Effects of NAA and shading on young fruit abscission and ethylene production by young fruit and leaves
A comparison of the relative effectiveness of NAA and shading as inducers of fruit abscission revealed significant treatment-specific differences in abscission rates and totals. For example, while both treatments promoted detectable increases in abscission rates within the first 7 d post-treatment, the NAA-induced abscission rate remained essentially unchanged from 7 to 13 d, whereas shading resulted in a relatively steady increase in the rate of fruit abscission for the same period (Figure ). By 15 d, however, similar rates of abscission were observed for both treatments and rates were near or below control rates by 19 d for both treatments. Ultimately, shading was significantly more effective in promoting fruit drop, causing 98% of the fruit to abscise within the 19-d period of the study, compared to a 75% loss in the same period following NAA treatment. Interestingly, the pattern of abscission exhibited by controls roughly mirrored that of treated trees but resulted in less than 10% of the fruit being shed, indicating that the NAA and shading treatments were able to act additively or synergistically with the endogenous early fruit abscission program (Figure ).
Figure 1 Effect of NAA and shading treatments on fruit abscission of 'Golden Delicious' apples. (A) NAA and shading increase fruit abscission rate. (B) NAA and shading treatment also increase the percentage of total fruit drop. Results represent the mean (± (more ...)
Previous studies have documented the link between NAA- and shading-induced abscission and ethylene production. We confirmed that the major increases in ethylene production preceded the onset of fruit abscission (Figure compared with Figure ). The maximum level of NAA-induced fruit ethylene production was detected at 1 d and had decreased to control level by 7 d after treatment. Shaded fruit also released higher levels of ethylene relative to the control between 1 and 5 d, but levels reached only 50% of those from NAA-treated fruit (Figure ). A similar pattern of ethylene release was detected for leaves, where the ethylene production of both shaded and NAA-treated leaves peaked at 1 d, but decreased to control level by 5 and 7 d after treatment, respectively. However, at its peak rate, measured at 1 d, almost 70-fold more ethylene was released by leaves compared to fruit (Figure ).
Figure 2 Effect of NAA and shading treatments on ethylene production of 'Golden Delicious' apples. (A) NAA and shading induce ethylene in young fruit. (B) NAA and shading also induce ethylene in leaves. Results represent the mean (± SE) of three replicates. (more ...)
Expression profiling of young fruit abscission induced by NAA and shading
To identify genes whose expression patterns correlated with the fruit abscission induced by NAA and shading, gene expression profiling was performed using the FAZ at three time points (D1, D3 and D5), a period spanning the earliest phase of the treatment-dependent increase in abscission above control levels (Figure ). For each time point, labeled cDNA from the FAZ of NAA- or shading-treated trees was hybridized to reference cDNA from the FAZ of non-treated trees for the same time point, so as not to confound the changes in gene expression caused by treatments with those occurring during fruit development.
Seven-hundred-twenty-two genes from NAA-treated sample hybridizations and 1057 genes from shading-treated sample hybridizations showed statistically significant changes in expression (Additional file 1
, Table S1). Of these genes, 168 were differentially expressed in FAZs from both NAA- and shading-treated samples, and 86% (145) of those displayed similar expression patterns, indicating that NAA- and shading-induced abscission share some common signaling pathways.
Time points and selected genes were grouped according to expression pattern by hierarchical cluster analysis (Additional file 2
, Figure S1). Following NAA treatment, the largest number of differentially expressed genes was detected at 3 d, followed by 5 and 1 d after NAA treatment, suggesting that NAA caused a transient effect during the abscission induction. In contrast, shading led to a sustained increase in the number of differentially expressed genes from 1 to 5 d. For both treatments, there were approximately equal numbers of genes showing upregulation or downregulation by 1 d, but induced genes outnumbered repressed genes for 3 and 5 d.
The two array datasets were further analyzed using K-means clustering (KMC) for genes whose expression pattern was correlated with the induction of fruit abscission (Additional file 3
, Table S2). The cluster names were assigned upregulated (u
), unchanged (o
) or downregulated (d
) for each time point. NAA-responsive genes were classified into 8 main clusters. As shown in Additional file 4
(Figure S2), the largest group (comprised of clusters 2 and 6) of differentially regulated genes was detected at 3 d. Shading-responsive genes were divided into 10 clusters. Similar to the NAA dataset, most clusters reflected either up- or down-regulation at one or two time points. In contrast to the NAA treatment, two clusters in the shading dataset (2 and 7) showed a persistent repression or induction at all three time points, and one cluster (5) comprised 51 genes that were first downregulated at 1 d, but later upregulated at 5 d after treatment (Additional file 4
, Figure S2).
