Report

Asymmetric Mating Interactions Drive Widespread Invasion and Displacement in a Whitefly

See allHide authors and affiliations

Science  14 Dec 2007:
Vol. 318, Issue 5857, pp. 1769-1772
DOI: 10.1126/science.1149887

Abstract

The role of behavioral mechanisms in animal invasions is poorly understood. We show that asymmetric mating interactions between closely related but previously allopatric genetic groups of the whitefly Bemisia tabaci, a haplodiploid species, have been a driving force contributing to widespread invasion and displacement by alien populations. We conducted long-term field surveys, caged population experiments, and detailed behavioral observations in Zhejiang, China, and Queensland, Australia, to investigate the invasion process and its underlying behavioral mechanisms. During invasion and displacement, we found increased frequency of copulation leading to increased production of female progeny among the invader, as well as reduced copulation and female production in the indigenous genetic groups. Such asymmetric mating interactions may be critical to determining the capacity of a haplodiploid invader and the consequences for its closely related indigenous organisms.

Biological invasions threaten agricultural and natural systems throughout the world (1). Invasive animals often thrive at the expense of indigenous, closely related organisms, and insight into the causes of animal invasions often hinges on detailed assessments of behavioral mechanisms (2). Closely related species often have incompletely isolated mate recognition systems conducive to reproductive interactions and interference (3), but rarely have such behavioral mechanisms been isolated and tested in an experimental setting to reveal their contribution to invasion biology (1). Here, we combined long-term field monitoring, caged population studies, and detailed behavioral observations to investigate the mechanisms underlying the widespread, rapid invasion by a genetic group of the whitefly Bemisia tabaci (Gennadius) (Hemiptera: Aleyrodidae).

The whitefly B. tabaci, a haplodiploid species, is a genetically diverse group including many morphologically indistinguishable populations that differ in biological characteristics but display clear geographic distributions, with indigenous populations that span the globe between 30°N and 30°S (46). A recent phylogenetic analysis of B. tabaci mitochondrial CO1 and ribosomal ITS1 DNA sequences supports the existence of 12 “major genetic groups” (5). The available data indicate reproductive compatibility within each of the major groups but reproductive isolation between them (7). As species-level distinctions among the members of B. tabaci remain elusive, we use the term “biotypes” to refer to individuals representing the different genetic groups involved in our description of between-group interactions.

A major recent event associated with B. tabaci has been the invasion by individuals commonly known as the B biotype [also called B. argentifolii (8)] from their presumed origin in the Mediterranean–Asia Minor region to much of the rest of the world, with a resulting impact that has led to it being described as one of the top 100 invasive species (46, 9, 10). The invasive B biotype has caused considerable damage to crops through phloem feeding, transmission of plant viruses, induction of phytotoxic disorders, and excretion of honeydew (4, 6). Circumstantial evidence indicates that in many regions, introduction of the B biotype has led to the displacement of some relatively innocuous, indigenous B. tabaci belonging to different genetic groups (4, 1113). However, the mechanisms for displacement of indigenous populations are not clear, although the role of some form of mating interference has been inferred (11, 14, 15).

The B biotype entered China in the mid-1990s (16) and Australia in the early 1990s (17) (Fig. 1). We conducted regular sampling of the field whitefly populations in Zhejiang, China, from 2004 to 2006 and in Queensland, Australia, from 1995 to 2005 to monitor the process of invasion and displacement by B in the two regions (18). The data indicate that before invasion, the ZHJ1 biotype, which is indigenous to China (5), was widely distributed in Zhejiang, and the AN biotype, which is indigenous to Australia (5), was widely distributed in Queensland (Fig. 1). Phylogenetic analysis shows that AN, B, and ZHJ1 belong to different major genetic groups (i.e., Australia, Mediterranean–Asia Minor–Africa, and Asia II, respectively) (5). Moreover, mating experiments between B and ZHJ1 and between B and AN indicated that neither could interbreed to produce fertile female offspring (14, 19).

Fig. 1.

Changes of the mean proportions of the exotic B biotype and indigenous biotypes of Bemisia tabaci after introduction of B. (A) ZHJ1 versus B on cotton at seven locations in Zhejiang, China, from 2004 to 2006. The map covers the area 122°E to 119°E from east to west and 27°30′N to 30°N from south to north. Of the seven locations, Jiande is in a western mountainous area with less transport activity relative to the other six locations along the east coast, where transport of vegetables and ornamental plants is frequent. (B) AN versus B on Sonchus oleraceus at 17 locations in Queensland, Australia, from 1995 to 2005. The map covers the area 153°30′E to 142°E from east to west and 34°S to 14°40′S from south to north. Data from the other 17 locations are not depicted because of space.

