Pollen dearth and poor-quality pollen affect longevity and nursing ability of honeybees
in intensive agricultural landscape

 La nutrition caractérise une fonction clé pour tous les organismes. Pour l’abeille domestique, Apis mellifera L., le pollen représente l’élément limitant pour son développement, sa santé, et sa survie. En effet, de nombreuses études ont démontré l’importance de l’apport en pollen pour l’abeille (Haydack, 1970; Rinderer and Elliott, 1977 ; Wahl and Ulm, 1983 ; Brodschneider et Crailsheim, 2010 ; Alaux et al., 2010a; DeGrandi-Hoffman et al., 2010 ; Wang et al., 2014). Les colonies sont de plus en plus fréquemment placées à proximité de grandes cultures de manière à ce qu’elles soient approvisionnées par de grandes quantités de pollens. Mais ce type de ressources varie en abondance et qualité dans le temps et l’espace (Decourtye et al., 2010). En effet, des périodes de pénuries en pollens suivent la floraison des grandes cultures (Maurizio,1950 ; Louveaux, 1959 ; Odoux et al., 2012 ; Requier, 2013). L’effet de ces périodes de pénuries, fournissant à l’abeille du pollen en plus faible quantité, n’a été que très peu étudié. Dans un premier temps nous avons donc testé l’hypothèse selon laquelle ces périodes de déplétion pollinique pouvaient apporter un stress à l’abeille. Pour cela, nous avons nourri des abeilles en laboratoire avec du pollen de culture récolté par des ruches sous serre et nous avons identifié les effets des diverses carences sur la physiologie d’abeilles nourrices (glandes hypopharyngiennes et taux de vitellogénine), et sur leur survie. Durant ces périodes de pénuries, la disponibilité des ressources florales de plantes sauvages prend toute son importance. Or, ces ressources varient en qualité et diversité en fonction du lieu et de la période de butinage. De nombreuses études ont démontré que malgré l’importance de l’apport protéique pour l’abeille, ces dernières ne sélectionnent pas des pollens contenant de plus fort taux de protéines (Levin et Bohart, 1955 ; Pernal et Currie, 2002). En revanche, d’après Schmidt (1984), elles consomment préférentiellement des mélanges plutôt que du pollen monofloral. Les pollens utilisés dans la seconde partie de notre étude sont de vrais échantillons de pollens récoltés par des colonies situées dans un paysage agricole à l’ouest de la France. Ces mélanges recueillis supportent le postulat de Shmidt et al. (1984) et plus récemment Odoux et al., (2012) et Requier (2013), que les abeilles sélectionnent des mélanges quand elles en ont le choix, même en présence d’une culture à floraison massive à proximité. Nous avons donc testé dans la seconde partie de cette étude les effets des différentes variétés de pollens (qualité et diversité) récoltées par des abeilles dans un environnement d’agriculture intensive, au cours d’une saison de butinage (de Mai à Septembre) sur la santé des abeilles.  

