Expression and Distribution of the Auxin Response Factors in Sorghum bicolor During Development and Temperature Stress

Auxin response factor (ARF) is a transcription factor that can specifically bind to the promoter of auxin-responsive genes in plants and plays an important regulatory role in plant growth and development. The previous studies have predicted 25 ARF genes in Sorghum bicolor (SbARFs) and indicated that SbARFs play complex roles in salt and drought stresses. In this study, we reclassified and analyzed the structures of ARFs in three plants, including sorghum, rice, and Arabidopsis. Phylogenetic analyses categorized 73 ARF into five classes. By studying the characterization of the structures, it was found that SbARFs from the same evolutionary branches showed similar motif patterns. Furthermore, the expression patterns of SbARF genes during development and temperature stress were investigated in sorghum. Quantitative transcription-quantitative polymerase chain reaction (qRT-PCR) results suggested that they had different expression patterns in vegetative and reproductive organs at various developmental stages. High and low-temperature treatments and qRT-PCR demonstrated some of them changed dramatically along with the increase of treatment time. Additionally, in situ hybridization results displayed that SbARF genes were accumulated in vascular tissues under temperature stress. These findings provide evidence that SbARFs may play important roles in sorghum vegetative development, reproductive development, and auxin response to temperature stress.


Introduction
Auxin is known to be involved in the regulation of growth and development and abiotic stress response. As an important family of proteins in auxin-mediated response, ARF can recognize the auxin-response elements (AuxREs) in the promoters of auxin-responsive genes and then activate or repress their expression [1,2]. Most ARF proteins consist of three parts: An N-terminal B3-type DNA binding domain (DBD) that may recognize the AuxREs, a variable middle region that may play as an activation domain (AD) or repression domain (RD), and two C-terminal domains (CTD: domain III/IV). The CTD is involved in protein-protein interactions by dimerizing with auxin/indole-3-acetic acid (Aux/IAA) family genes or even dimerizing between ARFs [1,3,4].
According to the characteristic of molecular structures, ARFs have been identified from different plant species, and their roles in auxin-related plant growth and development have been studied intensively. The first ARFwas identified from Arabidopsis (AtARF1) and, subsequently, 23 ARF members were found in Arabidopsis [5]. Since then, based on the genome-wide analysis, ARF gene family has been identified in more than 20 species of higher plants, such as 25 ARFs from rice [6], 36 ARFs from

Identification of SbARF Genes
The previous studies have predicted 25 ARF genes in Sorghum bicolor and divided into four distinct classes. To verify the SbARF genes, BLASTP approaches were employed for the mining of all putative SbARF members in the Sorghum bicolor genome. Finally, a total of 25 potential SbARF genes were identified which have the same sequence ID as described before (Table 1). Further analysis of the 25 SbARF genes revealed that the ORF length of this family ranged between 1557 (SbARF16) and 3432 (SbARF7/25) bp, with a mean length of 2389 bp, while the transcript length ranged between 1869 (SbARF2) and 3836 (SbARF7) bp, with a mean length of 2763 bp. The predicted polypeptide length ranged between 518 (SbARF16) and 1143 (SbARF7/25) amino acids, with a mean length of 3 of 15 795 amino acids. The isoelectric point of the ARF proteins ranged between 5.7 (SbARF8) and 8.7 (SbARF2), with a mean value of 6.8. The predicted molecular mass of the proteins ranged between 56 (SbARF16) and 127 (SbARF7) kDa, with a mean molecular mass of 88 kDa.
A phylogenetic tree without roots was generated based on multiple alignments of SbARFs with their orthologs from rice and Arabidopsis ( Figure 1). All 73 ARFs (including 25 SbARFs, 25 OsARFs, 23 AtARFs) fell into five broad groups: Class I, II, III, IV, and V, which contain 23, 25, 10, 11, and four members, respectively. Obviously, most ARFs of the three species were clustered together in Class I and II while Class V contained the fewer members of ARFs. Thus, the 25 SbARFs could also be divided into five separate clusters as well. Class I included SbARF (25), Oryza sativa (25), and Arabidopsis (23) were clustered into five classes (I-V). Class I, II, III, IV, and V are marked with red, fuchsia, green, yellow, and blue lines, respectively.

