Quantitative Trait Loci that Influence the Expression of Guarding and Stinging Behaviors of Individual Honey Bees

Transcription

1 Behavior Genetics, Vol. 33, No. 3, May 2003 ( 2003) Quantitative Trait Loci that Influence the Expression of Guarding and Stinging Behaviors of Individual Honey Bees Miguel E. Arechavaleta-Velasco, 1,2 Greg J. Hunt, 1 and Christine Emore 1 Received 11 Sept. 2002Final 11 Dec This study was conducted to test for the effect of three stinging behaviors QTLs (sting-1, sting- 2 and sting-3) on the expression of guarding and stinging behavior of individual honey bees, and to determine if results of defensive behavior QTLs found in studies with Africanized honey bees could be extended to other populations of bees. Samples of guards, stingers, foragers and nurse bees were taken from two backcross colonies derived from a defensive colony and a gentle colony. The genotype of each bee for both types of colonies was determined for two sequence tagged site (STS) markers linked to sting-1 and for another two STSs, one linked to sting-2 and one linked to sting-3. Results showed that sting-1 had an effect on the expression of both stinging and guarding behaviors, sting-2 and sting-3 influenced the expression of guarding behavior. These results indicate that division of labor is influenced by specific QTLs. Results also show that QTLs mapped in a population of Africanized honey bees using colony level phenotypes also influenced the expression of guarding and stinging behavior of individual bees of other populations. KEY WORDS: Honey bees; Apis mellifera; QTLs; guarding behavior; stinging behavior; defensive behavior; division of labor. INTRODUCTION Honey bee colonies are composed of one queen, which is the only reproductive female, thousands of workers, which are non-reproductive females and hundreds of drones that are males. Female honey bees are diploid, but males are haploid that develop from un-fertilized eggs through parthenogenesis. Under natural conditions, queens mate with 10 to 17 drones (Adams et al., 1977) and as a consequence, colonies are composed of subfamilies of workers that share the same queen mother but different drone fathers. All workers in a subfamily inherit the same genetic material from their father, since all sperm cells produced by a drone are genetically identical (Page and Laidlaw, 1988). Workers perform most of the tasks in a colony. Division of labor in honey bee colonies is primarily 1 Department of Entomology, Purdue University, West Lafayette, Indiana 47904, USA. 2 To whom all correspondence should be addressed at Department of Entomology, Purdue University, West Lafayette, Indiana Tel: (765) miguel@entm.purdue.edu based on temporal polyethism. Workers perform different jobs in the colony depending on their age (Robinson and Page, 1991; Seeley, 1985; Winston, 1987). Typically young bees work close to the center of the brood nest, middle-age bees work at the outer sections of the nest and older bees forage outside (Moritz and Southwick, 1992; Seeley, 1985; Winston, 1987). Although each bee performs a set of tasks in the colony according to her age, not all bees of the same age group perform all the tasks that are typical of their group. Some workers are specialists that perform only one of the tasks (Robinson and Page, 1988; Trumbo et al., 1997). Several studies have shown that this task specialization is at least partly genetic in origin and that genetic variability among individuals in a colony is a factor that influences the division of labor among the workers of a colony (Breed et al., 1990; Frumhoff and Baker, 1988; Guzmán-Novoa et al., 2002; Page et al., 2000; Robinson and Page, 1988, 1991; Trumbo et al., 1997; Tugrul et al., 2000). An important adaptive characteristic of insect societies is the ability to defend their nest. Division of labor occurs in honey bee colony defense, but it is not /03/ / Plenum Publishing Corporation

