Mechanisms that Favor the Continuity of the African Honeybee Genome in the Americas

StanleyS. Schneider1 and Gloria DeGrandi-Hoffman2

1 Department of Biology, University of North Carolina, Charlotte, NC 28223 USA,

2 Carl Hayden Bee Research Center, Tucson, AZ 85719 USA,

Since its introduction into Brazil in 1956, the African race of honeybees, Apis mellifera scutellata, has spread throughout Central and South America and the southwestern U.S. The colonization of much of the western hemisphere in less than 50 years by a single race of insect is one of the most rapid and spectacular biological invasions known. One of the most remarkable features of the African bee is its ability to displace European honeybee subspecies in the New World. Although African and European bees will hybridize, gene flow between the populations is highly asymmetrical. There is substantial introgression of African alleles into European populations, but comparatively little reciprocal gene flow. Despite almost 50 years of contact between African and European bees, the frequency of European mitochondrial and nuclear alleles in colonized areas rarely exceed 20% and 35%, respectively, in managed colonies and are typically much lower in feral populations (10). Although the introduced bee is no longer genetically identical to the honeybees of south Africa, it has remained predominantly African in its nesting biology, foraging behavior, diet selection, swarming and absconding behavior and genetic characteristics (10). For brevity, we therefore refer to the descendents of A. m. scutellata in the Americas as African bees.

No single factor determines the ability of African bees to displace European bees. The continuity of the African genome in the Americasarises from a complex interaction of population dynamics, genetic phenomena, and physiological and behavioral mechanisms. In this presentation, I focus on four factors that we have studied recently: 1) African-patriline advantages during queen replacement; 2) differential use of African and European sperm by queens; 3) nest usurpation by African swarms; and 4) decreased developmental stability in hybrid workers. Each of these factors may be involved in the asymmetrical gene flow between African and European populations and contribute to the prevalence of African bees in areas that were once dominated by European honeybees.

African-patriline Advantages During Queen Replacement: When African bees colonize an area, queens mate with a combination of African and European drones, which results in colonies composed of African- and European-patriline workers. During queen replacement, these colonies will rear virgin queens (VQs) from both patrilines. Any advantages associated with African paternity could contribute to African-patriline VQs becoming the new laying queens of their colonies, and a loss of European paternal alleles. There are at least three ways in which African paternity could influence queen success: 1) the preferential rearing of African-patrline VQs by workers; 2) greater fighting success by African-patriline queens after they emerge; and 3) preferential interactions of workers with emerged African-patriline VQs in a manner that promotes their survival.

We examined VQ behavior and success in “hybrid” colonies created by instrumentally inseminating African and European queens with equal mixed volumes of semen from one African and one European drone (7,8). This resulted in African (A)- and European (E)-matriline colonies, each of which contained African- and European-patriline workers. The European queens and drones used in the inseminations expressed the “light blond” cordovan (cd) coloration. Workers and queens arising from the cd drones (which were of European paternity) showed yellow cuticular coloration patterns that allowed us to easily distinguish them from the darker workers and queens of African paternity. We established our colonies in observation hives and then removed the inseminated queens, which initiated queen replacement in which VQs of both patrilines were reared. We then monitored queen rearing activity, VQ behavior and worker-queen interactionsthroughout the queen replacement process in each of 7 A-matriline and 6 E-matriline observation colonies.

We found no evidence that workers preferentially reared African-patriline queens (7). Both the A- and E-matriline colonies reared similar numbers of African- and European-paternity VQs to emergence (A-matriline: 12 African- and 6 European-patriline VQs; E-matriline: 5 African- and 7 European-patriline VQs; X2 = 0.53; df = 1; P>0.05). Also, in both matrilines, workers visited and incubated cells containing African- and European-patriline queens at similar rates (Table 1). Workers may influence the development and emergence of queen cells by performing “vibration signals” on them, but in our colonies workers did not preferentially perform vibration signals on African-paternity cells (Table 1).

Table 1. The mean + SE rate at which workers visited, incubated and performed vibration signals on queen cells that gave rise to African- and European-paternity VQs. A total of 16 African-paternity cells and 12 European-paternity cells were monitored. Rates refer to the number of worker interactions occurring per 30-min period.

African-paternityEuropean-paternity

VQsVQsP

Visitation Rate4.52 + 0.414.31 + 0.51>0.05

Incubation Rate7.24 + 0.607.81 + 0.69>0.05

Vibration Rate0.98 + 0.211.16 + 0.23>0.05

Thus, there were no advantages associated with African paternity during the queen-

rearing phase of queen replacement. However, paternity did influence the order in which VQs emerged. Of the 13 first-emerging VQs in our observation colonies, 9 were of African and 4 of European paternity. Although this difference was not significant (X2 = 1.92; df = 1; P >0.05), our results suggest that African-patriline VQs develop faster and emerge sooner than their European-paternity sister queens. Because first-emerging queens have a competitive advantage during queen rivalry (8,9), this could promote the success of African-patriline VQs.