Gene Structure and Protein Structure of SbARF Genes
To better understanding, the structure of SbARF genes, their intron/exon number, positions, and functional domains were analyzed. As shown in Table 1 and Figure 2, the number of exons in DNA sequences varied from 3 to 16, while the number of coding exons varied from 2 to 14. All SbARFs had two characteristic regions: a highly conserved region corresponding to the B3-type DNA-bind domain (DBD) in the N-terminal portion, and a middle region that functions as activation or repression domain (MR). Thirteen of which contained the C-terminal Aux/IAA domains (CTD). Nineteen of which contained the type I/II Phox and Bem1p protein-protein interaction domains (PB1). In general, the SbARF genes from the same clusters shared a similar structure. For example, polypeptide length of Class I was the longest, with a mean length of 981 amino acids. Members of Class I and Class II had 13 or 14 coding exons. ARF genes from other classes contained fewer exons  (25), and Arabidopsis (23) were clustered into five classes (I-V). Class I, II, III, IV, and V are marked with red, fuchsia, green, yellow, and blue lines, respectively.

Gene Structure and Protein Structure of SbARF Genes
To better understanding, the structure of SbARF genes, their intron/exon number, positions, and functional domains were analyzed. As shown in Table 1 and Figure 2, the number of exons in DNA sequences varied from 3 to 16, while the number of coding exons varied from 2 to 14. All SbARFs had two characteristic regions: a highly conserved region corresponding to the B3-type DNA-bind domain (DBD) in the N-terminal portion, and a middle region that functions as activation or repression domain (MR). Thirteen of which contained the C-terminal Aux/IAA domains (CTD). Nineteen of which contained the type I/II Phox and Bem1p protein-protein interaction domains (PB1). In general, the SbARF genes from the same clusters shared a similar structure. For example, polypeptide length of Class I was the longest, with a mean length of 981 amino acids. Members of Class I and Class II had 13 or 14 coding exons. ARF genes from other classes contained fewer exons and encoded proteins of shorter lengths. All proteins from Class I had four characteristic regions (DBD, MR, CTD, PB1), while proteins from Class III, IV, V had only two characteristic regions (DBD, MR). A conserved motif analysis of 25 SbARF protein sequences using the MEME online tool predicted 20 motifs ( Figure 3). The results displayed, the closer they were in the evolution tree, the more similar location of the motifs would be. Motif 1 was the most conserved motifs with an e-value of 9.6E-1367, and 23 SbARFs had this motif. Twenty-five SbARFs had motif 2, while this motif was also very conservative with the e-value of 7.2 × 10 −544 (Table S3). Besides these, 24 SbARFs contained motifs 3/4/5/8/9/10. Three characteristic domains (DBD, MR, CTD) of SbARFs were divided into 10 motifs. Motif 1 and 9 constituted the DBD. The MR domain consisted of motifs 2, 4, 8, and 10. The CTD corresponded to motifs 6, 7, and 11. Additionally, all the SbARFs were predicted to be nuclear localized using UniProt online analysis tool and subCELlular LOcalization predictor. A conserved motif analysis of 25 SbARF protein sequences using the MEME online tool predicted 20 motifs (Figure 3). The results displayed, the closer they were in the evolution tree, the more similar location of the motifs would be. Motif 1 was the most conserved motifs with an e-value of 9.6E-1367, and 23 SbARFs had this motif. Twenty-five SbARFs had motif 2, while this motif was also very conservative with the e-value of 7.2 × 10 −544 (Table S3). Besides these, 24 SbARFs contained motifs 3/4/5/8/9/10. Three characteristic domains (DBD, MR, CTD) of SbARFs were divided into 10 motifs. Motif 1 and 9 constituted the DBD. The MR domain consisted of motifs 2, 4, 8, and 10. The CTD corresponded to motifs 6, 7, and 11. Additionally, all the SbARFs were predicted to be nuclear localized using UniProt online analysis tool and subCELlular LOcalization predictor. 5 also very conservative with the e-value of 7.2 × 10 −544 (Table S3). Besides these, 24 SbARFs contained motifs 3/4/5/8/9/10. Three characteristic domains (DBD, MR, CTD) of SbARFs were divided into 10 motifs. Motif 1 and 9 constituted the DBD. The MR domain consisted of motifs 2, 4, 8, and 10. The CTD corresponded to motifs 6, 7, and 11. Additionally, all the SbARFs were predicted to be nuclear localized using UniProt online analysis tool and subCELlular LOcalization predictor.  Table S3. The length of proteins can be estimated using the scale at the bottom. The clusters' information is marked on the left.  Table S3. The length of proteins can be estimated using the scale at the bottom. The clusters' information is marked on the left.

