Honey bees are exposed to phytochemicals through the nectar, pollen and propolis consumed to sustain the colony. They may also encounter mycotoxins produced by Aspergillus fungi infesting pollen in beebread. Moreover, bees are exposed to agricultural pesticides, particularly in-hive acaricides used against the parasite Varroa destructor. They cope with these and other xenobiotics primarily through enzymatic detoxificative processes, but the regulation of detoxificative enzymes in honey bees remains largely unexplored.
We used several approaches to ascertain effects of dietary toxins on bee susceptibility to synthetic and natural xenobiotics, including the acaricide tau-fluvalinate, the agricultural pesticide imidacloprid, and the naturally occurring mycotoxin aflatoxin. We administered potential inducers of cytochrome P450 enzymes, the principal biochemical system for Phase 1 detoxification in insects, to investigate how detoxification is regulated. The drug phenobarbital induces P450s in many insects, yet feeding bees with phenobarbital had no effect on the toxicity of tau-fluvalinate, a pesticide known to be detoxified by bee P450s. Similarly, no P450 induction, as measured by tau-fluvalinate tolerance, occurred in bees fed xanthotoxin, salicylic acid, or indole-3-carbinol, all of which induce P450s in other insects. Only quercetin, a common pollen and honey constituent, reduced tau-fluvalinate toxicity. In microarray comparisons no change in detoxificative gene expression was detected in phenobarbital-treated bees. However, northern blot analyses of guts of bees fed extracts of honey, pollen and propolis showed elevated expression of three CYP6AS P450 genes. Diet did not influence tau-fluvalinate or imidacloprid toxicity in bioassays; however, aflatoxin toxicity was higher in bees consuming sucrose or high-fructose corn syrup than in bees consuming honey.
Citation: Johnson RM, Mao W, Pollock HS, Niu G, Schuler MA, Berenbaum MR (2012) Ecologically Appropriate Xenobiotics Induce Cytochrome P450s in Apis mellifera. PLoS ONE 7(2): e31051. https://doi.org/10.1371/journal.pone.0031051
Editor: Frederic Marion-Poll, AgroParisTech, France
Received: June 28, 2011; Accepted: December 31, 2011; Published: February 3, 2012
Copyright: © 2012 Johnson et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: United States Department of Agriculture, Agriculture and Food Research Initiative (USDA-AFRI). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Apis mellifera, the western honey bee, is the premier managed pollinator in the United States; the value of its contribution to agriculture in the form of providing pollination services to over 90 crop species exceeds $14 billion annually . Stresses experienced by this species in the form of environmental toxins therefore have impacts across the agricultural spectrum. The ability of the honey bee to forage across the landscape leaves it vulnerable to exposure to a wide range of agricultural chemicals. For decades, pesticides used for control of crop pests and human disease vectors have caused honey bee mortality and morbidity , . Despite labeling restrictions and a trend toward reduced use of pesticides in agriculture and forestry, pesticide applications continue to kill nontarget honey bee colonies . Sublethal effects of pesticides exposure (e.g., neonicotinoids used for seed pretreatment) are suspected of causing reductions in hive viability . Moreover, for the last two decades, the presence of Varroa destructor, a devastating parasitic mite that infests honey bee colonies, has led to additional xenobiotic stresses in the form of in-hive acaricide use, exposing bees to synthetic pesticides for the entire duration of their life cycle. The deliberate introduction of chemical pesticides to the hive environment has occurred largely without detailed knowledge of how honey bees process and thus tolerate these toxic compounds.
Like most other insects, honey bees rely in part on a suite of detoxification enzymes to metabolize naturally occurring xenobiotics and pesticides. Chief among these enzymes are the cytochrome P450 monooxygenases (P450) . P450s play a role in the detoxification of phytochemicals  present in the nectar, honey and pollen that bees consume –. Additionally, the beehive, with its stores of pollen and beebread, provides a hospitable environment for fungi in the genus Aspergillus, which produce mycotoxins ,  that are detoxified by P450s in the honey bee .
Synthetic pesticides are also metabolized by P450s in honey bees. Tau-fluvalinate, a pyrethroid acaricide that is used in-hive by beekeepers to control Varroa mites, is metabolized by P450s , , as are the pyrethroid lambda-cyhalothrin  and the organophosphate in-hive acaricide coumaphos , . Indeed, tolerance of these acaricides is attributable in part to rapid P450-mediated detoxification by bees and is the reason these pesticides can be safely used in the hive by beekeepers for Varroa control. Many other organophosphate and pyrethroid pesticides are highly toxic to honey bees, although the toxicity varies according to the specific pesticide .
The apparent sensitivity of honey bees to some pesticides became the focus of discussion following the sequencing of the honey bee genome . P450s are central to tolerance and evolved resistance to pesticides in many pest insects  but the A. mellifera genome encodes only 46 P450s, far fewer than most other insect genomes. Moreover, the carboxylesterase and the glutathione-S-transferase gene families, the other major detoxification genes in aerobic organisms, are similarly reduced in size . It has been suggested that this reduced diversity of detoxification enzymes may contribute to the sensitivity of honey bees to certain pesticides  (but see ). Along with insights into honey bee biology, sequencing of the honey bee genome has also provided a wealth of new tools for investigating honey bee regulation of xenobiotic detoxification, a critical yet hitherto unexplored dimension of how this pollinator copes with a wide array of phytochemicals in its diet as well as synthetic pesticides and other xenobiotics. The paucity of genes in families associated with detoxification in combination with the existence of behavioral mechanisms of reducing toxin intake (e.g., ) suggests that regulation of these genes may differ in honey bees in comparison to nonsocial species and to herbivores that feed on chemically defended foliage.
Induction, the phenomenon whereby the production of a detoxification enzyme increases in response to toxin exposure , is thought to be widely associated with induced transcription of detoxification genes because it minimizes resource investment in superfluous metabolic capability and protects organisms from the oxidative damage that can accompany P450 activity . Because P450 enzymes are frequently inducible by their substrates, induction has served as a useful tool in identifying specific P450s associated with pesticide tolerance and xenobiotic response . To date, only one study has demonstrated induction of P450 activity in bees. Benzo(α)pyrene monooxidase activity in honey bee guts was induced by exposure to benzo(α)pyrene itself and by the in-hive acaricides tau-fluvalinate and cymiazole hydrochloride .
