Tolerance to crowding stress has played a crucial role in improving agronomic productivity in field corn; however, commercial sweet corn hybrids vary greatly in crowding stress tolerance. The objectives were to 1) explore transcriptional changes among sweet corn hybrids with differential yield under crowding stress, 2) identify relationships between phenotypic responses and gene expression patterns, and 3) identify groups of genes associated with yield and crowding stress tolerance. Under conditions of crowding stress, three high-yielding and three low-yielding sweet corn hybrids were grouped for transcriptional and phenotypic analyses. Transcriptional analyses identified from 372 to 859 common differentially expressed genes (DEGs) for each hybrid. Large gene expression pattern variation among hybrids and only 26 common DEGs across all hybrid comparisons were identified, suggesting each hybrid has a unique response to crowding stress. Over-represented biological functions of DEGs also differed among hybrids. Strong correlation was observed between: 1) modules with up-regulation in high-yielding hybrids and yield traits, and 2) modules with up-regulation in low-yielding hybrids and plant/ear traits. Modules linked with yield traits may be important crowding stress response mechanisms influencing crop yield. Functional analysis of the modules and common DEGs identified candidate crowding stress tolerant processes in photosynthesis, glycolysis, cell wall, carbohydrate/nitrogen metabolic process, chromatin, and transcription regulation. Moreover, these biological functions were greatly inter-connected, indicating the importance of improving the mechanisms as a network.
Citation: Choe E, Drnevich J, Williams MM II (2016) Identification of Crowding Stress Tolerance Co-Expression Networks Involved in Sweet Corn Yield. PLoS ONE 11(1): e0147418. doi:10.1371/journal.pone.0147418
Editor: Jin-Song Zhang, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, CHINA
Received: September 29, 2015; Accepted: January 4, 2016; Published: January 21, 2016
This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.
Data Availability: All field data are within the paper and its Supporting information files. Microarray results have been submitted in NCBI’s Gene Expression Omnibus database and are accessible through GEO Series accession number GSE73435 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE73435).
Funding: These authors have no support or funding to report.
Competing interests: The authors have declared that no competing interests exist.
Modern corn (Zea mays) yields are largely the result of the interaction between improved genetics and agronomy [1–3]. Regarding genetics, tolerance to biotic and abiotic stresses has been significant . Improvements in crowding stress tolerance have enabled the use of increasingly higher plant population density (‘plant density’) over the last 80 years . Therefore yield of today’s hybrids are in part the result of high plant density made possible by crowding stress tolerance.
Tolerance to high plant density varies greatly among modern commercial sweet corn hybrids [4, 5]. Evaluation of 26 hybrids showed distinct yield differences when grown at high plant density . Under conditions of crowding stress, high-yielding hybrids demonstrate greater crowding stress tolerance than low-yielding hybrids. Understanding and utilizing the genetic mechanism(s) that sustain yield under high plant density may provide a key target to future sweet corn improvement.
As plant density increases, individual plants experience crowding stress due to resource competition. Plant stress tolerance mechanisms can be defined as biological processes that reduce stress via resource capture and/or utilization . Although plant biological responses to multiple abiotic stresses such as shade, water, and nutrients are critical components to crowding stress tolerance, plant responses to abiotic stress may not always influence yield. Studies found morphological changes of corn plants under crowding stress [6–8]. Increases in physiological processes such as nitrogen use efficiency , leaf photosynthesis rate , and post-flowering source-sink ratio  are possible factors in crowding stress tolerance. However, agronomically meaningful processes have a positive influence in crop yield potential . Although shoot height and canopy leaf area index increase with crowding stress , these plant traits are not always associated with yield improvement. Moreover, manipulating biological processes to improve water-use efficiency are associated with reductions in biomass accumulation [13, 14]. In contrast, yield-related traits such as ear barrenness  or grain number per cob [16, 17] may be more direct indicators to crowding stress tolerance and yield. Identifying stress tolerance mechanisms linked with improving yield-related traits will be critical for agronomic utilization.
Transcriptional profiling has been a useful tool for understanding plant response to various abiotic stress factors in corn [18–20]. However, limited information is available on the relationships among transcriptional and phenotypic responses, particularly yield. Recent advances in transcriptional co-expression network analysis enables exploration of gene expression patterns with phenotypic and physiological traits in higher plants such as corn and Arabidopsis [21, 22]. Exploring transcriptional patterns among hybrids differing in yield under crowding stress, and connecting phenotypic responses with transcriptional responses, could identify mechanisms of crowding stress tolerance.
The goal of this work was to elucidate mechanisms conferring tolerance to crowding stress in sweet corn. Specifically, the objectives were to 1) explore transcriptional changes among sweet corn hybrids with differential yield under crowding stress, 2) identify relationships among phenotypic responses and gene expression patterns, and 3) identify groups of genes associated with yield and crowding stress tolerance.
Materials and Methods
Plant materials and field experiments
Yield data from twenty-six sweet corn hybrids evaluated in 2012 under crowding stress were used to identify crowding stress tolerant and sensitive germplasm . Based on kernel mass per hectare, the three highest-yielding hybrids and three lowest-yielding hybrids were grouped into high- and low-yielding groups, respectively (Table 1). Average yield difference between groups was ~3 Mt per hectare. The high- and low-yielding groups represented crowding stress tolerant and sensitive groups, respectively.
In 2013, a field experiment was conducted at the University of Illinois Vegetable Crop Research Farm near Urbana, IL. The six hybrids were grown at 71,700 plants per hectare, the optimal plant density of the most crowding stress-tolerant hybrid from the previous study and 13,700 plants per hectare above normal . Production practices common to the region, such as tillage and pest control, were used. All plots were fertilized with 202 kg N per hectare prior to planting and sprinkler irrigated as needed to ensure crop establishment and maintain homogeneity of water and fertility throughout the field. The experimental design was a randomized complete block with 4 replications. Each plot consisted of four rows of each hybrid 9 m long with 76 cm spacing between rows. The experiment was planted at 65 plants per row and hand-thinned to 50 plants per row at 3-collar corn to achieve the target plant density. Stand counts were done after thinning and at harvest to confirm plant density.
Plant data were collected from the center two rows of each plot. Thermal time to mid-silk was determined as cumulative growing degree days (GDD) from crop emergence to mid-silk date. Plant height was measured on 6-collar corn (V6) and at silking (R1), taken from the soil surface to uppermost leaf apex. Leaf greenness was measured using a chlorophyll meter (SPAD 502 plus chlorophyll meter, Konica Minolta) for relative chlorophyll content on the mid-length of the oldest emerging leaf at V6 and R1. Leaf area index (LAI) was measured at R1 using a linear ceptometer (LP-80 AccuPAR, Decagon Devices, Pullman, WA). Total leaf nitrogen content was assessed from a sample of six primary ear leaves per plot. All measurements taken from multiple plants per plot were averaged by each plot for statistical analyses.