Pollen dearth and poor-quality pollen affect longevity and nursing ability of honeybees in intensive agricultural landscapes To submit to Plos One Garance Di Pasquale1,2, Cédric Alaux1,3, Yves Le Conte1,3, Jean-François Odoux4 , Maryline Pioz1,3, Bernard E. Vaissière1,3, Luc P. Belzunces1,3, Axel Decourtye1,2,5 1UMT PrADE, CS 40509, 84914 Avignon, France 2ACTA, Avignon, France 3 INRA, UR 406 Abeilles et Environnement, CS 40509, 84914 Avignon, France 4 INRA, Unité expérimentale d’entomologie Le Magneraud, 17700 Surgères, France 5 ITSAP-Institut de l’abeille, Avignon, France Abstract Intensive agricultural systems often expose honey bees (Apis mellifera L.) to large temporal variations in the availability, diversity and quality of nutritional resources, with long periods of dearth following the flowering of a main resource-rich mass-flowering crop. Such nutritional irregularity is expected to affect worker health. The positive effects of pollen consumption on workers have largely been documented, but the extent to which the nutritional quality of the pollen mix available in agrosystems can modify honey bee physiology is not clear. We tested the effect of natural pollen shortage and different seasonal pollen mixes on honey bee survival and nursing capacities (hypopharyngeal gland development and vitellogenin expression) by feeding them with field harvested pollen diets in different amounts. A drop in pure oilseed rape (Brassica napus L.) pollen availability (60 % and more) resulted in a drastic reduction of survival and performance of nursing in a quantitative-dependent manner. Despite some variation in taxonomic diversity and nutritional quality, the pollen mixes harvested over the season had a similar positive influence on bee health, except for the one collected in the late July which induced poor survival and nursing capacities. This period coincided with the massflowering of maize (Zea mays L.), which produces poor-quality pollen. In intensively farmed agricultural landscapes, periods between the blooming of mass-flowering crops can be stressful 43 for honey bees due not only to resources depletion but also to poor pollen quality as the pollen available ad libitum during the mass flowering of some crops can fail to provide workers with an adequate diet for their development. 1- Introduction Bee species can be classified into two broad categories regarding their pollen diet: specialists, who feed on a few or even a single plant species, and generalists, who forage on a large array of phylogenetically unrelated plant species (Cane and Sipes 2006). Compared to specialists, generalists usually have a better resilience to environmental changes (Biesmeijer et al. 2006). Honey bees (Apis mellifera L.), which are extreme generalist, are thus expected to adapt to the human impact on landscape, notably in intensive agricultural landscapes. However, beekeepers have frequently cited starvation and poor foraging conditions as a significant driver of colony losses (Otis, 2007; Van Engelsdorp et al., 2007). In addition, Naug (2009) suggested that nutritional stress due to habitat loss plays an important role in honeybee colony losses. The intensification of agriculture often causes a decrease in the diversity of floral resources due to the destruction of natural habitats and the use of monocultures over large areas, but it also affects the quantity of resources available since the flowering period of crops is usually short (Decourtye et al., 2010). Such changes in the landscape and thus in spatial and temporal availability of foraging ressources can impact colonies as it has long been recognized that a lack of food, in particular pollen dearth, contributes to weaken colonies (Mattila and Otis, 2006; Maurizio, 1950). Indeed, pollen nutrients (proteins, lipids, vitamins and minerals) are essential to colony development and survival (Brodschneider and Crailsheim, 2010). Several studies have shown in details how the lack of pollen can decrease the longevity (Haydack, 1970; Rinderer and Elliott, 1977, Smeets and Duchateau, 2003; Wang et al., 2014), metabolism (Toth, 2005; Fischer and Grozinger , 2008; Willard et al., 2011; Alaux et al., 2011), immuncompetence (Alaux et al., 2010a; DeGrandi-Hoffman et al., 2010), and the tolerance threshold of individual honey bee workers to pathogens and pesticides ( Rinderer et al., 1974; Rinderer and Elliott, 1977; Wahl and Ulm, 1983; Bowen-Walker and Gunn, 2001; Mayack and Naug, 2009; DeGrandi-Hoffman et al., 2010). Pollen shortage also hinders the development of hypopharyngeal glands required for the production of brood food (Maurizio, 1950; Loper and Berdel, 1980a, b; Pernal and Currie, 2000; Degrandi-Hoffman et al., 2010). A reduction of brood rearing combined with a shorter lifespan of adults can directly impact the colony 44 population (Blaschon et al, 1999; Schmickl et al., 2003; Keller et al, 2005; Mattila and Otis, 2006, 2007; Girard et al., 2012). Pollen resources in an agricultural farmland are continuously available, but with considerable variation over time during a season (Maurizio, 1950; Louveaux, 1959; Odoux et al., 2012; Requier, 2013). For example, amounts of pollen collected by colonies in an agrosystem in western France follow a very irregular pattern, with clear shortage periods (Steffan-Dewenter and Kuhn, 2003; Decourtye et al., 2010; Odoux et al., 2012; Requier, 2013). These temporal variations in the harvest of pollen also depend a lot on the apiary location (Louveaux, 1959; Duelli and Obrist, 2003). Furthermore, pollen quality and diversity can also affect the size of hypopharyngeal glands and worker longevity (Schmidt et al, 1984, 1987; Tasei and Aupinel, 2008; Di Pasquale et al., 2013). And, the diversity of pollen species harvested by honey bee colonies shows also considerable temporal variation (Dimou and Thrasyvoulou, 2007; Decourtye et al., 2010; Wratten et al., 2012). In order to better understand how the availability of pollen resources (quantity, quality and diversity) in an agricultural landscape can influence honey bee health, we fed workers with environmentally-relevant diet. We first tested the effect of pollen shortage by providing bees with controlled amount of pollen encountered in the environment during a period of low food supply. Whether a threshold amount of pollen is required for developing nurse physiology or the response is gradual could also be determined. Then, by feeding workers with representative pollen mixes harvested in an intensive agricultural area across different seasons, we assessed the influence of diet diversity and quality over time and determined whether some time frame are critical for honey bees or, conversely, if there is no stress period. The impact of the different diet on the survival and nurse physiology (Toth and Robinson, 2005; Toth et al., 2005) was determined. For that purpose, we measured the development of hypopharyngeal glands and the level of vitellogenin. Nurses secrete 60 to 80 % of the brood diet from their hypopharyngeal glands, providing a secretion rich in protein for larvae (in Winston, 1987). Vitellogenin is the main storage protein in the haemolymph and precursor for many other proteins in the honey bee (Amdam et al., 2003). This lipoprotein synthesized by the fat body was found to act as an antioxidant to promote longevity in both queen and worker bees (Seehuus et al., 2006; Corona et al., 2007). In addition, vitellogenin is used by nurses for the production of brood food (Amdam et al., 2003, Seehuus et al., 2007). 