Phenobarbital, a synthetic barbiturate drug, is a potential inducer of P450 activity in honey bees in that it induces P450s in a wide range of organisms , . In insects, phenobarbital induction increases enzymatic P450 activity in Diptera –, Lepidoptera –, and Blattodea . Induction of P450 enzymatic activity has been measured either in vitro using pesticide metabolism assays, or in vivo, using pesticide toxicity as an indicator of detoxificative P450 activity –. Although phenobarbital is a reliable inducer of P450 activity in many insects, only a single study using phenobarbital has been performed in Hymenoptera; phenobarbital feeding had no effect on the toxicity or metabolism of carbaryl in alfalfa leafcutter bees (Megachile rotundata) .
In addition to inducing enzymatic activity, phenobarbital also induces transcription of P450 genes in many insects, with studies demonstrating elevated transcription of CYP6, CYP9 and CYP4 family P450s in both Lepidoptera – and Diptera –. Microarray studies with Drosophila melanogaster demonstrated induced expression of as many as 29 P450 genes following phenobarbital treatment –. To date, no P450 induction studies based on gene expression have been reported in honey bees or other hymenopterans treated with phenobarbital.
Natural phytochemicals that honey bees encounter in nectar, pollen and propolis may also serve as inducers of P450-mediated detoxification. Indeed, honey is known to be an effective P450 inducer in humans; elevated P450 enzyme activity was observed in humans after eating honey , although the specific components responsible for induction were not identified. Flavonoids, compounds important to plant resistance to insect herbivory , that are present in both pollen , , and honey , may induce P450s in bees. Lepidopteran larvae that consumed quercetin, a common flavonoid in foliage as well as honey and pollen, experienced increased P450 gene expression  as well as elevated P450 enzymatic activity against model substrates , . Propolis, a resinous material collected by honey bees for use as a structural sealant and as an antibiotic , is rich in flavonoids and phenolic compounds and induces P450s involved in mycotoxin detoxification in this species .
Several classes of phytochemicals, which may not be present in nectar, pollen and propolis, act as inducers of P450-mediated metabolism in foliage-feeding insect herbivores. Xanthotoxin, a furanocoumarin produced by plants in the families Apiaceae and Rutaceae, is an effective inducer of xenobiotic-metabolizing P450s in several species of lepidopterans , , ,  as are indole-3-carbinol , , , a derivative of the toxic glucosinolates produced by plants in the Brassicaceae, and salicylic acid , a ubiquitous phytohormone active in initiating plant defensive response to herbivory.
We examined the phenomenon of P450 induction in honey bees using two different approaches after the administration of chemicals that induce P450s in other organisms—by testing for functional evidence of induction by assaying tolerance of toxic compounds, and by measuring changes in P450 transcript abundance in response to candidate inducers. In the toxicity assays, we examined adult workers for the in vivo effects of putative inducers on the LD50 of pesticides known to interact with P450s. The toxicity of two pyrethroid pesticides detoxified by P450s in bees, tau-fluvalinate  and lambda-cyhalothrin , and two pesticides bioactivated by P450 activity in honey bees, imidacloprid  and aldrin , was assessed using this approach.
To determine whether as-yet unidentified phytochemicals in honey function as inducers, an additional set of toxicity bioassays was conducted. High-fructose corn syrup (HFCS) and sucrose syrup are commonly used in commercial apiculture . Neither supplemental carbohydrate source contains the suite of plant secondary compounds that are present in nectar and honey and that may be important in P450 regulation. Susceptibility of adult workers to tau-fluvalinate and imidacloprid was compared on diets of honey, sucrose, and HFCS. In addition, longevity of adult worker bees in the presence of the naturally occurring mycotoxin aflatoxin B1, known to be metabolized by P450s , was compared on diets of honey, sucrose, and HFCS. We selected aflatoxin B1 for longevity assays as a toxin naturally present in the honey bee's environment to contrast with synthetic pesticides.
To evaluate effects of phenobarbital on P450 transcription, we analyzed bees exposed to phenobarbital using specialty honey bee microarrays . To identify potential natural sources of P450 inducers, we also conducted northern blot analysis on expression levels of the CYP6AS subfamily of P450 genes in bees consuming extracts of honey, pollen and propolis. Mao et al.  demonstrated using heterologous expression that one P450, CYP6AS3, contributes to metabolizing quercetin, a flavonol that occurs widely in plant nectars, pollen and honey. The bee-specific expansion of a group of CYP6AS P450s in the honey bee genome, in contrast with the genome of the parasitoid wasp Nasonia vitripennis , suggests that other members of this clade of CYP6AS P450s may be involved in metabolism of diet constituents unique to the honey bee, including those found in honey, pollen and propolis.
Functional P450 induction measured using LD50 bioassays
Of all of the potential inducers assayed, only one, quercetin, significantly decreased the toxicity of tau-fluvalinate to bees (Table 1). Pretreatment with indole-3-carbinol or salicylic acid did not alter the toxicity of tau-fluvalinate, while xanthotoxin and phenobarbital pretreatment actually increased the toxicity of tau-fluvalinate to bees. Phenobarbital feeding also increased the toxicity of lambda-cyhalothrin, although to a lesser degree than tau-fluvalinate. Toxicity of aldrin and its P450-bioactivated metabolite dieldrin increased similarly in bees fed phenobarbital.
In assays conducted to determine the ability of different diets (sucrose, HFCS, or honey) to induce P450s, one-way analysis of variance revealed no significant differences in tau-fluvalinate or imidacloprid toxicity to bees based on diet. Thus, there are apparently no constituents of honey that induce P450s that either detoxify or bioactivate these two pesticides.
Functional P450 induction measured using longevity bioassays
A Kaplan-Meier Survival Analysis was conducted to examine differential sensitivity to aflatoxin based on diet. The median time for survival was calculated for each treatment independently. No significant differences in median survival time were found among bees consuming control-treated sucrose, HFCS, or honey (Table 2). However, in the presences of aflatoxin B1, bees maintained on a diet containing honey had a significantly longer median survival time (55.0 h) than bees fed diets containing HFCS (47.3 h) or sucrose (40.9 h) (Wilcoxon, p = 0.001). These findings suggest that honey contains one or more constituents that allow bees to better tolerate aflatoxin exposure, possibly through induction of P450s capable of detoxifying this mycotoxin.
