Transcriptomic approach to uncover dynamic events in the development of mid-season sunburn in apple fruit

Abstract

Apples grown in high heat, high light, and low humidity environments are at risk for sun injury disorders like sunburn and associated crop losses. Understanding the physiological and molecular mechanisms underlying sunburn will support improvement of mitigation strategies and breeding for more resilient varieties. Numerous studies have highlighted key biochemical processes involved in sun injury, such as the phenylpropanoid and reactive oxygen species (ROS) pathways, demonstrating both enzyme activities and expression of related genes in response to sunburn conditions. Most previous studies have focused on at-harvest activity of a small number of genes in response to heat stress. Thus, it remains unclear how stress events earlier in the season affect physiology and gene expression. Here, we applied heat stress to mid-season apples in the field and collected tissue along a time course-24, 48, and 72 h following a heat stimulus-to investigate dynamic gene expression changes using a transcriptomic lens. We found a relatively small number of differentially expressed genes (DEGs) and enriched functional terms in response to heat treatments. Only a few of these belonged to pathways previously described to be involved in sunburn, such as the AsA-GSH pathway, while most DEGs had not yet been implicated in sunburn or heat stress in pome fruit.

Keywords: Malus; 3′ RNA-Seq; QuantSeq; apple; heat injury; heat stress.

Fig. 1.

Sunburn development and experimental design. a) Experimental design for timing of heat stimulus application, tissue collection, and physiological assays. b) Sunburn severity ratings using a scale of SB0 (no sunburn) to SB4 (severe with necrosis). “Honeycrisp” and “Granny Smith” were treated with temperatures below the FST threshold and above the FST threshold for sunburn, and sunburn severity was assayed 72 h later. Experiments took place in late June and early July in 2018 and 2019. Three replicates of 5–10 fruits per treatment were averaged, and average frequency is shown. Solid bars represent data from 2019, while striped bars represent 2018. Error bars indicate ± standard error of the mean (SEM) of the three replicates. c) Representative images of sunburn scale for “Honeycrisp” developed for mid-season fruits.

Fig. 2.

Pigment accumulation and spectrometry readings under increasing heat treatment. a) Anthocyanins, chlorophyll a, chlorophyll b content, and chlorophyll a/b ratios were measured in fruit peels collected from the sunned side of heat-treated and control “Honeycrisp” and “Granny Smith” fruits, 72 h post-treatment in both 2018 and 2019. Peels from 5 to 10 fruits per treatment per replicate (collection day) were pooled and pigments extracted. Error bars indicate ± SEM of the three replicates. One-way ANOVA with Tukey HSD test was used to determine significant differences, and letters represent significance with a P-value of <0.05. b) Spectrometry measurements were taken on the sunned side of fruit in the field 72 h post-treatment, prior to peeling. Indices measured were Anthocyanin Reflective Index (ARI), Carotenoid Reflective Index (CRI), Chlorophyll Normalized Difference Vegetation Index (CNDVI), and Photochemical Reflective Index (PRI). Error bars indicate ± SEM of n = 15–30 fruits for each treatment category. One-way ANOVA with Tukey HSD test was used to determine significant differences, and letters represent significance with a P-value of <0.05.

Fig. 3.

Transcript mapping to different genomes revealed annotation effect on read counts. a) Comparison of the frequency of the transcript reads mapped vs reads counted when using either the “Gala” or GDDH13 genome as a reference revealed differences between the genome used. Error bars indicate ± SEM average of the 27 sequenced libraries. Nonparametric (Wilcoxon) test results are reported. b) Comparison of expression levels using a random set of apple genes mapped to “Gala” and GDDH13 references indicated higher expression values for genes mapped to GDDH13, as well as a low R² value, indicating gene specificity. Dotted line represents y = x. c) 3′ UTRs in both the “Gala” and GDDH13 annotations were identified, counted, and binned according to length, demonstrating a difference in frequency and distribution between genome annotations. Pearson’s Chi-squared test with Yates’ continuity correction was used to determine significance. d) Cartoon demonstrating the common read distribution resulting from QuantSeq, explaining a model for the differences in mapping and read counts between genomes. e) Visual representation of read pile-ups in an example gene of interest, using JBrowse 2.