2 358 Arechavaleta-Velasco, Hunt, and Emore well known how it is organized (Breed et al., 1990; Guzmán-Novoa et al., 2002). Colony defense primarily involves two distinctive behaviors: guarding and stinging. are workers with a mean age of about 15 days that patrol the entrance of the hive in search of bees, insects, animals or any other object that approaches the nest. are bees with a mean age of about 19 days that respond to major disturbances to the colony by flying out, stinging and pursuing intruders (Breed et al., 1990; Moore et al., 1987). are specialists. Only a small proportion of the middle-age bees in a colony behave as guards. In fact, about 15% of the bees in a colony will ever perform guarding behavior during their life (Moore et al., 1987). Robinson and Page (1988) found that the subfamily membership of the workers influences the probability that a bee behaves as a guard in the colony. Breed et al. (1990) reported that guards and responders were behaviorally differentiated groups of bees. In another study Breed and Rogers (1991) using defensive and gentle colonies, found that the expression of guarding behavior is influenced by both genotypic and environmental effects. Some studies have identified quantitative trait loci (QTLs) that influence the expression of foraging and defensive behavior in honey bee colonies (Hunt et al., 1995, 1998; Page et al., 1995, 2000). Recently, QTLs have been identified that influence specific individual honey bee behaviors (Guzmán-Novoa et al., 2002; Page et al., 2000). In the study conducted by Hunt et al. (1998), QTLs that affect stinging behavior were mapped as a colony trait in a population of 172 colonies derived from Africanized and European bees. This study identified five potential QTLs with LOD scores of 1.4 to 3.5 influencing the mean colony defensive response, measured in terms of number of stings. In subsequent studies, new crosses between European and Africanized bees were performed that showed that one of these QTLs, called sting-1, influences the tendency of individuals to perform guarding behavior and stinging behavior (Guzmán-Novoa et al., 2002). Most of the genetic studies of defensive behavior in the honey bee have been done at the colony level. Only a few studies have been able to establish a link between whole colony behavioral phenotype and the behavioral patterns of individual bees (Guzmán-Novoa et al., 2002; Page et al., 2000). It is important to try to find evidence of this relationship in order to understand how genetically influenced behaviors of individual honey bees determine the whole colony phenotype. The objectives of this study were 1) to test for the effect of three defensive behavior QTLs (sting-1, sting-2 and sting-3) on the expression of guarding and stinging behavior of individual honey bees and 2) to determine if results of defensive behavior QTLs found in studies with Africanized honey bees could be extended to other populations of honey bees. MATERIALS AND METHODS Experimental Colonies Two honey bee colonies were used as parental sources, one classified as high defensive and one classified as low defensive. A queen was reared from the defensive colony and was artificially inseminated with the semen of three of her brothers. A daughter queen reared from this cross was inseminated with the semen of a drone from the gentle colony. From this daughter queen, twelve hybrid queens were reared and divided into two groups. Six queens were single-drone artificially inseminated with drones from the defensive colony and six queens were single-drone artificially inseminated with drones of the gentle colony in order to produce two types of colonies composed of backcross workers (Fig. 1). The F1 queens that headed the backcrossed colonies shared the same queen mother and drone father. In consequence, the queens were super-sisters that had an average genetic relationship of at least Therefore, all the F1 queens inherited the same alleles from their low defensive drone father. The inbreeding step performed in the defensive source lineage was done to help insure more genetically uniform F1 queens. Each queen was introduced into a small colony made of three frames of brood, two frames of honey, four frames of empty comb and approximately 1.5 kg of bees. The colonies were kept in single deep standard Langstroth hives in the same apiary. All colonies were managed the same way for a period of 60 days to allow time for workers in the colony to be replaced by daughters of the inseminated queens. Two colonies were selected, one from each backcross, based on their relatively high defensive behavior. Behavior Assays To collect samples of guards, each of the two selected colonies was observed for a period of 30 minutes and bees performing guarding behavior at the entrance of the hive were marked with a dot of enamel paint on their thorax. A bee was identified as a guard if it patrolled the entrance of the hive and actively ap-