Expression of SbARF Genes in Response to Temperature Stress
As transcription factors, ARFs could modulate auxin signaling pathway, and then controlled a wide array of plant behaviors including developmental processes and abiotic stress response. To further determine the role of SbARF genes in temperature stress response, expression characteristics of SbARF genes were investigated in seedling under 4 • C and 40 • C treatment. As shown in Figure 5 and Figure S1, the expressions of 9 SbARF genes (SbARF4/7/9/15/17/19/21/22/24) were always up-regulated after cold and heat stress, while SbARF13 remained down-regulated during the whole treatments. Three SbARF genes (SbARF1/3/16) showed low expression levels during 4 • C treatment, but 9 SbARF genes (SbARF5/6/8/10/11/12/18/20/25) displayed different expression patterns. For example, SbARF11 and SbARF25 were significantly down-regulated after 1 h of 4 • C treatment and then were rapidly up-regulated after a 3 h treatment. Eleven SbARF genes (SbARF1/5/6/8/10/11/12/16/18/20/25) showed high expression levels during 40 • C treatment, but SbARF3 was obviously suppressed after 1 h treatment and then slightly accumulated after 6 h treatment. It was worth mentioning that the expression of the SbARF23 gene was obvious changed not only at 4 • C but also at 40 • C treatments.
Most of the SbARF genes from Class I and Class III displayed transcript accumulation after temperature stress. Instead, the expression levels of SbARF1/10/13 (Class IV) and SbARF16 (Class V) were significantly suppressed after treatment. Furthermore, SbARF13 was the only gene that always maintained a low level. 7 then were rapidly up-regulated after a 3 h treatment. Eleven SbARF genes (SbARF1/5/6/8/10/11/12/16/18/20/25) showed high expression levels during 40 °C treatment, but SbARF3 was obviously suppressed after 1 h treatment and then slightly accumulated after 6 h treatment. It was worth mentioning that the expression of the SbARF23 gene was obvious changed not only at 4 °C but also at 40 °C treatments.
Most of the SbARF genes from Class I and Class III displayed transcript accumulation after temperature stress. Instead, the expression levels of SbARF1/10/13 (Class IV) and SbARF16 (Class V) were significantly suppressed after treatment. Furthermore, SbARF13 was the only gene that always maintained a low level.

Distribution Patterns of the SbARF Genes under Temperature Stress
To investigate the tissue site of auxin signal response under temperature stress in Sorghum bicolor, in situ hybridization analysis was used to detect the distribution patterns of SbARF genes in seedling leaves under temperature stress. According to the significant expression change of SbARFs, SbARF3/15/17/24 genes were selected for in situ hybridization analysis. As shown in Figure 6, these SbARF mRNAs displayed similar accumulation sites in the leaves. The hybridization signals were accumulated in the vascular bundles of the main and secondary veins, especially in the vascular bundle sheath cells and the surrounding parenchyma cells. The other cells were almost non-detected signals. Although the stain signals were weak after cold stress, they followed a similar distribution pattern no matter heat stress or cold stress.  To investigate the tissue site of auxin signal response under temperature stress in Sorghum bicolor, in situ hybridization analysis was used to detect the distribution patterns of SbARF genes in seedling leaves under temperature stress. According to the significant expression change of SbARFs, SbARF3/15/17/24 genes were selected for in situ hybridization analysis. As shown in Figure 6, these SbARF mRNAs displayed similar accumulation sites in the leaves. The hybridization signals were accumulated in the vascular bundles of the main and secondary veins, especially in the vascular bundle sheath cells and the surrounding parenchyma cells. The other cells were almost non-detected signals. Although the stain signals were weak after cold stress, they followed a similar distribution pattern no matter heat stress or cold stress.