Approximately 21 days after mid-silk date of the plots, green ears (> 4.5 cm in diameter) were hand-harvested from the center two rows in 6.1 m length. Ear traits which are indirectly related with yield were measured on green ears. Ear length and filled ear length were measured from five random ears per plot and fill percentages were calculated by dividing mean filled ear length by mean ear length. Kernel moisture was calculated by the quantity fresh kernel mass minus dried kernel mass divided by fresh kernel mass times 100.
Yield traits were measured on harvested ears and kernels. Twelve random ears per plot were husked (A&K Development, Eugene, OR) and kernels were cut from the cob with an industry-grade hand-fed corn cutter (A&K Development, Eugene, OR). Green ear mass, husked ear mass, cob mass and kernel mass of 12 ear samples were measured. Based on these measurements, yield traits such as ear number per plant, ear number per hectare, ear mass per plant, ear mass per hectare, kernel mass per hectare, and kernel mass per plant was calculated.
Statistical analysis of field data
Phenotypic data from field experiments were analyzed using PROC MIXED in SAS version 9.2 (SAS Institute Inc. Cary, NC, USA). Yield group and hybrids nested within group were considered fixed factors. Replications were considered random factors. Data complied with ANOVA assumptions of homogeneity of variance based on the modified Levene’s test  and normality based on diagnostic test of residuals.
Plant materials and microarray experiment
Plant tissue samples were collected by bulking 4 ear leaves per plot at the R1 growth stage. The reproductive stage of the plant was selected because corn is most susceptible to stress at flowering, when silk growth, pollination, and kernel set occur . Moreover, leaf photosynthesis after anthesis influences biomass accumulation and allocation in corn under crowding stress . Two to three biological replications per hybrid were frozen in liquid nitrogen immediately after removal from the plant and stored at -80 C until RNA extraction.
Total RNA was extracted from 16 samples (H1, H2, L1 and L2 with 3 replications; H3 and L3 with 2 replications) using the RNeasy mini kit (Qiagen, Hilden, North Rhine-Westphalia, Germany). Quantity and quality of total RNA was checked using the Agilent 2100 Bioanalyzer (Agilent Technologies, Inc. USA). A microarray experiment was performed at the Roy J. Carver Biotechnology Center at University of Illinois. The microarray design was based on field corn inbred B73 coding sequences from MaizeGDB (http://www.MaizeGDB.org). A set of gene representations was created by retaining the longest transcript from each gene. A custom microarray was designed using Agilent eArray (Agilent Amadid # 060449). Out of 39,653 coding sequences, 39,091 unique probes were designed. The array contained 39,091 unique probes, of which 34,379 were single-spotted and 4,712 were double-spotted, plus 1,264 positive controls and 153 negative controls. Seventy-five ng of total RNA was labeled using the Agilent 2-color Low Input Quickamp Whole Transcriptome Labeling kit (Agilent Technologies, Santa Clara, CA) according to manufacturer protocol. Labeled samples were hybridized on the custom microarray (4x44K format) and scanned on an Axon 4000B microarray scanner (Molecular Devices, Sunnyvale, CA) at 5 um resolution. GenePix 6.1 image analysis software (Molecular Devices, Sunnyvale, CA) was used for spotfinding.
Statistical analyses of microarray data
Pre-processing and statistical analyses of microarray data were performed in R (version 3.2.1) using the Limma package  (version 3.24.15). Median foreground values from the 8 arrays were read into R and any spots that had been manually flagged (-100 values) were given a weight of zero . The individual Cy5 and Cy3 values were all normalized together using the quantile method and then log2-transformed . Correlations between replicate spots for the double-spotted probes were high; therefore, expression values were averaged to obtain a single value for the double-spotted probes for each sample. The positive and negative control probes were used to assess the minimum expression level considered ‘detectable above background noise’ (6.25 on the log2 scale) and then discarded.
A mixed effects statistical model  was fit on the 39,091 unique probes to estimate the mean expression level for each of the 6 hybrid groups while accounting for the random effect of array . After model fitting, 8,901 probes were discarded because they did not have expression values > 6.25 in at least 2 samples. For ease of discussion, the remaining 30,190 probes will be called genes. Principal component analysis (PCA) was performed on 30,190 genes to examine the group and hybrid effects on overall gene expression.
From the group means model, a contrast of all high-yielding hybrids vs. all low-yielding hybrids was made. Since hybrid effect was significant, pairwise hybrid comparisons were made between each hybrid from high- versus low-yielding groups to identify individual hybrid differences. Fold change (FC) from pairwise comparisons were calculated to show up- or down-regulation of genes between high- and low-yielding hybrids. Raw p-values were adjusted separately for each comparison using the False Discovery Rate (FDR) method . Differentially expressed genes (DEGs) were identified from pairwise comparisons when FDR p<0.05 and |FC|>1.5.
Additionally, a one-way ANOVA across all 6 hybrids was computed to identify genes that changed among the 6 hybrids. From the result, 7,670 genes were identified that had FDR p-value < 0.05 and |FC|>1.5 between any two hybrid comparisons. Weighted Gene Correlation Network Analysis (WGCNA) [31, 32] was performed using these 7,670 genes to cluster genes using a distance metric that separates them into different ‘modules’ which share a consistent expression pattern associated with phenotypic responses. The genes were clustered using WGCNA (version 1.47) package using the default values of the blockwiseModules() function except for: soft thresholding power β = 30, networkType = ‘signed’, minModuleSize = 20, and mergeCutHeight = 0.25. This resulted in 40 modules ranging from 30 to 679 genes, plus the ‘module 0’ consisting of 6 genes that did not fit any of the other patterns. The overall expression pattern of each module can be represented by the eigengene values, which are the first principal component score for each sample from the expression values of the genes in the module. The eigengene values for each module were then subjected to an overall high-yielding versus low-yielding group comparison in order to determine modules related to group differences.
Functional analysis and visualization of groups of DEGs and modules were conducted using AgriGO , REVIGO , Cytoscape (version 3.2.1) , and Mapman software . Over-representation analysis was conducted using Fisher’s exact test based on corn gene ontology (GO) and Mapman annotation. All p-values were adjusted using Benjamini Hochberg FDR method with significance level at 0.05 .
Validation of gene expression using RT-qPCR
Quantitative reverse transcription-polymerase chain reaction (RT-qPCR) was performed to validate microarray results. Four transcripts were selected based on their importance to crowding stress response from this microarray experiment and previous studies (S1 File) [19, 37]. Ubiquitin conjugating enzyme was selected as the endogenous control. Invitrogen Superscript First-Strand Synthesis System (Invitrogen) was used to synthesize cDNA from the same total RNA samples. Primer Express Software Version 3.0 (Applied Biosystems, Foster, CA) was used to design primers. Power SYBR Green Master Mix (Applied Biosystems, Foster, CA) and ABI 7900 real time PCR machine were used for performing RT-qPCR. SDS2.4 software (Applied Biosystems, Foster, CA) was used to analyze the RT-qPCR result and identify threshold values. Three technical replications were used for each sample and averaged for the analysis when PCR efficiency was between 90 and 100% and R2 close to 0.99 . The cycle threshold values were normalized to the expression of control genes and the ∆∆Ct method was used for comparing the gene expression values involved in crowding stress .
