Genome-wide characterization and expression analyses of the auxin/indole-3-acetic acid (Aux/IAA) gene family in barley (Hordeum vulgare L.)

Aux/IAA genes are early auxin-responsive genes and essential for auxin signaling transduction. There is little information about Aux/IAAs in the agriculturally important cereal, barley. Using in silico method, we identified and subsequently characterized 36 Aux/IAAs from the barley genome. Based on their genomic sequences and the phylogenic relationship with Arabidopsis and rice Aux/IAA, the 36 HvIAAs were categorized into two major groups and 14 subgroups. The indication of the presence or absence of these domains for the biological functions and acting mechanisms was discussed. The cis-element distributions in HvIAA promoters suggests that the HvIAAs expressions may not only regulated by auxin (the presence of AuxREs and TGA-element) but also by other hormones and developmental and environmental cues. We then studied the HvIAAs expression in response to NAA (1-Naphthaleneacetic acid) using quantitative real-time PCR (qRT-PCR). Like the promoter analysis, only 14 HvIAAs were upregulated by NAA over two-fold at 4 h. HvIAAs were clustered into three groups based on the spatiotemporal expression data. We confirmed by qRT-PCR that most HvIAAs, especially HvIAA3, HvIAA7, HvIAA8, HvIAA18, HvIAA24 and HvIAA34, are expressed in the developing barley spike compared within seedling, suggesting their roles in regulating spike development. Taken together, our data provide a foundation for further revealing the biological function of these HvIAAs.

www.nature.com/scientificreports www.nature.com/scientificreports/ less than e-5. As some ARF protein family proteins share conserved C-terminals with Aux/IAA family proteins (discussed in the introduction), the initial found 42 barley proteins consist some ARF proteins with conserved B3 DNA binding domain (pfam02362) and AUX/RESP domain (pfam06507) [40][41][42] . We eliminated these proteins as well as the redundant proteins and retained 36 proteins for further study.
Analysis of gene structure and conserved motifs. For exon/intron structure of HvIAAs analysis, the CDS and protein sequences corresponding to each predicted gene were downloaded from Barley Genome Database Annotation on the Phytozome website. Multiple sequences alignments with barley (Hordeum vulgare L.) were conducted by Clustal W and Clustal X version 2.0 program of Jalview 2.11.0 software with Defaults 47,48 . The Multiple Expectation Maximization for motif Elicitation (MEME, http://meme-suite.org/tools/meme) tool was used to predict conserved motifs of HvIAA proteins 49 . The gene structure and conserved motifs were generated with TBtools software 50 .

Analysis of chromosome locations and cis-acting elements. Chromosome physical position of
HvIAAs was obtained from Phytozome 36 , using MapGene2Chrom web v2 (http://mg2c.iask.in/mg2c_v2.0/) to draw chromosome physical map. The identified CDS sequence of the HvIAAs was downloaded from Phytozome, and translated into protein sequences, using Muscle program alignment, and then introduced into the DnaSP 6 software to calculate the Ka and Ks values among the sequences of paralogous genes 51 . Super Circos was generated using TBtools software. The cis-elements in the 2,000 bp sequences upstream of the coding sequences were analyzed by Plant CARE databases (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/). expression analysis of barley Aux/IAA genes. Raw datasets were obtained from the BARLEX (Barley Genome Explorer) 36 (https://apex.ipk-gatersleben.de/apex/f?p = 284:10:1811619314937) 52 . These data were applied to analyze HvIAAs expression profiles in different tissues. Heatmap was generated with the gplots package in R (https://www.r-project.org/). It showed the expression differences based on the FPKM values, which were normalized by log 2 (FPKM+1) transform. Hierarchical clustering algorithms was used for recognition of similar patterns in expression files 53 . Plant growth, tissue collection and treatment. The spring barley cultivar of Gold Promise was grown in the greenhouse of Shanghai Jiao Tong University under 16 °C/14 °C day/night, 16 h/8 h light/dark, and 50% relative humidity. For NAA (1-Naphthaleneacetic acid) treatment, one-week-old seedling was sprayed with 5 nM NAA, and then the seedling was sampled at 0.5 h, 1 h, 2 h and 4 h after spraying. Seedlings were sprayed with DMSO (dimethyl sulfoxide) as a control (CK). To detect HvIAAs expression during the barley spike development, five stages were collected: two-week-old seedling (SD), the double ridge stage (DR), the lemma primordium stage (LP), the stamen primordium stage (SP), the awn primordium stage (AP) and the white anther stage (WA) 54,55 . The developmental stages were determined by dissection under a stereomicroscope. All samples were frozen in liquid nitrogen and stored at −80 °C until RNA extraction.
Quantitative RT-PCR analysis. Barley tissues total RNA were extracted with TRIZOL reagent (Invitrogen), then reverse transcription reaction was carried out using a PrimeScript RT reagent kit with gDNA eraser (Takara), according to the manufacturer's instructions. SYBR Green SuperReal PreMix Plus (TIANGEN) kit was applied for quantitative RT-PCR(qRT-PCR) experiments using CFX96 Real-time PCR machine (Bio-Rad). HvACTIN (HORVU5Hr1G039850.3) was used as an internal control 56 . Three biological repeats with three technical repeats were performed (primers used are listed in Supplementary Table S5).