Transcriptional P450 induction from phenobarbital measured using microarrays
Only a single gene, tetraspanin 16, was differentially expressed (p≤0.05, FDR) in response to phenobarbital; tetraspanin 16 showed a 1.4-fold induction in phenobarbital-treated bees. No P450s were differentially expressed, nor were any genes in other gene families associated with detoxification.
Transcriptional P450 induction measured by northern blot
Feeding on honey extract brought about substantial induction of CYP6AS3 and moderate induction of CYP6AS1 and CYP6AS4, in a dose-dependent manner (Figure 1). Expression of these P450 genes in the absence of honey extract was low. CYP6AS10 and CYP6AS15 were expressed at consistently high levels and their expression was not responsive to honey extract ingestion.
Expression of selected P450 genes, as measured by northern blot, in guts of bees fed five g candy containing only sucrose or sucrose plus extract from the given quantity of honey, pollen or propolis. The neighbor-joining tree is rooted with Homo sapiens CYP3A4 and was created using CLUSTALW  alignment and PHYLIP  with 1000 bootstrap replicates. Branches with greater than 50% bootstrap support are indicated with an asterisk. Branch length in the final tree was corrected for multiple substitutions with TREE-PUZZLE .
Pollen and propolis extract also caused slight induction of CYP6AS1 and CYP6AS3. Only propolis induced CYP6AS4. The induction brought about by low and high doses of pollen extract appeared the same, while only the highest dose of propolis extract resulted in induction of these genes. CYP6AS10 and CYP6AS15 were expressed at high levels and were not responsive to ingestion of either pollen or propolis extract.
Visible effects of honey extract on the gut
Dissection of guts for extraction of RNA for microarray analysis suggested that the nature of the diet consumed by bees affects the morphology of the guts. Midguts of sucrose-fed bees, compared to those consuming honey extract, appeared fragile, flaccid and generally smaller. In order to quantify this apparent difference, a separate bioassay was conducted for the express purpose of quantifying morphological attributes of guts of bees fed different diets. Midguts of bees fed honey extract measured at their broadest point had a statistically greater diameter than did guts of bees fed plain “bee candy” (Figure 2; ANOVA, p<0.01, N = 41). Midguts of bees fed unaugmented candy (1.63 +/− 0.14 (SD) mm) were smaller in diameter than midguts from bees fed candy with a high dose of honey extract (1.82 +/− 0.17 mm; Tukey's HSD, p<0.01), while midguts of bees fed a low dose of honey extract were intermediate in width (1.75 +/− 0.15 mm).
Midguts of bees fed on (a) plain sucrose candy were narrower than midguts of bees fed candy fortified with (b) a low dose (extract of 3 ml honey in 5 g sucrose candy) or (c) high dose of honey extract (10 ml honey). Scale bar = 1 mm.
In contrast with numerous studies finding phenobarbital induction of P450s in other insects, no P450s or any other detoxification genes showed a change in expression after phenobarbital feeding. Moreover, none of the pesticide toxicity bioassays showed evidence of P450 induction after phenobarbital exposure. Rather, phenobarbital treatment elevated the toxicity of all pesticides, suggesting that pesticides compete with phenobarbital for P450-mediated detoxification , . Although phenobarbital has failed to induce P450-mediated detoxification in other insects (e.g., xanthotoxin detoxification in Papilio polyxenes ), phenobarbital enhancement of pesticide toxicity has not been previously reported.
Manipulation of honey bee diet yielded a dichotomy of responses in tolerance and toxicity, consistent with differential ability of dietary components to induce P450-mediated detoxification. Non-honey diets significantly decreased the ability of honey bees to tolerate the natural toxin aflatoxin B1 yet had no measurable effect on toxicity of the synthetic toxins tau-fluvalinate and imidacloprid. Although the number of substrates assayed is limited, this finding is consistent with the suggestion that regulation of honey bee P450s is more specialized than has been found in other insects. Such specialization may reflect the fact that honey bees for the most part consume only foods that they have processed—honey, beebread, or royal jelly. When a bee encounters a novel xenobiotic, such as the in-hive acaricides or imidacloprid, with which it has not coevolved, these synthetic compounds may not activate the same molecular pathways as are activated by naturally occurring xenobiotics in hive products.
Although honey bees tolerate a variety of synthetic pesticides , , , many others are extremely toxic to honey bees, including other pyrethroids similar to tau-fluvalinate . Thus, metabolism of pesticides by honey bee P450s is most likely an incidental convergence of molecular structure and not an indicator of molecular evolution in response to selection. That aflatoxin B1 is detoxified by honey bee P450s  and not bioactivated, as it is by P450s in many other organisms, also is consistent with a coevolutionary history of adaptation.
The presence of substances in honey that induce or upregulate detoxificative P450s in honey bees raises the possibility that the longstanding practice of feeding bees sucrose or HFCS  may have unintended adverse impacts. beyond those already documented. Fructose in HFCS can be converted into a toxic by-product, 5-hydroxymethylfurfural, which can cause dysentery-like symptoms and mortality . This study suggests that the survival of honey bees fed on HFCS and sucrose may be compromised when bees are also exposed to the fungal toxin aflatoxin B1, possibly because of reduced P450 activity and a resulting decreased capacity to tolerate aflatoxin. Because bees do not seem to be capable of inducing P450s in response to ecologically inappropriate xenobiotics, prophylactic induction of P450s through consumption of pollen and honey flavonoids may enhance bee survival.
Our findings are also consistent with the interpretation that honey bees, possibly due to the reduction in P450 genomic inventory, may rely on a small number of enzymes to detoxify both natural and synthetic xenobiotics. Because both aflatoxin  and acaricides such as tau-fluvalinate  are metabolized by P450s, there is potential for synergism between natural and synthetic xenobiotics, given that such synergism has been demonstrated between pyrethroid and organophosphate acaricides  and between pyrethroid insecticides and fungicides . Recent dramatic declines in honey bee abundance associated with a suite of unusual attributes, collectively characterized as Colony Collapse Disorder , have led to speculation that pesticide exposures may be causing or contributing to bee losses . Conducting an extensive survey across 23 states and a Canadian province over the 2007–2008 season, Mullin et al.  found “unprecedented levels of miticides and agricultural pesticides” in colonies. Residues of 118 different pesticides were recovered, with an average 6.5 pesticide detections per sample across wax, pollen, beebread, adult bees, and brood. Multiple exposures were typical; over 90% of the 749 samples analyzed contained at least two pesticides. These authors conclude their report with the statement that “the high frequency of multiple pesticides in bee collected pollen and wax indicates that pesticide interactions need thorough investigation before their roles in decreasing bee health can be either supported or refuted.” These levels of exposure, in the context of our findings that the ability of the honey bee to upregulate P450 detoxification genes in response to toxic exposure may be constrained and dependent in part on diet constituents, suggest that understanding precisely how honey bees process toxins, either individually or in combination, is a pressing necessity for maintaining the vitality of the U.S. apicultural enterprise.