Results and Discussion
Phenotypic response under crowding stress
For most yield traits, higher values were observed in the high-yielding group than the low-yielding group (Table 2). For instance, ear mass per hectare and kernel mass per hectare were 26% and 90% higher in the high-yielding group, respectively. Hybrids within a yield group were similar for all yield traits. All hybrids produced less than 1 ear per plant, indicating plants experienced crowding stress.
Effects of yield group and hybrid were significant for most plant and ear traits, but not all (Table 2). Hybrid L1 was the tallest plant at R1 (223.0 cm) and latest maturity (822 GDD to silk) hybrid. In contrast, hybrid L3 was the shortest plant at R1 (161.0 cm) and earliest maturity (693 GDD to silk) hybrid. Of ear traits, H3 had the largest ear length but the smallest fill percentage. Leaf nitrogen and LAI were similar among hybrids, averaging 2.89% and 5.53, respectively (S3 File).
Kernel mass per hectare was positively correlated with all yield traits (ρ ≥ 0.46) and plant height at V6 (ρ = 0.29). Kernel moisture and fill percentage were negatively correlated (ρ≤ -0.52) with kernel mass per hectare (Table 3).
Most phenotypic responses varied among sweet corn hybrids, indicating hybrids were genetically different in their response to crowding stress. In contrast, yield trait differences among hybrids within yield group were consistent (Table 2) indicating common tolerance mechanism(s) may be present in high-yielding hybrids.
Agronomically useful crowding stress tolerance mechanisms should be associated with greater yield potential. Plant responses such as plant height and SPAD were significant at V6, but not significant at R1 between yield groups. Difference in plant height and SPAD at V6 may be important indicators to plant competition during vegetative growth stage. However, no differences at R1 suggest plant height and SPAD lack consistent connections with yield at the flowering stage. Difference in GDD to silk between yield groups was due largely to late maturity in L1 and L2. Hybrid L3 had the earliest maturity. Therefore GDD to silk may not be suitable for discriminating yield groups. Identifying internal genetic mechanisms that explain yield group differences would be critical for better understanding of crowding stress tolerance.
Transcriptional difference between crowding stress tolerant and sensitive groups
The microarray experiment was performed to compare gene expression patterns among hybrids. Results have been submitted in NCBI’s Gene Expression Omnibus database  and are accessible through GEO Series accession number GSE73435 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE73435). The microarray result was validated using RT-qPCR (S1 and S2 Files).
Initial investigation of gene expression patterns showed a strong hybrid effect, yet little similarity within the same yield group. The heatmap showed varied expression patterns across hybrids (S1 Fig). Principal component analysis (PCA) and hierarchical clustering indicated L1 was most different from all other hybrids (S2 Fig). However, there were no clear visual patterns or clusters of transcripts to differentiate yield groups.
Out of 39,091 transcripts assayed, 199 up-regulated and 207 down-regulated DEGs in the high-yielding group were identified from the contrast of all high-yielding hybrids vs. all low-yielding hybrids (Data not shown). Nonetheless, understanding individual hybrid gene expression patterns to crowding stress may be more important than averaging expression across hybrids.
Pairwise comparisons between hybrids were conducted to identify common DEGs. For each hybrid, groups of DEGs were first identified from pairwise comparisons. Common DEGs among three pairwise comparisons of one high-yielding hybrid to all low-yielding hybrids (e.g. H1 vs. L1, L2 and L3) were identified and vice versa. These common DEGs represented unique genes influenced by crowding stress of each high-yielding hybrid. There were 372 (225 up- and 147 down-regulated), 374 (215 up- and 159 down-regulated), and 669 (364 up-and 305 down-regulated) common DEGs in H1, H2 and H3 compared to all low-yielding hybrids, respectively (Fig 1). Likewise 859 (393 up- and 466 down-regulated), 408 (236 up- and 172 down-regulated) and 387(213 up- and 174 down-regulated) DEGs were identified in L1, L2 and L3 compared to all high-yielding hybrids. There were 14 (up-regulated in high-yielding hybrids) and 12 (down-regulated in high-yielding hybrids) common DEGs across all pairwise comparisons, accounting for only 3 to 7% of common DEGs (Fig 1). The small number of common DEGs was identified across all pairwise comparisons implies each hybrid has a unique mechanism(s) in crowding stress response.
a Up- and down- regulated DEGs in the first hybrid from listed pairwise comparison identified at FDR p-value<0.05 and |FC|>1.5 were labeled + and -, respectively.
Functional analysis of common DEGs revealed distinct functional differences among hybrids (Table 4). A previous transcriptional study on density stress on corn seedlings found a small number of genes commonly expressed among different genotypes, indicating genotype is an important factor to crowding stress response . Plants under crowding stress experience growth alterations due to differences in individual plant resource requirement and/or resource utilizing ability . Different genotypes may experience varying levels of crowding tolerance even in similar environmental conditions (i.e. same plant density). Therefore, high-yielding hybrids have genetic mechanisms in tolerating crowding stress and resulting in better kernel and ear mass than low-yielding hybrids. For example, biological functions identified from high-yielding hybrids such as genes involved in steroid biosynthetic process (H1) , fatty acid/carbohydrate metabolic process (H2) [40, 41] or GTP/UTP/CTP biosynthetic process (H3)  may have associations with yield gain (Table 4). In contrast, low-yielding hybrids are more sensitive to crowding stress than high-yielding hybrids that resulted in reduction in kernel and ear mass. Biological functions such as genes involved in sexual reproduction/cell wall organization (L1) , electron transport chain/cytochrome complex assembly (L2) , or apoptotic process/oligopeptide transport (L3)  may be associated with crowding stress sensitivity (Table 4). Collectively, this genomic information indicates that each hybrid has a unique crowding stress response mechanism. Significant biological functions identified in all hybrids may be involved in crowding stress response, but the function identified in high-yielding hybrids will have a positive association with yield. These mechanisms of crowding stress tolerance in high-yielding hybrids may be exploited for further agronomic improvement.
Of 26 DEGs significant among all pairwise hybrid comparisons, 15 DEGs related to known biological functions. Up-regulated DEGs in high-yielding hybrids were involved in amino acid degradation/polyamine metabolism, cell wall degradation, glycolysis cytosolic branch, hormone metabolism (Auxin), RNA transcription regulation, signaling receptor kinases and tetrapyrrole synthesis. Whereas down-regulated DEGs in high-yielding hybrids were involved in flavonoids secondary metabolism, miscellaneous enzyme families, protein folding, protein post-translational modification and development (Table 5). These genes may be important candidate genes linked to crowding stress tolerance.
Connecting phenotypic traits and transcript expression
In addition to individual gene differences, constructing a co-expression network of transcriptome data is increasingly used to understand complex gene interactions and to correlate gene expression patterns to phenotypic traits of crops [21, 22, 46–48]. We performed WGCNA in order to identify relationships among phenotypic responses and gene expression patterns of groups of genes associated with yield and crowding stress tolerance. The eigengene values of 40 modules were tested for group difference, which identified 9 significant modules. These 9 modules consisted of 51 to 264 genes. On average, modules 10, 13, 32, 36 and 38 showed up-regulation and modules 9, 14, 17 and 22 showed down-regulation in the high-yielding group (S3 Fig). Only modules 22 and 36 showed the same expression patterns in all hybrids within a group. Remaining modules showed the same expression in two out of three hybrids in a group.
