Journal of Insect Physiology 58 (2012) 327–334
Contents lists available at SciVerse ScienceDirect
Journal of Insect Physiology journal homepage: www.elsevier.com/locate/jinsphys
Additive and interactive effects of nutrient classes on longevity, reproduction, and diet consumption in the Queensland fruit fly (Bactrocera tryoni) Benjamin G. Fanson ⇑, Phillip W. Taylor Department of Biological Sciences, Macquarie University, Sydney, NSW 2109, Australia
a r t i c l e
i n f o
Article history: Received 14 September 2011 Received in revised form 3 November 2011 Accepted 3 November 2011 Available online 9 November 2011 Keywords: Amino acids CAFE assay Lifespan Minerals Nutrition Vitamins
a b s t r a c t Insect lifespan is often closely linked to diet, and diet manipulations have been central to studies of ageing. Recent research has found that lifespan for some flies is maximised on a very low yeast diet, but once all yeast is removed, lifespan drops precipitously. Although effects of yeast availability on lifespan are commonly interpreted in terms of protein, yeast is a complex mix of nutrients and provides a rich source of vitamins, minerals and sterols. Elucidating which components of yeast are involved in this lifespan drop provides insights into more specific nutritional requirements and also provides a test for the commonplace interpretation of yeast in terms of protein. To this end, we fed Queensland fruit flies (Bactrocera tryoni) one of eight experimental diets that differed in the nutrient group(s) found in yeast that were added to sucrose: none, vitamins, minerals, amino acids, cholesterol, vitamin + mineral + cholesterol (VMC), vitamin + mineral + cholesterol + amino acids (VMCA), and yeast. We measured survival rates and egg production in single sex and mixed sex cages, as well as nutrient intake of individual flies. We found that the addition of minerals increased lifespan of both male and female flies housed in single sex cages by decreasing baseline mortality. The addition of just amino acids decreased lifespan in female flies; however, when combined with other nutrient groups found in yeast, amino acids increased lifespan by decreasing both baseline mortality and age-specific mortality. Flies on the yeast and VMCA diets were the only ones to show significant egg production. We conclude that the drop in lifespan observed when all yeast is removed is explained by missing micronutrients (vitamins, minerals and cholesterol) as well as the absence of protein in females, whereas minerals alone can explain the pattern for males. These results indicate a need for caution when interpreting effects of dietary yeast as effects of protein. Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction Nutritional environment is a potent mediator of an organism’s lifespan. In particular, dietary restriction (DR: reduction in nutrient intake without malnutrition) has been consistently found to extend lifespan across a vast range of animal taxa, including yeast (Lin et al., 2002), fruit flies (Chippindale et al., 1997; Good and Tatar, 2001), mice (Weindruch and Walford, 1982), and rhesus monkeys (Roth et al., 1999; Ramsey et al., 2000; for general DR reviews, see Masoro, 2002, 2005). Effects of DR on lifespan have most often been attributed to caloric restriction (CR) (Masoro, 2005; Partridge and Brand, 2005). However, in insects, growing evidence suggests that nutrients, rather than calories, are responsible for effects of DR on lifespan (Drosophila melanogaster – Mair et al., 2005; Lee et al., 2008; field crickets Teleogryllus commodus – Maklakov et al., 2008; Queensland fruit fly Bactrocera tryoni (‘Q-fly’) – Fanson et al., 2009).