Phenobarbital, xanthotoxin, quercetin, indole-3-carbinol and salicylic acid (Sigma, St. Louis, MO) were incorporated into confectioners sugar using a mortar and pestle, which was then used to make bee “candy.” A variety of doses were initially tested and a sublethal dose for each compound was chosen for the bioassays. Technical grade tau-fluvalinate (95%; Chem Service, West Chester, PA), aldrin and dieldrin were dissolved in chromatography-grade acetone for LD50 determination.
Tau-fluvalinate (95%) and imidacloprid (99.5%) were purchased from Chem Service (West Chester, PA). Aflatoxin B1 was purchased from Sigma Chemical Co. (St. Louis, MO). Tau-fluvalinate and imidacloprid stocks were dissolved in acetone purchased from Fisher Scientific (Pittsburgh, PA). Aflatoxin B1 stocks were dissolved in analytical grade dimethyl sulfoxide (DMSO) purchased from Fisher Scientific (Pitttsburgh, PA).
Honey, pollen and propolis extract
Methanolic extracts were made from hive products collected from University of Illinois apiaries in 2007–2008. Honey, predominantly from soy and wildflower sources, was collected from University of Illinois apiaries in 2007. First, the honey was dissolved in distilled water to make a 10% solution. Diluted honey was filtered through a paper filter (Whatman, Kent, England) to remove particulate matter and then processed through a C18 silica column under vacuum. The column was washed with 5% methanol in water and then eluted with pure HPLC-grade methanol. Methanol was removed using a rotary evaporator (Büchi-Brinkmann, Flawil, Switzerland) and the remainder was resuspended in 1 ml methanol for each 20 ml of honey in the original solution. Ground pollen (Betterbee, Greenwich, NY) extract was made by extracting 10 g pollen in 100 ml of 90% methanol in water for 1 h at 25°C and then centrifuging and removing the liquid. This procedure was repeated three times. Raw pollen extract was then processed over a C18 silica column as was honey, with a final concentration of 2 g pollen for each 1 ml methanol. Propolis was scraped from frames and boxes of University of Illinois colonies located in a forested area in summer 2008. Propolis extract was made by freezing 3 g propolis with liquid nitrogen and grinding with a mortar and pestle. Propolis powder was dissolved in 30 ml methanol and heated just to boiling.
Incorporation of xenobiotics into diet
Chemicals or extracts were administered in “bee candy” made from equal parts powdered sugar and heavy sucrose syrup with a ratio of 2∶1 sucrose to water (w/w). Sucrose (granulated table sugar) was processed in a blender to make starch-free powdered sugar. Approximately 5 g fresh liquid candy was poured into 2 oz (56 ml) plastic cups (Solo, Urbana, IL) and the candy was allowed to harden for at least 30 min before feeding to bees. Any candy not used immediately was stored at 4°C.
Phenobarbital, xanthotoxin, quercetin, salicylic acid, and indole-3-carbinol were incorporated into the powdered sugar component using a mortar and pestle. Final concentrations of treated candy fed to bees were the maximum concentration that did not cause increased mortality over control after 3 days and were as follow: 5 mg/g phenobarbital, 1 mg/g xanthotoxin, 10 mg/g quercetin, 2.5 mg/g salicylic acid, and 1 mg/g indole-3-carbinol.
Methanolic honey, pollen and propolis extracts were applied to powdered sugar, as was pure methanol as a control, and the solvent was allowed to evaporate overnight, prior to addition of heavy sucrose syrup. Bees were fed either a high or low dose of honey, pollen, or propolis extract, containing the extract of 3 or 10 ml honey, 2 or 4 g pollen, or 150 or 300 mg propolis per gram of candy.
Five microliters of aflatoxin B1 (20 µg/µl), or a DMSO control, was incorporated directly into the wet candy after addition of heavy sucrose syrup. Honey and HFCS were administered in the form of candy as well by using honey (University of Illinois apiaries) or high-fructose corn syrup (55% fructose, Archer Daniels Midland) in place of heavy sucrose syrup.
Frames of late-stage sealed worker brood were taken from healthy colonies in the University of Illinois apiaries near Urbana, IL in July-August 2006, August-September 2008, and September-October 2009 and placed in a dark humid (∼80% RH) incubator maintained at 32–34°C. Newly eclosed adults were brushed from the frames at 24 h intervals and placed in screen-topped wooden boxes (330 cm3) in groups of 150–250. Newly emerged bees were immediately fed treated or control candy and maintained in the incubator for 3 days.
Full LD50 trials (Table 1) included an acetone control and doses causing 0% and 100% mortality, as well as at least four doses causing >0% and <100% mortality. Three- to four-day-old bees were anesthetized with CO2 in groups of 20 and 1 µl of tau-fluvalinate, lambda-cyhalothrin, aldrin, or dieldrin dissolved in acetone was applied to the thorax of each bee with a microliter syringe fitted in a Hamilton PB-600 repeating dispenser (Reno, Nevada).
Bees were also fed sucrose, HFCS, or honey candy and treated topically with three doses of tau-fluvalinate  (3, 5 and 10 µg) or imidacloprid  (0.005, 0.01 and 0.03 µg) ranging between the LD25 and the LD50. All trials included an acetone control, and no mortality was observed in any control bees.
Following treatment, bees were placed in wax-coated paper cups (177 cm3; Sweetheart, Owings Mills, MD) that were covered with cotton cheesecloth secured by two rubber bands. Sucrose water (1∶1 sucrose and water) was provided in a punctured 1.5 ml plastic tube. Bees were maintained in a dark 32–34°C incubator until mortality was assessed 24 h after treatment. Bees incapable of righting themselves inside the cup were scored as dead.
