SEARCH

SEARCH BY CITATION

Keywords:

  • Apis mellifera;
  • Cucurbita pepo;
  • genetically modified organism;
  • GMO;
  • Peponapis;
  • plant virus;
  • Xenoglossa;
  • Zucchini yellow mosaic virus

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

1. Two ecological risks associated with the use of transgenic crops are transgene movement into wild populations and effects on non-target organisms, such as pollinators. Despite the importance of pollinators, and their contribution to the global food supply, little is known about how they are affected by transgenic crops. Pollinator preferences affect plant mating patterns; thus understanding the effects of transgenic crops on pollinators will aid in understanding transgene movement.

2. Honey bee and squash bee visit number and duration were recorded on conventional and transgenic virus-resistant squash Cucurbita pepo planted in a randomized block design. Floral characters were measured to explain differences in pollinator behaviour. The effect of Zucchini Yellow Mosaic Virus infection on pollinator behaviour was also examined.

3. Honey bees visited female conventional flowers more than female transgenic flowers. Conventional flowers were generally larger with more nectar than transgenic flowers, although floral traits did not account for differences in pollinator visitation.

4. Squash bees visited male transgenic flowers more than male conventional flowers; squash bees also spent more time in female transgenic flowers than in female conventional flowers. Transgenic flowers were significantly larger with greater amounts of sweeter nectar and they were present in greater number. Floral traits accounted for some of the variation in pollinator visitation.

5. Squash bee visit number and duration did not differ between virus-infected and healthy plants, but this may be because pollinator behaviour was observed early in the virus infection.

6.Synthesis and applications. Pollinator behaviour controls patterns of plant mating thus non-target effects of transgenic resistance, such as those observed here, may influence transgene movement into wild populations. These results suggest that transgenic crops should not be planted within the native range of wild relatives because pleiotropic effects may affect crop-wild hybridization and transgene introgression into wild populations.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Transgenic crops have been grown commercially since 1992 (James 1998). Since their introduction the number of hectares planted with transgenic varieties has increased from 1·7 million hectares in five countries in 1996 (James 1998) to more than 125 million hectares in 25 countries in 2008 (James 2008). Although many countries, including the United States and Argentina, rapidly adopted transgenic crops, others, such as those in the European Union and Africa, have been hesitant to accept this technology because of potential ecological and economic risks (Devos et al. 2006; Wafula, Persley, & Karembu 2008).

There are two ecological risks associated with the use of transgenic crops. First, transgenes may move into wild populations (following transgenic crop-wild hybridization). Crop-wild hybridization and subsequent introgression provides wild plants with novel traits, and these traits could potentially alter the size and dynamics of wild plant populations. Crop-wild hybridization is common (Wilson 1990; Ellstrand 2003), indicating that transgene introgression from transgenic crops to wild populations is likely when a crop and wild relative co-occur. In fact, transgene introgression into wild and feral populations has been identified in canola (Brassica napus L., Hall 2000) and creeping bentgrass (Agrostis stolonifera L., Watrud et al. 2004). However, in these cases the effects of the transgene on plant population dynamics are unknown (although see Claessen et al. 2005).

The second ecological risk associated with the use of transgenic crops is the potential effects on non-target organisms (Pilson & Prendeville 2004; O’Callaghan et al. 2005; Felber et al. 2007). Non-target effects can occur when organisms that do not reduce yield are negatively affected by the product of the transgene itself (Pilson & Prendeville 2004). For instance, a transgenic crop that produces an insecticide may negatively affect beneficial insects (Groot & Dicke 2002). Non-target effects can also be caused by a pleiotropic effect of the transgene. For example, transgenic herbicide-resistant canola produced fewer flowers in comparison with conventional canola (Pierre et al. 2003). Reduced flower production because of a transgene could, in turn, affect pollinator behaviour. Non-target effects and transgene introgression have usually been considered distinct ecological risks (but see Cresswell & Osborne 2004; Hoyle, Hayter, & Cresswell 2007). However, in crops that require insect-mediated pollination, patterns of pollinator visitation may differ between transgenic and conventional varieties (Picard-Nizou et al. 1997; Pham-Delègue et al. 2000; Malone 2004). Such preference differences could affect the frequency of crop-wild hybridization and transgene introgression into wild populations.

Insect-mediated pollination is an important ecosystem service that contributes to the production of more than one-third of our global food supply (Klein et al. 2007). Despite the importance of pollinators and concern about pollinator declines (Allen-Wardell et al. 1998; Kearns, Inouye, & Waser 1998; Pauw 2007; Goulson, Lye, & Darvill 2008), little is known about non-target effects of transgenic crops on native or introduced pollinators. Laboratory toxicity tests suggest that Bt toxin has no effect on honey bees and bumble bees (Morandin & Winston 2003; Malone 2004; Malone et al. 2007; Duan et al. 2008), but that protease inhibitors reduce honey bee survival and development (Burgess, Malone, & Christeller 1996; Malone et al. 1998; Brødsgaard et al. 2003). The results of field studies have been variable, indicating for example, no effect of transgenic crops on pollinator abundance, diversity (Pierre et al. 2003), and development time (Huang et al. 2004); lower pollinator abundance in transgenic fields (Haughton et al. 2003; Morandin & Winston 2005); and higher pollinator abundance in transgenic crop margins (Roy et al. 2003). In addition, transgenic canola produced a greater volume of sweeter nectar than conventional canola in one study (Picard-Nizou et al. 1995) but not in another (Pierre et al. 2003). Pollinator visitation to conventional and transgenic canola was similar in both of these studies. These idiosyncratic results indicate the need for further investigation of non-target effects of transgenic crops on pollinators.