⇑ Corresponding author. Tel.: +61 (0)2 9850 1312; fax: +61 (0)2 9850 4299. E-mail address:
[email protected] (B.G. Fanson). 0022-1910/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jinsphys.2011.11.002
Protein has emerged as a strong candidate as a nutritional mediator of lifespan in flies. Studies of D. melanogaster (Lee et al., 2008), Mexican fruit flies (Anastrepha ludens, Carey et al., 2008) and Q-flies (Fanson et al., 2009) have reported monotonic increase in lifespan when systemically decreasing the protein:carbohydrate (P:C) ratio (i.e., reducing protein content) of adult diet. In these studies, changing P:C ratios was accomplished by altering sucrose:yeast ratios. However, removing all protein (yeast) from the adult diet caused lifespan to precipitously drop instead of increase. When switched from a very low P:C diet to a diet containing no protein, mean lifespan decreased by 20% in D. melanogaster, 38% in Mexican fruit flies, and 50% in Q-flies. Using less detailed experimental designs than above, other studies with flies have found a similar pattern of a reduction in lifespan when flies are fed sugar without yeast (e.g., B. cucurbitae Haq and Hendrichs in press; Anastrepha suspensa ‘Caribbean fruit fly’ Pereira et al., 2010; Ceratitis capitata ‘medfly’ Barry et al., 2007; San Andrés et al., 2009). Since yeast contains a variety of nutrients, it is not clear which nutrient(s) are responsible for this drop in lifespan. Understanding the cause of this drop in lifespan should provide insights into the role of nutrients, especially protein, as mediators of ageing.
328
B.G. Fanson, P.W. Taylor / Journal of Insect Physiology 58 (2012) 327–334
Yeast is a complex mixture of carbohydrates, protein, vitamins, minerals and sterols. Though high levels of dietary protein may be causing the decrease in lifespan, it may be that adults nonetheless need to acquire at least a small amount for basic biological function. Hence the precipitous drop in lifespan observed in flies fed sucrose without yeast may be due mainly to absence of protein. However, micronutrients (e.g. vitamins, minerals, and sterols) are required for many biological functions during the larval stage in insects, such as energy metabolism, growth and development (vitamin A, Claret and Volkoff, 1992; vitamin Bs, Dadd, 1961; vitamin C, Vanderzant et al., 1962; sterols, Clayton, 1964; phosphorus, Perkins et al., 2004). The role of adult-acquired micronutrients for insects is less known, especially in relation to survival, and failure to obtain these nutrients during the adult stage may result in increased mortality rates. To understand the relative roles of protein and micronutrients as mediators of lifespan there is a need for experimental approaches that manipulate these dietary components separately. Here we examine the role of protein and micronutrients in mediating the sharp reduction in lifespan experienced by Q-flies maintained without yeast. We used a chemically defined diet (Fanson and Taylor, in press) that allowed us to add individual nutrient classes found in yeast (amino acids, vitamins, minerals, and sterols) to a sucrose diet and create select nutrient combinations. We measured lifespan, reproduction, and consumption rates on each diet. Consumption rates were measured as caloric intake affects lifespan positively (Lee et al., 2008; Maklakov et al., 2008; Fanson et al., 2009; Fanson and Taylor, in press). Mortality patterns of flies maintained in single sex and mixed sex cages can often differ, possibly due to physiological consequences of mating (Carey et al., 2002a; Piper and Partridge, 2007; Papadopoulos et al., 2010). We therefore recorded mortality patterns for males and females in both single sex and mixed sex cages.
2. Materials and methods 2.1. Study species Q-flies were obtained as pupae from the Fruit Fly Production Facility at Elizabeth Macarthur Agricultural Institute (EMAI, New South Wales, Australia). The EMAI fly stock was maintained on a larval diet of lucerne chaff, cane sugar, and torula yeast, and adult diet of cane sugar and yeast hydrolysate. For all experiments, temperature and relative humidity were maintained at ca. 25 °C and ca. 70%, respectively, and on a light schedule of 14L:10D including 1 h dawn and dusk periods.