In both regions, the process of invasion and displacement agrees in general with earlier circumstantial evidence collected from other regions of the world (4). Thus, in Zhejiang, the earliest invasions by B occurred in locations with the most frequent transport of ornamentals (i.e., Cixi, Ningbo, and Wenzhou), whereas in Australia the incursion was first associated with a major importing wholesale nursery in New South Wales and from there to major wholesale nurseries in Cairns and Brisbane in Queensland (Fig. 1). Rapid and widespread displacement of the indigenous population has occurred in both regions. In Zhejiang, displacement of ZHJ1 was complete by 2005 in three locations and is currently occurring in other locations; in Queensland, displacement of AN was complete in all locations by 2005 or earlier, and in any particular locationitusually took 3 to 5 years (Fig. 1).

In parallel with the displacement, we detected significant changes in sex ratio in both the indigenous and alien populations in both regions. In Zhejiang, when populations of either B or ZHJ1 occurred alone, B usually had female ratios of 60 to 70%, which were higher than the 50 to 60% female ratios in ZHJ1 (Fig. 2A and fig. S1). This difference in sex ratio between the two was enhanced by the relative increase of females in B and decrease of females in ZHJ1 when B and ZHJ1 occurred together during displacement (Fig. 2A and fig. S1). Similar differences in the sex ratio between AN and B when each of them occurred alone, as well as a similar shift in sex ratio when they co-occurred, were observed in Queensland (Fig. 2B and fig. S2).

Fig. 2.

Changes in mean sex ratios in field populations of the exotic B biotype and indigenous biotypes before, during, and after displacement of the indigenous biotypes by B. (A) Female ratios of ZHJ1 and B on cotton from 2004 to 2006; for locations, see Fig. 1. The data set includes 13 samples for ZHJ1 before invasion, 26 samples for ZHJ1 and B during invasion and displacement, and 28 samples for B after displacement. (B) Female ratios of AN and B on spurge from 1995 to 2005; for locations, see Fig. 1. The data set includes 74 samples for AN before invasion, 66 samples for AN and B during invasion and displacement, and 77 samples for B after displacement. One-way analysis of variance (ANOVA) and Fisher protected least significant difference (LSD) tests were applied to each of the data sets, based on proportion data transformed by arcsine square root. For (A), F3,89 = 224.14, P < 0.001; for (B), F3,280 = 81.96, P < 0.001. Different letters above bars indicate significant differences; error bars indicate SD.

We then simulated the process of displacement in controlled laboratory and field cage experiments (18). In Zhejiang, we reared mixed cohorts of B and ZHJ1, as well as control cohorts of either B or ZHJ1 alone, on caged cotton plants for 225 days (about 10 generations for the earliest-born individuals in every generation) and monitored the changes in relative proportion as well as sex ratio. The mixed cohorts began with 13% B and 87% ZHJ1. The relative proportion of B increased steadily with time and had totally supplanted ZHJ1 after 225 days (fig. S3). Each cohort was initiated with 50% females, and divergences in sex ratio among the different cohorts occurred with time. In cohorts with B only, the female ratio increased in the first 50 days and remained at 60 to 70% thereafter; in cohorts with ZHJ1 only, the female ratio remained at 50 to 60% throughout. In contrast, the female ratio of B in the mixed cohorts increased and reached 70 to 80% from days 50 to 150 when displacement was actively taking place, while that of ZHJ1 decreased to mostly 30 to 40% (fig. S3). In Queensland, we reared mixed cohorts of AN and B, where B made up 10% of the females, as well as control cohorts of either AN or B alone, on spurge plants (Euphorbia cyathophora) for five generations in field cages (18). In the AN-only cohorts, females made up 56% of the population, whereas 62% of the B-only cohorts were females. However, in the mixed cohort, AN females declined to 22% while B females increased to 76% by the fifth generation (table S1).