The R statistical package  with MASS libraries  was used to perform log-probit analyses of mortality data represented in Table 1. Fieller's method was used for calculation of LD50 values and 95% confidence intervals, with correction for heterogeneity where appropriate . Non-overlapping 95% confidence intervals at the LD50 level were considered significantly different. SPSS 17.0 was used for bioassay analyses related to the effects of HFCS, sucrose and honey diets (SPSS Inc., Chicago, IL). One-way analysis of variance (ANOVA) was used to determine whether susceptibility to tau-fluvalinate and imidacloprid changes when bees are fed HFCS, sucrose or honey candy.
To test the effect of diet on longevity in the presence of aflatoxin, newly emerged bees were transferred in groups of 20 to wax-coated paper cups and fed one of seven treatments: sucrose candy, HFCS candy or honey candy, with the addition of either 0.1% DMSO as a control or 20 µg/g aflatoxin B1. A group fed on pure sucrose candy was also included. After treatment, bees were placed in an incubator and monitored in 6 h intervals for 72 h.
A Kaplan-Meier Survival Analysis was conducted to examine differential survivorship in the presence of aflatoxin based on diet. The median time for survival was calculated for each treatment independently (Table 2).
Experiments were designed to meet Minimum Information About a Microarray Experiment (MIAME) standards and all microarray data obtained in these studies were deposited in NCBI-GEO (http://www.ncbi.nlm.nih.gov/projects/geo Accession Number GSE34029). A custom microarray  was constructed by The Functional Genomics Unit of the W. M. Keck Center. Included on the array were probes specific for 45 P450 genes, 10 carboxylesterase genes and 7 glutathione-S-transferase genes along with 206 chemosensory-related genes and 17 tetraspanins, as well as houskeeping genes and controls, for 313 genes in total, using the A. mellifera assembly 2 as the basis for probe design .
Microarray RNA isolation
Frames of brood from five different colonies were collected and newly emerging adults were fed phenobarbital (2.5 µg/g candy) or plain candy for three days as described. Total RNA was then isolated from 10 whole honey bee workers by first grinding in liquid nitrogen using a mortar and pestle and then extracting RNA with Trizol (Invitrogen, Carlsbad, CA), following the manufacturer's instructions. RNA concentration was quantified on a spectrophotometer and visually assessed on agarose gels.
cDNA synthesis for microarrays
Fifteen µg of RNA from each treatment was reverse-transcribed into cDNA overnight at 46°C using SuperscriptIII (Invitrogen, Carlsbad, CA) with an oligo-dT16 primer and amino-allyl dNTP. cDNA was purified over a Qiaquick PCR purification kit column (Qiagen, Valencia, CA), substituting a phosphate wash for the provided kit buffers and then dried in a SpeedVac. The cDNA was labeled with either Cy-3 or Cy-5 mono-reactive dyes (GE Healthcare, Piscataway, NJ) in sodium carbonate buffer for 1 h in complete darkness and then purified over a PCR purification kit column; the cDNA concentration and labeling efficiency were quantified by spectrophotometer.
Spotted oligonucleotides were rehydrated by passing microarray slides through steam and then cross-linked by UV light exposure. Slides were then vigorously washed in 0.2% SDS, placed in prehybridization buffer for 1 h at 42°C, washed in ultrapure water followed by isopropanol, and spun dry. Labeled cDNA was resuspended in water and denatured on a 95°C block. Hybridization buffer (2×) was added to the probes and pipetted under a Lifterslip (Fisher, Pittsburgh, PA) covering the array. Arrays were hybridized in Corning (Lowell, MA) hybridization chambers overnight at 42°C in complete darkness. Arrays were washed in successively less stringent wash buffers, then spun-dry and stored in darkness until scanned on an Axon Instruments 4000B Scanner using GenePix Pro (Molecular Devices, Sunnyvale, CA) software.
Microarray statistical analysis
The LIMMA/Bioconductor/R statistical package was used for statistical analysis of the intensity data from the arrays , , . NORMEXP was used for background correction , followed by LOESS correction within arrays . Specialty microarrays pose special problems during normalization; the LOESS normalization procedure has been found to be valid for arrays with as many as 20% of probes showing differential expression . Detoxification genes make up approximately 20% of the genes spotted on the array, so LOESS normalization was used. DUPCOR from the LIMMA package was used to estimate the correlation between duplicate spots on the arrays . Filtered corrected intensity values were fitted to a linear model and then ranked in order of evidence for differential expression using EBAYES. An intensity filter was applied to remove spots with average intensity values less than the intensity of negative control spots. A p-value<0.05, after false-discovery-rate correction , was established as the cutoff for genes differentially expressed.
Gut dissection and measurement
The hindgut and midgut were dissected from 20 three-day-old bees fed on candy containing honey, pollen or propolis extracts, as were guts from bees fed unaugmented candy, by pulling on the sting with forceps. Dissected midguts were immediately separated on a chilled glass Petri dish and frozen in liquid nitrogen. Frozen guts were ground in a mortar using a pestle and RNA was extracted using Trizol. Parallel bioassays were set up to provide guts for morphological characterization; these midguts were dissected, stretched on a glass Petri dish by dragging with forceps, and measured across their widest girth to the nearest 0.1 mm with a dissecting microscope fitted with an ocular micrometer.
Northern blot analysis
Expression of five CYP6AS subfamily gene transcripts, chosen based on their expression in the microarray experiments, was assayed with a northern blot. Probes specific for the entire P450 transcript were labeled with [α-32P]dATP (Amersham Biosciences) and purified using G-50 packed column. Total RNA (20 µg of each sample) was heated in loading buffer at 65°C for 15 min and separated on 1% formaldehyde-agarose gel by electrophoresis. After transferring RNA to Hybond-XL nylon membranes (Amersham Biosciences), the membranes were hybridized with probes following the manufacturers' procedures for these membranes. Northern blots were visualized using x-ray film exposed for 4 h, a duration chosen to minimize signal saturation.
We thank Arthur Zangerl for assistance making honey and pollen extracts, Alex Wild and Josephine Rodriguez for help photographing bee guts, Hugh Robertson and Kim Walden for access to and assistance with the honey bee specialty microarrays, Karen Pruiett for help with hive-work, Gene Robinson for access to UIUC apiaries and Archer Daniels Midland for providing high-fructose corn syrup.
