Strigolactone Signaling Genes Showing Differential Expression Patterns in Arabidopsis max Mutants

Strigolactone (SL) is a recently discovered class of phytohormone that inhibits shoot branching. The molecular mechanism underlying SL biosynthesis, perception, and signal transduction is vital to the plant branching phenotype. Some aspects of their biosynthesis, perception, and signaling include the role of four MORE AXILLARY GROWTH genes, MAX3, MAX4, MAX1, and MAX2. It is important to identify downstream genes that are involved in SL signaling. To achieve this, we studied the genomic aspects of the strigolactone biosynthesis pathway using microarray analysis of four max mutants. We identified SL signaling candidate genes that showed differential expression patterns in max mutants. More specifically, 1-AMINOCYCLOPROPANE-1-CARBOXYLATE SYNTHASE 4 (ACC4) and PROTEIN KINASE 3 (PKS3) displayed contrasting expression patterns, indicating a regulatory mechanism in SL signaling pathway to control different phenotypes apart from branching phenotype.


Introduction
Strigolactones (SLs) are known for stimulating seed germination in species of the genera Orobanche, Phelipanche, and Striga [1][2][3]. Thus, the release of SLs from roots into the rhizosphere under nutrient-deprived conditions is the first response for the symbiotic association with host arbuscular mycorrhizal fungi to boost its capacity to uptake sugars, minerals, and water [1,4]. This process validates the importance of SLs in root exudates as a chemical signal that initiates germination and induces branching to promote the pre-symbiotic growth of these fungi for access to phosphorus, water, and other minerals, and in return provides photosynthetic products [1,[5][6][7]. During phosphate starvation, plants release a large amount of SLs [8,9].

Microarray Analysis Revealed Differentially Expressed Genes (DEGs) in max Mutants and the Role of Those Genes in the Biological Process
Transcriptomic analysis was performed on four max mutant plants and compared with wildtype plants, using the Roche NimbleGen microarray system [46]. The experiment was performed on each independent biological sample. Preprocessed data were exported to the Limma package in R language [47]. An adjusted P-value was estimated using empirical Bayesian statistics implemented using eBayes. Significant DEGs were identified as those with log2 FC (fold-change) > 1 for upregulation and log2 FC (fold-change) < -1 for down-regulation with an adjusted p-value < 0.05. Five comparisons were performed with: wild type-control versus wild type-control to normalize the expression within the control group; max1, max2, max3, and max4 mutants versus wild-type to determine differentially expressed transcripts in the max mutants ( Figure S1). Among the 39,042 genes identified in the NimbleGen array, 12,664 genes were differentially expressed (with p-value < 0.05) in the max mutants versus wild-type control comparison, corresponding to 34.1% of the total genes ( Figure S1; Table S2). It is comprised of 6951 (18.7%) up-regulated genes and 5713 (15.4%) down-regulated genes ( Figure 2A). In the max3-9 mutant, a total of 400 genes were differentially expressed, and among those genes, 46 were down-regulated, and 354 were up-regulated. In the max4-1 mutant, a total of 93 genes were differentially expressed, and among those genes, 34 were downregulated, and 59 were up-regulated. Surprisingly in the max1-1 mutant, total 12,012 genes were differentially expressed, and among those genes 5550 were down-regulated, and 6462 were upregulated ( Figure 2A). The reason for a large number of genes differentially expressed in max1 mutant might be due to the accumulation of a large amount of carlactone (CL) [48]. Since CL acts as a signal molecule, it might invoke the expression of many genes [43]. Finally, in the max2-1 mutant, a total of 159 genes were differentially expressed, and among those genes, 83 were down-regulated, and 76 were up-regulated ( Figure 2A). Venn diagram explains commonly expressed genes between mutant