Correlation analysis among these 9 modules and phenotypic traits showed positive associations of up-regulated modules with yield traits but negative associations of down-regulated modules with yield traits (Table 6). In contrast, most plant and ear traits were not correlated with up-regulated modules. Notably, GDD to silk, fill percentage, and SPAD at V6 were negatively correlated with module 13, 32, and 36, respectively. These three plant and ear traits were positively correlated with down-regulated modules 9, 14 and 22, respectively (Table 6). Hierarchical clustering of modules and phenotypic traits showed two large branches. One cluster consisted of the up-regulated modules and yield traits, and the other cluster contained the down-regulated modules, plant and ear traits (Fig 2). Positive correlation and clustering of up-regulated modules with yield traits also support that these modules have positive genetic effects on yield.
Phenotypic traits were labeled as ‘Height.V6’, height at V6; ‘Ear.No.plant’, number of ears per plant; ‘Ear.No.ha’, number of ears per hectare; ‘Kernel.mass.plant’, kernel mass per plant; ‘Kernel.mass.ha’, kernel mass per hectare; ‘SPAD.R1’, SPAD at R1; ‘Leaf.N’, leaf nitrogen; ‘Fill.percentage’, fill percentage; ‘SPAD.V6’, SPAD at V6; ‘LAI’, leaf area index; ‘Kernel.moisture’, kernel moisture; ‘Height.R1’, height at R1; ‘GDD.silk’, GDD to silk; ‘Ear.length’, ear length; ‘Fill.length’, fill length.
Correlation coefficients with bold text significant at FDR p-value<0.05.
Enrichment analysis identified over-represented GO terms for each module. The most significant GO biological processes or molecular functions identified on up-regulated modules were cellular component organization (module 10), cellular nitrogen compound metabolic process/carbohydrate metabolic process (module 13), transcription regulation (module 32), and RNA splicing (module 36) (Table 7). The most significant GO terms identified on down-regulated modules were transporter activity/binding (module 9), translation (module 17), and metabolic process/Chitin catabolic process (module 22). Biological functions of genes and modules for each hybrid or group in this research provide useful genetic information on yield response to crowding stress.
Biological functions for modules were significant at FDR p-value<0.05.
Photosynthesis related metabolism
Reduction in photosynthate supply after anthesis was a critical cause of ear or kernel abortion under crowding stress [49, 50]. From pairwise comparisons, genes involved in electron transport chain were significant in L2. Most of down-regulated genes in electron transport chain in L2 are involved in light reaction with the highest FC in multiple ferredoxin electron transport related genes such as ferredoxin 1 and 2. Photosynthetic electron transport has an important function in metabolizing energy sources NADPH and ATP. Ferredoxin is a light-sensitive electron carrier  that has an important role in the multiple metabolic pathway regulations and transduction of redox signals into the regulatory network  and significantly associated with linear electron flow and photosynthesis capacity [53, 54]. Down-regulation or knockout of ferredoxin in tobacco, potato and Arabidopsis has resulted in growth inhibition and photosynthesis inactivation [53–56]. Studies also identified ferredoxin plays an important role in tolerance to many stress factors such as water deprivation, oxidative stress, and heat or cold stress . Relative down-regulation of photosynthesis electron transport in L2 to all high-yielding hybrids may be part of a mechanism responsible for low productivity of L2 under crowding stress.
Genes involved in tetrapyrrole biosynthesis, especially in chlorophyll and heme, were up-regulated in all high-yielding hybrids. Heme oxygenase (grmzm2g043277_t02), required for phytochrome chromophore biosynthesis as a plant photoreceptor, was up-regulated in all high-yielding hybrids. Arabidopsis HY1 gene encoding heme oxygenase has been characterized and requires electron transfer from ferredoxin in vivo [57, 58]. These results suggest the critical and close interaction of photosynthesis-related biological processes, specifically electron transport and tetrapyrrole synthesis, may contribute to yield difference under crowding stress. The down-regulation of both ferredoxin and terapyrrole activities in L2 may provide the connection for significantly lower leaf chlorophyll content (SPAD) at R1 in L2 than other hybrids.
Important crowding stress tolerant mechanisms may be plant biological processes influencing yield. Stress tolerance mechanisms have to be evaluated for yield effect since mechanisms can be negatively associated with biomass accumulation [12–14]. Comparing DEGs among hybrids under crowding stress would not capture all crowding stress response genes since some stress response genes might be expressed similarly across all hybrids. For example, plant phytochrome photoreceptor genes such as phyA and phyB are important candidate genes involved in shade avoidance response due to reduction in the ratio of red:far-red light [37, 59]. Due to plant competition for light under crowding stress, these genes also may be involved in crowding stress response. However, five phyA and phyB related genes screened in the present work were expressed similarly across all hybrids. Therefore, although these genes may be involved in stress response mechanisms, they may not be important for crowding stress tolerance mechanisms influencing yield.
Cell wall metabolism
Cell wall metabolism is a complex network responsible for various abiotic stress responses . Transcript grmzm2g151257_t01, known for cellulases/endoglucanase (cell wall degradation), was up-regulated in all high-yielding hybrids. Cellulases are cell-wall-loosening protein genes involved in cell expansion and root elongation in Arabidopsis and believed to be important for drought tolerance  and corn yield . Cell wall degradation activity is often closely regulated by plant hormones. Plant hormones have an important role in plant growth such as regulating the expression of root cell-wall-loosening protein genes [63, 64]. Plant hormone auxin is believed to regulate cellulase activity . Up-regulation of auxin and jasmonate synthesis-related genes along with up-regulation of cellulase-related genes in all high-yielding hybrids suggests their role in crowding stress tolerance. Understanding the relationships of genes with different functions may be valuable for further investigation, because complex traits like corn yield are associated with a large chain of metabolic responses .
Other cell wall degradation and modification-related genes showed mixed expressions among hybrids. For example, many genes in cell wall degradation and modification were significantly down-regulated in L1. Out of 52 genes involved in cell wall degradation in L1, 20 common DEGs were identical or similar to corn, rice or Arabidopsis expansin, a protein involved in cell wall-loosening and cell growth . Expansin showed mixed expression depending on plant species, varying in response to abiotic factors such as temperature, light, and water . Expression of expansin was associated with plant growth and flowering date in rice . Hybrid L1 showed a unique phenotypic response with the tallest plant at R1 and the latest maturity. Significant expression of expansin in L1 may have an influence in the phenotypic response.
Primary and secondary metabolic pathway
Primary and secondary metabolic pathways such as glycolysis metabolism and polyamine synthesis were important in crowding stress tolerance. Glycolysis had a positive association with corn yield . Most enzymes involved in cytosolic branch of glycolysis and mitochondrial electron transport were up-regulated in high-yielding hybrids, and the highest up-regulation was observed on pyrophosphate-fructose-6-P-phsphotransferase (ac205100.3_fgt001). This enzyme is involved in conversion of fructose-6-phosphate to fructose-1,6-bisphosphate. Increased activity of the enzyme has been observed in anaerobiosis condition in rice  and was positively associated with corn yield .