Conceived and designed the experiments: RMJ WM HSP GN MAS MRB. Performed the experiments: RMJ WM HSP GN. Analyzed the data: RMJ WM HSP GN. Wrote the paper: RMJ HSP MRB.
- 1. Morse RA, Calderone NW (2000) The value of honey bee pollination the United States. Bee Cult 128: 1–15.
- 2. Johansen CA, Mayer DF (1990) Pollinator protection: A bee and pesticide handbook. Cheshire, CT: Wicwas Press. 212 p.
- 3. Johnson RM, Ellis MD, Mullin CA, Frazier M (2010) Pesticides and honey bee toxicity – USA. Apidologie 41: 312–331.
- 4. Schell R (2005) A big win for beekeepers. Am Bee J 145: 392–394.
- 5. Cresswell J (2011) A meta-analysis of experiments testing the effects of a neonicotinoid insecticide (imidacloprid) on honey bees. Ecotoxicology 20: 149–157.
- 6. Feyereisen R (2005) Insect cytochrome P450s. Biochemistry and molecular biology. Comprehensive molecular insect science. Vol. 4. Oxford: Elsevier Pergamon. pp. 1–77.
- 7. Mao W, Rupasinghe SG, Johnson RM, Zangerl AR, Schuler MA, et al. (2009) Quercetin-metabolizing CYP6AS enzymes of the pollinator Apis mellifera (Hymenoptera: Apidae). Comp Biochem Physiol B Biochem Mol Biol 154: 427–434.
- 8. Wiermann R (1968) Phenylpropanoid metabolism by pollen I. Survey of flavonoid components isolated from gymnosperms and angiosperms. Ber Deut Bot Ges 81: 3–16.
- 9. Campos M, Markham KR, Mitchell KA, Cunha AP da (1997) An approach to the characterization of bee pollens via their flavonoid/phenolic profiles. Phytochem Anal 8: 181–185.
- 10. Kenjeric D, Mandic ML, Primorac L, Bubalo D, Perl A (2007) Flavonoid profile of Robinia honeys produced in Croatia. Food Chem 102: 683–690.
- 11. Medina A, Gonzalez G, Saez JM, Mateo R, Jimenez M (2004) Bee pollen, a substrate that stimulates ochratoxin A production by Aspergillus ochraceus Wilh. Syst Appl Microbiol 27: 261–267.
- 12. Gonzalez G, Hinojo MJ, Mateo R, Medina A, Jimenez M (2005) Occurrence of mycotoxin producing fungi in bee pollen. Int J Food Microbiol 105: 1–9.
- 13. Niu G, Johnson RM, Berenbaum MR (2011) Toxicity of mycotoxins to honeybees and its amelioration by propolis. Apidologie 42: 79–87.
- 14. Johnson RM, Wen Z, Schuler MA, Berenbaum MR (2006) Mediation of pyrethroid insecticide toxicity to honey bees (Hymenoptera: Apidae) by cytochrome P450 monooxygenases. J Econ Entomol 99: 1046–1050.
- 15. Mao W, Schuler MA, Berenbaum MR (2011) CYP9Q-mediated detoxification of acaricides in the honey bee (Apis mellifera). Proc Natl Acad Sci U S A.
- 16. Pilling ED, Bromley-Challenor KAC, Walker CH, Jepson PC (1995) Mechanism of synergism between the pyrethroid insecticide lambda-cyhalothrin and the imidazole fungicide prochloraz, in the honeybee (Apis mellifera L.). Pestic Biochem Physiol 51: 1–11.
- 17. Johnson RM, Pollock HS, Berenbaum MR (2009) Synergistic interactions between in-hive miticides in Apis mellifera. J Econ Entomol 102: 474–479.
- 18. Hardstone MC, Scott JG (2010) Is Apis mellifera more sensitive to insecticides than other insects? Pest Manag Sci 66: 1171–1180.
- 19. The Honeybee Genome Sequencing Consortium (2006) Insights into social insects from the genome of the honeybee Apis mellifera. Nature 443: 931–949.
- 20. ffrench-Constant RH, Daborn PJ, Le Goff G (2004) The genetics and genomics of insecticide resistance. Trends Genet 20: 163–170.
- 21. Claudianos C, Ranson H, Johnson RM, Biswas S, Schuler MA, et al. (2006) A deficit of detoxification enzymes: Pesticide sensitivity and environmental response in the honeybee. Insect Mol Biol 15: 615–636.
- 22. Liu F-L, Zhang X-W, Chai J-P, Yang D-R (2006) Pollen phenolics and regulation of pollen foraging in honeybee colony. Behav Ecol Sociobiol 59: 582–588.
- 23. Nebert DW, Gonzalez FJ (1987) P450 genes: structure, evolution, and regulation. Annu Rev Biochem 56: 945–993.
- 24. Lewis DF (2002) Oxidative stress: the role of cytochromes P450 in oxygen activation. J Chem Technol Biotech 77: 1095–1100.
- 25. Li X, Schuler MA, Berenbaum MR (2007) Molecular mechanisms of metabolic resistance to synthetic and natural xenobiotics. Annu Rev Entomol 52: 231–253.
- 26. Kezic N, Lucic D, Sulimanovic D (1992) Induction of mixed function oxidase activity in honey bee as a bioassay for detection of environmental xenobiotics. Apidologie 23: doi:10.1051/apido:19920304. 217–223.
- 27. Remmer H, Merker HJ (1963) Drug-induced changes in the liver endoplasmic reticulum: Association with drug-metabolizing enzymes. Science 142: 1657–1658.
- 28. Denison MS, Whitlock JP (1995) Xenobiotic-inducible transcription of cytochrome P450 genes. J Biol Chem 270: 18175–18178.
- 29. Terriere LC, Yu SJ (1974) The induction of detoxifying enzymes in insects. J Agric Food Chem 22: 366–373.
- 30. Agosin M (1976) Insect cytochrome P-450. Mol Cell Biochem 12: 33–44.
- 31. Amichot M, Brun A, Cuany A, De Souza G, Le Mouel T, et al. (1998) Induction of cytochrome P450 activities in Drosophila melanogaster strains susceptible or resistant to insecticides. Comp Biochem Physiol C Comp Pharmacol 121: 311–319.
- 32. Fisher T, Crane M, Callaghan A (2003) Induction of cytochrome P-450 activity in individual Chironomus riparius Meigen larvae exposed to xenobiotics. Ecotoxicol Environ Saf 54: 1–6.