Microarray Analysis Revealed Differentially Expressed Genes (DEGs) in max Mutants and the Role of Those Genes in the Biological Process
Transcriptomic analysis was performed on four max mutant plants and compared with wild-type plants, using the Roche NimbleGen microarray system [46]. The experiment was performed on each independent biological sample. Preprocessed data were exported to the Limma package in R language [47]. An adjusted P-value was estimated using empirical Bayesian statistics implemented using eBayes. Significant DEGs were identified as those with log2 FC (fold-change) > 1 for up-regulation and log2 FC (fold-change) < -1 for down-regulation with an adjusted p-value < 0.05. Five comparisons were performed with: wild type-control versus wild type-control to normalize the expression within the control group; max1, max2, max3, and max4 mutants versus wild-type to determine differentially expressed transcripts in the max mutants ( Figure S1). Among the 39,042 genes identified in the NimbleGen array, 12,664 genes were differentially expressed (with p-value < 0.05) in the max mutants versus wild-type control comparison, corresponding to 34.1% of the total genes ( Figure S1; Table S2). It is comprised of 6951 (18.7%) up-regulated genes and 5713 (15.4%) down-regulated genes ( Figure 2A). In the max3-9 mutant, a total of 400 genes were differentially expressed, and among those genes, 46 were down-regulated, and 354 were up-regulated. In the max4-1 mutant, a total of 93 genes were differentially expressed, and among those genes, 34 were down-regulated, and 59 were up-regulated. Surprisingly in the max1-1 mutant, total 12,012 genes were differentially expressed, and among those genes 5550 were down-regulated, and 6462 were up-regulated ( Figure 2A). The reason for a large number of genes differentially expressed in max1 mutant might be due to the accumulation of a large amount of carlactone (CL) [48]. Since CL acts as a signal molecule, it might invoke the expression of many genes [43]. Finally, in the max2-1 mutant, a total of 159 genes were differentially expressed, and among those genes, 83 were down-regulated, and 76 were up-regulated ( Figure 2A). Venn diagram explains commonly expressed genes between mutant lines ( Figure 2B). Twenty-nine genes were commonly expressed in all four mutants (Table S3). Sixty-three genes were commonly expressed between max1 and max2 mutants (Table S4). Five genes were commonly expressed in the max3 and max2 mutants (Table S5). Three-hundred-and-four genes were commonly expressed in the max3 and max1 mutants (Table S6). Twenty-two genes were commonly expressed in the max4 and max1 mutants (Table S7). In max3, max1, and max2 mutants, the number of genes differentially expressed were 1, 11, 516, and 5, respectively (Tables S8-S10). The max4 mutant did not show any max4-specific differentially expressed genes ( Figure 2B).
Differentially down-or up-regulated genes enriched in GO terms were functionally classified in the max mutants ( Figure 2C). Terms containing up-regulated genes in max3 were GO:0090627 (plant epidermal cell differentiation) and GO:0010411 (xyloglucan metabolic process). GO:0022611 (dormancy process) displayed significantly down-regulated genes. Marginally down-regulated genes were found in max4 and are associated with GO:0022611 (dormancy process) and GO:0050896 (response to stimulus). In max1, GO:0015851 (nucleobase transport) was significantly enriched while many genes were significantly down-regulated in GO terms. In max2, enrichment was the same as max3, and many biological terms were down-regulated, but the degree of extent was less than those of max1. lines ( Figure 2B). Twenty-nine genes were commonly expressed in all four mutants (Table S3). Sixtythree genes were commonly expressed between max1 and max2 mutants (Table S4). Five genes were commonly expressed in the max3 and max2 mutants (Table S5). Three-hundred-and-four genes were commonly expressed in the max3 and max1 mutants (Table S6). Twenty-two genes were commonly expressed in the max4 and max1 mutants (Table S7). In max3, max1, and max2 mutants, the number of genes differentially expressed were 1, 11, 516, and 5, respectively (Tables S8-S10). The max4 mutant did not show any max4-specific differentially expressed genes ( Figure 2B). Differentially down-or up-regulated genes enriched in GO terms were functionally classified in the max mutants ( Figure 2C). Terms containing up-regulated genes in max3 were GO:0090627 (plant epidermal cell differentiation) and GO:0010411 (xyloglucan metabolic process). GO:0022611 (dormancy process) displayed significantly down-regulated genes. Marginally down-regulated genes were found in max4 and are associated with GO:0022611 (dormancy process) and GO:0050896 (response to stimulus). In max1, GO:0015851 (nucleobase transport) was significantly enriched while many genes were significantly down-regulated in GO terms. In max2, enrichment was the same as max3, and many biological terms were down-regulated, but the degree of extent was less than those of max1.

