Testing the Effects of dl-Alpha-Tocopherol Supplementation on Oxidative Damage, Total Antioxidant Protection and the Sex-Specific Responses of Reproductive Effort and Lifespan to Dietary Manipulation in Australian Field Crickets (Teleogryllus commodus)

The oxidative stress theory predicts that the accumulation of oxidative damage causes aging. More generally, oxidative damage could be a cost of reproduction that reduces survival. Both of these hypotheses have mixed empirical support. To better understand the life-history consequences of oxidative damage, we fed male and female Australian field crickets (Teleogryllus commodus) four diets differing in their protein and carbohydrate content, which have sex-specific effects on reproductive effort and lifespan. We supplemented half of these crickets with the vitamin E isoform dl-alpha-tocopherol and measured the effects of nutrient intake on lifespan, reproduction, oxidative damage and antioxidant protection. We found a clear trade-off between reproductive effort and lifespan in females but not in males. In direct contrast to the oxidative stress theory, crickets fed diets that improved their lifespan had high levels of oxidative damage to proteins. Supplementation with dl-alpha-tocopherol did not significantly improve lifespan or reproductive effort. However, males fed diets that increased their reproductive investment experienced high oxidative damage to proteins. While this suggests that male reproductive effort could elevate oxidative damage, this was not associated with reduced male survival. Overall, these results provide little evidence that oxidative damage plays a central role in mediating life-history trade-offs in T. commodus.

. The nutrient content of the diets used in this experiment. Points represent the protein and carbohydrate composition of the four diets used. The ratio of protein to carbohydrate within each diet is represented by black lines or "nutrient rails". We plot all the nutrient rails used in an earlier study of T. commodus [1] to allow comparison. In this study [1] diets were numbered, 4 (red), 8 (blue), 10 (green) and 14 (yellow).  Figure S3. The DL-alpha-tocopherol content of female crickets fed each of the diets used in this experiment (4,8,10,14) and supplemented (+, closed circles) or not (−, open circles) with DL-alpha-tocopherol. Diets are as in Figure S1, error bars are standard errors around the mean.  Figure S4. The DL-alpha-tocopherol content of male crickets fed each of the diets used in this experiment (4,8,10,14) and supplemented (+, closed circles) or not (−, open circles) with DL-alpha-tocopherol. Diets are as in Figure S1, error bars are standard errors around the mean.
Text S1. Confirming that Our DL-Alpha-Tocopherol Supplementation Treatment Was Successful To measure DL-alpha-tocopherol, 5 mL of pyrogallol was added to 300 μL of whole cricket homogenate in a 40 mL glass vial. Pyrogallol prevented any oxidation of DL-alpha-tocopherol. 1 mL of KOH was added to each sample for saponification. Samples were recapped under N2 gas and placed in a water bath at 70 °C for 30 min. 5 mL of hexane was then added to extracted DL-alpha-tocopherol. 10 mL of Milli-Q was added to the samples, all samples were recapped under N2 gas and put on a bench-top mixing plate for five minutes. They were then centrifuged at 10,000 RPM at 4 °C for 4 min. 3 mL of the hexane layer was extracted and dried down in an evaporating rotary chamber for 20 min. Once completely evaporated, 100 uL of 0.05% butylated hydroxytouene (BHT) in ethanol was added to dissolve the DL-alpha-tocopherol. Samples were then vortexed for 30 s at 15 Hz. 100 uL of sample was injected into a Dionex HPLC system fitted with a Waters Spherisorb 3 um ODS2 column (4.6 × 150 mm, Hertfordshire, UK). The mobile phase was methanol to water (97:3), run isoratically for 15 min. DL-alpha-tocopherol was detected using a fluorescence detector, with an excitation wavelength of 330 nm and an emission wavelength of 480 nm. Peak area was quantified using a standard curve prepared from alpha-tocopherol standard. The peaks allowed for quantitative amounts of μg of vitamin per gram of cricket to be calculated.
To test if DL-alpha-tocopherol levels were higher in supplemented animals, we used the "glm" function in R. Sex, diet and supplementation level were included as explanatory variables, as were all the interactions between them, and DL-alpha-tocopherol content per mg of cricket was our response variable. DL-alpha-tocopherol levels were log transformed prior to analyses. Significance of terms was We used a sequential model building approach to determine if the linear and nonlinear effects of protein and carbohydrate consumption were different across our response variables [2,3]. Because our response variables (e.g., lifespan versus female fecundity) were measured in different scales (e.g., days versus eggs laid), we standardized them for statistical comparison to make sure that any differences we see in the linear or nonlinear effects of nutrient intake are not driven simply by differences in scale. To do this, we used a Z-transformation to standardize each response variable and nutrient intake to a mean of zero and standard deviation of one. We then included a dummy variable, response type (RT), in a reduced model containing only the standardized linear terms: where R is our standardized response variables, Ni refers to the intake of the ith nutrient, n represents the number of nutrients contained in the model and ε is the unexplained error.
From Equation (1), the unexplained (i.e., residual) sums of squares for this reduced model (SSr) was compared to the same quantity (SSc) from a second (complete) model that included all of the terms in Equation (1) with the addition of the terms αiNiRT which represents the linear interaction of RT and the ith nutrient.
To compare SSr and SSc from (Equation (1)) and (Equation (2)) respectively we used a partial F-test [4]: where a is the number of terms that differ between the reduced and complete model while b is the error degrees of freedom for SSc.
To test whether the quadratic effect of nutrient intake differed across response variables, the SSr from the reduced model: Antioxidants 2015, 4 S5 was compared to the SSc of the complete model: using (Equation (3)).
To test if the correlational effects of nutrient intake on response variables differed, the SSr from the reduced model: was compared to the SSc of the complete model: using (Equation (3)). This approach means that the comparison of model (Equation (1)) versus (Equation (2)), (Equation (4)) versus (Equation (5)), and (Equation (6)) versus (Equation (7)) tests for the overall significance of the interaction between response type and the linear, quadratic and correlational effects of nutrient intake, respectively. Significant differences in these model comparisons (identified with a partial F-test) therefore show that the linear, quadratic and/or correlational effects of nutrient intake on the response variables differ. Finally, we also considered the interaction of different nutrients with the response variable terms from the full model (Equation (7)) to determine if intake of protein, carbohydrate or both nutrients were responsible for the significance of the overall partial F-test.
Text S3. The Annotated R Code That Was Used to Estimate the Angle ( ), and 95% CIs, between Linear Vectors for the Effects of Nutrients on Lifespan, Reproductive Effort, Oxidative Damage and Antioxidant Protection. The package MCMCglmm, developed by Jarrod Hadfield (https://cran.rproject.org/web/packages/MCMCglmm/index.html) is required to run this code.