2.2. Experimental diets Increasing yeast of the diet from 0% to just 5% has been found to double Q-fly mean lifespan (Fanson et al., 2009). To investigate which nutrient(s) in yeast are responsible for this lifespan extension, we developed eight experimental diets that included one or more nutrient classes found in yeast. All diets contained sucrose (300 g/l), agar (0.5% w/v), distilled water, and either cholesterol, Vanderzant vitamin mixture, amino acids, Wesson salt mixture, yeast hydrolysate or a combination of these ingredients (Table 1, S1). Sucrose, cholesterol, vitamins and amino acids were obtained from SigmaAldrich (St. Louis, MO, USA) and salt mixture and yeast hydrolysate from MP Biomedicals (Aurora, Ohio, USA). Cholesterol is the dominant sterol in most insects and is essential for many fundamental physiological processes (Chapman, 1998). Yeast contains a variety of sterols types (Quail and Kelly, 1996) and these are converted to cholesterol by insects (Ikekawa
Table 1 List of ingredients and amounts in each diet. Distilled water was added to make 100 ml solution. V = vitamins, M = minerals, C = cholesterol, A = amino acids. Diet
Ingredient Sucrose (g)
Sucrose Vitamins Minerals Cholesterol Amino acids Yeast V+M+C V+M+C+A a
30 30 30 30 30 30 30 30
Vitamina (mg)
Minerala (mg)
Cholesterola (mg)
Amino acida (g)
Yeast (g)
25 60 50 0.65 1.5 25 25
60 60
50 50
0.65
Estimated from MP Biomedical Nutritional Analysis of hydrolysed yeast. Cho-
et al., 1993). We therefore used cholesterol as the source of sterols in our experimental diets. For the demographic studies (2.3 below), water was added to dry ingredients, brought to boil to activate the agar, poured into plastic dishes (35 35 10 mm), and then refrigerated until used. The individual consumption experiment (below) just used the liquid form of the diets (including the agar but not boiled), as well as 0.01% (v/v) blue food dye (Queen Fine Foods Pty Ltd, Alderly, Queensland, Australia). 2.3. Demographic study 2.3.1. Mixed sex cages This experiment explored the effect of experimental diets on longevity and egg production in mixed sex cages. Within 24 h of emergence, 35 male and female flies were transferred to each of 8 demographic cages with each cage provisioned with 1 of the 8 experimental diets. We ran three replicates, with each replicate being a new fly batch from EMAI fly stock. The demographic cages were 5-litre rectangular plastic containers in which one side had been replaced by fibreglass insect screening for ventilation. Each cage was provisioned with a 70 ml container of distilled water with a cotton wick, two 55 mm plastic Petri dishes containing 15 ml of diet, and an ovipositing dish. The ovipositing dish was a parafilmcovered 55 mm plastic Petri dish containing 7 ml of 0.7% lemon essence solution (Queen Fine Foods Pty Ltd, Alderly, Queensland, Australia) (following Perez-Staples et al., 2007). The parafilm was pierced several times with an entomological pin in order to release chemical cues. Mortality was checked daily, removing any dead flies; ovipositing dishes and food were replaced every 3 days; water was replaced every 6 days. The experiment was terminated when all cages contained less than 10% survivorship of the original cohort and any remaining flies were then censored in the survival analyses (Allison, 1995). 2.3.2. Single sex cages This experiment followed the same protocol as the mixed sex cage experiment, except that 70 flies of the same sex were transferred to each of 16 cages (8 diets 2 sexes = 16). We ran two replicates of this study, using different fly batches from EMAI. No ovipositing dish was provided in this experiment. Again, the experiment was terminated when all cages contained less than 10% survivorship of the original cohort and any remaining flies were then censored in the survival analyses. 2.4. Individual consumption To explore the effects of diet composition on overall food intake, we measured diet consumption on the eight diets. Sixty-four flies
329
