Omega-3 deficiency impairs honey bee learning

Abstract

Deficiency in essential omega-3 polyunsaturated fatty acids (PUFAs), particularly the long-chain form of docosahexaenoic acid (DHA), has been linked to health problems in mammals, including many mental disorders and reduced cognitive performance. Insects have very low long-chain PUFA concentrations, and the effect of omega-3 deficiency on cognition in insects has not been studied. We show a low omega-6:3 ratio of pollen collected by honey bee colonies in heterogenous landscapes and in many hand-collected pollens that we analyzed. We identified Eucalyptus as an important bee-forage plant particularly poor in omega-3 and high in the omega-6:3 ratio. We tested the effect of dietary omega-3 deficiency on olfactory and tactile associative learning of the economically highly valued honey bee. Bees fed either of two omega-3-poor diets, or Eucalyptus pollen, showed greatly reduced learning abilities in conditioned proboscis-extension assays compared with those fed omega-3-rich diets, or omega-3-rich pollen mixture. The effect on performance was not due to reduced sucrose sensitivity. Omega-3 deficiency also led to smaller hypopharyngeal glands. Bee brains contained high omega-3 concentrations, which were only slightly affected by diet, suggesting additional peripheral effects on learning. The shift from a low to high omega-6:3 ratio in the Western human diet is deemed a primary cause of many diseases and reduced mental health. A similar shift seems to be occurring in bee forage, possibly an important factor in colony declines. Our study shows the detrimental effect on cognitive performance of omega-3 deficiency in a nonmammal.

Keywords: Apis mellifera; alpha-linolenic acid; associative conditioning; fatty acids; proboscis extension response.

Fig. 1.

(A) Hypopharyngeal gland perimeter and diameter (short axis) of bees fed four treatment diets. Horizontal lines compare between diets rich (Om, oil mixture; Pn, pollen) or poor (Cn, corn; Se, sesame) in omega-3. Sample sizes were 10 (8 for Om diet) bees per treatment. (B) Total fatty acid (TFA) percent of sample dry weight of bee body and brain. Numbers at bottom of bars are sample size; for brains, each sample is a pool of 12 bee brains. Data not available for TFA% of brains in experiment 1. (C and E) Percent essential fatty acids of total fatty acids (TFAs) in bee bodies and (D and F) brains. Sample sizes were 10 or 24 bees per treatment for body analyses, and five (four for Om diet) samples of 20 bees each or two samples of 12 bees each for brain analyses, in experiments 1 and 2, respectively. Different letters represent statistically significant differences between treatments (Tukey’s test, P < 0.05). **P < 0.01, ***P < 0.001. Error bars represent SE.

Fig. S1.

Mean (± SE) weight of treatment diet collected per colony per day. Sample sizes were seven (six for oil mixture diet) colonies per treatment, in the two experiments combined.

Fig. 2.

Performance in olfactory (A–C, experiment 1; D–F, experiment 2) and tactile (G–I) PER conditioning of bees fed four treatment diets. The first figure in each row shows learning curves to a positively rewarded conditioned stimulus (CS+; full lines). Different letters represent statistically significant differences between treatments (Tukey’s test, P < 0.05). The olfactory assay included a negatively rewarded conditioned stimulus (CS−; dashed lines). Bees quickly learned to respond to the CS+ and not to the CS−. In all three experiments, bees fed omega-3–rich diets (oil mixture and pollen) learned better than those fed omega-3–poor diets (corn and sesame). The second figure in each row shows the response to the sucrose reward (US+; dotted lines). Response percentages were high in all treatments but were statistically significantly lower in the omega-3–poor treatments. The third figure in each row shows performance in a memory test 24 h after conditioning; numbers in bars are sample sizes. *P < 0.05, ***P < 0.001.

Fig. S2.

Sucrose sensitivity scores of bees in experiment 2 fed four different diets before olfactory (A) or tactile (B) PER conditioning. Sucrose sensitivity did not differ significantly between treatments in either of the two groups (χ²₃ = 1.57, P = 0.67, and χ²₃ = 1.51, P = 0.68, respectively). There were no significant differences between diets rich (oil mixture and pollen) or poor (corn and sesame) in omega-3 (χ²₁ = 0.87, P = 0.35, and χ²₁ = 1.14, P = 0.29, respectively). Because sensitivity was tested before PER conditioning, the data could be combined into a single model; still, no significant differences were found between the omega-3–rich and the omega-3–poor diets (χ²₁ = 2.08, P = 0.15, n = 470). Box plots represent medians and interquartiles. n = 51–66 bees per diet in each group.

Fig. 3.

Performance in olfactory PER conditioning of bees in experiment 3 fed four treatment diets as in experiments 1 and 2 (A) or pollen pellets of Eucalyptus or of a bee-collected mixture (B). Lines coded as in Fig. 2.