3 QTL Effects on Honey Bee Behaviors 359 the suede by their stingers. The plastic bag was marked and immediately placed in dry ice and later stored at 80 C. Five repetitions of this procedure were performed with a 48 h period between repetitions. Samples of pollen foragers and nurse bees were taken as controls from each of the two backcrossed colonies as soon as guards and stingers were collected. In general nurse bees are younger than guards and foragers are older. Nurse bees were collected by removing frames containing open brood from the center of the colony and bees were removed from the center of the nest and placed in dry ice. To collect pollen foragers, a screen of wire mesh (4 4 mm) was placed over the entrance of each colony to slow the flow of foragers into the interior of the hive. that carried a load of pollen on their legs were collected with a pair of forceps and introduced into 1.5 ml plastic tubes and placed in dry ice. The samples of nurse bees and pollen foragers were kept at 80 C until DNA analysis was performed. Fig. 1. Mating scheme used to produce colonies composed of backcross workers derived from a defensive colony and a gentle colony. proached and inspected incoming foragers. Different colors were used for each colony. Twenty-four hours later, marked bees that continued to guard were collected with a pair of forceps and introduced into a 1.5 ml plastic tube. The sampled guards were immediately placed in dry ice and kept at 80 C until the DNA analysis was performed. This procedure was repeated on five different occasions, allowing a 24 h period between each repetition. To collect samples of stingers from the two backcross colonies, a cm piece of black suede attached to the end of a one-meter stick called a flag was used. The piece of suede was impregnated with 5 l of 98% isopentyl acetate (Sigma 11,267-4), one of the principal honey bee alarm pheromone components and waved by hand in a rhythmic way approximately 10 cm from the entrance of the hive (Guzmán-Novoa and Page, 19; Villa, 1988). Bees were allowed to sting the flag for one minute after the first sting. After this period, the piece of suede was introduced into a plastic bag. Most of the bees that stung the flag were collected in the bag because they remained attached to Genetic Analysis DNA was extracted from individual guards, stingers, nurses and foragers from each of the two backcross colonies. This involved grinding the bees in lysis solution (1% CTAB, 50 mm Tris, ph 8.0, 10 mm EDTA, 1.1 M NaCl), followed by phenol extraction and ethanol precipitation of the DNA (Hunt, 1997). The DNA of each individual bee was quantified with a flourometer and diluted to a final concentration of 14 ng/ l in double distilled water. Sequence tagged sites (STSs) were used to test for the effect of each of the three QTLs (sting-1, sting-2, sting-3) on the expression of guarding and stinging behaviors of individual honey bees. STSs markers were derived from amplified fragment length polymorphism (AFLP) markers linked to the three QTLs by cloning the DNA fragments and sequencing them. To clone the DNA fragments, the bands containing the markers were cut from the polyacrylamide gels. The DNA fragment was amplified by PCR with the original AFLP pre-amplification primers and the product was run on a 2.5% agarose gel to confirm that the product corresponded to the bands of the AFLP marker. Then the PCR product was cloned into E. coli using the vector pcr 4-TOPO and the TOPO TA Cloning kit for sequencing (Invitrogen, Carlsbad, CA) and the DNA inserts were sequenced. Since the cloned DNA fragments were relatively small, unknown genomic DNA sequence adjacent to