Whiteflies, including B. tabaci, exhibit haplodiploidy, producing males from unfertilized eggs and females from fertilized eggs (20, 21). In AN, B, and ZHJ1, both females and males start mating soon after emergence and both sexes mate repeatedly (19, 21). We found that for both B and ZHJ1, continuous availability of males to females was essential if the usual level of egg fertilization was to be maintained, and that deprivation of males from mated females resulted in rapid reduction in fertilization and an increase in the proportion of unfertilized eggs leading to male progeny (18) (fig. S4). When males and females of B and one of the indigenous biotypes were placed together, they frequently exhibited courtship but never copulated. Furthermore, when two males from different biotypes and a female of a given biotype were placed together, the female was frequently courted by both males and courting and copulation could be interrupted by the second male (19, 21). We thus suspected that changes of sex ratio were related to changes in male availability and copulation events due to interactions between the invader and indigenous individuals.

We next conducted a series of observations to identify the behavioral mechanisms behind the changes in sex ratios. This was done by placing a newly emerged adult female and male of a given biotype on a leaf in a clip cage and supplementing the pair with none, one, or three males of the same or different biotype. Using a video recording system (21), we were then able to record the movement and every copulation event that occurred over the following 72 hours (18). In parallel with these observations, we examined the progeny production by females of each biotype over the 5-day period after emergence, using the same numbers and combinations of males as in the observations on copulation events (18).

When a pair of B individuals on cotton was supplemented with one male of either B or ZHJ1, or with three males of B, the numbers of copulation events of the B females increased, as did the percentage of female progeny, but the numbers of progeny produced remained unchanged; however, the addition of three ZHJ1 males did not affect either copulation events or progeny sex ratio (Fig. 3, A and B). When a pair of ZHJ1 individuals was supplemented with one or three ZHJ1 males, the number of copulation events of the ZHJ1 females remained unchanged; in contrast, when a pair of ZHJ1 was supplemented with one or three B males, the numbers of copulation events as well as the percentage of females in the progeny decreased (Fig. 3, C and D). The observations made with B and AN on spurge showed that the responses of B to the addition of either B or AN males were similar to those of B on cotton when interacting with ZHJ1, in terms of both number of copulation events and progeny sex ratio (Fig. 4, A and B). Likewise, the responses of AN to the addition of either AN or B males were similar to those of ZHJ1 interacting with B (Fig. 4, C and D). Identification of the progeny produced using nuclear DNA markers detected no hybrids, demonstrating reproductive isolation between B and ZHJ1 and between B and AN (18). These results showed that B interacting with an indigenous biotype could increase production of female progeny by increasing its frequency of copulation, and at the same time could reduce the production of female progeny by the indigenous females through reducing copulation by the latter.

Fig. 3.

Changes in the mean number of copulation events during the first 72 hours after emergence and production of progeny for the first 5 days after emergence when a pair of B biotype ♂×♀ was supplemented with one or three ♂ of the B or the indigenous ZHJ1 (Z) biotype (A and B), or when a pair of ZHJ1 biotype ♂×♀ was supplemented with one or three ♂ of the ZHJ1 or B biotype (C and D). We conducted 15 to 50 replicates for each of the treatments. One-way ANOVA and LSD tests were applied to each of the four data sets, and the data in (B) and (D) were transformed by arcsine square root before analysis. For (A), F4,106 = 5.67, P < 0.001; for (B), F4,199 = 3.02, P = 0.019; for (C), F4,95 = 8.02, P < 0.001; and for (D), F4,201 = 15.28, P < 0.001. Means ± SEM of the number of progeny produced by B females in the five treatments in (B) were 20.8 ± 1.5, 24.1 ± 1.5, 20.3 ± 1.4, 18.4 ± 1.3, and 19.1 ± 1.7, respectively, and the means did not differ significantly (F4,199 = 2.31, P = 0.059). Means ± SEM of the number of progeny produced by ZHJ1 females in the five treatments in (D) were 19.6 ± 1.9, 22.7 ± 1.7, 17.5 ± 1.6, 18.1 ± 1.7, and 17.7 ± 1.5, respectively, and the means did not differ significantly (F4,201 = 1.74, P = 0.143). Different letters above bars indicate significant differences; error bars indicate SE.

Fig. 4.