B.G. Fanson, P.W. Taylor / Journal of Insect Physiology 58 (2012) 327–334
(8 diets 2 sexes 4 replicates) were transferred to individual 70 ml plastic containers. Each individual container had a 200 ll micropipette tip filled with 100 ll distilled water and two 10 lL microcapillary tubes (Brand Blaubrand intraend) filled with the same diet solution. Containers were arranged on a shelving apparatus with clear plexiglas shelves (Fig. S1). Diet consumption was measured by taking still pictures using a mounted 8MP Canon IXUS 80IS camera that was programmed to photograph the containers every morning at 8 am. We then corrected pictures for barrel distortion using Adobe Photoshop CS4 (San Jose, CA, USA) and measured the distance that diet extended up the microcapillary tubes (visible as blue dye) using ImageTool (v2.0 University of Texas Health Science Centre, San Antonio, Texas). Diet consumption over each 24 h period was calculated by measuring the change in diet volumes from the previous day and correcting for evaporation (see below). Correlation between photograph method and measurement using callipers was r = 0.99 (N = 84). Microcapillaries were refilled every four days, or sooner if depleted (checked regularly through each day). To correct for evaporation, an additional 8 containers with no flies containing 4 diets were added evenly throughout the shelving unit, giving 4 replicates per diet. We then fitted a regression model for each diet using solution concentration to predict daily evaporative loss. Next, we corrected for evaporation using the following algorithm: (1) calculate the start solution concentration for each day; (2) adjust consumption volume by subtracting predicted evaporation amount based on diet concentration; (3) calculate total nutrients consumed by multiplying start solution concentration by adjusted consumption volume; (4) calculate new start solution concentration for next 24 h period by calculating total nutrients remaining (start nutrient – total consumed) and divide by volume of liquid remaining. 2.5. Statistical analyses All analyses were conducted using SAS 9.1.3 (Cary, NC, USA). For the mixed and single sex cages, we conducted a two-step survival analysis process in order to extract information about both (1) changes in baseline mortality rates (the scale factor) and (2) rates of senescence over time (the shape factor) (see Pletcher et al., 2000). First, we analysed each replicate separately using a parametric survival analysis assuming a Weibull distribution. Linear log–log survival plots suggested Weibull distribution was appropriate (Allison, 1995). Second, for each sex, we conducted a general linear model separately on the scale factor and Weibull shape parameters. Increasing Weibull shape values indicate higher rates of mortality risk with age. For this analysis, we included housing type (mixed or single sex) and an indicator variable for each nutrient class (minerals, amino acids, cholesterol, and vitamins) that the diet contained. Additionally, we included a yeast indicator variable, so that yeast was treated as a unique ingredient. For the mixed sex experiment, we described fecundity patterns using graphs and statistical comparison of total egg production. Similar to the survival analysis, we included indicator variables for each nutrient class (minerals, amino acids, cholesterol, and vitamins). For this model, we treated each cage as the individual unit. Because some cages produced no eggs and there was very low variation in total egg production from cages of flies provided yeast, no simple variable transformation would accommodate homoscedasticity and normality assumptions across the treatment groups (Fig. S2). Consequently, we conducted a general linear mixed model in which we modelled the variance estimate for each diet (Littell et al., 2006). Both AICc (Akaike information criterion corrected) and BIC (Bayseian information criteria) values were improved in comparison to the simple model and residuals met normality assumptions within each diet group.