We found strong evidence that African paternity was associated with increased VQ fighting success (8). This was especially true in our E-matriline colonies, in which African-paternity VQs killed more rivals and produced more bouts of “piping” then did their European-patriline sister queens (F1,15 > 7.81; P < 0.0136 for both comparisons; Fig. 1). Piping is a pulsed sound that may inhibit the emergence of rivals and contribute the VQ success (9). In contrast, in the A-matriline colonies, African- and European-patriline VQs did not differ in the number of rivals eliminated or piping rates (F 1,15 < 4.39; P > 0.05 for both comparisons; Fig. 1). The weight of African- and European-paternity VQs did not differ in either matriline (P>0.05). Thus, any fighting advantages associated with African paternity in the E-matriline colonies did not arise because of size differences.

African paternity also influenced the degree to which workers interacted with VQs after they emerged (8). Worker often perform vibration signals on emerged VQs, sometimes at rates of several hundreds of signals per hour (9). VQs that are vibrated at higher rates produce more bouts of piping, eliminate more rivals, survive longer and are more likely to become the new laying queens of their colonies (9). The vibration signal may therefore give workers a degree of control over VQ fighting and may be used to promote the success of certain queens. In our E-matriline colonies, African-paternity VQs were vibrated at a rate 4-6 times greater than that of their European-paternity sister queens (F 1,15 = 9.03; P = 0.0089; Fig. 1). In contrast, in the A-matriline colonies, there was no difference in the vibration rates experienced by the two VQ types (F 1,15 = 3.32; P > 0.05; Fig. 1).

In both our A- and E-matriline colonies, more African- than European-paternity VQs survived to become laying queens (A-matriline: 5 African- vs. 1 European-patrline VQs; E-matriline: 5 African- vs. 2 European-patriline VQs). This difference was significant in the E-matriline colonies (Fisher’s exact test; P < 0.05), and approached significance in the A-matriline colonies (P > 0.05). Thus, the combination of earlier emergence, increased fighting success and greater vibration rates experienced by African-patriline VQs may have resulted in an increased likelihood that they would become the new queens of their colonies. We do not fully understand why the effects of African paternity were more pronounced in the E-matriline colonies. Perhaps African traits are dominant for the queen characteristics that we examined. In our A-matriline colonies, all queens may have been African-like, because the genome of each was at least 50% African. If so, then the variability in fighting success, piping activity and vibration rates observed in the A-matriline colonies may have reflected behavioral differences between essentially African queens, rather than differences between patrilines. In contrast, in the E-matriline colonies, VQs were either pure European or half African. A dominance of African traits would therefore have conveyed an advantage to only the African-patriline queens.

Our results suggest that queen replacement in colonized areas may result in the inevitable loss of European paternal characteristics. Hybrid colonies are likely to become headed by an African-patriline VQ when they raise new queens. These VQs will then mate disproportionately with African drones, because of high African colony density and greater drone production compared to European colonies (10). There will therefore be an increasing introgression of African paternal alleles over consecutive queen replacement events. An African-patriline advantage for queens would therefore promote the “Africanization” of managed European colonies and could act as a barrier to the introgression of European paternal alleles into African colonies.

Differential Use of African and European Sperm by Queens: When African bees become established in an area, queens mate disproportionately with African drones, resulting in the rapid displacement of European paternal alleles. This occurs in part because of the numerical superiority of African drones (10). However, even if African and European drones occur in equal numbers in mating areas, an African-drone advantage could persist if queens preferentially used African sperm. We tested this idea by instrumentally inseminating African and European queens with equal, mixed volumes of semen from African and European drones and then determining the patriline of the workers produced, using the cuticular color markers described above. We examined sperm usage in a total of 11 African and 10 European inseminated queens over a 6-month period.

If the queens used the sperm of both drone types equally, then we expected that 50% of the resulting workers would be of African paternity. However, when viewed over the entire 6-mo. study period, the African and European queens produced 65% and 61% African-paternity workers, respectively, and this difference was highly significant for both matrilines (X2 > 119.0; df = 1; P < 0.001 for each comparison)(1). The brood-rearing patterns in our study colonies indicated that there was little larval mortality. Thus, it is unlikely that the greater production of African-patriline workers resulted from subviability of European-paternity brood or the preferential rearing of African-patriline larvae by workers. Furthermore, the unequal use of African sperm did not occur during the first and last months of the study, during which the proportions of European- and African-patriline workers did not differ from the expected 50% (1). It is therefore unlikely that the overrepresentation of African-patriline workers resulted from reduced fertilization success of European sperm or sperm clumping in the queen’s spermathecae. Taken together, our results suggest that when queens mate with a combination of African and European drones, they may use African sperm disproportionately, which would further contribute to an African-drone mating advantage and the displacement of European paternal alleles in invaded areas.