Strigolactone Metabolism-Related Genes Showed Differential Expression Pattern in the max Mutant
A close look at the expression levels of known SLs biosynthesis and signaling related genes in max mutant plants revealed a more significant role of MAX2 (Figure 3). The expression of most of the SLs signaling related genes were induced or repressed in the max2 mutant ( Figure 3D). There was some exception in other max mutants where few SL signaling-related genes showed differential expression patterns, such as repression of SUPPRESSOR OF MAX2 LIKE 8 (SMXL8) in the max3 mutant, increased expression of MAX1 and LOB1 in the max4 mutant, and repression of SUPPRESSOR OF MAX2 LIKE 6 (SMXL6) in the max1 mutant. However, it was the max2 mutant that showed significant differential expression of the SLs signaling related genes. LATERAL ORGAN BOUNDARIES (LOB1) negatively regulates brassinosteroid accumulation to limit growth in organ boundaries [49]. LOB1 expression was up-regulated in all max mutants except max1. It might be due to the absence of MAX genes; there is no controlled regulation of LOB1 for the growth limit. In our

Strigolactone Metabolism-Related Genes Showed Differential Expression Pattern in the max Mutant
A close look at the expression levels of known SLs biosynthesis and signaling related genes in max mutant plants revealed a more significant role of MAX2 (Figure 3). The expression of most of the SLs signaling related genes were induced or repressed in the max2 mutant ( Figure 3D). There was some exception in other max mutants where few SL signaling-related genes showed differential expression patterns, such as repression of SUPPRESSOR OF MAX2 LIKE 8 (SMXL8) in the max3 mutant, increased expression of MAX1 and LOB1 in the max4 mutant, and repression of SUPPRESSOR OF MAX2 LIKE 6 (SMXL6) in the max1 mutant. However, it was the max2 mutant that showed significant differential expression of the SLs signaling related genes. LATERAL ORGAN BOUNDARIES (LOB1) negatively regulates brassinosteroid accumulation to limit growth in organ boundaries [49]. LOB1 expression was up-regulated in all max mutants except max1. It might be due to the absence of MAX genes; there is no controlled regulation of LOB1 for the growth limit. In our analysis, SUPPRESSOR OF MAX2 1 (SMAX1), SMAX-LIKE2 (SMXL2) and SMAX-LIKE8 (SMXL8) expressions were significantly down-regulated in the max2 mutant, which makes sense since there is no MAX2 in the mutant and SMAXLs may not be able to be regulated by MAX2. This result is in correlation with the previously published result where increased branching phenotype of max2 rescued in the line where SMAX1, SMXL2, and SMXL8 were constitutively targeted by artificial microRNAs [50].