Discussion
As a worldwide food and feed cereal crop, sorghum has the characteristics of rapid growth, high-temperature tolerance, and drought tolerance. Here, a total of 25 potential SbARF genes were identified, which have the same sequence ID as described before (Table 1). There was the same number of SbARF genes as ARF genes encoded in Oryza sativa (25) [6], but more than Arabidopsis (23) [5]. The previous studies have also predicted 25 ARF genes in Sorghum bicolor [34]. It is worth

Discussion
As a worldwide food and feed cereal crop, sorghum has the characteristics of rapid growth, high-temperature tolerance, and drought tolerance. Here, a total of 25 potential SbARF genes were identified, which have the same sequence ID as described before (Table 1). There was the same number of SbARF genes as ARF genes encoded in Oryza sativa (25) [6], but more than Arabidopsis (23) [5]. The previous studies have also predicted 25 ARF genes in Sorghum bicolor [34]. It is worth mentioning that effective bioinformatics methods were used to further analyze the sequence feature in our study, such as Ensembl Plants, MEGA X, MEME, SMART, InterPro, and so on.
We found that 73 ARFs (25 SbARFs, 25 OsARFs, and 23 AtARFs) could be categorized into five classes in phylogenetic analyses (Figure 1). Sorghum is an ideal C 4 grass model and can be a good complement to C 3 rice for biofuel crops, animal feed, and forage. In the evolutionary tree of ARFs, there were also showed a more close relationship between monocotyledon Sorghum bicolor and Oryza sativa. Furthermore, the protein typical domains and conserved motifs analysis showed that SbARFs from the same class might have similar regions (Figures 2 and 3). The typical N-terminal B3-type DNA binding domain (DBD) of ARF protein may recognize and bind to AuxREs in the promoters of auxin response genes [35]. In Sorghum bicolor, all 25 SbARFs had this typical DBD domain, and very conservative. As showing in Figure 3, 23 of those DBD domains were constituted by motif 1 and 9. Only ARF2 and ARF13 (members of Class IV) were different because their DBD domains were constituted by motif 9 or motif 2, respectively. Besides the DBD domain, 25 SbARFs also had an Auxin_resp middle region (MR), and most of the MR regions were constituted by motif 2, 4, 8, and 10. Only SbARF16 was different because its MR region was constituted by motif 2. The MR region may act as an activation domain (AD) or repression domain (RD). Although there may be still some confusion on the mechanisms of activation and repression, the ADs are found to be enriched in glutamine (Q), while the RDs are enriched in serine (S), proline (P), and threonine (T) [10,35,36]. Interestingly, motifs 2, 4, 8, and 10 in SbARFs contained amino acids S, T, and P, but not Q. The amino acid composition of MR may contribute to further SbARF functional annotation.
C-terminal domain (CTD) was the third characteristic domain in ARF proteins. It has consisted of a dimerization domain (domain III/IV) and involved in protein-protein interactions by dimerizing with proteins encoded by genes from Aux/IAA family or even dimerizing between ARFs [1,3,4]. However, only 13 of 25 SbARFs (members of Class I and II) contain the typical AUX_IAA domains, while 19 of 25 SbARFs contain a type I/II Phox and Bem1p protein-protein interaction domain (PB1) ( Table 1 and Figure 2). Furthermore, The PB1 domains overlapped with the AUX_IAA domains in those 13 SbARFs. Recently researches suggest that PB1 domain can interact with other PB1 via electrostatic interactions, and provides positive and negative interfaces. The underlying protein interaction between ARFs and Aux/IAA also contains a PB1 domain [35,36].
The transcript expression profiles can help us screen for candidate SbARF genes with potentially distinct functions. The transcriptional abundance of the SbARF genes varied greatly in different vegetative organs and reproductive organs (Figure 2), suggesting that the SbARF genes had multiple functions in sorghum growth and development. According to the evolutionary tree analysis, Class II members (SbARF11/18, SbARF9/12) with the highest homology showed the most similar expression patterns, followed by Class I (SbARF7/25, SbARF8/23), and other classes had lower similarity, indicating potential functional redundancy in Class I and Class II members. Eleven of 25 SbARF genes were