- 33. Brattsten LB, Wilkinson CF (1975) Properties of 5-aminolaevulinate synthetase and its relationship to microsomal mixed-function oxidation in the southern armyworm (Spodoptera eridania). Biochem J 150: 97–104.
- 34. Yu SJ, Ing RT (1984) Microsomal biphenyl hydroxylase of fall armyworm larvae and its induction by allelochemicals and host plants. Comp Biochem Physiol C 78: 145–152.
- 35. Yu SJ (1995) Allelochemical stimulation of ecdysone 20-monooxygenase in fall armyworm larvae. Arch Insect Biochem Physiol 28: 365–375.
- 36. Natsuhara K, Shimada K, Tanaka T, Miyata T (2004) Phenobarbital induction of permethrin detoxification and phenobarbital metabolism in susceptible and resistant strains of the beet armyworm Spodoptera exigua (Hübner). Pestic Biochem Physiol 79: 33–41.
- 37. Anspaugh DD, Roe RM (2005) Regulation of JH epoxide hydrolase versus JH esterase activity in the cabbage looper, Trichoplusia ni, by juvenile hormone and xenobiotics. J Insect Physiol 51: 523–35.
- 38. Brown D, Zhang L, Wen Z, Scott JG (2003) Induction of P450 monooxygenases in the German cockroach, Blattella germanica L. Arch Insect Biochem Physiol 53: 119–124.
- 39. Wilkinson CF, Brattsten LB (1973) Microsomal drug metabolizing enzymes in insects. In: Davidson IWF, editor. Drug metabolism reviews. Vol. 1. Marcel Dekker, Inc. pp. 153–227.
- 40. Riskallah MR, Dauterman WC, Hodgson E (1986) Host plant induction of microsomal monooxygenase activity in relation to diazinon metabolism and toxicity in larvae of the tobacco budworm Heliothis virescens (F.). Pestic Biochem Physiol 25: 233–247.
- 41. Zeng R, Wen Z, Niu G, Schuler M, Berenbaum M (2007) Allelochemical induction of cytochrome P450 monooxygenases and amelioration of xenobiotic toxicity in Helicoverpa zea. J Chem Ecol 33: 449–461.
- 42. Guirguis G, Brindley W (1976) Effect of chlorcyclizine, aminopyrine, or phenobarbital on carbaryl metabolism in alfalfa leafcutting bees (Hymenoptera-Megachilidae). Environ Entomol 5: 590–594.
- 43. Snyder MJ, Stevens JL, Andersen JF, Feyereisen R (1995) Expression of cytochrome P450 genes of the CYP4 family in midgut and fat body of the tobacco hornworm, Manduca sexta. Arch Biochem Biophys 321: 13–20.
- 44. Ranasinghe C, Hobbs AA (1999) Induction of cytochrome P450 CYP6B7 and cytochrome b5 mRNAs from Helicoverpa armigera (Hubner) by pyrethroid insecticides in organ culture. Insect Mol Biol 8: 443–447.
- 45. Stevens JL, Snyder MJ, Koener JF, Feyereisen R (2000) Inducible P450s of the CYP9 family from larval Manduca sexta midgut. Insect Biochem Mol Biol 30: 559–568.
- 46. Li X, Berenbaum MR, Schuler MA (2000) Molecular cloning and expression of CYP6B8: A xanthotoxin-inducible cytochrome P450 cDNA from Helicoverpa zea. Insect Biochem Mol Biol 30: 75–84.
- 47. Scott JG, Lee SS (1993) Tissue distribution of microsomal cytochrome P-450 monooxygenases and their inducibility by phenobarbital in the insecticide resistant LPR strain of house fly, Musca domestica L. Insect Biochem Mol Biol 23: 729–738.
- 48. Carino FA, Koener JF, Plapp FW, Feyereisen R (1994) Constitutive overexpression of the cytochrome P450 gene CYP6A1 in a house fly strain with metabolic resistance to insecticides. Insect Biochem Mol Biol 24: 411–418.
- 49. Dunkov BC, Guzov VM, Mocelin G, Shotkoski F, Brun A, et al. (1997) The Drosophila cytochrome p450 gene Cyp6a2: Structure, localization, heterologous expression, and induction by phenobarbital. DNA Cell Biol 16: 1345–1356.
- 50. Brun A, Cuany A, Le Mouel T, Berge J, Amichot M (1996) Inducibility of the Drosophila melanogaster cytochrome P450 gene, CYP6A2, by phenobarbital in insecticide susceptible or resistant strains. Insect Biochem Mol Biol 26: 697–703.
- 51. Danielson PB, MacIntyre RJ, Fogleman JC (1997) Molecular cloning of a family of xenobiotic-inducible drosophilid cytochrome p450s: Evidence for involvement in host-plant allelochemical resistance. Proc Natl Acad Sci U S A 94: 10797–802.
- 52. Kasai S, Scott JG (2001) Expression and regulation of CYP6D3 in the house fly, Musca domestica (L.). Insect Biochem Mol Biol 32: 1–8.
- 53. King-Jones K, Thummel CS (2005) Nuclear receptors – a perspective from Drosophila. Nat Rev Genet 6: 311–323.
- 54. Le Goff G, Hilliou F, Siegfried BD, Boundy S, Wajnberg E, et al. (2006) Xenobiotic response in Drosophila melanogaster: Sex dependence of P450 and GST gene induction. Insect Biochem Mol Biol 36: 674–682.
- 55. Sun W, Margam VM, Sun L, Buczkowski G, Bennett GW, et al. (2006) Genome-wide analysis of phenobarbital-inducible genes in Drosophila melanogaster. Insect Mol Biol 15: 455–464.
- 56. Willoughby L, Chung H, Lumb C, Robin C, Batterham P, et al. (2006) A comparison of Drosophila melanogaster detoxification gene induction responses for six insecticides, caffeine and phenobarbital. Insect Biochem Mol Biol 36: 934–42.
- 57. Tushar T, Vinod T, Rajan S, Shashindran C, Adithan C (2007) Effect of honey on CYP3A4, CYP2D6 and CYP2C19 enzyme activity in healthy human volunteers. Basic Clin Pharmacol Toxicol 100: 269–272.
- 58. Treutter D (2005) Significance of flavonoids in plant resistance and enhancement of their biosynthesis. Plant Biology 7: 581–591.