We examined gene expression patterns to identify functional categories correlated with fruit abscission. Genes probed by the apple array have both annotation and gene ontology (GO) information. However, as some annotations and GO categories do not provide detailed information on the biological mechanisms, additional manual annotations and literature validations were conducted for the entire list of differentially expressed genes. The resulting 15 functional categories included eight categories (photosynthesis, metabolism, membrane/cellular trafficking, cell cycle, hormone response, cell wall modification, protein metabolism and transcription factors) that accounted for over 70% of all the differentially expressed genes for both treatments (Figure ). After further gene categorization, we found that some categories and their subcategories showed trends where most members were either up- or downregulated by one or both treatments. To determine if such expression trends were statistically significant or just occurred by chance, χ2
and Fisher's exact tests were performed on these categories and their subcategories, showing most, but not all categories and subcategories with non-random expression trends. A similar approach was used by Dardick et al (2007) for the enrichment analysis [24
]. Those statistically significant categories and subcategories were shown in Figure . All the resulting classifications were displayed in Additional file 5
(Table S3) and used for the subsequent analysis.
Figure 3 Functional categories of statistically significant genes. (A) Differentially expressed genes are categorized from NAA-treated FAZ. (B) Differentially expressed genes are categorized from shading-treated FAZ. The functional categorization is based on the (more ...)
The chloroplast is the site of the energy transduction and Calvin Cycle phases of photosynthesis and starch metabolism. Reductions in chloroplast function in response to shading have been reported previously [7
] and were consistent with our transcript profiles from shading-treated trees. We found that NAA treatment also led to strong reductions in photosynthesis-related gene expression, which supported a previous report [4
]. Although shading downregulated a larger number of photosynthesis-related genes than NAA, in both treatments over 90% of the differentially expressed genes related to chloroplast function were repressed (Additional file 5
, Table S3). The affected genes function in light-harvesting, oxygen evolution, electron transport and carbon fixation. Shading-repressed genes were also involved with chlorophyll biosynthesis, chloroplast DNA binding, thylakoid formation and carbon utilization. However, only a small overlap was observed between this latter group of shading-repressed genes and those repressed by NAA, indicating that NAA and shading repress photosynthesis-related processes through partially distinct mechanisms.
Carbohydrate metabolism and sugar sensing
Not surprisingly, the repression of photosynthesis-related gene expression caused by both treatments was linked with changes in the expression of genes in the metabolism category, with the largest subset of differentially expressed genes belonging to carbohydrate metabolism. Affected genes within this category include those associated with glycolysis, the cleavage of glycosidic bonds, sugar phosphorylation and signal transduction. Thirty-eight carbohydrate metabolism genes showed significantly altered expression in response to NAA, while 149 genes were regulated by shading (Additional file 5
, Table S3). NAA-induced genes included those involved with glycolysis and starch degradation, such as pyruvate kinase, alcohol dehydrogenase, amylase and limit dextrinase. Similarly, shading treatment induced genes associated with glycolysis, but also led to changes in the expression of genes for carbohydrate active enzymes, i.e., induction of beta-glycosidases and glycosyltransferase and repression of alpha-glycosidases. Genes related to sucrose metabolism, e.g., sucrose phosphate synthase (SPS
) and sucrose phosphate phosphatase (SPP
), were inversely regulated by the two treatments: NAA repressed SPS
and induced SPP
, while shading induced SPS
and repressed SPP
. A group of genes identified as cytosolic and cell wall invertases were induced by shading in the FAZ, indicating a possible increase in sucrose breakdown in the FAZ. These same invertases were not differentially regulated by NAA, however. The expression of three distinct putative alkaline/neutral invertase genes was reduced after both NAA and shading treatments. Sorbitol dehydrogenase (SDH
) was repressed by both NAA and shading, while NADP-dependent D-sorbitol-6-phosphate dehydrogenase (S6PDH
) was induced by NAA but repressed by shading. Many ADP/UDP-glucose pyrophosphorylase genes responsible for starch synthesis were downregulated by shading, but none was differentially regulated by NAA.