Protein (P)
Carbohydrate (C) P + C P:C Diet Number        Table S5. Sequential model comparing the linear and nonlinear effects of protein (P) and carbohydrate (C) intake on protein carbonylation (PC) and total antioxidant capacity (TAC) across the sexes in T. commodus with and without DL-alpha-tocopherol supplementation. SS refers to sums of squares of the reduced (SSr) and the complete (SSc) models. DF quantifies the degrees of freedom and F is the test statistic.      Antioxidants 2015, 4 S13 Table S8. Results of some studies examining how dietary manipulation affects oxidative protection (antioxidant levels), ROS production, or oxidative damage. This is not intended to be a comprehensive list but simply to illustrate the complex associations between dietary manipulation, oxidative damage and protection in a range of taxa. Abbreviations: Sex: NA-not appropriate, F = female, M = male. Dietary manipulation: AAR-amino acid restriction, DR-total food or energy restriction, GlucR-Glucose Restriction, PRprotein restriction, MetR-methionine restriction. Results: CAT-antioxidant catalase, GSH-antioxidant glutathione, GPX-antioxidant glutathione peroxidase, mtDNAmitochondrial DNA, nDNA-nuclear DNA, SOD-antioxidant superoxide dismutase. Symbols: ↑-increased, "−"-no significant difference, ↓-reduced. All comparisons are relative to control animals e.g., DR increases damage relative to controls. If two symbols are provided, results differed across tissue types or over time.