Predicting transgene movement from transgenic crops to wild populations, as well as changes in transgene frequency within wild populations requires an understanding of pollinator behaviour on transgenic crops. In this study, we examined pollinator behaviour on transgenic virus-resistant and conventional squash Cucurbita pepo L. In particular, we compared: (1) pollinator visit number and duration on virus-resistant transgenic and conventional squash and (2) pollinator visit number and duration on virus-infected and healthy squash. In addition, to explain observed differences in pollinator behaviour, we (3) compared floral traits on conventional and transgenic plants, and on virus-infected and healthy plants.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Natural history

Cultivated summer squash Cucurbita pepo L. is a monoecious annual and requires insect-mediated pollination for reproduction. Squash plants produce flowers for several weeks; however individual flowers last for <1 day, opening at dawn and closing around noon, depending upon environmental conditions. Squash is visited by a variety of pollinators. Two common and efficient pollinators of squash are honey bees (Apis mellifera, L.; a generalist social bee) and, in North America, squash bees (Peponapis spp. and Xenoglossa spp.) (Tepedino 1981). Squash bees are oligolectic solitary bees that provision their nests only with cucurbit pollen (Hurd & Linsley 1964).

Wild squash is native to south-central US and northern Mexico, and this range overlaps with commercial squash production (Wilson 1993). Cultivated squash readily interbreeds with its wild progenitor (also C. pepo), and non-transgenic cultivated alleles have been identified in wild squash populations (Wilson 1990, 1993; Decker-Walters et al. 2002). Wild and cultivated squash are susceptible to mosaic viruses, which are transmitted by aphids and beetles. In the experiment described here, plants in the virus-infected treatment were inoculated with Zucchini Yellow Mosaic Virus, which is found in wild squash populations (Quemada et al. 2008), is common in cultivated squash production fields, and is transmitted by aphids (Lecoq, Pitrat, & Clement 1981).

Virus infection typically affects plant metabolism and transpiration (Hull 2002) and may also affect floral traits, including pigmentation (Kruckelmann & Seyffert 1970), nectar volume and nectar sugar concentration. In squash, mosaic virus infection causes mottling and deformity of fruits, leaves, and flowers, and can drastically reduce yield (Fuchs & Gonsalves 1995; Gianessi et al. 2002). Because of the effect on yield, virus-resistant transgenic squash was developed, deregulated in the US, and has been available for commercial use since 1994 (USDA/APHIS 1994).

In these experiments we used four commercially available varieties (Declaration II, Patriot II, Prelude II, and Independence II) resistant to two potyviruses (Zucchini yellow mosaic virus and Watermelon mosaic virus) and three varieties (Destiny III, Liberator III, and Conqueror III) resistant to the two potyviruses and a cucumovirus (Cucumber mosaic virus; Tricoli et al. 1995). We paired these seven transgenic varieties with seven conventional varieties (Dixie, Sensation, Lemondrop, Superpik, Seneca, Prelude and Spineless Beauty) with similar agronomic properties (fruit type, size, growth form and time to flower). For three of these conventional-transgenic pairs (Dixie/Destiny III: Lin et al. 2003; Lemondrop/Liberator III; Prelude/Prelude II, Rowell, Nesmith, & Snyder 1999), breeders backcrossed the original transgenic line into the conventional line for multiple generations such that these two lines are nearly isogenic except for the presence or absence of the virus-resistance transgene construct.

Growing season temperature and precipitation were within the normal range in 2004 and 2005, when we performed our experiments. However, the 2004 growing season was generally cooler, wetter, and with less solar radiation in comparison with the 2005 growing season (Table S1, Supporting Information).

Experiment 1: effect of transgenic virus-resistance on honey bee behaviour (2004)

Field methods

On 5 May 2004 greenhouse-reared seedlings from each of the seven transgenic and seven conventional varieties were transplanted without replication into 15 spatial blocks in a common garden at the University of Nebraska-Lincoln, for a total of 210 plants. Flowering began on 4 June and peak flowering coincided with honey bee activity. We observed honey bees from 29 June to 20 July. Squash bees were present, but only became abundant after peak flowering in the experimental plot.

Spatial blocks were haphazardly assigned to observation days and times, such that each block was observed three times during the experiment, once in each of three observation periods (6:00–7:30am, 7:30–9:00am and 9:00–10:30am). During each observation period an observer walked through the block and recorded all observed pollinators on each male and female flower on each plant. When a pollinator was observed landing in a flower, visit duration (seconds) was recorded. On each day that pollinators were observed the number of male and female flowers on each plant in the observed blocks was also recorded. Because of herbivore damage or natural variation in flowering phenology, not all plants had open flowers on all observation days.

We used digital calipers to measure corolla length and width on a single male and female flower on each of the 210 plants in the experiment. Nectar volume was measured with Drummond microcapillary tubes and nectar sugar concentration was quantified with a hand-held Brix refractometer on a single male and female flower on each plant. Nectar measurements were made between 7:30 and 11:30am when nectar volume is near its peak and before reabsorption occurs (Nepi, Pacini, & Willemse 1996). So that nectar could not be removed by pollinators or other insects prior to nectar measurements, flowers were bagged with pollination bags on the day before flowers opened. As a result of natural variation in flowering some plants did not have an open flower on days floral traits were measured. Thus, the data set contains missing values.

Statistical analyses

Before analysis, visits to male and female flowers on each plant were summed by flower sex over the three observation periods (i.e. total visits per plant over 4·5 h of observation per block). Thus, the unit of observation in our analyses is the plant. Similarly, visit duration on male and female flowers was averaged over all observed visits on each plant. To evaluate the effect of transgenic virus resistance on honey bee visit number and visit duration we used a generalized linear mixed model (GLMM; SAS 9·1; SAS Institute 2003) with error distributions chosen as appropriate for each response variable. In these analyses (as well as all that follow) block was considered a random effect and transgene (conventional and transgenic) and variety nested within transgene were considered fixed effects.