4 360 Arechavaleta-Velasco, Hunt, and Emore each of the inserts was obtained using the Universal GenomeWalker kit (Clontech, Palo Alto, CA). The final PCR product containing the extended DNA sequence for each of the original AFLP markers, was cloned into E. coli using the TOPO TA Cloning kit for sequencing. The DNA inserts were sequenced and primers were designed for each of the STSs. The STSs developed were STSA linked to sting-1, STSA linked to sting-2 and STSA linked to sting-3. A second marker linked to sting-1 was obtained using DNA amplified by primers for STSN4-2 (Guzmán-Novoa et al., 2002; Hunt et al., 1998) to probe a genomic bacterial artificial chromosome (BAC) library of the honey bee. End sequence from one of the resulting clones was used to design a pair of primers that provided three polymorphic markers. Linkage analysis in haploid drone brothers indicated that one of these markers, STSN4P88K14b was tightly linked ( 1 cm) to the original STSN4-2. The two markers linked to sting-1 flank the LOD score peak for this QTL. STSA is located approximately at 12 cm from the LOD score peak for sting-1 and STSN4P88K14b is located approximately 9 cm away from the LOD score peak for sting-1. The marker STSA linked to sting-2 is located approximately at 5cM from the LOD peak score for the QTL. The marker linked to sting-3 is located approximately 3cM away from the LOD peak score for the position of the QTL. Polymerase chain reaction (PCR) was used to amplify the STSs linked to each of the QTLs. The PCR was performed under the following conditions. The DNA was initially denatured at C for 30 s; followed by five cycles of denaturing at C for 1 minute, annealing at 60 C for 1 minute and extension at 72 C for 2 minutes; followed by 35 cycles of C for 30 s, 55 C for 1 minute and 72 C for 2 minutes; followed by a final extension of 72 C of 8 minutes. The primer sequences were STS A17.080, 5 TGG TGG AAG GTT TGT ATA TTC G and AAG TTT CTT ACC ACG AGC CTG T; STS A11.310, 5 ACT TTT GAG GCG AAG AGG AAT AC and CTT GTC CAC GAC GAT TAC TTT TC; STS A64.084, 5 ATC CAG AGG ATT GAT CTC GAT G and TGC AAC ATT TGT CTC TGT GAT G. The primer sequences for STSN4P88K14b were 5 GAA ATT GTT GAC GCG AAA GAC and GTT GTA ACG GAA GAT TGG AAG G. Samples of 91 to 96 guards and stingers from each of the two types of backcross colonies were screened with each of the STSs linked to the three defensive behavior QTLs and the allelic frequencies were determined. For those STSs where the genotypic frequencies were significantly different from the expected 1:1 ratio, samples of 90 to 96 foragers and nurse bees of each backcross colony were genotyped as controls. In these cases, the PCR and marker analysis was also repeated at least once for the guards and stingers. To determine which alleles were inherited from the defensive source and which were inherited from the gentle source, the phase of the markers was determined by analyzing the genotype of the drones used to produce each backcrossed colony and the father of the F1 queens. Statistical Analysis To determine the effect of each of the three QTLs on the expression of guarding and stinging behaviors of individual bees, a chi-square goodness of fit test was used to look for deviations from the 1:1 segregation of the genotypes that would be expected for a colony composed of backcross workers. Significant deviations from the expected proportion in the marker allele genotypes of the guards and stingers, and no deviations in the samples of nurses and foragers used as controls would indicate an effect of the linked QTL on guarding or stinging behaviors. RESULTS Guarding Behavior Sting-1 A deviation from the expected 1:1 segregation of the marker alleles was found in the genotypic frequencies of guards of the defensive backcross for the marker STSA linked to sting-1 ( ; n 96; df 1; P 0.014). The marker allele inherited from the defensive source was over represented in the sample. No deviations from the expected frequencies were found in the foragers ( ; n ; df 1; P 0.30) or in the nurse bees ( ; r 90; df 1; P 1.00) used as controls (Table I). However the genotypic frequencies of STSA in the guards of the gentle backcross were not significantly different from the expected ratio ( ; n ; df 1; P 0.3) (Table I). There were no deviations from the expected genotypic segregation pattern for the other marker linked to sting-1, STSN4P88K14b, in guards from either the defensive backcross colony ( ; n 91; df 1; P 0.25) or the gentle backcross colony ( ; n ; df 1; P 0.76) (Table I).