Expression Analysis Revealed Candidate Genes Involved in Downstream SL Pathway
We focused on the highly differentially regulated genes in the max mutant, and we narrowed down to four genes. For the validation of microarray data, real-time PCR analysis was performed, and specific primers were designed for it (Table S1). Three genes, namely, 1-AMINOCYCLOPROPANE-1-CARBOXYLATE SYNTHASE 4 (ACC4), LEUCINE-RICH REPEAT/EXTENSIN 1 (LRX1), and LANOSTEROL SYNTHASE 1 (LAS1) gene, showed induced expression, whereas one gene, PROTEIN KINASE 3 (PKS3) expression was down-regulated in the max mutant (Figure 4). These genes reportedly played a different set of role such as the expression of the ACC4, a key regulatory enzyme in the biosynthesis of the plant hormone ethylene repressed by ABA INSENSITIVE4 (ABI4) by directly binding to their promoters [51]. Additionally, atacc4 showed reduced branching, increased plant height, and hypocotyl length, which is precisely antagonistic to the phenotype of max mutant [52][53][54]. It also explains the cooperation between strigolactone signaling and ACC4. For root hair elongation in Arabidopsis, it is also reported that ethylene interacts with strigolactone [15]. PKS3 is a phytochrome kinase substrate-like protein related to control of hypocotyl growth orientation [55], and max2 hypocotyls were also more elongated compared with wild-type [56]. It indicated that to control hypocotyl growth, and there is some working mechanism going on between SL signaling and PKS3. Similarly, lrx1 have short root hair and might be regulated by MAX2 since its expression is high in max2 [57]. LAS1 is involved in the lanosterol biosynthesis, but las1 mutant did not display any phenotype [58,59]. There is no report of any relation between LAS1 and MAX. It is the first time we have reported that expression of LAS1 increase in max mutant and further Relative expression of SL metabolism-related genes in max3-9, max4-1, max1-1, and max2-1 mutant respectively compared with wild-type based on microarray results. Leaf of the mature plant was used to isolate RNA. Each data point represents the mean (±SE) from 2 different plants. The star indicates a significant difference between mutant and wild-type. * p < 0.05, ** p < 0.05, *** p < 0.001 (Two-way ANOVA test).

Expression Analysis Revealed Candidate Genes Involved in Downstream SL Pathway
We focused on the highly differentially regulated genes in the max mutant, and we narrowed down to four genes. For the validation of microarray data, real-time PCR analysis was performed, and specific primers were designed for it (Table S1). Three genes, namely, 1-AMINOCYCLOPROPANE-1-CARBOXYLATE SYNTHASE 4 (ACC4), LEUCINE-RICH REPEAT/EXTENSIN 1 (LRX1), and LANOSTEROL SYNTHASE 1 (LAS1) gene, showed induced expression, whereas one gene, PROTEIN KINASE 3 (PKS3) expression was down-regulated in the max mutant ( Figure 4). These genes reportedly played a different set of role such as the expression of the ACC4, a key regulatory enzyme in the biosynthesis of the plant hormone ethylene repressed by ABA INSENSITIVE4 (ABI4) by directly binding to their promoters [51]. Additionally, atacc4 showed reduced branching, increased plant height, and hypocotyl length, which is precisely antagonistic to the phenotype of max mutant [52][53][54]. It also explains the cooperation between strigolactone signaling and ACC4. For root hair elongation in Arabidopsis, it is also reported that ethylene interacts with strigolactone [15]. PKS3 is a phytochrome kinase substrate-like protein related to control of hypocotyl growth orientation [55], and max2 hypocotyls were also more elongated compared with wild-type [56]. It indicated that to control hypocotyl growth, and there is some working mechanism going on between SL signaling and PKS3. Similarly, lrx1 have short root hair and might be regulated by MAX2 since its expression is high in max2 [57]. LAS1 is involved in the lanosterol biosynthesis, but las1 mutant did not display any phenotype [58,59]. There is no report of any relation between LAS1 and MAX. It is the first time we have reported that expression of LAS1 increase in max mutant and further study needed to confirm if LAS1 is functionally involved in SL signaling. Based on our findings, we proposed a MAX dependent model ( Figure 5). study needed to confirm if LAS1 is functionally involved in SL signaling. Based on our findings, we proposed a MAX dependent model ( Figure 5).