Arginine decarboxylase (grmzm2g374302_t02) was up-regulated in high-yielding hybrids. Arginine decarboxylase in Arabidopsis is a key enzyme involved in the first step of polyamine synthesis, converting arginine to agmatine. The enzyme influences early seed development during embryo growth. Arginine decarboxylase in Arabidopsis showed close relationship with the production of putrescine, one of the most abundant polyamines in plants and animals . Polyamines have critical functions in senescence, cell proliferation, differentiation biotic and abiotic stress response [71–73]. Overexpression of arginine decarboxylase in tobacco and tomato was involved in drought tolerance by reactive oxygen species detoxification . Arginine decarboxylase may be an important key to crowding stress tolerance in sweet corn.
Flavonoids are color pigments in plants, protecting leaf cells from photooxidative damages and improving the efficiency of nutrient retrieval during senescence . Flavonol 3-O-glucosyltransferases is an enzyme involved in anthocyanin biosynthesis by providing sugars through UDP-glucose. Gene (Gdi-15), homologous to flavonol 3-O-glucosyltransferases, in groundnut responded to heat and moisture stress during early growth stage . Three forms of flavonol 3-O-glucosyltransferases were down-regulated in high-yielding hybrids and all belonged to module 22, which was negatively correlated to yield traits and positively correlated to SPAD at V6. This relationship indicates the enzyme may have an important stress tolerance function during early vegetative stage, but may not be associated with yield.
All up-regulated genes identified from pairwise comparisons belonged to either module 13 or 36 (Table 7). By performing WGCNA, additional relationships of these genes with other important biological mechanisms were identified. Module 13 consisted of multiple independent primary metabolisms with a major cluster of nitrogen compound metabolism connected with biological processes such as transcription regulation and homeostatic process. Module 36 related to RNA metabolism and biosynthesis with protein metabolism, phosphorylation, post-translational protein modification and homeostatic process (Fig 3). These complex inter-relationships among biological processes will be useful crowding stress tolerance networks influencing sweet corn crop yield.
Significance (FDR p-values) of GO term is color labeled from red to white. Similar GO terms are linked by edges, where the degree of similarity represented by the line width. (A) Module 13, (B) Module 36.
Module 10 and 32 were clusters of genes not identified from pairwise comparisons, because expression of genes showed up-regulation only in H2 and H3, and in H1 and H3, respectively (S3 Fig). Yet, strong up-regulation of these modules in two of three high-yielding hybrids may indicate an important role in crowding stress tolerance. Module 10 has an over-represented function of cellular component organization. The plant genome is packaged by coiling to histone, creating the nucleosome, and further condensed to chromatin. Genome packaging is a crucial system for gene expression and very responsive to stress factors such as heat shock and oxidative stress by acting as a transcript regulator [77–81]. Chemical modifications, such as histone acetylation or methylation, directly affect chromatin structure changes by loosening or condensing the structure to alter gene transcription. Since chromatin modification is influenced by abiotic stress and modification impacts gene expression, the role of chromatin modification in stress response has been discussed extensively [82, 83]. For example, acetylation of histone H4 increased in response to cold, high salinity, and ABA application in Arabidopsis and tobacco cells . Expression of histone protein, H2B and H4, was involved in module 10 indicating up-regulation of these genes might be involved in crowding stress tolerance in sweet corn.
The most over-represented biological function in module 32 was transcription regulation. Two bZIP transcription factors, bZIP71 and bZIP118, belonged to module 32. Expression of bZIP71 in rice was associated with drought, polyethylene glycol, salinity, and ABA applications . Auxin responsive protein IAA17 also belonged to module 32. Salt stress condition reduced root elongation by increasing IAA17 stabilization and repressing auxin signaling in Arabidopsis . These transcriptional factors may interact closely to confer crowding stress tolerance in sweet corn. One of Arabidopsis bZIP family, bZIP11, was associated with histone acetylation to influence the expression of auxin responsive genes such as IAA proteins . Therefore, the interaction among these transcriptional factors in module 10 and 32 would be an important crowding stress tolerance mechanism in sweet corn. Further investigation on exploiting the interaction may be helpful.
Sweet corn productivity can be improved with greater utilization of crowding stress tolerance. The most useful mechanisms of crowding stress tolerance will be those influencing crop yield. Genomic information from this research identified several potential crowding stress tolerance mechanisms linked to yield improvement. Gene expression patterns revealed each high-yielding hybrid had unique mechanisms to tolerate crowding stress. However, we also identified networks of related genes associated with crowding stress tolerance. Specifically, genes involved in biological functions such as photosynthesis, glycolysis, cell wall, carbohydrate/nitrogen metabolic process, chromatin and transcription regulation related processes were identified as possible mechanisms of crowding stress tolerance. Crowding stress tolerance genes were inter-connected with many biological functions, indicating complex nature of crowding stress tolerance mechanisms and the importance of improving mechanisms as a network.
S1 Fig. Heatmap of 7,670 genes that had a one-way ANOVA FDR p-value<0.05 and at least 1.5 FC between any two hybrid comparisons.
S2 Fig. PCA plot of hybrids in respect to PC1 and PC2 using normalized probe expression values.
S3 Fig. Bar plots of mean eigengene values (+/- 1 SEM) of each hybrid for the modules with significant expression difference between yield groups.
S1 File. Genes and primers used for RT-qPCR validation.
S2 File. Microarray result and RT-qPCR validation of selected transcripts.
S3 File. Leaf nitrogen and LAI of sweet corn hybrids and yield groups measured in a field experiment near Urbana, IL in 2013.
Authors would like to thank Dr. Mark Band at Roy J. Carver Biotechnology Center at University of Illinois for technical support on conducing microarray experiment. Authors also appreciate Jim Moody, Nick Hausman, Laura Crawford, Joseph Kibiwott, and undergraduate students for proving technical support and assistance on the field experiment. Mention of a trademark, proprietary product, or vendor does not constitute a guarantee or warranty of the product by the U.S. Dept. of Agriculture and does not imply its approval to the exclusion of other products or vendors that also may be suitable.
Conceived and designed the experiments: EC MMW. Performed the experiments: EC. Analyzed the data: EC JD. Contributed reagents/materials/analysis tools: EC JD MMW. Wrote the paper: EC.
- 1. Duvick DN. What is yield? In: Edmeades GO, Bänziger B, Mickelson HR, Pena-Valdivia CB, Editors. Developing drought and low N-tolerant maize. Mexico: CIMMYT, El Batan; 1997. pp. 332–335.
- 2. Tollenaar M, Lee EA. Yield potential, yield stability and stress tolerance in maize. Field Crops Res. 2002;75: 161–169. doi: 10.1016/s0378-4290(02)00024-2
- 3. Duvick DN. Genetic progress in yield of United States maize (Zea mays L.). Maydica. 2005;50: 193–202.