accumulated in the roots than other tissues, with five (SbARF7/8/17/19/23) in Class I and four (SbARF6/9/11/18) in Class II, SbARF5 in Class III and SbARF24 in Class IV. Seven of 25 SbARF genes were accumulated in the leaves, with SbARF15 in class I, SbARF3/12 in Class II, SbARF20/21 in Class III, SbARF13 in Class IV, SbARF16 in Class V. SbARF4 (Class III) and SbARF25(Class I) genes were highly induced in the mature inflorescence. SbARF10 gene (Class IV) was remarkably increased in the post-flowering inflorescence. SbARF14 and SbARF22 genes (both members of Class I) were highly expressed in the seeds than other tissues. SbARFs shared remarkable homology with OsARFs as showed in Figure 1. It has been reported that the transgenic antisense-OsARF1 rice grew slowly and leaves were short and curly [30], while its sorghum homolog, SbARF3, was prominently higher expressed in the leaves, implying its possible function in vegetative development. The knockout mutant of OsARF12 and double mutant of OsARF12/25 display unusual growth of seedling, especially in root growth [37,38]. Thus, their sorghum homologs SbARF15 and SbARF19 genes may be important for vegetative growth in sorghum, along with obviously higher accumulation in both leaves and roots, but lower in reproductive organs. However, different from its rice homolog OsARF16 gene, which is required for auxin responses in the roots [39], the expression of the SbARF22 gene peaked in the mature seeds, and then in the roots. The evidence of the similarities and differences between the transcription levels may imply the evolution and divergence of Sorghum bicolor and Oryza sativa in some way.
Sorghum is a typical C 4 grass model crop adapted to heat stress [33]. Auxin is well-known for its numerous functions in plant morphogenesis and plant tolerance to stresses, including heat stress and cold stress [31,32]. It was reported that ARF10 and ARF17 genes are induced in anthers of heat stress-insensitive cotton and suppressed in heat stress-sensitive cotton [31]. In rice, microarray and quantitative PCR analyses find that most OsARF genes are suppressed by cold and heat stresses, while OsARF4/14/18/19 and OsARF11/13/16 are induced by cold and heat stresses, respectively [32]. It's worth mentioning that their sorghum homologs SbARF6/24/25 and SbARF16/22 also were significantly induced by cold and heat stresses, respectively. Furthermore, SbARF genes with the highest homology in Class I, Class II, and Class III showed similar accumulation pattern in normal growth and development, except members in Class IV and Class V. In temperature stress, members of Class I and Class III showed similar accumulation pattern, while some members of Class II and Class IV had different expression patterns. The finding also supported the potential functional redundancy in Class I members and no only that, those genes (such as members in Class I/II/IV/V) that the expression levels changed significantly may have special roles in temperature stress. Clearly, most of SbARF genes were induced by heat stress, which possibly indicated the participation of auxin in the high temperature resistant of sorghum.
Since ARFs are important proteins in auxin-mediated response, which effect the activation of Aux/IAA and PIN, the in situ hybridization might point out the auxin response site in tissues under temperature stress. We found that no matter heat or cold stress, the distribution patterns of SbARF mRNAs were similar in the seedling leaves ( Figure 6). They mainly expressed in the vascular bundle sheath cells and the surrounding parenchyma cells. The distribution pattern of vascular tissue accumulation is consistent with the accumulation of PINs and auxin in the vascular strands of Arabidopsis [40][41][42][43]. It is reported that in the loss-of-function and gain-of-function AtARF5 mutation, not only PIN1 expression is changed, but also the vascular tissue formation is abnormal [22,23]. Auxin efflux carrier PIN3/4/6/7/8 can be also detected in the vascular tissues of Arabidopsis, a green signal of PIN3-GFP presents primarily in the xylem parenchyma cells, PIN4/7 expressed in the cambium, PIN6 expressed in the cambial cells adjacent to the xylem, and PIN8 expressed in the cambial cells near PIN6 [41]. It is hypothesized that these SbARF genes have participated in auxin response to temperature stress in the vascular tissues.