Microarray Hybridization and Data Analysis
The 135k microarray was prepared from 39,042 genes (TAIR v.9). Four, sixty-nt long probes per gene were outlined. NimbleGen Inc. (Madison, WI, USA) manufactured the microarray (http: //www.nimblegen.com/). Random 40,000 GC probes were included to assist with overlaying the grid on the image for the monitoring of the hybridization efficiency and four corner fiducial controls (225). We repeated the experiment 3 times and independently prepared total RNA from the leaf of the mature plant for the reproducibility of microarray analysis [61]. Ten micrograms of Cy3-labeled target DNA fragments were used for microarray hybridization. For hybridization, an MAUI chamber (Biomicro, Salt Lake, UT, USA) was used at 42 • C for 16-18 h. The data were processed and normalized by cubic spline normalization using quantiles to adjust signal variations between chips [62]. For producing call files, a probe-level summarization by Robust Multi-Chip Analysis (RMA) using a median polish algorithm was implemented in NimbleScan. Limma package in the R computing environment was used for multiple analysis [47]. Genes having p-value below 0.05 or false discovery were collected. Multivariate statistical tests such as clustering, multidimensional scaling, and principal component analysis were performed by Acuity 3.1 (Axon Instruments, Union City, CA, USA) [63].
Biological term enrichment was assessed using GoMiner TM [64], as described previously [65]. The microarray contained 39,042 genes; these genes were used as the total gene set in GoMiner. Among the 39,042 genes, 545 were associated with biological processes, 96 with molecular functions, and 68 with cellular components. Top terms in each clad were selected to reduce redundancy, and hierarchical clustering was performed with the Bioconductor R package 3.9 (http://master.bioconductor.org/). GO terms with FDR values less than 0.05 were considered as significant. For color presentation, up-regulated terms were scaled 0.5 to 5 and indicated as red code in Figure 2C. Similarly, down-regulated terms were scaled -0.5 to -5 and indicated as blue code [66]. Other non-significant FDR values were transformed to 0 with Perl scripting language [65].

Quantitative Real-Time RT-PCR Analysis
Total RNA from plant leaf was isolated using the RNAeasy Mini Kit-Qiagen (Hilden, Germany). Isolated RNA was treated with DNAse I (Takara Bio, Shiga, Japan) to avoid contamination with genomic DNA. RNA was reverse transcribed using a Reverse Transcription kit (Takara Bio, Shiga, Japan). Two micrograms of RNA were used for cDNA synthesis according to manual instructions. Ten-fold diluted cDNA was used as a template for real-time quantitative PCR using SYBR Mix (Roche Diagnostics). Gene-specific forward and reverse primers were used, five µM of each. Gene-specific qRT-PCR primers were designed using NCBI primer blast tool and synthesized (Bioneer, Daejeon, Korea). The primers are listed in Supplementary Table S1. The genes were amplified using a StepOnePlus Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). For normalizing expression levels of a constitutively expressed gene, EIF4A1 was used as a reference gene. The thermal profile used consisted of an initial denaturation step at 95 • C for 5 min, followed by 40 cycles of 95 • C for 10 s, annealing for 15 s, and 72 • C for 15 s followed by one cycle of 95 • C for 10 s.

Conclusions
SL signaling-related knowledge has been rapidly increasing. Recently, a study indicated that a transcription factor, IDEAL PLANT ARCHITECTURE1 (IPA1), is directly involved to SL signaling and regulates shoot branching [67] (Figure 5). However, this pathway still lacks basic answer, particularly, the molecular mechanism which identifies its downstream genes, and those genes are the direct or indirect target. Our study identifies DEGs that may act downstream of the SL signaling network for various functions.
Supplementary Materials: The following are available online at http://www.mdpi.com/2223-7747/8/9/352/s1, Figure S1: Comparison of signal distribution between biological repeats of samples, Figure S2: Hierarchical clustering of significantly changed genes of expression max mutants compared with wild-type and the log2 ratio used for normalizing the expression difference, Table S1: List of primers used in this study for qRT-PCR analysis, Table S2: The list of the genes that significantly increased or decreased in max mutants compared to wild-type, Table S3: Commonly expresses genes in max3, max4, max1, and max2, Table S4: Commonly expressed genes in max1 and max2, Table S5: Commonly expressed genes in max3 and max2, Table S6: Commonly expressed genes in max3 and max1, Table S7: Commonly expressed genes in max4 and max1, Table S8: A list of max3-specific changed gene, Table S9: A list of max1-specific changed genes, Table S10: A list of max2-specific changed genes.