1 INTRODUCTION

In nature, food resources are usually scarce, scattered, and greatly differ in nutrient composition. To face this nutritional challenge, animals are usually able to balance their nutrient intake to obtain an optimum mixture of nutrients, or intake target, required for growth, development and reproduction. When they cannot reach the intake target, animals have to over- or under-consume some nutrients, which may induce fitness costs (Simpson & Raubenheimer, 2012). The geometric framework is an integrative approach to model animal nutrition. This approach is used to study how animals regulate different nutrients simultaneously to maximize their fitness (Simpson & Raubenheimer, 2012). Many studies using this approach have shown how the intake of nutritionally unbalanced diets influences a variety of physiological, behavioural and life-history traits. For instance, an unbalanced protein to carbohydrate (P:C) ratio can affect the lifespan (Le Couteur et al., 2016; Lee et al., 2008), reproduction (Jensen et al., 2015; Solon-Biet et al., 2015), immunity (Ponton et al., 2011) or even personality traits of individuals (Bouchebti, Cortés-Fossati, et al., 2022; Han & Dingemanse, 2017).

Many amino acids have direct and indirect roles in the nervous system and can affect brain functions (Wu, 2009). Thus, dietary proteins are essential for cognitive abilities. In vertebrates, protein deprivation in prenatal and/or postnatal stages decreases brain development and impairs neuronal activities, affecting motor and cognitive functions (Laus et al., 2011). These effects are irreversible in adulthood. Yet in invertebrates, although the quality of the diet is known to influence learning performances and memory (Arien et al., 2015, 2018; Kishani Farahani et al., 2021), only few studies, focused on Drosophila, investigated the effect of dietary proteins on cognitive abilities. A high percentage of dietary protein improves visual learning and memory in adult Drosophila (Guo et al., 1996; Xia et al., 1997), whereas a balanced P:C ratio induces better learning abilities in larvae (Lihoreau et al., 2016).

The honey bee is a well-established model to study cognition in insects (Giurfa, 2003). Honey bees feed on flowers; nectar provides them carbohydrates and pollen provides them proteins, lipids, vitamins and minerals (Wright et al., 2018). When honey bees have a choice between complementary diets, they are able to balance their intake of proteins and carbohydrates (Altaye et al., 2010), as well as their intake of free amino acids (Hendriksma et al., 2014). This intake target changes over their life stage, younger bees requiring higher P:C ratio than older forager bees (Paoli et al., 2014). An unbalanced intake of P:C ratio can affect survival and ovarian development in adults (Pirk et al., 2010) and growth in larvae (Helm et al., 2017). Pollen (therefore including protein) deprivation also shortens the lifespan of individuals and induces a decrease of foraging activity and waggle dance (Scofield & Mattila, 2015), but does not seem to have an impact on learning or memory (Gage et al., 2020; Mattila & Smith, 2008). However, pollen ingestion does modify the amino acids content in the brain (Gage et al., 2020). The quality of pollen greatly differs from one plant species to another and can affect the physiology of honey bees (Pasquale et al., 2013). Moreover, pollen protein content is highly variable and ranges from 2.5% to 61% (Roulston et al., 2000). Therefore, to investigate the role of dietary protein on honey bee cognitive abilities, artificial diets are necessary.

After a honey bee emerges, its brain undergoes several changes thorough a maturation process (Fahrbach & Robinson, 1996). Thus, we hypothesize that the diet consumed in the early stage of bee adulthood is critical for cognitive abilities. By using the geometric framework approach, we fed newly emerged adult honey bees with artificial diets varying in the P:C ratio and studied their effects on survival and cognitive performance, using an olfactory proboscis extension response (PER) conditioning assay. The learning, short-term memory (STM), and early long-term memory (eLTM) were assessed after 1 week. Since appetitive learning and memory in honey bees are bound to the sucrose sensitivity threshold of individuals (Scheiner et al., 2004, 2005), we also investigated the effects of P:C ratio on sucrose sensitivity.