- 59. Liu X, Liang P, Gao X, Shi X (2006) Induction of the cytochrome P450 activity by plant allelochemicals in the cotton bollworm, Helicoverpa armigera (Hübner). Pestic Biochem Physiol 84: 127–134.
- 60. Yu SJ (1983) Induction of detoxifying enzymes by allelochemicals and host plants in the fall armyworm. Pestic Biochem Physiol 19: 330–336.
- 61. Simone-Finstrom M, Spivak M (2010) Propolis and bee health: the natural history and significance of resin use by honey bees. Apidologie 41: 295–311.
- 62. Cohen MB, Schuler MA, Berenbaum MR (1992) A host-inducible cytochrome P-450 from a host-specific caterpillar: Molecular cloning and evolution. Proc Natl Acad Sci U S A 89: 10920–10924.
- 63. Mao W, Berhow M, Zangerl A, McGovern J, Berenbaum M (2006) Cytochrome P450-mediated metabolism of xanthotoxin by Papilio multicaudatus. J Chem Ecol 32: 523–536.
- 64. Li X, Schuler MA, Berenbaum MR (2002) Jasmonate and salicylate induce expression of herbivore cytochrome P450 genes. Nature 419: 712–715.
- 65. Suchail S, Debrauwer L, Belzunces LP (2004) Metabolism of imidacloprid in Apis mellifera. Pest Manag Sci 60: 291–296.
- 66. Yu SJ, Robinson FA, Nation JL (1984) Detoxication capacity in the honey bee, Apis mellifera L. Pestic Biochem Physiol 22: 360–368.
- 67. Furgala B, McCutcheon DM (1992) Wintering productive colonies. In: Graham JM, editor. The hive and the honey bee. Hamilton, IL: Dadant & Sons, Inc. pp. 829–868.
- 68. Wanner KW, Nichols AS, Walden KKO, Brockmann A, Luetje CW, et al. (2007) A honey bee odorant receptor for the queen substance 9-oxo-2-decenoic acid. Proc Natl Acad Sci U S A 104: 14383–14388.
- 69. Oakeshott JG, Johnson RM, Berenbaum MR, Ranson H, Cristino AS, et al. (2010) Metabolic enzymes associated with xenobiotic and chemosensory responses in Nasonia vitripennis. Insect Mol Biol 19: 147–163.
- 70. Bull DL, Ivie GW, Beier RC, Pryor NW (1986) In vitro metabolism of a linear furanocoumarin (8-methoxypsoralen, xanthotoxin) by mixed-function oxidases of larvae of black swallowtail butterfly and fall armyworm. J Chem Ecol 12: 885–892.
- 71. Tanaka E (1999) Clinically significant pharmacokinetic drug interactions between antiepileptic drugs. J Clin Pharm Ther 24: 87–92.
- 72. Berenbaum MR, Cohen MB, Schuler MA (1990) Cytochrome P450 in plant-insect interactions: Inductions and deductions. Molecular insect science. New York: Plenum Press. pp. 257–262.
- 73. Brodschneider R, Crailsheim K (2010) Nutrition and health in honey bees. Apidologie 41: 278–294.
- 74. LeBlanc BW, Eggleston G, Sammataro D, Cornett C, Dufault R, et al. (2009) Formation of hydroxymethylfurfural in domestic high-fructose corn syrup and its toxicity to the honey bee (Apis mellifera). J Agric Food Chem 57: 7369–7376.
- 75. Pilling ED, Jepson PC (1993) Synergism between EBI fungicides and a pyrethroid insecticide in the honeybee (Apis mellifera). Pestic Sci 39: 293–297.
- 76. VanEngelsdorp D, Underwood R, Caron D, Hayes J (2007) An estimate of managed colony losses in the winter of 2006–2007: A report commissioned by the apiary inspectors of America. Am Bee J 147: 603.
- 77. Mullin CA, Frazier M, Frazier JL, Ashcraft S, Simonds R, et al. (2010) High levels of miticides and agrochemicals in North American apiaries: implications for honey bee health. PLoS ONE 5: e9754.
- 78. Suchail S, Guez D, Belzunces LP (2000) Characteristics of imidacloprid toxicity in two Apis mellifera subspecies. Environ Toxicol Chem 19: 1905.
- 79. R Development Core Team (2011) R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. Available: http://www.R-project.org.
- 80. Venables WN, Ripley BD (2002) Modern Applied Statistics with S. 4th ed. Springer, New York 495.
- 81. Finney DJ (1971) Probit Analysis. 3rd ed. Cambridge: University Press. 333 p.
- 82. Gentleman R, Carey V, Bates D, Bolstad B, Dettling M, et al. (2004) Bioconductor: Open software development for computational biology and bioinformatics. Genome Biol 5: R80.
- 83. Smyth GK (2005) Limma: Linear models for microarray data. In: Gentleman R, Carey V, Dudoit S, Irizarry R, Huber W, editors. Bioinformatics and computational biology solutions using R and Bioconductor. Springer, New York. pp. 397–420.
- 84. Ritchie ME, Silver J, Oshlack A, Holmes M, Diyagama D, et al. (2007) A comparison of background correction methods for two-colour microarrays. Bioinformatics 23: 2700–2707.
- 85. Yang YH, Dudoit S, Luu P, Lin DM, Peng V, et al. (2002) Normalization for cDNA microarray data: A robust composite method addressing single and multiple slide systematic variation. Nucleic Acids Res 30: e15.
- 86. Oshlack A, Emslie D, Corcoran LM, Smyth GK (2007) Normalization of boutique two-color microarrays with a high proportion of differentially expressed probes. Genome Biol 8: R2.
- 87. Reiner A, Yekutieli D, Benjamini Y (2003) Identifying differentially expressed genes using false discovery rate controlling procedures. Bioinformatics 19: 368–375.
- 88. Thompson JD, Higgins DG, Gibson TJ (1994) CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res 22: 4673–4680.
- 89. Felsenstein J (n.d.), PHYLIP (Phylogeny Inference Package) Department of Genome Sciences, University of Washington, Seattle. Distributed by the author.
- 90. Schmidt HA, Strimmer K, Vingron M, von Haeseler A (2002) TREE-PUZZLE: maximum likelihood phylogenetic analysis using quartets and parallel computing. Bioinformatics 18: 502–504.