2.1 Plant Materials
An F5 RIL population was derived from reciprocal crosses of the parents Ailsa Craig and Kentucky Beefsteak (TR00021), hereafter abbreviated with “A ~ K”. Among the 148 individuals in the population 75 genotypes have Ailsa Craig as a mother and 73 plants Kentucky Beefsteak as a mother. The F4 generation was genotyped by Nunhems Netherlands B.V. The genotyping was done with 185 KASP markers (2.9% of marker calls scored as heterozygous and were excluded from the QTL analysis).
2.2 Plant growth conditions
For the phenotyping of the RIL populations, the plants were grown in glasshouses of Unifarm (Wageningen University and Research) between September 2019 and January 2020. The plants were sown, measured and harvested in eight batches spread over two weeks. The plants were pre-grown on ebb and flood tables for four and a half weeks under natural lighting (supplemented with approximately 100 µmol m-2 s-1 HPS light 16h before sunset unless the outside global radiation was above 250W m-2). Afterwards the plants were divided onto two compartments (12 m x 12 m) planted in double rows (ca. 2.2 plants per m2) with drip irrigation. In each compartment the same set of 148 genotypes was grown with one replicate each as well as six replicates of each parent of the RIL population. A border row of plants was grown that was not measured (to avoid a border effect). The plants were kept at a height of approximately 2.5 m to ensure near equal supplemental light distribution for all genotypes. The two compartments differed in the supplemental lighting system, with an average of about 220 µmol m-2 s-1 supplemental lighting (measured at 2.5 m over ground on several spots per row within each compartment) of either 95% red and 5% blue LEDs (Green Power LED top lighting module DR/B LB, 190W, Philips) or high pressure sodium lamps (HPS; Master green power, cgt 400W, Philips). The lamps hung at ca. 3.5m over ground. Supplemental lighting was provided for 16 hours, ending at sunset. The plants were grown on stone-wool (Rockwool Grodan B.V., Roermond, The Netherlands), irrigated with nutrient solution: NH4 1.2 mM, K 7.2 mM, Ca 4 mM, Mg 1.82 mM, NO3 12.4 mM, SO4 3.32 mM, P 1 mM, Fe 35 μM (Fe‐DTPA/ Fe‐EDDHSA), Mn 8 μM, Zn 5 μM, B 20 μM, Cu 0.5 μM and Mo 0.5 μM. The macronutrients were obtained from a mixture of fertilizers of Yara Benelux BV (Rotterdam–Vlaardingen, The Netherlands) and the micronutrients were from the Agrispoor product line of Horticoop BV (Bleiswijk, The Netherlands). The final pH was adjusted with KOH to be between 5.5 and 5.8 with an EC of 2.0 dS m-1.The EC of the nutrient solution in the stone wool slab was on average 3.8 ds m-1 under HPS versus 3.2 ds m-1 under LED. If the air temperatures in the compartments dropped below ca. 22°C (in a later stage of the experiment 24°C) during the supplemental lighting period or 19°C outside the supplemental lighting time, the compartments were heated (in case of the LED supplemental lighting with heating pipes below the plants and above the lamps to compensate for the heat irradiance from the HPS lamps; the HPS compartment was only heated by pipes below the plants). The Photosynthetically Active Photon Flux Density (PPFD) from the sun was estimated from measured solar radiation outside the greenhouse. For the estimation of the sunlight inside the greenhouse, a transmissivity factor of 0.62 was used based on measurements. The measured solar radiation was corrected for the fact that ca 47% of the solar radiation is in the PPF range and W m-2 was converted into µmol m-2 s-1 by assuming that 1 W m-2 solar light equals 4.57 µmol m-2 s-1(McCree 1972; Britton and Dodd 1976). Average environmental conditions over the experimental growth period are given in Supplemental Table 1. Pollination was warranted by the placement of bumblebees (Koppert, biological systems) in the greenhouse compartments.
Blossom-end rot was scored as the number of fruit with visual outside symptoms (Figure 1) compared to the number of total mature fruits of each plant over the whole growth period.
Brix was determined on one healthy, mature fruit per genotype and light condition (with no visual symptoms of rot outside or inside the fruit) using a refractometer (PR-32α, ATAGO, Tokio, Japan). For the parents 6 to 11 fruits (of up to 6 genotypic replicates) were measured per light condition for Brix and acidity with the PAL-BX|Acid3 (ATAGO, Tokio, Japan).
The fruit shape was scored visually with a numeric score from 0 to 3, where 0 are round fruit and 3 are highly facetted fruits (Supplemental Figure 1). In case fruits were not clearly scorable in those four categories half score points (e.g. 0.5 or 1.5) were given.
At the end of the growth period initial symptoms of tobacco mosaic virus (TMV) were observed (curled leaves at the newly developing leaves in the top). However, there was no significant difference neither in BER (p-value = 0.47) nor in Brix (p-value = 0.89) between plants with and without first TMV signs.
2.4 Statistical Analysis
If not otherwise indicated, the statistical analysis was done using the basic statistics of R (version 3.5.0; www.r-project.org). Normality was tested with a Shapiro test and consecutively equal variance between both light conditions with an F-test. The Brix values of the RILs were normally distributed and showed equal variance between samples from the two light conditions. For this trait a paired t-test was used to test for light environment effects. For all other traits a Wilcox rank sum test was performed, as the difference of the pairs was not symmetrically distributed around the median of the pair-average and therefore did not fulfil the requirement for a signed rank sum test. The reciprocal cross group effect was tested per light-condition. To test the environmental effect on the LED-specific QTL for Brix, an ANOVA was performed with the Car-package in R (version 3.0-0; Fox and Weisberg 2019). In this model, the allele scores at the peak marker of the QTL on chromosome 6 and the light conditions were interacting factors and the marker alleles of the QTL on chromosome 2 as additive factor. A post hoc Tukey test on the four haplotype groups (two parental alleles at QTL on chr. 2 and chr. 6) was done with the package Agricolae (version 1.3-2; De Mendiburu 2020). Average values in the text are, if not otherwise indicated, shown with their standard error of the mean. All graphics were made in R, using the basic functions or with the package ggplot2 (version 3.0.0; Wickham 2016) or R/qtl (Broman et al. 2003). The genetic map was re-calculated for the RIL population (based on map information provided by the company) and the QTL analysis was done with the package R/qtl (Broman et al. 2003), following Bowman and Sen (2009). This QTL analysis was done separately for the plants grown under HPS or LED light. For fruit form and percentage BER nonparametric interval mapping function was used with a LOD score threshold determined based on 1000 permutations. For Brix, the parametric QTL analysis was done with the in R/qtl implemented multiple imputation method (Sen and Churchill 2001). The LOD score threshold was calculated based on 1000 permutations with 300 imputations for the estimation of the genotype probabilities. The explained variance of the QTL was estimated for all traits with the multiple imputation method. The size of the QTL regions were determined by an approximate 95% Bayes credible interval (expanded to the closest markers). The number of candidate genes was predicted based on the SL2.5 tomato gene model using the tool Jbrowse of the Sol genomics network (solgenomics.net).