- 4. Williams MMII. Agronomics and economics of plant population density on processing sweet corn. Field Crops Res. 2012;128: 55–61. doi: 10.1016/j.fcr.2011.12.007
- 5. Williams MMII. Identifying crowding stress-tolerant hybrids in processing sweet corn. Agron J. 2015;107: 1782–1788. doi: 10.2134/agronj15.0011
- 6. Cox WJ. Whole-plant physiological and yield responses of maize to plant density. Agron J. 1996;88: 489–496. doi: 10.2134/agronj1996.00021962008800030022x
- 7. Hashemi A, Herbert S, Putnam D. Yield response of corn to crowding stress. Agron J. 2005;97: 839–846. doi: 10.2134/agronj2003.0241
- 8. Mansfield BD, Mumm RH. Survey of plant density tolerance in US maize germplasm. Crop Sci. 2014;54: 157–173. doi: 10.2135/cropsci2013.04.0252
- 9. McCullough DE, Aguilera A, Tollenaar M. N-uptake, N-partitioning, and photosynthetic N-use efficiency of an old and a new maize hybrid. Can J Plant Sci. 1994;74: 479–484. doi: 10.4141/cjps94-088
- 10. Dwyer LM, Tollenaar M, Stewart DW. Changes in plant-density dependence of leaf photosynthesis of maize (Zea-Mays L) hybrids, 1959 to 1988. Can J Plant Sci. 1991;71: 1–11. doi: 10.4141/cjps91-001
- 11. Borras L, Otegui ME. Maize kernel weight response to postflowering source-sink ratio. Crop Sci. 2001;41: 1816–1822. doi: 10.2135/cropsci2001.1816
- 12. Tardieu F, Tuberosa R. Dissection and modelling of abiotic stress tolerance in plants. Curr Opin Plant Biol. 2010;13: 206–212. doi: 10.1016/j.pbi.2009.12.012. pmid:20097596
- 13. Tardieu F. Virtual plants: modelling as a tool for the genomics of tolerance to water deficit. Trends Plant Sci. 2003;8: 9–14. pmid:12523994 doi: 10.1016/s1360-1385(02)00008-0
- 14. Blum A. Effective use of water (EUW) and not water-use efficiency (WUE) is the target of crop yield improvement under drought stress. Field Crops Res. 2009;112: 119–123. doi: 10.1016/j.fcr.2009.03.009
- 15. Bunting ES. Plant density and yield of grain maize in England. J Agric Sci. 1973;81: 455–463. doi: 10.1017/s0021859600086512
- 16. Baenziger PS, Glover DV. Effect of reducing plant-population on yield and kernel characteristics of Sugary-2 and normal maize. Crop Sci. 1980;20: 444–447. doi: 10.2135/cropsci1980.0011183x002000040005x
- 17. Karlen DL, Camp CR. Row spacing, plant-population, and water management effects on corn in the Atlantic coastal-plain. Agron J. 1985;77: 393–398. doi: 10.2134/agronj1985.00021962007700030010x
- 18. Humbert S, Subedi S, Cohn J, Zeng B, Bi Y, Chen X, et al. Genome-wide expression profiling of maize in response to individual and combined water and nitrogen stresses. BMC Genomics. 2013;14: 3. doi: 10.1186/1471-2164-14-3. pmid:23324127
- 19. St Pierre S, Springer NM, Muehlbauer GJ. Density stress has minimal impacts on the barley or maize seedling transcriptome. Plant Genome. 2011;4: 47–54. doi: 10.3835/plantgenome2010.08.0020
- 20. Hayano-Kanashiro C, Calderon-Vazquez C, Ibarra-Laclette E, Herrera-Estrella L, Simpson J. Analysis of gene expression and physiological responses in three Mexican maize landraces under drought stress and recovery irrigation. PLOS One. 2009;4: e7531. doi: 10.1371/journal.pone.0007531. pmid:19888455
- 21. Downs GS, Bi Y, Colasanti J, Wu W, Chen X, Zhu T, et al. A developmental transcriptional network for maize defines coexpression modules. Plant Physiol. 2013;161: 1830–1843. doi: 10.1104/pp.112.213231. pmid:23388120
- 22. Weston DJ, Gunter LE, Rogers A, Wullschleger SD. Connecting genes, coexpression modules, and molecular signatures to environmental stress phenotypes in plants. BMC Syst Biol. 2008;2: 16. doi: 10.1186/1752-0509-2-16. pmid:18248680
- 23. Neter J, Kutner MH, Nachtsheim CJ, Wasserman W. Applied linear statistical models. 4th ed. Chicago: Irwin; 1996.
- 24. Shaw RH. Water use and requirements of maize—a review. World Meteorological Organization: Agrometeorology of the maize (corn) crop. 1977; 119–134.
- 25. Allison JCS. Effect of plant population on production and distribution of dry matter in maize. Ann Appl Biol. 1969;63: 135–144. doi: 10.1111/j.1744-7348.1969.tb05474.x
- 26. Smyth GK. Limma: linear models for microarray data. In: Gentleman R, Carey V, Dudoit S, Irizarry R, Huber W, editors. Bioinformatics and Computational Biology Solutions using R and Bioconductor. New York: Springer; 2005. pp. 397–420.
- 27. Smyth GK, Speed TP. Normalization of cDNA microarray data. Methods. 2003;31:265–273. pmid:14597310 doi: 10.1016/s1046-2023(03)00155-5
- 28. Smyth GK. Linear models and empirical bayes methods for assessing differential expression in microarray experiments. Stat Appl Genet Mol. 2004;3:1–25. doi: 10.2202/1544-6115.1027
- 29. Smyth GK, Michaud J, Scott HS. Use of within-array replicate spots for assessing differential expression in microarray experiments. Bioinformatics. 2005;21: 2067–2075. pmid:15657102 doi: 10.1093/bioinformatics/bti270
- 30. Benjamini Y, Hochberg Y. Controlling the false discovery rate—a practical and powerful approach to multiple testing. J R Stat Soc Series B Stat Methodol. 1995;57: 289–300.