We examined pleiotropic effects of transgenic virus resistance on corolla length and corolla width of male and female flowers using manova. The transgene effect was significant in the manova thus we used univariate anova to examine the effects of transgenic virus resistance on each of these floral traits. Because of non-normality male and female flower number, nectar sugar concentration and nectar volume were analysed using GLMM (with appropriate error distributions) rather than manova. We used a sequential Bonferroni correction (Holm 1979) to control for multiple tests (one manova and five GLMM) of differences in floral characters between conventional and transgenic plants.

When the transgene effect was significant in analyses of all varieties, we also conducted post-hoc comparisons of each of the three near-isogenic line pairs (Dixie/Destiny III, Lemondrop/Liberator III and Prelude/Prelude II). We made post-hoc comparisons of honeybee visits to female flowers (GLMM) and corolla width on male and female flowers (anova). We used a sequential Bonferroni test (Holm 1979) to correct for the three within line pair comparisons.

Experiment 2: effect of transgenic virus-resistance and virus infection on squash bee behaviour (2005)

Field methods

In 2005 we planted a month later so that peak flowering would coincide with squash bee activity. Flowering began on 24 June. Honey bees were present but were not abundant during observations. On 9 June 2005 greenhouse reared seedlings of the same seven transgenic and seven conventional varieties (from the same seedlots) used in experiment 1 were transplanted into eight spatial blocks in a common garden experiment at the University of Nebraska-Lincoln. Each block was planted with two seedlings from each of the 14 varieties (224 plants total). One seedling from each variety in each block was later inoculated with Zucchini Yellow Mosaic Virus; the other was not experimentally infected.

Plants in the virus-infected treatment were inoculated on 1 July by rubbing one to two new leaves with 800 μl of phosphate buffer with celite and homogenized squash leaf tissue infected with Zucchini Yellow Mosaic Virus, a standard inoculation method (Hull 2002). One or two new leaves on plants in the healthy treatment were rubbed with 800 μl of phosphate buffer with celite and homogenized healthy squash leaf tissue, a standard control for plant handling. Virus inoculations were verified with reverse-transcriptase polymerase chain reaction (using methods similar to Pfosser & Baumann 2002).

Pollinator observations occurred from 18 July to 1 August, when virus symptoms began to appear on squash leaves and all plants were blooming. Observations were performed as previously described except that the observation period was just 1 h (6:00–7:00am, 7:00–8:00am or 8:00–9:00am) and two blocks were observed during each observation period (one by each of two observers). Thus, each block was observed twice in each time period for a total of six times during the experiment. Corolla length, corolla width, nectar volume and nectar sugar concentration on male and female flowers were measured as described above.

Statistical analyses

Squash bee visit number and duration on male and female flowers were calculated as above (i.e. total visits per plant was summed over the 6 h of observation of each block). Analyses included virus treatment, transgene, variety (transgene) and virus treatment × transgene as a fixed effects, and block as a random effect.

We analysed the effect of transgenic virus resistance on floral characters (corolla length, corolla width and nectar sugar concentration of male and female flowers) first by manova, and then, because transgene, virus treatment, and the transgene × virus treatment interaction were all significant, by univariate anova. Because of non-normality, flower number and nectar volume of male and female flowers were analysed using GLMM. We used a sequential Bonferroni correction (Holm 1979) to control for multiple tests (one manova and three GLMM) of differences in floral characters between conventional and transgenic varieties.

When the main effect of the transgene was significant we performed post-hoc comparisons within each of the three near-isogenic line pairs (Dixie/Destiny III, Lemondrop/Liberator III, Prelude/Prelude II). We used GLMM (with appropriate error distributions) to make post-hoc comparisons of visit number; visit duration; male and female nectar sugar concentration, flower number and nectar volume; and male corolla width. We used a sequential Bonferroni correction (Holm 1979) for multiple tests of differences between conventional and transgenic varieties within near-isogenic line pairs.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Experiment 1: effect of transgenic virus-resistance on honey bee behaviour (2004)

Honey bees visited female flowers on conventional plants more often than female flowers on transgenic plants (< 0·0001, d.f. = 133, Fig. 1a). However, honey bee visits to male flowers did not differ between transgenic and conventional plants (= 0·7940, d.f. = 115, Fig. 1a). In addition, honey bee visit duration on male and female flowers did not differ between conventional and transgenic plants (male: P = 0·1317, d.f. = 108; female: P = 0·4055, d.f. = 103, Fig. 1b).

image

Figure 1.  Mean pollinator visit number, visit duration, and floral traits for male and female flowers in 2004 and 2005 (with standard error bars; ***< 0·0001, **< 0·007, *< 0·05).

Download figure to PowerPoint

Floral traits differed between conventional and transgenic varieties (manova including corolla width and corolla length in male and female flowers as the multivariate response, P = 0·0030, d.f. = 121). Subsequent univariate anova found that corolla widths were greater on both male and female conventional flowers (P = 0·0209 d.f. = 161, P = 0·0132 d.f. = 136 respectively; Fig. 1c). However, honey bees only visited female conventional flowers more often, suggesting that traits not measured in this experiment may be responsible for differences in honey bee preference. No other measured floral traits (flower number, nectar sugar concentration, and nectar volume in male and female flowers) differed between conventional and transgenic plants (GLMM analyses; Fig. 1d–g).

Post-hoc comparisons of near-isogenic pairs showed a similar pattern as the full analysis. Honey bees visited female flowers on all three conventional lines more often than female flowers on the paired transgenic lines, although none of these comparisons was significant after Bonferroni correction (Table 1). Corolla width was greater on both male and female conventional flowers in each of the three near isogenic lines pairs, but these differences were not significant only for male flowers from one of three pairs (Table 1).

Table 1.   Pollinator visit number, visit duration, and floral traits of male and female flowers on conventional and transgenic near-isogenic line pairs (least square mean ± 1 SE)
 d.f.DixieDestiny IIIP-valued.f.LemondropLiberator IIIP-valued.f.PreludePrelude IIP-value
  1. Bold indicates significance at < 0·05 using a sequential Bonferroni.