For individual consumption, we investigated whether the amount of diet consumed differed among the experimental diets. We calculated total nutrient consumption of sucrose and amino acids for each fly and then conducted a linear model with sex, diet treatment, replicate, shelf row and column. Row and column were included to control for potential effects of different shelf location in the experimental apparatus and for residual picture distortion. Higher order interactions were removed if p > 0.1. Homoscedasticity and normality assumptions were valid without any variable transformation. 3. Results 3.1. Survival Lifespan of Q-flies varied with diet, sex and housing (Table 2). The two-step survival analysis approach allowed us to analyse how baseline mortality (scale factor) and age-specific mortality rate (Weibull shape) contributed to these differences in mean lifespan. Both sexes had higher baseline survival in mixed cages (Table 2 and 3). For males, the addition of minerals to the diet caused a marginally significant decrease in baseline mortality (Table 3; Fig 1b and c), thus increasing lifespan. For females the response was more complex. Minerals extended lifespan in single-sex housing of females by decreasing baseline mortality, whereas in mixed sex housing, females fed mineral diets had lifespan similar to females fed sucrose only. Thus, mixed housing negated the benefit of the mineral-rich diets (Table 3; Fig. 1A and C). In contrast, adding only amino acids increased baseline mortality in females (Fig. 1A and C). For males, amino acids did not strongly affect baseline mortality but it did increase the shape parameter when housed in mixed cages. In general, vitamins and cholesterol had little or no effect on survival patterns for both sexes (Table 3). Females also show a more complex scenario than males in regards to combination diets. The VMCA diet illustrates that that individual ingredients are not strictly additive (Table 3). Adding amino acids to a diet already containing cholesterol, vitamins, and minerals resulted in extended lifespan by decreasing baseline mortality and slowing ageing rates. In contrast, adding just amino acids to a diet containing sucrose only resulted in decreased lifespan. For males, yeast increased lifespan through a slightly greater decrease in baseline mortality compared to minerals ( 0.0050 vs. 0.0035). However, for females, yeast only increased lifespan in single sex housing, having little or no effect on lifespan in the mixed sex housing. 3.2. Egg production Lifetime egg production varied greatly among the diets (Table 4; Fig 2, S2). Flies fed yeast had the highest egg production rates and
Table 2 Mean lifespan for Q-flies on different diets. Flies were housed in either single sex or mixed sex cohorts. Diet
Sucrose Vitamins Minerals Amino acids Yeast Cholesterol V+M+C V+M+C+A
Female
Male
Single
Mixed
Single
Mixed
25.74 ± 1.55 27.60 ± 1.51 34.71 ± 1.95 24.31 ± 1.22 48.46 ± 3.43 27.29 ± 1.72 34.10 ± 1.90 37.07 ± 2.64
39.11 ± 2.04 41.06 ± 2.06 41.72 ± 2.55 29.29 ± 1.12 37.67 ± 1.70 41.08 ± 2.05 39.91 ± 1.92 41.71 ± 1.93
26.49 ± 1.24 27.29 ± 1.45 30.76 ± 1.88 28.32 ± 1.61 32.63 ± 1.83 29.40 ± 1.47 31.74 ± 2.03 33.76 ± 2.29
30.43 ± 1.53 30.72 ± 1.47 35.60 ± 1.96 35.04 ± 1.25 34.47 ± 1.65 31.03 ± 1.46 33.75 ± 1.82 35.38 ± 1.80
330
B.G. Fanson, P.W. Taylor / Journal of Insect Physiology 58 (2012) 327–334
Table 3 Statistical results from two-step survival analysis process. First, Weibull survival curves were fitted to each replicate. Then, a general linear model was conducted separately on the scale (baseline mortality) and Weibull shape (age-specific mortality) parameters from each replicate. The estimated differences are the change in scale and Weibull shape when the specific ingredient or condition is added. For example, adding minerals to females housed in single sex cage is 0.0088. Housing type is the effect of mixed housing compared to single-sex. For V + M + C and V + M + C + A, the estimated effect represents the interactive effects of the diets (see Section 2 for more details). Significant effects are bolded (P < 0.05). Sex
Predictor variable
Scale Diff
Female
Male
Housing (Mixed) Vitamins Minerals Minerals Housing Amino acids Yeast Yeast Housing Cholesterol V+M+C V+M+C+A Housing (Mixed) Vitamins Minerals Amino acids Amino acids Housing Yeast Cholesterol V+M+C V+M+C+A
Shape T
P-value
0.0107 0.0017 0.0088 0.0058 0.0055 0.0159 0.0156 0.0018 0.0049 0.0066 0.0034 0.0004 0.0035 0.0029
9.10 1.02 4.50 3.24 3.36 7.00 5.91 1.10 1.50 2.85 3.66 0.24 1.87 1.57