- 31. Zhang B, Horvath S. A general framework for weighted gene co-expression network analysis. Stat Appl Genet Mol. 2005;4: 17. doi: 10.2202/1544-6115.1128
- 32. Langfelder P, Horvath S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics. 2008;9: 559. doi: 10.1186/1471-2105-9-559. pmid:19114008
- 33. Du Z, Zhou X, Ling Y, Zhang Z, Su Z. agriGO: a GO analysis toolkit for the agricultural community. Nucleic Acids Res. 2010;38: W64–W70. doi: 10.1093/nar/gkq310. pmid:20435677
- 34. Supek F, Bosnjak M, Skunca N, Smuc T. REVIGO Summarizes and Visualizes Long Lists of Gene Ontology Terms. PLOS One. 2011;6: e21800. doi: 10.1371/journal.pone.0021800. pmid:21789182
- 35. Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, et al. Cytoscape: A software environment for integrated models of biomolecular interaction networks. Genome Res. 2003;13: 2498–2504. pmid:14597658 doi: 10.1101/gr.1239303
- 36. Usadel B, Poree F, Nagel A, Lohse M, Czedik-Eysenberg A, Stitt M. A guide to using MapMan to visualize and compare Omics data in plants: a case study in the crop species, Maize. Plant Cell Environ. 2009;32: 1211–1229. doi: 10.1111/j.1365-3040.2009.01978.x. pmid:19389052
- 37. Sheehan MJ, Farmer PR, Brutnell TP. Structure and expression of maize phytochrome family homeologs. Genetics. 2004;167: 1395–1405. pmid:15280251 doi: 10.1534/genetics.103.026096
- 38. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(T)(-Delta Delta C) method. Methods. 2001;25: 402–408. pmid:11846609 doi: 10.1006/meth.2001.1262
- 39. Edgar R, Domrachev M, Lash AE. Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res. 2002;30: 207–210. pmid:11752295 doi: 10.1093/nar/30.1.207
- 40. Divi UK, Krishna P. Brassinosteroid: a biotechnological target for enhancing crop yield and stress tolerance. N Biotechnol. 2009;26: 131–136. doi: 10.1016/j.nbt.2009.07.006. pmid:19631770
- 41. Schraudner M, Ernst D, Langebartels C, Sandermann H. Biochemical-plant responses to ozone .3. Activation of the defense-related proteins beta-1,3-glucanase and chitinase in tobacco-leaves. Plant Physiol. 1992;99: 1321–1328. pmid:16669039 doi: 10.1104/pp.99.4.1321
- 42. Smith AM. Prospects for increasing starch and sucrose yields for bioethanol production. Plant J. 2008;54: 546–558. doi: 10.1111/j.1365-313X.2008.03468.x. pmid:18476862
- 43. Geilfus C, Zoerb C, Muehling KH. Salt stress differentially affects growth-mediating beta-expansins in resistant and sensitive maize (Zea mays L.). Plant Physiol. Biochem. 2010;48: 993–998. doi: 10.1016/j.plaphy.2010.09.011. pmid:20970350
- 44. Tognetti VB, Palatnik JF, Fillat MF, Melzer M, Hajirezaei M, Valle EM, et al. Functional replacement of ferredoxin by a cyanobacterial flavodoxin in tobacco confers broad-range stress tolerance. Plant Cell. 2006;18: 2035–2050. pmid:16829589 doi: 10.1105/tpc.106.042424
- 45. Doyle SM, Diamond M, McCabe PF. Chloroplast and reactive oxygen species involvement in apoptotic-like programmed cell death in Arabidopsis suspension cultures. J Exp Bot. 2010;61: 473–482. doi: 10.1093/jxb/erp320. pmid:19933317
- 46. Ficklin SP, Luo F, Feltus FA. The association of multiple interacting genes with specific phenotypes in rice using gene coexpression networks. Plant Physiol. 2010;154: 13–24. doi: 10.1104/pp.110.159459. pmid:20668062
- 47. Ficklin SP, Feltus FA. Gene coexpression network alignment and conservation of gene modules between two grass species: Maize and Rice. Plant Physiol. 2011;156: 1244–1256. doi: 10.1104/pp.111.173047. pmid:21606319
- 48. Schlueter U, Colmsee C, Scholz U, Braeutigam A, Weber APM, Zellerhoff N, et al. Adaptation of maize source leaf metabolism to stress related disturbances in carbon, nitrogen and phosphorus balance. BMC Genomics. 2013;14: 442. doi: 10.1186/1471-2164-14-442. pmid:23822863
- 49. Iremiren G, Milbourn G. Effects of plant-density on ear barrenness in maize. Exp Agric. 1980;16: 321–326. doi: 10.1017/s001447970001108x
- 50. Boyle MG, Boyer JS, Morgan PW. Stem infusion of liquid culture-medium prevents reproductive failure of maize at low water potential. Crop Sci. 1991;31: 1246–1252. doi: 10.2135/cropsci1991.0011183x003100050033x
- 51. Haslett BG, Cammack R, Whatley FR. Quantitative studies on ferredoxin in greening bean-leaves. Biochem J. 1973;136: 697–703. pmid:4360717 doi: 10.1042/bj1360697
- 52. Scheibe R, Dietz K. Reduction-oxidation network for flexible adjustment of cellular metabolism in photoautotrophic cells. Plant Cell Environ. 2012;35: 202–216. doi: 10.1111/j.1365-3040.2011.02319.x. pmid:21410714
- 53. Hanke GT, Hase T. Variable photosynthetic roles of two leaf-type ferredoxins in Arabidopsis, as revealed by RNA interference. Photochem Photobiol. 2008;84: 1302–1309. doi: 10.1111/j.1751-1097.2008.00411.x. pmid:18673322
- 54. Voss I, Koelmann M, Wojtera J, Holtgrefe S, Kitzmann C, Backhausen JE, et al. Knockout of major leaf ferredoxin reveals new redox-regulatory adaptations in Arabidopsis thaliana. Physiol Plantarum. 2008;133: 584–598. doi: 10.1111/j.1399-3054.2008.01112.x
- 55. Holtgrefe S, Bader KP, Horton P, Scheibe R, von Schaewen A, Backhausen JE. Decreased content of leaf ferredoxin changes electron distribution and limits photosynthesis in transgenic potato plants. Plant Physiol. 2003;133: 1768–1778. pmid:14645726 doi: 10.1104/pp.103.026013
- 56. Blanco NE, Ceccoli RD, Segretin ME, Poli HO, Voss I, Melzer M, et al. Cyanobacterial flavodoxin complements ferredoxin deficiency in knocked-down transgenic tobacco plants. 2011;65: 922–935.