2004
Honey bee visit no. to female flowers2410·33 ± 3·373·28 ± 1·070·0205236·71 ± 1·604·18 ± 1·190·2159216·91 ± 2·512·18 ± 0·900·0492
Female corolla width (mm)2458·29 ± 2·1059·60 ± 1·950·65042356·24 ± 2·4859·30 ± 2·480·39361956·07 ± 2·3854·19 ± 2·270·5759
Male corolla width (mm)2858·91 ± 1·7657·17 ± 1·760·49002657·39 ± 1·3357·44 ± 1·240·98172757·24 ± 1·1056·05 ± 1·140·4580
2005
Squash bee visit no. to male flowers2616·61 ± 2·2822·46 ± 2·880·1211307·06 ± 1·2312·68 ± 1·230·0031257·84 ± 2·2217·42 ± 2·140·0047
Male flower nectar sugar concentration (%)2540·61 ± 0·7543·00 ± 0·720·03172738·35 ± 0·6440·86 ± 0·610·00892241·70 ± 0·8244·21 ± 0·690·0289
Male flower nectar volume (μl) 2448·18 ± 7·2373·39 ± 7·230·02642921·54 ± 4·7736·60 ± 4·620·03102324·99 ± 5·7533·79 ± 5·100·2644
Average male flower no. day−1269·30 ± 1·518·20 ± 1·240·5734308·18 ± 0·688·56 ± 0·710·7069256·69 ± 1·3310·21 ± 1·950·1383
Male corolla length (mm)2561·12 ± 1·3364·19 ± 1·250·10712958·75 ± 1·1063·28 ± 1·530·08752357·53 ± 4·6359·54 ± 5·410·6385
Squash bee visit duration on female flowers (s)1613·62 ± 2·4512·70 ± 2·290·7874237·45 ± 5·8410·89 ± 5·650·0832912·40 ± 9·5915·86 ± 8·500·6344
Female flower nectar volume (μl)2444·31 ± 8·0472·67 ± 12·210·05712925·23 ± 0·9445·32 ± 0·990·01962347·90 ± 0·7654·08 ± 0·680·0194
Female flower nectar sugar concentration (%)2438·66 ± 1·1336·71 ± 1·050·21932734·92 ± 0·9437·80 ± 0·990·04522339·27 ± 0·7641·85 ± 0·680·0194

Experiment 2: effect of transgenic virus-resistance and virus infection on squash bee behaviour (2005)

Squash bees visited male transgenic flowers significantly more often than male conventional flowers (< 0·0001, d.f. = 173), but visited female conventional and transgenic flowers equally (P = 0·1614, d.f. = 124, Fig. 1h). Visit duration was longer on female transgenic flowers (P = 0·0259, d.f. = 105), but did not differ on male flowers (P = 0·0945, d.f. = 165; Fig. 1i).

manova (of nectar sugar concentration, corolla length, and corolla width of male and female flowers) indicated that floral traits differed between transgenic and conventional plants (P = 0·0062, d.f. = 118). In univariate anova male and female nectar sugar concentration (< 0·0001, d.f. = 156; P = 0·0002, d.f. = 145 respectively) and male corolla length (= 0·0068, d.f. = 168) were each greater on transgenic plants than on conventional plants Fig. 1k–l). In GLIMM male and female nectar volume (< 0·0001, d.f. = 153; < 0·0001, d.f. = 153 respectively) and male flower number (P = 0·0007, d.f. = 179) were also greater on transgenic plants in comparison with conventional plants (Fig. 1m–n). Overall, flowers on transgenic plants were more rewarding and were preferred by squash bees.

In all three post-hoc comparisons of near-isogenic lines squash bees visited male flowers on transgenic plants more often, and this difference was significant in two of the comparisons (Table 1). Squash bee visit duration on female flowers did not differ between conventional and transgenic plants within any of the three pairs of near-isogenic lines (Table 1).

In post-hoc comparisons of floral traits in the near-isogenic lines (Table 1) nectar volume in male flowers was greater in the three transgenic lines, but in only one comparison was this difference significant. Similarly, male corolla width was greater in all three of the transgenic varieties, but none of these differences were significant. Average male flower number per day, female nectar sugar concentration and nectar volume did not differ between conventional and transgenic members of near-isogenic pairs.

In addition, there was no effect of virus treatment or the virus treatment × transgene interaction on squash bee visits to female (P = 0·1456, 0·3955, d.f. = 124 respectively) or male flowers (P = 0·9196, 0·7762, d.f. = 173 respectively). Similarly, neither the virus treatment nor the virus treatment × transgene interaction affected squash bee visit duration on female (P = 0·7566, 0·3444, d.f. = 105 respectively) or male flowers (P = 0·2179, 0·8428, d.f. = 165 respectively). Although virus treatment and the virus treatment × transgene interaction did affect floral traits analysed by manova (transgene: < 0·0001, virus treatment: P = 0·0169, virus treatment × transgene: P = 0·0062, d.f. = 118), only male nectar sugar concentration was significant in analyses of individual traits (by univariate anova or GLMM). Male nectar sugar concentration was greater in healthy (P = 0·0024, d.f. = 123) and transgenic (< 0·0001) plants. The virus × transgene interaction was also significant (healthy conventional plants produced nectar with a slightly higher sugar concentration than did virus-infected conventional plants, whereas nectar sugar concentration in transgenic virus-resistant varieties did not differ between virus treatments (Table 2; P = 0·0006). Because floral traits were mostly not affected by virus infection it is perhaps not surprising that squash bee visit number and duration were also not affected by virus.

