their ovaries at the same stage after emergence. Ovaries were stained with 1% bromocresol green buffer for 20 sec, washed three times, each for 1 min in deionized water, and were then transferred to a slide containing one drop of deionized water. The morphology of the ovaries was evaluated according to the different starvation and feeding periods, and photographed using a binocular stereomicroscope (MZS0870; Leica Microsystems, Wetzlar, Germany).
Effects of starvation on ovarian development. Ovaries were removed from live aphids at the end of each treatment (number of treatments: 30 in 0, 24, 48, 72, 96, 120, or 144 h of feeding, respectively; and 30, 30, 29, 28, 18, and 8, in 24, 48, 72, 96, 120, or 144 h of starvation, respectively). A single drop of 1% bromocresol green was used to facilitate visualization. The number of ovarioles, total and mature embryos, and the average volume of the largest embryos in ovarioles were recorded for each dissected aphid, as these were considered to be the parameters that best describe the status of ovarian development. Embryos were considered as mature if the eyes of the embryo were pigmented 39 . Length (l) and breadth (b) of the largest embryo in each ovariole were measured under a binocular microscope equiped with an ocular micrometer. The accuracy for the length (l) and breadth (b) measures was 0.01 mm. The volume of the largest embryo in each ovariole was calculated using (1 × b 2 mm 3 ) according to Newton and Dixon (1990) 41 .