Sugar signals are generated from various source organs in response to stresses and changes in metabolic fluxes [25
]. Hexokinase (HXK
) senses glucose levels and SNF-related protein kinases (SnRKs
) are important to metabolic reprogramming in response to changes in carbohydrate levels [26
]. Shading altered the expression of various carbohydrate kinases, but HXK
was upregulated by NAA only in the FAZ.
Trehalose serves as a storage carbohydrate and stress protectant, which usually accumulates during starvation conditions [25
]. Trehalose metabolism genes were downregulated by shading but not by NAA treatment. Specifically, genes encoding trehalose-6-phosphate synthase (TPS
) and trehalose-6-phosphate phosphatase (TPP
) were repressed in the shading-treated FAZ, suggesting a decreased trehalose level in the FAZ resulted from shading.
A large group of transporters for sugars, lipids, amino acids and metal ions were differentially expressed in response to both treatments. The expression of all sorbitol/sucrose transporter genes was consistently repressed by both treatments, while shading downregulated more genes related to general sugar transport, such as hexose transporters. A class of genes related to membrane and cytoskeleton function, including microtubule, vesicle-mediated membrane transporter and cell adhesion genes, were found exclusively repressed by NAA. In all, twice as many transport-related genes were affected by shading compared to NAA, among which several ion transporters, especially for calcium and potassium, were significantly upregulated by shading while others for water transport were downregulated. Another group of transporters, ATP-binding cassette transporters (ABC transporters), were induced by both treatments.
Cell cycle-related genes
Similar numbers of cell cycle genes were identified in the two datasets of differentially expressed genes, including two classes of regulatory genes, cyclin and cyclin-dependent kinase (CDK), being repressed by both NAA and shading. Several cell division control proteins were also downregulated while one CDK inhibitor was upregulated in the FAZ.
Hormone synthesis and signaling
Many genes involved in different hormone synthesis and signaling pathways showed significant expression changes in response to both treatments. ABA has been implicated as a regulator of stress-induced senescence [28
]. In this study, NAA appeared to have limited effect on ABA-related genes in that it only upregulated three 9-cis-epoxycarotenoid dioxygenase (NCED
) genes and a zeaxanthin epoxidase gene, which encode key enzymes in ABA biosynthesis. In contrast, shading altered 26 ABA-related genes involved in biosynthesis, including NCED
, short chain dehydrogenase/reductase (SDR
) and abscisic aldehyde oxidase (AAO
), and several genes related to ABA signaling, including protein phosphatase type 2C and ABA responsive elements-binding factors.
A divergence in auxin-related gene expression was noted for shading- versus NAA-treated trees. Only two auxin-induced SAUR-like and two auxin transport genes showed significant changes in response to shading. However, 21 auxin-related genes were differentially altered by NAA and these genes included IAA-amido synthase, auxin-amidohydrolase, AUX/IAA proteins and various auxin response factors (ARFs
). Genes related to auxin polar transport were also affected by NAA, with auxin influx carriers induced and efflux carriers largely repressed. The latter effect diverged from that observed for shading-treated trees, where the same auxin efflux carrier genes were induced (Additional file 5
, Table S3).
In response to both treatments, genes for ethylene biosynthesis and perception were upregulated, including 1-aminocyclopropane-1-carboxylate synthase (ACS) and oxidase (ACO) and two classes of ethylene receptors (ERS and ETR). Coinciding with the increased ethylene biosynthesis (Figure ), the expression of spermidine synthase gene, a key gene related to polyamine biosynthesis, was consistently reduced by both treatments.
Genes involved with cytokinin and gibberellic acid (GA) signaling pathways were downregulated by shading and NAA. Also, shading increased the expression of a GA2-oxidase gene, which is responsible for GA catabolism. Regarding brassinosteroid (BR)-related genes affected by shading, a BR oxidase gene was repressed, while a BR-signaling kinase gene was induced 3 d after shading. In contrast, expression of BR-related genes was not affected by NAA treatment.
Cell wall modification
A shared set of 11 genes associated with cell wall biosynthesis, loosening and degradation was responsive to both treatments, with most exhibiting changes at 3 and 5 d after treatment. Specifically, cellulose synthase genes were repressed while other genes related to cell wall loosening and hydrolysis, including β-1,3-glucanase, polygalacturonase and expansin, were all induced.