Table 2.   Male and female floral characters in 2005 (least square mean ± 1 SE)
 FemaleMale
ConventionalTransgenicConventionalTransgenic
HealthyInfectedHealthyInfectedHealthyInfectedHealthyInfected
  1. Bold signifies P-value <0.005, only male nectar sugar concentration differed between healthy and infected conventional plants (univariate anova following significant manova).

Corolla width (mm)54·43 ± 1·4553·87 ± 1·3554·97 ± 1·3456·29 ± 1·2646·87 ± 1·5047·03 ± 1·3943·39 ± 1·3844·11 ± 1·31
Corolla length (mm)58·81 ± 1·2258·77 ± 1·1454·92 ± 1·1358·97 ± 1·0761·96 ± 0·9760·17 ± 0·9063·54 ± 0·8963·82 ± 0·84
Nectar sugar concentration (%)37·71 ± 0·6636·51 ± 0·6139·03 ± 0·6139·31 ± 0·5842·00 ± 0·4639·22 ± 0·4243·39 ± 0·4243·53 ± 0·40
Nectar volume (μl)36·63 ± 4·9445·18 ± 5·9760·73 ± 7·8272·06 ± 9·2629·02 ± 4·6621·11 ± 4·5647·03 ± 4·6345·10 ± 4·60
Average flower no. day−13·36 ± 0·473·30 ± 0·463·44 ± 0·482·87 ± 0·416·98 ± 0·677·08 ± 0·688·87 ± 0·868·89 ± 0·86

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Honey bees and squash bees showed opposite preferences for conventional and transgenic squash. In 2004, honey bees visited female conventional flowers more than female transgenic flowers. By contrast, in 2005 squash bees visited male transgenic flowers more than male conventional flowers, and spent more time in female transgenic flowers than on female conventional flowers. These data suggest that the presence of transgenic virus-resistance has different effects on honey bees and squash bees. However these data were collected in different years with different weather conditions (Table S1, Supporting Information), and floral traits known to affect pollinator preference (Young & Stanton 1990; Schemske & Bradshaw 1999) also varied between years. In particular, in 2004 female conventional flowers, preferred by honey bees, were generally larger and contained more nectar than transgenic flowers (although these differences were small). In contrast, in 2005 transgenic flowers, preferred by squash bees, were generally larger, produced a larger volume of sweeter nectar, and were in greater number.

From our data it is difficult to determine whether honey bees and squash bees have different responses to transgenic resistance, or if they have similar responses to pleiotropic effects of transgenic resistance on floral (or other unmeasured) characters that varied with the environment. In an attempt to separate the effects of transgene status and measured floral characters on pollinator preference we performed analyses of covariance on preference traits (results not reported). These analyses were performed as described above, but included principal components derived from floral traits as covariates, as well as transgene (conventional vs. transgenic), block, and virus (in 2005) as a categorical variables. If the floral traits we measured (but not the transgene) determine preference, then in these analyses the covariate(s) (but not the main effect of transgene) will be significant. In the analyses of both honey bee visits to female flowers in 2004 and squash bee visits to male flowers in 2005 the floral traits did not explain visit number, but the transgene effect was significant (as shown in Fig. 1). However, squash bee visit duration on female flowers was explained by floral characters and not by transgene status. Taken together, these analyses suggest that both honey bees and squash bees respond to transgenic status, but that squash bees respond to floral characters (independent of transgenic status) as well. However, the conclusions from these analyses are tentative because the sample size is small when analyzing all floral traits simultaneously (reduced by missing values for many of the floral traits), and thus these analyses have little statistical power. In addition, we cannot rule out the possibility that both pollinators prefer unmeasured character(s) which show transgene × environment effects in their expression.

Although it is not known which particular traits honey bees and squash bees prefer, it is not surprising that these pollinators have different preferences. Honey bees forage on many species and usually only collect nectar from squash. By contrast, squash bees are specialists that provision their nests only with cucurbit pollen and use cucurbits as their primary nectar source. For this reason, it could be that subtle changes in floral characters have a larger effect on the specialist squash bee than the generalist honey bee.

If preference does differ between pollinator species, then in some years there could be assortative mating, such that transgenic × transgenic and conventional × conventional matings are more frequent than expected under random mating. Such assortative mating could also reduce the frequency of transgenic crop-wild hybridization. In addition, assuming that the phenotypic effects of the transgene are similar in wild and cultivated plants, assortative mating could potentially slow the rate of introgression of transgenic virus resistance into wild populations. On the other hand, if virus resistance is beneficial in wild populations, theoretical work suggests that assortative mating could speed the fixation of a resistance gene (Winterer & Weis 2004). Which outcome is most likely will depend on the degree of assortative mating, as well as the strength of selection for virus resistance (both of which are unknown).

Alternatively, if pollinator preference is similar in honey bees and squash bees, and is determined by floral (or other) characters affected by the virus-resistance transgene, then in some years transgenic plants will receive more visits, and potentially have higher fitness (Proctor, Yeo, & Lack 1996; Jones & Reithel 2001; Roldan-Serrano & Guerra-Sanz 2005), while in other years conventional plants will receive more visits. Thus, natural selection by pollinators will act to increase transgene frequency in some years and decrease transgene frequency in others. Net selection on transgenic resistance in any particular year will be determined by the sum of selection by viruses to (presumably) increase resistance and selection by pollinators to either increase or decrease resistance.

In both scenarios pollinator preferences are influenced by pleiotropic effects of the virus-resistance transgene. Such non-target effects can be caused by the position of the transgene within the plant genome (e.g. Alla et al. 2003). However, in the data presented here the pattern of pollinator response was similar in each of the three near-isogenic pairs. Because these three pairs were derived from two separate insertion events, position effects seem unlikely. Alternatively, non-target effects could be caused by the effect of transgenic products on the function of biochemical pathways. A third possibility involves RNA silencing, a natural defense mechanism which can also affect developmental pathways (Eamens et al. 2008). Because transgenic virus resistance involves the expression of virus coat protein genes (Tricoli et al. 1995), it is possible that viral protein produced from the transgene initiates the RNA silencing pathway, thus affecting flower development (Bazzini et al. 2007). Future investigation of non-target effects of transgenic crops should focus on transgenic phenotypes, such as disease resistance, that initiate RNA silencing or otherwise have the potential to affect developmental pathways.