Identification and Phylogenetic Analysis of SbARF Genes
The Sorghum bicolor genome was downloaded from the Resources for Plant Comparative Genomics (PlantGDB; Available online: http://www.plantgdb.org/). The BLASTP method was adopted to search for the SbARF genes and to determine the maximum number of genes. The sequences of 25 Oryza sativa ARF and 23 Arabidopsis ARF genes were downloaded from the Plant Transcription Factor Database (PlantTFDB; Available online: http://planttfdb.cbi.pku.edu.cn/) [5,6]. These ARFs were used as bait to identify SbARF, via a BLASTP search at the score value of ≥100 and e-value ≤1 × 10 −5 . Then all identified potential ARFs were further validated and analyzed by Ensembl Plants (available online: http://plants.ensembl.org/) and PFAM database (available online: https://www.pfam.xfam.org/). Finally, 25 SbARFs genes were identified for further analysis. The protein sequences were investigated by multiple alignment analyses using Clustal W. The resultant sequence alignment then served as input data for a neighbor-joining phylogenetic tree constructed using the MEGA X software [44] with a bootstrap analysis of 1000 replicates.

Gene Structure and Protein Structure Analysis
The mRNA sequences of SbARF genes were obtained from PlantTFDB. The DNA sequence information and protein information were predicted online using Ensembl Plants and SMART (available online: http://smart.embl-heidelberg.de/). The conserved motifs were predicted by the MEME website (available online: http://alternate.meme-suite.org/tools/meme) [45] with the following parameters: optimum motif width was 6 to 200, and the maximum number of motifs was 20. The InterPro online program (available online: https://www.ebi.ac.uk/interpro/) was used to further analyze and annotate the protein sequences. To predict the function and subcellular location of SbARFs, the UniProt online analysis tool (available online: https://www.uniprot.org/) and CELLO v.2.5: subCELlular LOcalization predictor (available online: http://cello.life.nctu.edu.tw/) were used.

Plant Materials and Treatment Methods
The tissues and organs from the adult Sorghum bicolor used for gene expression analysis were collected in the experiment field at Xidian University (Xi'an, China). Leaves (L), roots (R), stems (St), immature inflorescence (one week before flowering, F1), mature inflorescence (flowers at full bloom, F2), post-flowering inflorescence (one week after flowering, F3), and mature seeds (Se) were collected from seven-week-old plants. The samples were performed according to the previous description [46]. For the temperature stress treatments, sorghum seeds were grown in Murashige and Skoog medium in an illumination incubator at 24 ± 2 • C with 16-h light/8-h dark cycles. Fourteen-day-old seedlings were treated with 40 • C as heat stress or 4 • C as cold stress for 1 h, 3 h, and 6 h, and then were collected immediately. Each experiment was repeated three times and all the materials were stored at −80 • C until RNA extraction.

Quantitative RT-PCR Analysis
Total RNA was extracted using RNAiso Plus (TaKaRa, Tokyo, Japan) according to the manufacturer's instructions, cDNA was synthesized using the Reverse Transcription System (TaKaRa, Tokyo, Japan). Quantitative RT-PCR (qRT-PCR) analysis was performed in a 7300 Real-Time System (ABI, NewYork, American) using the SYBR Green RealMasterMix (TIANGEN, Beijing, China). The qRT-PCR primers were designed by Primer 5 software (Table S1). The housekeeping gene, Actinβ, was used as internal control. The quantification of each sample of cDNA was performed in triplicate, each PCR was replicated three times for verification, and the 2 −∆∆CT method was used to analyze the relative changes in gene expression from the qPCR experiments.

Paraffin Section of Leaves and in Situ Hybridization
For fine localization of SbARF mRNA signals in leaves under temperature stress, paraffin section and in situ hybridization analysis were used after 4 • C and 40 • C treatments. The 14-day old seedlings of Sorghum bicolor were obtained after 1 h and 6 h temperature stress treatments. The leaf tissues were collected and fixed with FAA (70% alcohol: glacial acetic acid: formalin = 18:1:1, made with diethylpyrocarbonate-treated water) for 1 h. Then the samples were dehydrated and embedded paraffin. The leaf sections were cut at a 10 mm thickness under a rotary microtome (Leica, Wetzlar, Germany) mounted on poly-Lys-coated glass slides. The gene-specific probe was amplified with the primers Probe-SbARF-F and Probe-SbARF-R (Table S2), and the mouse anti-digoxigenin-labeled alkaline phosphatase (anti-DIG-AP) was added dropwise. Pretreatment, hybridization and immunological detection of sections were performed as described [47].

Conclusions
The systematic characterization of the ARF gene family in sorghum has revealed key features about their structures and the potential functions of the SbARF gene family in vegetative and reproductive organs. High and low-temperature treatments and qRT-PCR demonstrated some SbARF genes changed dramatically along with the increase of treatment time. Additionally, in situ hybridization results indicated the location and distribution of SbARF genes under temperature stress. Our findings provide evidence for future understanding of ARFs in sorghum development and temperature stress.