Changes in the mean number of copulation events during the first 72 hours after emergence and production of progeny for the first 5 days after emergence when a pair of B biotype ♂×♀ was supplemented with one or three ♂ of the B or the indigenous AN biotype (A and B), or when a pair of the AN biotype ♂×♀ was supplemented with one or three ♂ of the AN or B biotype (C and D). We conducted 15 to 30 replicates for each of the treatments. One-way ANOVA and LSD tests were applied to each of the four data sets, and the data in (B) and (D) were transformed by arcsine square root before analysis. For (A), F2,42 = 3.28, P= 0.048; for (B), F4,115 = 15.76, P < 0.001; for (C), F2,57 = 3.24, P = 0.047; and for (D), F4,133 = 7.16, P < 0.001. Means ± SEM of the number of progeny produced by B females in the five treatments in (B) were 36.7 ± 2.7, 37.8 ± 2.9, 35.0 ± 2.8, 38.4 ± 2.7, and 36.2 ± 2.6, respectively, and the means did not differ significantly (F4,115 = 0.248, P = 0.910). Means ± SEM of the number of progeny produced by AN females in the five treatments in (D) were 44.2 ± 2.7, 44.8 ± 1.7, 41.5 ± 3.0, 41.0 ± 3.2, and 47.7 ± 3.0, respectively, and the means did not differ significantly (F4,133 = 0.83, P = 0.505). Different letters above bars indicate significant differences; error bars indicate SE.

Detailed analysis of the movement of whitefly adults revealed behavioral elements related to the increase and decrease in copulation events (18). In B and ZHJ1, both females and males of the two biotypes differed in mating behavior and interactions in three important ways (tables S2 and S3). First, in the absence of interference by a second male, males of the two biotypes courted at a similar frequency, but B females accepted courtships leading to copulation more frequently than did ZHJ1 females. Second, when a second male of the same biotype was present in the same arena, males of both biotypes increased their frequency of uninterrupted courtships despite some interference between the males, but the responses of females differed: B females increased their acceptance of courtships, leading to more events of copulation, whereas ZHJ1 females did not. Third, when males of two biotypes were present with a female of a given biotype in the same arena, the B male and female responded by increasing their frequency of courtships, leading to more copulation events, whereas the ZHJ1 male and female did not do so. Moreover, although courtships between the two biotypes occurred, copulation never resulted; this confirmed that both B and ZHJ1 share incompletely isolated mate recognition systems. Further, B males not only courted females of either biotype more frequently than did ZHJ1 males, they also interfered more frequently with the courtships initiated by ZHJ1 males than did ZHJ1 males with courtships initiated by B males (tables S2 and S3). The mating behavior and interactions between B and AN differed in ways similar to what we found for B and ZHJ1, although details varied between the two combinations (18) (tables S4 and S5).

These results help to explain the underlying basis of the B biotype's capacity to invade and displace indigenous populations. The strong competitive ability of B results partly from its capacity to adjust sex ratio in favor of its population increase, and partly from its capacity to interfere with the mating of indigenous individuals. When the proportion of males is increased, B adults respond by increasing the frequency of copulation and consequently increasing the proportion of female progeny. Critical to this is that B responds independently of whether the males are all B or a mix. This interaction is extraordinary because the indigenous males actually help to promote copulation among the invaders and consequently increase the invaders' competitive capacity. In contrast, the indigenous females do not respond to increased numbers of adult males. Moreover, copulation by indigenous individuals is partly blocked by B males that readily attempt to court with females of either biotype—a behavior not reciprocated by the indigenous males. These asymmetric mating interactions have obvious population-level implications because the increase in the proportion of B females and the concomitant decrease in the proportion of indigenous females results in an immediate higher population growth rate for B and a lower growth rate for the indigenous population. As the abundance of B increases relative to the indigenous individuals, the increased allocation of eggs to female progeny and the active interference of mating of indigenous males by B males combine to drive the indigenous population to local extinction.

Mating interactions between closely related but reproductively isolated genetic groups are likely a common phenomenon (3, 2224) and are expected given the widespread existence of hybridization and introgression (25). Although examples of asymmetric competition are well known (2628), asymmetric mating interactions are less well described (28). The rarity of examples may be, as illustrated by this study, the consequence of such interactions leading to the rapid displacement of the disadvantaged organisms. Biological invasions offer opportunities to gauge and characterize the potential magnitude and form of asymmetric mating interactions before species are lost through competitive exclusion or before the importance of competition is reduced over evolutionary time through niche partitioning and character displacement.

Allopatric species often demonstrate greater similarity in mating signals than do sympatric species, even when they have been diverging for a similar length of time (3). As a consequence of biological invasions, previously allopatric species are brought together and their partially similar mate recognition systems may promote asymmetric mating interactions between them. As we have shown, these interactions may play a critical role in determining the capacity of the invader to establish itself and the consequences for indigenous species.

Supporting Online Material

www.sciencemag.org/cgi/content/full/1149887/DC1

Materials and Methods

SOM Text

Figs. S1 to S4

Tables S1 to S5

References

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article