It seems likely that pollinator preferences will often differ between healthy and virus-infected plants because of the well documented effects of virus infection on plant physiology, plant morphology (Hull 2002) and flowering phenology (Mackenzie 1985; H. R. Prendeville, unpublished data). However, in our 2005 experiment squash bee preferences were not affected by virus infection. In this experiment we inoculated plants after flowering had begun, and pollinator observations were completed before virus symptoms were severe. Specifically, leaf mottling was present, but structures were not malformed. It is possible that if we had inoculated plants earlier, so that infections were more severe at the time of pollinator observations, pollinator behaviour may have been affected. Thus, these data suggest that the timing of virus infection relative to flowering could have important effects on pollinator behaviour.

We have shown that pleiotropic effects of transgenic virus resistance can affect pollinator visit number and duration. Moreover, if pollinator behaviour determines plant mating patterns, then we expect that crop-wild hybridization and transgene introgression into wild populations will depend on direct fitness benefits of transgenic virus resistance as well as on how patterns of non-random mating affect response to selection. We recommend that future risk assessments examine pleiotropic effects of transgenes on native and introduced pollinators in different environments. In addition, these studies should evaluate the effects of pollinator behaviour (as well as pollinator composition, abundance and efficiency) on plant mating patterns. The results from these studies will allow us to better predict the evolution of transgenic resistance in wild populations and guide policy decisions on the use and deregulation of transgenic crops.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Work was funded by USDA-BRAG #05-03806 to DP and T.J. Morris, University of Nebraska SBS Special Funds to HRP, and a US-Department of Education GAANN (#P200A040126 to UNL) fellowship to HRP. The authors thank L. Fiedler, K. Carlson, K. Elgersma, and K. Bradley for field assistance; S. Adams, L. Hodges, and R. French for guidance in plant and virus cultivation; T.J. Morris, C. Brassil, Nick Waser, and an anonymous reviewer provided insightful comments.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  • Alla, S., Cherqui, A., Kaiser, L., Azzouz, H., Sangwann-Norreel, B.S. & Giordanengo, P. (2003) Effects of potato plants expressing nptII-gus fusion marker genes on reproduction, longevity, and host-finding of the peach-potato aphid, Myzus persicae. Entomologia Experimentalis et Applicata, 106, 95102.
  • Allen-Wardell, G., Bernhardt, P., Bitner, R., Burquez, A., Buchmann, S., Cane, J., Cox, P.A., Dalton, V., Feinsinger, P., Ingram, M., Inouye, D., Jones, C.E., Kennedy, K., Kevan, P., Koopowitz, H., Medellin, R., Medellin-Morales, S., Nabhan, G.P., Pavlik, B., Tepedino, V., Torchio, P. & Walker, S. (1998) The potential consequences of pollinator declines on the conservation of biodiversity and stability of food crop yields. Conservation Biology, 12, 817.
  • Bazzini, A.A., Hopp, H.E., Beachy, R.N. & Asurmendi, S. (2007) Infection and coaccumulation of tobacco mosaic virus proteins alter microRNA levels, correlating with symptom and plant development. Proceedings of the National Academy of Science, 104, 1215712162.
  • Brødsgaard, H.F., Brødsgaard, C.J., Hansen, H. & Lovei, G.L. (2003) Environmental risk assessment of transgene products using honey bee (Apis mellifera) larvae. Apidologie, 34, 139145.
  • Burgess, E.P.J., Malone, L.A. & Christeller, J.T. (1996) Effects of two proteinase inhibitors on the digestive enzymes and survival of honey bees (Apis mellifera). Journal of Insect Physiology, 42, 823828.
  • Claessen, D., Gilligan, C.A., Lutman, P.J.W. & Van Den Bosch, F. (2005) Which traits promote persistence of feral GM crops? Part 1: implications of environmental stochasticity. Oikos, 110, 2029.
  • Cresswell, J.E. & Osborne, J.L. (2004) The effect of patch size and separation on bumblebee foraging in oilseed rape: implications for gene flow. Journal of Applied Ecology, 41, 539546.
  • Decker-Walters, D.S., Staub, J.E., Chung, S.M., Nakata, E. & Quemada, H.D. (2002) Diversity in free-living populations of Cucurbita pepo (Cucurbitaceae) as assessed by random amplified polymorphic DNA. Systematic Botany, 27, 1928.
  • Devos, Y., Reheul, D., De Waele, D. & Van Speybroeck, L. (2006) The interplay between societal concerns and the regulatory frame on GM crops in the European Union. Environmental Biosafety Research, 5, 127149.
  • Duan, J.J., Marvier, M., Huesing, J., Dively, G. & Huang, Z.Y. (2008) A meta-analysis of effects of Bt crops on honey bees (Hymenoptera: Apidae). PLoS ONE, 3, e1415.
  • Eamens, A., Wang, M.-B., Smith, N.A. & Waterhouse, P.M. (2008) RNA silencing in plants: yesterday, today, and tomorrow. Plant Physiology, 147, 456468.
  • Ellstrand, N.C. (2003) Dangerous Liaisons? When Cultivated Plants Mate with Their Wild Relatives. Johns Hopkins University Press, Baltimore, MD.
  • Felber, F., Kozlowski, G., Arrigo, N. & Guadagnuolo, R. (2007) Genetic and ecological consequences of transgene flow to the wild flora. Advances in Biochemical Engineering/Biotechnology, 107, 173205.