Proteolysis and programmed cell death
A number of genes putatively involved in the ubiquitylation pathway were upregulated. Several upregulated genes within the NAA dataset encode F-box proteins and other members of ubiquitin E3 ligase complex, including cullin and ubiquitin-conjugating enzymes. In comparison, shading caused a more widespread induction of genes responsible for protein ubiquitylation and degradation. Shading also had greater impact on the expression of 26S proteasome subunit genes. Another group of genes co-induced by both treatments included those possibly involved in programmed cell death, such as clp and cysteine proteases and autophagy genes. Similar to the pattern observed for cell wall degrading genes, the induction of almost all genes identified in cell death category was detected at 3 and 5 d.
Transcription factors (TFs)
Several classes of TFs exhibited significant changes in expression. Ten TFs were co-regulated by shading and NAA, including ERF/AP2 transcription factors, bZIP proteins, MADS-box and MYB domain proteins. The differentially expressed ERF/AP2 TFs were co-expressed with the genes for biosynthesis and signaling of ethylene and ABA, consistent with their roles in these two hormone signaling pathways [30
]. Interestingly, a homolog of the JOINTLESS
), which encodes a MADS-box TF and regulator of abscission zone formation [32
], was upregulated by both treatments. While there were both up- and down- regulated NAC domain genes in the shading dataset, NAC
genes were not differentially expressed in response to NAA. Distinct sets of WRKY TFs were induced by NAA and repressed by shading (Additional file 5
, Table S3).
Validation of array data in the FAZ and analysis of selected genes in other tissues via RT-qPCR
Subsets of genes from the above categories were selected for validation of array data in the FAZ by RT-qPCR (Additional files 6
, Table S4 and S5). cDNA samples derived from three additional time points (D0, D7 and D9 after treatment) were included to expand the expression pattern data for these genes. The relative expression levels measured by RT-qPCR were converted to fold change relative to the value obtained from the array data for reference control samples to enable direct comparison to the RT-qPCR results. Generally, the RT-qPCR results from the FAZ samples were consistent with the array data in terms of the overall expression pattern but variations were also observed (Figures , , and ). To further explore the effects of NAA and shading on source-to-sink relationships, we analyzed tissue-specific expression pattern of selected genes involved in photosynthesis, sugar metabolism and hormone metabolism and signaling using cDNA derived from leaf and fruit cortex (FC) (Figures and ).
Figure 4 Expression of genes related to ABA biosynthesis and signaling as determined by RT-qPCR. Left column, Gene expression in fruit abscission zone (FAZ) from 'Golden Delicious' apple trees after application of NAA and shading. Red lines indicate normalized (more ...)
Figure 5 Expression of genes related to ethylene biosynthesis and signaling as determined by RT-qPCR. Left column, Gene expression in fruit abscission zone (FAZ) from 'Golden Delicious' apple trees after application of NAA and shading. Red lines indicate normalized (more ...)
Figure 6 Expression of genes related to sugar transport and polar auxin transport as determined by RT-qPCR. Left column, Gene expression in fruit abscission zone (FAZ) from 'Golden Delicious' apple trees after application of NAA and shading. Red lines indicate (more ...)
Figure 7 Expression of genes related to abscission zone formation and cell wall degradation as determined by RT-qPCR. Gene expression in fruit abscission zone (FAZ) from 'Golden Delicious' apple trees after application of NAA and shading. Red lines indicate normalized (more ...)
Figure 8 Expression of genes related to photosynthesis and sugar availability as determined by RT-qPCR. Gene expression in leaf from 'Golden Delicious' apple trees after application of NAA. The values of transcript levels in the leaf from control trees were arbitrarily (more ...)
Figure 9 Expression of genes related to sugar metabolism as determined by RT-qPCR. Gene expression in fruit cortex (FC) from 'Golden Delicious' apple trees after application of NAA and shading. The value of transcript levels in the FC from control trees were arbitrarily (more ...)