- 57. Muramoto T, Kohchi T, Yokota A, Hwang IH, Goodman HM. The Arabidopsis photomorphogenic mutant hy1 is deficient in phytochrome chromophore biosynthesis as a result of a mutation in a plastid heme oxygenase. Plant Cell. 1999;11: 335–347. pmid:10072395 doi: 10.2307/3870864
- 58. Muramoto T, Tsurui N, Terry MJ, Yokota A, Kohchi T. Expression and biochemical properties of a ferredoxin-dependent heme oxygenase required for phytochrome chromophore synthesis. Plant Physiol. 2002;130: 1958–1966. pmid:12481078 doi: 10.1104/pp.008128
- 59. Sawers RJH, Sheehan MJ, Brutnell TP. Cereal phytochromes: targets of selection, targets for manipulation? Trends Plant Sci. 2005;10: 138–143. pmid:15749472 doi: 10.1016/j.tplants.2005.01.004
- 60. Hl Gall, Philippe F, Domon JM, Gillet F, Pelloux J, Rayon C. Cell wall metabolism in response to abiotic stress. Plants. 2015;4: 112–166. doi: 10.3390/plants4010112
- 61. Xu P, Cai X, Wang Y, Xing L, Chen Q, Xiang C. HDG11 upregulates cell-wall-loosening protein genes to promote root elongation in Arabidopsis. J Exp Bot. 2014;65: 4285–4295. doi: 10.1093/jxb/eru202. pmid:24821957
- 62. Fu J, Thiemann A, Schrag TA, Melchinger AE, Scholten S, Frisch M. Dissecting grain yield pathways and their interactions with grain dry matter content by a two-step correlation approach with maize seedling transcriptome. BMC Plant Biol. 2010;10: 63. doi: 10.1186/1471-2229-10-63. pmid:20385002
- 63. Jan A, Yang GX, Nakamura H, Ichikawa H, Kitano H, Matsuoka M, et al. Characterization of a xyloglucan endotransglucosylase gene that is up-regulated by gibberellin in rice. Plant Physiol. 2004;136: 3670–3681. pmid:15516498 doi: 10.1104/pp.104.052274
- 64. Zhao M, Han Y, Feng Y, Li F, Wang W. Expansins are involved in cell growth mediated by abscisic acid and indole-3-acetic acid under drought stress in wheat. Plant Cell Rep. 2012;31: 671–685. doi: 10.1007/s00299-011-1185-9. pmid:22076248
- 65. Fry SC. Cellulases, Hemicelluloses and auxin-stimulated growth—a possible relationship. Physiol Plantarum. 1989;75: 532–536. doi: 10.1111/j.1399-3054.1989.tb05620.x
- 66. Campos H, Cooper A, Habben JE, Edmeades GO, Schussler JR. Improving drought tolerance in maize: a view from industry. Field Crops Res. 2004;90: 19–34. doi: 10.1016/j.fcr.2004.07.003
- 67. Li ZC, Durachko DM, Cosgrove DJ. An oat coleoptile wall protein that induces wall extension in-vitro and that is antigenically related to a similar protein from cucumber hypocotyls. Planta. 1993;191: 349–356. doi: 10.1007/bf00195692
- 68. Choi DS, Lee Y, Cho HT, Kende H. Regulation of expansin gene expression affects growth and development in transgenic rice plants. Plant Cell. 2003;15: 1386–1398. pmid:12782731 doi: 10.1105/tpc.011965
- 69. Mertens E, Larondelle Y, Hers HG. Induction of pyrophosphate—fructose 6-phosphate 1-phosphotransferase by anoxia in rice seedlings. Plant Physiol. 1990;93: 584–587. pmid:16667507 doi: 10.1104/pp.93.2.584
- 70. Urano K, Yoshiba Y, Nanjo T, Ito T, Yamaguchi-Shinozaki K, Shinozaki K. Arabidopsis stress-inducible gene for arginine decarboxylase AtADC2 is required for accumulation of putrescine in salt tolerance. Biochem Biophys Res Commun. 2004;313: 369–375. pmid:14684170 doi: 10.1016/j.bbrc.2003.11.119
- 71. Kumar A, Altabella T, Taylor MA, Tiburcio AF. Recent advances in polyamine research. Trends Plant Sci. 1997;2: 124–130. doi: 10.1016/s1360-1385(97)01013-3
- 72. Walden R, Cordeiro A, Tiburcio AF. Polyamines: Small molecules triggering pathways in plant growth and development. Plant Physiol. 1997;113: 1009–1013. pmid:9112764 doi: 10.1104/pp.113.4.1009
- 73. Kakkar RK, Sawhney VK. Polyamine research in plants—a changing perspective. Physiol Plantarum. 2002;116: 281–292. doi: 10.1034/j.1399-3054.2002.1160302.x
- 74. Wang B, Zhang Q, Liu J, Li G. Overexpression of PtADC confers enhanced dehydration and drought tolerance in transgenic tobacco and tomato: Effect on ROS elimination. Biochem Biophys Res Commun. 2011;413: 10–16. doi: 10.1016/j.bbrc.2011.08.015. pmid:21871871
- 75. Feild TS, Lee DW, Holbrook NM. Why leaves turn red in autumn. The role of anthocyanins in senescing leaves of red-osier dogwood. Plant Physiol. 2001;127: 566–574. pmid:11598230 doi: 10.1104/pp.010063
- 76. Gopalakrishna R, Kumar G, KrishnaPrasad BT, Mathew MK, Kumar MU. A stress-responsive gene from groundnut, Gdi-15, is homologous to flavonol 3-O-glucosyltransferase involved in anthocyanin biosynthesis. Biochem Biophys Res Commun. 2001;284: 574–579. pmid:11396938 doi: 10.1006/bbrc.2001.4992
- 77. Smith KT, Workman JL. Chromatin proteins: Key responders to stress. PLOS Biol. 2012;10: e1001371. doi: 10.1371/journal.pbio.1001371. pmid:22859908
- 78. Huebert DJ, Kuan P, Keles S, Gasch AP. Dynamic changes in nucleosome occupancy are not predictive of gene expression dynamics but are linked to transcription and chromatin regulators. Mol Cell Biol. 2012;32: 1645–1653. doi: 10.1128/MCB.06170-11. pmid:22354995
- 79. Zanton SJ, Pugh BF. Full and partial genome-wide assembly and disassembly of the yeast transcription machinery in response to heat shock. Genes Dev. 2006;20: 2250–2265. pmid:16912275 doi: 10.1101/gad.1437506
- 80. Rando OJ, Winston F. Chromatin and transcription in yeast. Genetics. 2012;190: 351–387. doi: 10.1534/genetics.111.132266. pmid:22345607
- 81. Zaret KS, Carroll JS. Pioneer transcription factors: establishing competence for gene expression. Genes Dev. 2011;25: 2227–2241. doi: 10.1101/gad.176826.111. pmid:22056668
- 82. Gutzat R, Scheid OM. Epigenetic responses to stress: triple defense? Curr Opin Plant Biol. 2012;15: 568–573. doi: 10.1016/j.pbi.2012.08.007. pmid:22960026
- 83. Kouzarides T. Chromatin modifications and their function. Cell. 2007;128: 693–705. pmid:17320507 doi: 10.1016/j.cell.2007.02.005
- 84. Sokol A, Kwiatkowska A, Jerzmanowski A, Prymakowska-Bosak M. Up-regulation of stress-inducible genes in tobacco and Arabidopsis cells in response to abiotic stresses and ABA treatment correlates with dynamic changes in histone H3 and H4 modifications. Planta. 2007;227: 245–254. pmid:17721787 doi: 10.1007/s00425-007-0612-1
- 85. Liu C, Mao B, Ou S, Wang W, Liu L, Wu Y, et al. OsbZIP71, a bZIP transcription factor, confers salinity and drought tolerance in rice. Plant Mol Biol. 2014;84: 19–36. doi: 10.1007/s11103-013-0115-3. pmid:23918260
- 86. Liu W, Li R, Han T, Cai W, Fu Z, Lu Y. Salt stress reduces root meristem size by nitric oxide-mediated modulation of auxin accumulation and signaling in Arabidopsis. Plant Physiol. 2015;168: 343–U607. doi: 10.1104/pp.15.00030. pmid:25818700
- 87. Weiste C, Droege-Laser W. The Arabidopsis transcription factor bZIP11 activates auxin-mediated transcription by recruiting the histone acetylation machinery. Nat Commun. 2014;5: 3883. doi: 10.1038/ncomms4883. pmid:24861440