  • Fuchs, M. & Gonsalves, D. (1995) Resistance of transgenic hybrid squash ZW-20 expressing the coat protein genes of Zucchini Yellow Mosaic Virus and Watermelon Mosaic Virus 2 to mixed infections by both potyviruses. Bio/Technology, 13, 14661473.
  • Gianessi, L.P., Silvers, C.S., Sankula, S. & Carpenter, J.E. (2002) Plant Biotechnology: Current and Potential Impact for Improving Pest Management in U.S. Agriculture: An Analysis of 40 Case Studies. National Center for Food and Agricultural Policy, Washington, DC. http://www.ncfap.org/40casestudies.html (accessed 20 February 2009).
  • Goulson, D., Lye, G.C. & Darvill, B. (2008) Decline and conservation of bumble bees. Annual Review of Entomology, 53, 191208.
  • Groot, A.T. & Dicke, M. (2002) Insect-resistant transgenic plants in a multi-trophic context. Plant Journal, 31, 387406.
  • Haughton, A.J., Champion, G.T., Hawes, C., Heard, M.S. & Brooks, D.R. (2003) Invertebrate responses to the management of genetically modified herbicide-tolerant and conventional spring crops. II. Within-field epigeal and aerial arthropods. Philosophical Transactions of the Royal Society of London Series B, 358, 18631877.
  • Holm, S. (1979) A simple sequential rejective multiple test procedure. Scandinavian Journal of Statistics, 6, 6570.
  • Hoyle, M., Hayter, K. & Cresswell, J.E. (2007) Effect of pollinator abundance on self-fertilization and gene flow: application to GM Canola. Ecological Applications, 17, 21232135.
  • Huang, Z.Y., Hanley, A.V., Pett, W.L., Langenberger, M. & Duan, J.J. (2004) Field and semifield evaluation of impacts of transgenic canola pollen on survival and development of worker honey bees. Journal of Economic Entomology, 97, 15171523.
  • Hull, R. (2002) Matthews’ Plant Virology, 4th edn. Academic Press, San Diego.
  • Hurd, E. & Linsley, G. (1964) The squash and gourd bees genera Peponapis Robertson an Xenglossa Smith inhabiting America North of Mexico (Hymenoptera: Apoidea). Hilgardia, 35, 375477.
  • James, C. (1998) Global Review of Commercialized Transgenic Crops: 1998. ISAAA Briefs No. 8. ISAAA, Ithaca, NY.
  • James, C. (2008) Global Status of Commercialized Biotech/GM Crops: 2008. ISAAA Brief No. 39. ISSA, Ithaca, NY.
  • Jones, K.N. & Reithel, J.S. (2001) Pollinator-mediated selection on a flower color polymorphism in experimental populations of Antirrhinum (Scrophulariaceae). American Journal of Botany, 88, 447454.
  • Kearns, C.A., Inouye, D.W. & Waser, N.M. (1998) Endangered mutualisms: the conservation of plant-pollinator interactions. Annual Review Ecology, Evolution, and Systematics, 29, 83112.
  • Klein, A.M., Vaissie`re, B.E., Cane, J.H., Steffan-Dewenter, I., Cunningham, S.A., Kremen, C. & Tscharntke, T. (2007) Importance of pollinators in changing landscapes for world crops. Proceeding of the Royal Society of London Series B, 274, 303313.
  • Kruckelmann, H.W. & Seyffert, W. (1970) Interactions between a turnip mosaic virus and the genotype of the host. Theoretical and Applied Genetics, 40, 121123.
  • Lecoq, H., Pitrat, M. & Clement, M. (1981) Identification et caractersation d’un potyvirus provoquant la maladie du rabougrissment juane du melon. Agronomie, 1, 827834.
  • Lin, H.X., Rubio, L., Smythe, A., Jiminez, M. & Falk, B.W. (2003) Genetic diversity and biological variation among California isolates of Cucumber mosaic virus. Journal of General Virology, 84, 249258.
  • Mackenzie, S. (1985) Reciprocal transplantation to study local specialization and the measurements of the components of fitness. PhD thesis, University College of North Wales, North Wales, UK.
  • Malone, L.A. (2004) Potential effects of GM crops on bee health. Bee World, 85, 2936.
  • Malone, L.A., Burgess, E.P.J., Christeller, J.T. & Gatehouse, H.S. (1998) In vivo responses of honey bee midgut proteases to two protease inhibitors from potato. Journal of Insect Physiology, 44, 141147.
  • Malone, L.A., Scott-Dupree, C.D., Todd, J.H. & Ramankutty, P. (2007) No sub-lethal toxicity to bumblebees, Bombus terrestris, exposed to Bt-corn pollen, captan, novaluron. New Zealand Journal of Crop and Horticultural Science, 35, 435439.
  • Morandin, L.A. & Winston, M.L. (2003) Effects of novel pesticides on bumble bee (Hymenoptera: Apidae) colony health and foraging ability. Environmental Entomology, 32, 555563.
  • Morandin, L.A. & Winston, M.L. (2005) Wild bee abundance and seed production in conventional, organic, and genetically modified canola. Ecological Applications, 15, 871881.
  • Nepi, M., Pacini, E. & Willemse, M.T.M. (1996) Nectary biology of Cucurbita pepo: ecophysiological aspects. Acta Botanica Neerlandica, 45, 4145.
  • O’Callaghan, M., Glare, T.R., Burgess, E.P.J. & Malone, L.A. (2005) Effects of plants genetically modified for insect resistance on nontarget organisms. Annual Review of Entomology, 50, 271292.
  • Pauw, A. (2007) Collapse of a pollination web in small conservation areas. Ecology, 88, 17591769.
  • Pfosser, M.F. & Baumann, H. (2002) Phylogeny and geographical differentiation of zucchini yellow mosaic virus isolates (Potyviridae) based on molecular analysis of the coat protein and part of the cytoplasmic inclusion protein genes. Archives of virology, 147, 15991609.