5 QTL Effects on Honey Bee Behaviors 361 Table I. Number of Homozygous Defensive (D/D), Heterozygous (D/G) and Homozygous Gentle (G/G) Guard, Nurse and Forager Honey Bees (Apis mellifera L.) for STS Markers Linked to Three Defensive Behavior QTLs from a Defensive Backcross Colony and a Gentle Backcross Colony Genotypes Backcross QTL Marker Bee type N D/D D/G G/G 2 P Defensive backcross STING-1 STING-2 STING-3 STING-1 A N4P88K14b A A A N4P88K14b A Gentle backcross STING-2 STING-3 A Sting-2 A deviation from the expected segregation pattern was found in the genotypes of guards of the defensive backcross colony for the marker STSA linked to the QTL sting-2 ( ; n ; df 1; P 0.007). The allelic frequencies were skewed toward the marker allele inherited from the defensive source colony. No deviations were observed in the genotypic frequencies of the nurse bees ( ; n 92; df 1; P 0.68) or in the foragers ( ; n ; df 1; P 0.75) used as controls (Table I). In the gentle backcross, the genotypic frequencies deviated from the expected pattern ( ; n ; df 1; P 0.04). The allele inherited from the defensive source colony for STSA11.310, was also overrepresented in the guards. Again, no deviations from the 1:1 ratio were observed in the nurse bees ( ; n ; df 1; P 0.54) or in the foragers ( ; n ; df 1; P 0.17) (Table I). Sting-3 The expected segregation pattern found in the genotypes of guards of the gentle backcross colony for the marker STSA linked to sting-3 was biased towards the allele inherited from the defensive source ( ; n ; df 1; P 0.009). No deviations from the expected 1:1 segregation were found in nurse bees ( ; n 95; df 1; P 0.76) or in the pollen foragers ( ; n 96; df 1; P 0.84) or in the guards of the defensive backcross ( ; n ; df 1; P 0.84) (Table I). Stinging Behavior Sting-1 The segregation of genotypes of stingers in the defensive backcross colony was significantly deviated from the 1:1 pattern for the marker STSN4P88K14b linked to sting-1 ( ; n 92; df 1; P 0.007). The allelic frequency was skewed toward the marker allele inherited from the defensive source colony. No deviations from the expected pattern were observed in the genotypic frequencies of the nurse bees ( ; n 92; df 1; P 0.84) or in the foragers ( 2 0.; n 95; df 1; P 0.) used as controls (Table II). A deviation from 1:1 in the allelic frequencies of stingers of the gentle backcross also was found ( ; n ; df 1; P 0.039). The frequency was biased toward the marker allele inherited from the defensive source. No deviations from the expected pattern were observed in the nurse bees ( ; n 96; df 1; P 0.30) or in the foragers ( ; n ; df 1; P 0.15) (Table II). However, the marker linked to sting-1 that showed biased segregation in the guards, STSA17.080, showed

6 362 Arechavaleta-Velasco, Hunt, and Emore Table II. Number of Homozygous Defensive (D/D), Heterozygous (D/G) and Homozygous Gentle (G/G) Stinger, Nurse and Forager Honey Bees (Apis mellifera L.) for STS Markers Linked to Three Defensive Behavior QTLs from a Defensive Backcross Colony and a Gentle Backcross Colony Genotypes Backcross QTL Marker Bee type N D/D D/G G/G 2 P Defensive backcross STING-1 STING-2 STING-3 STING-1 A N4P88K14b A A A N4P88K14b Gentle backcross STING-2 STING-3 A A no deviations in the genotypes of stingers of both the defensive backcross ( ; n ; df 1; P 0.41) and the gentle backcross ( ; n 96; df 1; P 0.42). Linkage analysis using the Kosambi function showed that markers STSA and STSN4P88K14b were linked in all samples analyzed of guards, stingers, foragers and nurse bees in both backcross colonies (Table III). Sting-2 The genotypic frequencies of STSA did not deviate from the expected ratio in the stingers of either the defensive backcross ( ; n 95; df 1; P 0.61) and in the gentle backcross colony ( ; n ; df 1; P 0.61) (Table II). Table III. Genetic Distances (cm) between Markers STSA and N4P88K14b Linked to sting1 for Different Types of Honey Bees (Apis mellifera L.) Collected from a Defensive Backcross Colony and a Gentle Backcross Colony Backcross Bee type n cm Gentle backcross Defensive backcross In all the samples, the marker alleles were linked. Sting-3 STSA did not show biased ratios in the bees that stung in either the defensive backcross