Bee rearing

The experiment was performed on 1-day-old workers. They were obtained from three colonies by placing brood combs with late-stage pupae into an incubator at 34°C and 50 – 70 % of humidity. Bees that emerged within 10 hours were collected and mixed before placing them in cages (10.5 cm x 7.5 cm x 11.5 cm) at 34°C and 50 – 70 % of humidity in an incubator. They were provided ad libitum with candy (Apifonda® + powdered sugar) and water throughout the experiment. Fresh pollen diets were prepared according to environmentally-relevant diet and were supplied to the bees from day 1 for 9. Each day, pollen diets were weighed to determine the amount of pollen consumed per bee, per day and per modality. If a bee died during the pollen feeding period, it was removed and the amount of pollen was adjusted daily to the number of surviving bees. Influence of pollen shortage on nurse worker physiology and survival To assess the influence of pollen shortage experienced by colonies in agrosystems, we fed bees with controlled amount of pollens. Pollen quantities were estimated from a recent study, which analyzed pollen collection in an agricultural landscape of western France (Requier, 2013). From April to October, the weight of pollen collected by colonies followed a bimodal seasonal trend, marked by a low food supply period between two oilseed mass-flowering crops (May for oilseed rape and July for sunflower). This dearth period was characterized by a median value of pollen weight loss of 66 %. On average, the pollen supply of colonies was then reduced of 40 % compared to the peak during the oilseed rape blooming period. Considering this variability of pollen availability under field conditions, we determined bee nursing capacities and survival along a decreasing gradient of consumed pollen quantities: 100 %, 40 %, 30 %, 20 %, 15 %, 7 % and 0 %. The pollen quantity consumed under the ad libitum feeding condition was recorded as the “100 %” modality (average of 3.6 mg / bee / day). Groups of 44 one-day old bees were reared in cages and fed with the pollen diet. The trial with each diet modality was replicated ten times. The pollen of oilseed rape (Brassica napus, commercial F1 hybrid ‘NK Fair’®, Syngenta, Basel, Switzerland) was chosen because of its high occurrence in the cereal farmland systems (Odoux et al., 2012). The pollen was collected on plants grown under two 24 m long x 8 m wide plastic tunnels with the openings covered with insectproof screening to prevent any contamination by pesticides or other pollutants on two hives introduced in the tunnel at the onset of flowering and continuously fitted with bottom pollen traps that were emptied daily. The pollen pellets obtained were stored at – 20°C until use.

Influence of “seasonal” diversity and quality of pollen mixes on nurse worker physiology and survival In this experiment, we tested the effect on worker health of pollen mixes collected by colonies in an agricultural landscape of western France. The different pollen mixes found across different seasons were obtained from experiments conducted in 2006 (Odoux et al., 2012). Workers were fed with 6 pollen mixes collected by colonies from early May to late September in an apiary located at 46 ° 09′ 13″ N; 0 ° 41′ 20″ O. The pollen composition of each mix is presented in Figure 1. Each pollen mix was provided ad libitum to the bees (10 cages of 47 bees per diet). The protein and lipid levels of the different mixtures were determined after the pollen was dried for 24 h at 75°C (Louveaux, 1959) and the protein and lipid contents were expressed as percent of dry matter (Figure 1). The protein content was determined by Kjeldahl analysis (N x 6.25) using a Vapodest 45 (Gerhardt) and according to the procedure ISO 5983-2 norm (ISO 5983, 1997). After treating pollen with with hydrochloric acid (HCl 6N), total lipids were extracted with a chloroform / methanol mixture (2:1, v / v) (Folch et al., 1957). The presence of pesticide residues in the different pollen diets was determined by Phytocontrol laboratory (Nimes, France) with gas and liquid chromatography (limit of quantification of 0.01 mg / kg and limit of detection of 0.005 mg / k (AFNOR15662, 2009 ; Table 1). The samples were analyzed for 348 pesticides (Table S1).