  • Pham-Delègue, M.H., Girard, C., Le Métayer, M., Picard-Nizou, A.L. & Hennequet, C. (2000) Long-term effects of soybean protease inhibitors on digestive enzymes, survival and learning abilities of honeybees. Entomologia Experimentalis et Applicata, 95, 2129.
  • Picard-Nizou, A.L., Pham-Delègue, M.H., Kerguelen, V., Douault, P. & Marilleau, R. (1995) Foraging behaviour of honey bees (Apis mellifera L.) on transgenic oilseed rape (Brassica napus L. var. oleifera). Transgenic Research, 4, 270276.
  • Picard-Nizou, A.L., Grison, R., Olsen, L., Pioche, C., Arnold, G. & Pham-Delègue, M.H. (1997) Impact of proteins used in plant genetic engineering: toxicity and behavioral study in the honeybee. Journal of Economic Entomology, 90, 17101716.
  • Pierre, J., Marsault, D., Genecque, E., Renard, M., Champolivier, J. & Pham-Delègue, M.H. (2003) Effects of herbicide tolerant transgenic oilseed rape genotypes on honey bees and other pollinating insects under field conditions. Entomologia Experimentalis et Applicata, 108, 159168.
  • Pilson, D. & Prendeville, H.R. (2004) Ecological effects of transgenic crops and the escape of transgenes into wild populations. Annual Review of Ecology Evolution and Systematics, 35, 149174.
  • Proctor, M., Yeo, P. & Lack, A. (1996) The Natural History of Pollination. Timber Press, Portland, OR.
  • Quemada, H., Strehlow, L., Decker-Walters, D.S. & Staub, J.E. (2008) Population size and incidence of virus infection in free-living populations of Cucurbita pepo. Environmental Biosafety Research, 7, 185196.
  • Roldan-Serrano, A.S. & Guerra-Sanz, J.M. (2005) Reward attractions of zucchini flowers (Cucurbita pepo L.) to bumblebees (Bombus terrestris L.). European Journal of Horticultural Science, 70, 2328.
  • Rowell, B., Nesmith, W. & Snyder, J.C. (1999) Yields and disease resistance of fall-harvested transgenic and conventional summer squash in Kentucky. HortTechnology, 9, 282288.
  • Roy, D.B., Bohan, D.A., Haughton, A.J., Hill, M.O., Osborne, J.L., Clark, S.J., Perry, J.N., Rothery, P., Scott, R.J., Brooks, D.R., Champion, G.T., Hawes, C., Heard, M.S. & Firbank, L.G. (2003) Invertebrates and vegetation of field margins adjacent to crops subject to contrasting herbicide regimes in the Farm Scale Evaluations of genetically modified herbicide-tolerant crops. Philosophical Transactions of the Royal Society of London Series B, 358, 18791898.
  • Schemske, D.W. & Bradshaw, H.D. (1999) Pollinator preference and the evolution of floral traits in monkeyfowers (Mimulus). Proceedings of the National Academy of Science, 96, 1191011915.
  • Tepedino, C.J. (1981) The pollination efficiency of the squash bee (Peponapis pruinosa) and the honey bee (Apis mellifera) on summer squash (Cucurbita pepo). Journal of the Kansas Entomological Society, 54, 359377.
  • Tricoli, D.M., Carney, K.J., Russel, P.F., McMaster, J.R., Groff, D.W., Hadden, K.C., Himmel, P.T., Hubbard, J.P., Boeshore, M.L. & Quemada, H.D. (1995) Field evaluation of transgenic squash containing single or multiple virus coat protein gene constructs for resistance to Cucumber mosaic Virus, Watermelon Mosaic Virus 2, and Zucchini Yellow Mosaic Virus. Bio/Technology, 13, 14581465.
  • USDA/APHIS (1994) Determination of Nonregulated Status for ZW-20 and CZW-3 Squash. USDA Petition. Documents on deregulation of the squash transformants ZW-20 and CWZ-3 are, respectively, http://www.aphis.usda.gov/brs/aphisdocs2/92_20401p_com.pdf and http://www.aphis.usda.gov/brs/aphisdocs2/95_35201p_com.pdf (confirmed June 2, 2009).
  • Wafula, D., Persley, G. & Karembu, M. (2008) GMOs and exports: demystifying concerns in Africa. Biosafety Policy Brief. http://www.isaaa.org/Resources/publications/Downloads/Biosafety%20Policy%20Brief%20(June%202008).pdf (accessed 2 June 2009).
  • Watrud, L.S., Lee, E.H., Fairbrother, A., Burdick, C., Reichman, J.R., Bollman, M., Storm, M. & Van de Water, P.K. (2004) Evidence for landscape-level, pollen-mediated gene flow from genetically modified creeping bentgrass with CP4 EPSPS as a marker. Proceedings of the National Academy of Sciences, 101, 1453314538.
  • Wilson, H.D. (1990) Gene flow in squash species: domesticated Cucurbita species may not represent closed genetic systems. BioScience, 40, 449455.
  • Wilson, H.D. (1993) Free-Living Cucurbita Pepo in the United States: Viral Resistance, Gene Flow, and Risk Assessment. USDA Animal and Plant Health Inspection Service, Hyattsville, MD.
  • Winterer, J. & Weis, A.E. (2004) Stress-induced assortative mating and the evolution of stress resistance. Ecology Letters, 7, 785793.
  • Young, H.J. & Stanton, M.L. (1990) Influences of floral variation on pollen removal and seed production in wild radish. Ecology, 71, 536547.

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Table S1.High Plains Regional Climate Center recorded data from 1 May to 1 August 2004 and 1 May to 15 August 2005 from weather stations located near experimental plot. Means for each year were calculated and compared by t-tests. Means are presented with * indicating a P < 0.05 and *** indicating a P < 0.0001

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

FilenameFormatSizeDescription
JPE_1698_sm_suppInfo.rtf65KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.