colony ( ; n 95; df 1; P 0.61) or in the gentle backcross ( ; n ; df 1; P 0.25) (Table II). DISCUSSION In general, our results showed that the genotypic frequencies of guard bees of both backcross colonies differed from the expected ratio for the markers linked to sting-2 and sting-3 and for one of the two markers linked to sting-1. These results indicate that these three loci influenced the expression of guarding behavior of individual honey bees. The genotypic frequencies of stingers of both backcross colonies also deviated from the expected ratio for one of the markers alleles linked to sting-1, indicating that this locus also influences the expression of stinging behavior of individual honey bees. For both guards and stingers, the genotypic frequencies were biased towards the marker allele inherited from the defensive source colony. No effect was detected for the other two QTLs (sting-2 and sting-3) on the bees that stung during the defensive behavior tests. All of the markers alleles linked to these three QTLs segregated 1:1 in foragers and nurse bees in both of the two types of backcross colonies, indicating that these loci only influence the behavior of bees performing defensive tasks and suggesting that the genetic composition of individual bees influence the division of labor in honey bee colonies. The effects of sting-1 on the expression of guarding behavior found in this study partially corroborate the re-

7 QTL Effects on Honey Bee Behaviors 363 sults of the study conducted by Guzmán-Novoa et al. (2002). In their study, it was found that sting-1 also influences guarding behavior of individual honey bees in backcrossed colonies derived from a highly defensive Africanized colony and a gentle European colony. The authors suggested a non-additive effect of the QTL since the deviations from the expected genotypes were only found for the marker STSN4-.2 linked to sting-1 in the colony backcrossed to the gentle European source. In our study the marker STSN4-.2 was not polymorphic. We used two markers, STSA and STSN4P88K14b flanking the QTL, to detect the effect of sting-1. STSN4P88K14b is tightly linked ( 1 cm) to STSN4-.2, the marker used by Guzmán-Novoa et al. (2002). We were able to detect the effect of sting-1 on the expression of guarding behavior in the defensive backcross colony and the marker allele STSA inherited from the defensive source was the one over represented in the sample, however STSN4P88K14b showed a segregation pattern that was not different from the 1:1 expected ratio. Results of our study also indicate that sting-1 influenced the expression of stinging behavior in individual honey bees. In both backcross colonies, the genotypes of stingers deviated from the expected pattern for the marker allele STSN4P88K14b linked to the QTL, but not for STSA In contrast, Guzmán- Novoa et al. (2002) were able detect the effect of sting-1 only in the sample of the bees that were among the first to sting the flag. In our study, we were able to detect the effect in a sample of all the bees that responded by stinging during the defensive behavior tests. Our findings suggest an additive mode of action of sting-1 since the genotypic frequencies of stingers of both backcross colonies were skewed toward the defensive marker allele, and the effect of the QTL on guards was detected only on the defensive backcross. This differs from the non-additive effect proposed in the study of Guzmán-Novoa et al. (2002). The differences between our study and that conducted by Guzmán-Novoa et al. (2002) could be due to differences in the genetic composition of the colonies used in each study. In our study, we used bees derived from populations in the United States for sources of defensive and gentle colonies, while in the study of Guzmán-Novoa et al. (2002) bees of European race where used as gentle source and Africanized bees were used as defensive source. Some differences could also be attributed to the particular parental lines chosen because they were unrelated and they were not derived from inbred lines. Two marker alleles flanking sting-1 were used to detect the effect of the QTL on the expression of guarding and stinging behavior of individual honey bees. It is interesting to notice that