Nest Usurpation by African Swarms: Nest usurpation (colony takeover) is a form of social reproductive parasitism in which small African swarms invade European colonies, replace the resident queens and cause the complete and instantaneous loss of European matrilines. Nest usurpation has frequently been proposed to contribute to the predominance of African-matrilines in invaded areas. However, the role of nest usurpation in the spread of African bees remains controversial. Annual usurpation rates reported for managed European colonies in Latin America vary from 0%-40% and often show pronounced year-to-year and seasonal variability (10,12). We also have a poor understanding of the factors that contribute to the susceptibility of European colonies to usurpation. Colonies that are queenless, contain a failing queen, or contain a queen confined in an ‘excluder cage’ (a common management practice for introducing new queens) may have an increased risk of takeover (10,12). Thus, cues associated with compromised colony condition or queen performance may be used by usurpation swarms to locate susceptible hosts. Nevertheless, strong queenright European colonies can also be usurped, which suggests that additional factors may also influence the likelihood of takeover. Nest usurpation remains one of the least understood aspect of the African bee invasion process and has received little systematic study in the U. S., even though the southwestern states and parts of California now contain large, expanding African populations.

We examined nest usurpation in an apiary of European colonies maintained over a 2-year period in Tucson, AZ. The African bee arrived in Arizona in 1993 and has subsequently established large populations that contain few or no European matrilines (4). The study apiary therefore provided an excellent opportunity to examine the impact of nest usurpation on managed European colonies that are increasingly challenged by a growing feral African population. Our apiary contained 76 five-frame nucleus hives, each with a “Golden Italian” European colony. The Golden Italian line expresses the light yellow coloration of cordovan bees. Each European queen was marked with a dot of point on her thorax. The coloration patterns and paint marks allowed us to visually discriminate European host queens and workers from the much darker colored African bees in the Tucson area.

We observed a total of 32 takeovers during our 2-year study period, for a mean annual usurpation rate of 21.1 + 9.3% (12). We found pronounced seasonal patterns for usurpation activity (Fig. 2). Usurpations occurred throughout the year, except for January and February. Peak usurpation activity occurred from October-December. The fall-winter period accounted for 56% of the total usurpation events observed and had mean monthly usurpation rates of 3%-5%. A secondary peak of usurpation activity occurred from April-August. The spring-summer period accounted for 37.5% of colony takeovers and had mean monthly usurpation rates of 1%-2% (12). These seasonal patterns may have been related to the annual colony cycle of African bees in southern Arizona. The minor peak of usurpation activity in spring-summer coincided with the primary swarming season in the Tucson area. In contrast, the major peak in fall-winter may have coincided with seasonal absconding by African colonies. Seasonal absconding is a non-reproductive process that consists of an entire colony abandoning a nest in response to

deteriorating foraging conditions and traveling up to 100 km to relocate to an area with greater floral resources (5,13). The mountains surrounding the Tucson basin harbor a large population of feral African colonies (4) and forage availability at these higher elevations declines during the fall and winter months. This may result in large-scale absconding into the Tucson basin, where horticultural and agricultural activity result in increased floral abundance during this period. Many usurpation swarms may therefore be small reproductive or absconding African swarms that would have a low probability of surviving if they attempted to establish their own

nests, especially during fall and winter. Rather, these swarms may adopt a strategy of invasion and reproductive parasitism, and this tendency may be particularly pronounced during seasonal absconding. Nest usurpation may therefore be an important component of the annual colony cycle that contributes to the spread of African matrilines in the Tucson region (12).

We also found that queen condition had a strong influence on host colony susceptibility to takeover. We monitored colony strength and queen condition throughout the study period by conducting bi-weekly inspections. During the inspections, each colony was classified as queenless, thriving, weak or invaded. Colonyclassifications were based on worker population size and brood rearing patterns. Weak colonies were re-queened with Golden Italian queens confined under wire-mesh push-in cages. Colonies were considered to have been invaded if we found a dark African queen instead of (or in addition to) a light colored, paint-marked European queen. When viewed over the entire study period, the usurpation rates (expressed as % colonies invaded/mo.) of queenless and caged-queen colonies (12.78 + 6.25 and 5.50 + 2.70, respectively) were 2-8 times greater than those of the thriving (1.61 + 1.15) and weak (2.23 + 1.43)colonies (P = 0.031). The association between queen condition and host susceptibility was particularly strong in the fall-winter months, during which the usurpation rates of queenless and caged-queen colonies were 6-20 times greater than those of the other colony types (P = 0.0048)(12).

Thus, usurpation may play an important, but seasonally variable role in the loss of European matrilines in the southwestern U.S. The associations that we observed between queen condition, season, and usurpation activity may have implications for the methods used to maintain strong European colonies in invaded regions. The most common method used for maintaining European colonies is re-queening annually with mated European queens. Our study indicates that re-queening colonies during certain times of the year could actually increase the vulnerability to becoming African due to usurpation. Identifying when swarming and absconding by feral African colonies occurs in a given area and avoiding re-queening at those times could be critical to maintaining European matrilines.

Developmental Stability of Hybrid Workers: Even when hybridization occurs between European and African bees, European characteristics could be lost if hybrid workers have reduced viability and survival. Hybrid workers have been suggested to have lower mass specific metabolic rates compared to African bees and to be slightly less efficient foragers (2,6). Indeed, hybrid inferiority has repeatedly been proposed to contribute to the loss of European mitotypes in Latin America and the U. S. (2,10). Yet, hybrid inferiority remains one of the more controversial aspects of the African bee invasion process.