中文翻译：

---

大量营养素平衡对蜜蜂的认知和生存有相反的影响

1 简介

在自然界中，食物资源通常稀缺、分散，营养成分差异很大。为了应对这一营养挑战，动物通常能够平衡其营养摄入，以获得生长、发育和繁殖所需的最佳营养混合物或摄入目标。当它们无法达到摄入目标时，动物不得不过度或不足地摄入某些营养物质，这可能会导致健身成本（Simpson & Raubenheimer,  2012）。几何框架是一种模拟动物营养的综合方法。这种方法用于研究动物如何同时调节不同的营养物质以最大限度地提高它们的适应性（Simpson & Raubenheimer,  2012）。许多使用这种方法的研究表明，营养不均衡的饮食摄入如何影响各种生理、行为和生活史特征。例如，不平衡的蛋白质与碳水化合物 (P:C) 比例会影响寿命（Le Couteur 等人，  2016 年；Lee 等人，  2008 年）、繁殖（Jensen 等人，  2015 年；Solon-Biet 等人）。 ,  2015 )、免疫力 (Ponton et al.,  2011 ) 甚至个人的人格特征 (Bouchebti, Cortés-Fossati, et al.,  2022 ; Han & Dingemanse,  2017 )。

许多氨基酸在神经系统中具有直接和间接的作用，并且可以影响大脑功能（Wu，  2009）。因此，膳食蛋白质对于认知能力至关重要。在脊椎动物中，产前和/或产后阶段的蛋白质剥夺会降低大脑发育并损害神经元活动，影响运动和认知功能（Laus 等人，  2011 年）。这些影响在成年后是不可逆转的。然而，在无脊椎动物中，尽管已知饮食质量会影响学习表现和记忆力（Arien 等人，  2015 年，2018 年；Kishani Farahani 等人，  2021 年），但只有少数研究专注于果蝇，研究了膳食蛋白质对认知能力的影响。高比例的膳食蛋白质可提高成年果蝇的视觉学习和记忆力（Guo 等人，  1996 年；Xia 等人，  1997 年），而平衡的 P:C 比例可提高幼虫的学习能力（Lihoreau 等人，  2016 年） ）。

蜜蜂是研究昆虫认知的成熟模型（Giurfa，  2003 年）。蜜蜂以花为食；花蜜为它们提供碳水化合物，花粉为它们提供蛋白质、脂质、维生素和矿物质（Wright 等人，  2018 年）。当蜜蜂在补充饮食之间进行选择时，它们能够平衡蛋白质和碳水化合物的摄入量（Altaye 等人，  2010 年）以及游离氨基酸的摄入量（Hendriksma 等人，  2014 年）。这个摄入目标会随着它们的生命阶段而变化，年轻的蜜蜂比年长的觅食蜜蜂需要更高的 P:C 比（Paoli 等人，  2014 年）。P:C 摄入不均衡会影响成人的生存和卵巢发育（Pirk et al.,  2010) 和幼虫的生长 (Helm et al.,  2017 )。花粉（因此包括蛋白质）剥夺也会缩短个体的寿命，并导致觅食活动和摇摆舞减少（Scofield & Mattila，  2015 年），但似乎对学习或记忆没有影响（Gage 等人，  2020 年） ；马蒂拉和史密斯，  2008 年）。然而，花粉摄入确实会改变大脑中的氨基酸含量（Gage 等人，  2020 年）。花粉的质量因植物物种而异，并且会影响蜜蜂的生理机能（Pasquale 等人，  2013 年）。此外，花粉蛋白含量变化很大，范围从 2.5% 到 61%（Roulston 等，  2000）。因此，为了研究膳食蛋白质对蜜蜂认知能力的作用，人工饲料是必要的。

蜜蜂出现后，它的大脑在成熟过程中经历了几次变化（Fahrbach & Robinson，  1996）。因此，我们假设蜜蜂成年早期的饮食对认知能力至关重要。通过使用几何框架方法，我们用 P:C 比例不同的人工饲料喂养新出现的成年蜜蜂，并使用嗅觉长鼻延伸反应 (PER) 调节试验研究了它们对生存和认知能力的影响。1 周后评估学习、短期记忆 (STM) 和早期长期记忆 (eLTM)。由于蜜蜂的食欲学习和记忆与个体的蔗糖敏感性阈值有关（Scheiner et al.,  2004 , 2005)，我们还研究了 P:C 比对蔗糖敏感性的影响。