one of these markers, STSA was significantly skewed in the guards of the defensive backcross colony but not in the guards of the gentle backcross or in the stingers of either of the backcross colonies. The other marker, STSN4P88K14b was significantly skewed in the samples of stingers of both colonies but not in the samples of guards of either of the backcross colonies. The genetic distance between the two markers was 20cM on average in the samples of different types of bees of our study. One centimorgan in the honey bee genome represents approximately 75 kb. One hypothesis that could explain the previous findings is that the genomic region mapped as sting-1 by Hunt et al. (1998), represents two loci, one affect the expression of guarding behavior while the other influences stinging behavior of individual honey bees. Another possibility is that sting-1 is actually one locus with pleiotropic effects on guarding and stinging behaviors that is located between the two markers and by chance we were not able to detect the effect on both behaviors with both marker alleles. Our findings indicate that sting-2 has an effect on the expression of guarding behavior. We were able to detect the effect of the QTL in both backcross colonies. In both types of colonies, bees behaving as guards were skewed for the marker allele inherited from the defensive source colony. Homozygous bees for the defensive marker allele were over-represented in the guards of the defensive backcross colony. While heterozygous bees carrying one defensive marker allele were overrepresented in the guards of the gentle backcross. These results suggest an additive mode of action of the QTL. Effects of sting-3 on guarding behavior were detected only in the gentle backcross colony. The genotypic frequencies indicated that bees heterozygous for the marker alleles linked to sting-3 were over-represented in the guards. These findings suggest a dominant mode of action of the QTL, since both genotypes in the defensive backcross showed the same likelihood for guarding and in the gentle backcross colony bees carrying the defensive allele were more likely to guard than those that were homozygous for the gentle marker allele. But do not exclude the possibility that failure to detect the QTL effect in the defensive backcross was caused by chance segregation of the alleles in the F1 queens. The genotypic and allelic frequencies for the marker alleles linked to the QTLs of guards and stingers were different from the controls. Bees involved in the defense of a colony were genetically different from other bees in the colony. Each colony in our study was composed of a single family of bees. The genotype of an individual bee influenced the tasks that she performed

8 364 Arechavaleta-Velasco, Hunt, and Emore in the colony. Other studies with colonies composed by more that one family has shown that task specialization is influenced by sub-family membership and suggested that genetic variation plays a role in division of labor (Robinson and Page, 1988; Frumhoff and Baker, 1988; Breed et al., 1990; Robinson and Page, 1991; Trumbo et al., 1997; Tugrul et al., 2000; Page et al., 2000; Guzmán-Novoa et al., 2002). Our results demonstrate that specific QTLs influence division of labor in a colony, affecting specifically defensive behavior. Moreover, differences were found between guards and stingers for the effect of the three QTLs. Sting-2 and sting-3 influenced only guarding behavior and sting-1 influenced stinging behavior and guarding behavior. It is possible that sting-1 represents two distinct loci, one influencing guarding and the other stinging behavior. These results suggest that guards and stingers are two genetically differentiated groups with different allelic frequencies for the marker alleles linked to the QTLs of this study and suggests that some degree of division of labor occurs during the defensive response of a colony. Breed et al. (1990) reported that guards and stingers were two behaviorally differentiated groups. In another study, using colonies derived from the same parental colonies used for this study, we found that most of the guards (97%) did not responded by stinging during colony defense (Arechavaleta-Velasco and Hunt, 