In both FAZ and FC, the expression of MdNCED was induced by both shading and NAA from 3 d. In addition, genes encoding a SDR family protein and a transcription factor (AHAP), for the regulation of ABA signaling were induced in the FAZ from 1 to 5 d. In the FC, the expression of those genes was consistently increased by both treatments, especially on 3 and 5 d after treatment (Figure ). An upregulation of genes encoding ethylene biosynthesis and signaling (MdACS, MdACO, MdETR and MdERS) was confirmed by RT-qPCR for NAA- and shading-treated FAZ and mirrored in the FC. Overall, the induction of these ethylene-related genes in the FC was greater in response to NAA than shading, corresponding with the higher levels of ethylene released by fruitlets treated with NAA versus shading (Figure compared to Figure ). Consistent with the microarray data, both sorbitol and sucrose transporter genes (MdSOT and MdSUT) were repressed in the FAZ by NAA and shading from 1 to 5 d. In contrast, the expression of these transporters in the FC was increased from 3 through 7 d after both treatments. As for auxin polar transport, a PIN-like auxin transporter gene (PIN) and an auxin efflux carrier gene (AEC) both showed consistently decreased expression from 3 d in NAA-treated FAZ and FC (Figure ). Concerning AZ formation and cell wall degradation, the MdJNT gene expression in the FAZ was increased by both NAA and shading from 3 d after treatment and remained higher than the control. We also observed an increase of expansin (EXP) gene expression in both NAA- and shading-treated FAZ as early as 1 d after treatment, concurrent with the burst of fruit ethylene production (Figure compared to Figure ). MdPG2 expression in the FAZ was induced by NAA and shading from 5 d onward, corresponding with the increased rate of fruit abscission (Figure compared to Figure ).
Since a widespread repression of chloroplast-related genes in the FAZ was evident from the array data, we further tested leaves to see if photosynthetic organs were similarly affected, RT-qPCR results showed a sustained repression of the selected genes involved with light-harvesting (CAB), oxygen evolving enhancement (PSB) and Rubisco activation (RuBACT) in NAA-treated leaves as early as 1 d after treatment (Figure ). The expression of genes encoding transporters for both sorbitol and sucrose (MdSOT and MdSUT) was also found to be repressed in leaves. The expression of three other genes related to sugar metabolism was tested in the FC, which is a site of active carbohydrate metabolism. As shown in Figure , HXK expression was significantly induced by shading, with maximum levels detected at 7 d. HXK expression was most increased by NAA on 5 d, and remained higher than the control level thereafter. The expression of SDH gradually increased in the control fruit, but was significantly repressed by NAA and shading. Although our array data showed a consistent downregulation of TPS in the FAZ from shading-treated trees, and no effect on TPS expression due to NAA, both NAA and shading were shown to cause an early induction of TPS in the FC (Figure ), implicating TPS in fruit-specific aspects of abscission independent of the method of induction.
Effects of NAA on leaf photosynthesis
From our array data, a large group of genes related to photosynthesis were identified as strongly repressed by NAA in the FAZ at an early stage, implying that NAA might directly interfere with photosynthesis. Therefore, we measured the effect of NAA on the leaf by monitoring the FV/FM value which provides a useful relative measure of the maximum quantum yield of PSII primary photochemistry. The NAA-affected leaves displayed a unique pattern where the Fv/Fm readings were significantly decreased under fluorescence imaging system, indicating that the leaves were under stress. We also found that NAA at various rates (15, 150, 450 and 900 mg L-1) caused concentration-dependent impairment of PSII in the leaves of young seedlings in the growth chamber (Figure ). NAA at 15 mg L-1, the working concentration used in the thinning experiment, caused significant photoinhibition of leaf PSII efficiency (Figure ). Such inhibition was observed as early as 10 min post-treatment and lasted for 8 or more hours, from which the leaves typically recovered within 1 d. Next, a field trial on fruit-bearing trees was performed. More severe effects of NAA at 15 mg L-1 on leaf photosynthesis were found and these effects lasted longer than in young seedlings (Figure ). This increase in severity is not surprising given the higher light levels in the field (full sunlight) versus greenhouse conditions. It is also important to note that under field conditions, the leaves showed visible necrosis by 24 h post-treatment, specifically near the petiole where photoinhibition was most strongly observed. It is not clear if this spatial imbalance was caused by pooling of the NAA solution near the base of the leaf or whether this portion of the leaf is particularly sensitive to NAA. Taken together, these findings indicate that NAA application has rapid and severe impacts on leaf photosynthetic efficiency.
Figure 10 Effect of NAA at 15 mg L-1 on apple leaf. (A) White light image of the effect of NAA at various rates on the leaves of young seedlings in growth chamber. Necrosis was observed in a concentration-dependent manner. (B-D) CFI image of the leaves of young (more ...)