2003). This is the first study that confirms the effect of the loci sting-2 and sting-3 on the individual behavior of bees performing defensive tasks in a colony and partially corroborates the findings of Guzmán-Novoa et al. (2002) for the locus sting-1. Our study demonstrates the expression of these QTLs under different environmental conditions and in a different population. It also shows that defensive behavior QTLs found in studies with Africanized honey bees also influence behavior in other populations of honey bees. ACKNOWLEDGMENTS We thank Tom Glenn for his help with the instrumental inseminations and to Damon Hall and Carmen Camacho-Rea for their assistance in various ways. This work was funded with grants from NIH R29 GM54850 and USDA REFERENCES Adams, R. J., Kerr, W. E., and Paulino, Z. L. (1977). Estimation of sex alleles and queen matings from diploid male frequencies in a population of Apis mellifera. Genetics 86: Arechavaleta-Velasco, M. E., and Hunt, G. J. (2003). Genotypic variation in the expression of guarding behavior and the role of guards in the defensive response of honey bee colonies. Apidologie (In press). Breed, M. D., and Rogers, K. B. (1991). The behavioral genetics of colony defense in honeybees: Genetic variability for guarding behavior. Behav. Genet. 21: Breed, M. D., Robinson, G. E., and Page, R. E., Jr. (1990). Division of labor during honey bee colony defense. Behav. Ecol. Sociobiol. 27: Frumhoff, P. C., and Baker, J. (1988). A genetic component to division of labour within honey bee colonies. Nature 333: Guzmán-Novoa, E., and Page, R. E., Jr. (19). Backcrossing Africanized honey bee queens to European drones reduces colony defensive behavior. Ann. Etomol. Soc. Am. 86: Guzmán-Novoa, E., Hunt, G. J., Uribe J. L., Smith, C., and Arechavaleta-Velasco, M. E. (2002). Confirmation of QTL effects and evidence of genetic dominance of honeybee defensive behavior: Results of colony and individual behavioral assays. Behav. Genet. 32: Hunt, G. J. (1997). Insect DNA extraction protocol. In M. R. Michelli and R. Bova (eds.), Fingerprinting methods based on arbitrarily primed PCR, pp Berlin: Springer-Verlag. Hunt, G. J., Guzmán-Novoa, E., Fondrk, M. K., and Page, R. E., Jr. (1998). Quantitative trait loci for honey bee stinging behavior and body size. Genetics 148: Hunt, G. J., Page, R. E., Jr., Fondrk, M. K., and Dullum, C. J. (1995). Major quantitative trait loci affecting honey bee foraging behavior. Genetics 141: Moore, A. J., Breed, M. D., and Moor, M. J. (1987). The guard honey bee: Ontogeny and behavioural variability of workers performing a specialized task. Anim. Behav. 35: Moritz, R. F. A., and Southwich, E. E. (1992). Bees as superorganisms: An evolutionary reality. Berlin Heidelberg New York: Springer. Page, R. E., Jr., and Laidlaw, H. H. (1988). Full sisters and super sisters: A terminological paradigm. Anim. Behav. 36:4 5. Page, R. E., Jr., and Robinson, G. E. (1991). The genetics of division of labour in honeybee colonies. Adv. Insect. Physol. 23: Page, R. E., Jr., Fondrk, M. K., Hunt, G. J., Guzmán-Novoa, E., Humphries, M. A., Nguyen, K., and Greene A. S. (2000). Genetic dissection of honeybee (Apis mellifera L.) foraging behavior. J. Hered. 91: Page, R. E., Jr., Waddington, K. D., Hunt, G. J., and Fondrk, M. K. (1995). Genetic determinants of honey bee foraging behavior. Anim. Behav. 50: Robinson, G. E., and Page, R. E., Jr. (1988). Genetic determination of guarding and undertaking in honey-bee colonies. Nature 333: Seeley, T. D. (1985). Honey bee ecology. Princeton: Princeton University Press. Trumbo, S. T., Huang, Z. Y., and Robinson, G. E., (1997). Division of labor between undertaker specialists and other middleaged workers in honey bee colonies. Behav. Ecol. Sociobiol. 41: Tugrul, G., Guzmán-Novoa, E., Carol, A. W., Zelinsky, B., Fahrbach, S. E., and Robinson, G. E. (2000). Genetic variation in worker temporal polyethism and colony defensiveness in the honey bee, Apis mellifera. Behav. Ecol. 11: Villa, J. D. (1988). Defensive behavior of Africanized and European honeybees at two elevations in Colombia. J. Apic. Res. 27: Winston, M. L. (1987). The biology of the honey bee. Cambridge: Harvard University Press. Edited by Lee Ehrman and Stephen Maxson