Results
identification of the Aux/IAA gene family in barley genome. Supplementary Tablestified from the barley genome using Hidden Markov Model (HMM) methods 38,39 . HvIAAs were designated as HvIAA1-HvIAA36 according to their physical position on barley each chromosome. We provided the gene characteristics including physical position, ORF (Open Reading Frame) sequence size, amino acid length, molecular weight (Da), isoelectric points (PI) and intron numbers. The number of amino acid length of the predict HvIAA proteins ranged from 96 (HvIAA21) to 772 (HvIAA34). The molecular weight of the HvIAA proteins differed from 10819.43 (HvIAA21) to 86193.2 (HvIAA34) Da, and the PI of the HvIAA proteins varied from 4.5 (HvIAA21) to 9.56 (HvIAA23) (Supplementary Table S1).

Phylogenetic tree of the HvIAA proteins in barley.
To examine the phylogenetic relationships among the Aux/IAAs from barley, rice and Arabidopsis (based on 31 Aux/IAA protein sequences in rice and 29 Aux/ IAA protein sequences in Arabidopsis), phylogenetic evolutionary trees of 96 Aux/IAA protein sequences are constructed using MEGA7.0 software 46 . According to the classification of rice and Arabidopsis 13,57 , 96 Aux/IAAs are divided into A and B groups with the well-supported branch. Based on the evolutionary, the classification of Aux/IAAs in Arabidopsis and rice is consistent with previous reports. There are 17 HvIAAs distributed in group A, www.nature.com/scientificreports www.nature.com/scientificreports/ while another 19 are found in group B (Fig. 2). Group A and B could be further subdivided into 8 and 6 subgroups (A1-A8 and B1-B6), respectively, with varying degrees of bootstrap support.
Subgroup A3 and A4 may specific to dicots, as they do not contain rice and barley IAAs. Subgroup A3 contains AtIAA5, AtIAA6, and AtIAA19. The Arabidopsis iaa5/6/19 triple mutants line has a minor defect in stomatal movement under drought stress 58,59 . The missing of A3 subgroup in monocot indicate that the dumbbell stomata apparatus in monocot does not require the A3 subgroup IAAs paralog. Subgroup A4 contains AtIAA8 and AtIAA9, which were reported to regulate vasculature formation, adventitious root formation, and lateral root elongation [60][61][62] . As IAAs are essential for regulating lateral roots and adventitious roots both in monocot and dicots 63 , the absence of subgroups A4 IAA function may be compensated by other IAAs in other subgroups. For instance, OsIAA11 and OsIAA13 from subgroup A5 were reported in regulating lateral roots in rice 63,64 . We also identified subgroup A7, A8 and B3 IAAs as monocot specific. In our phylogeny tree, the subgroup A7 contains only two members, HvIAA21 and OsIAA26, which are also the smallest proteins in the Aux/IAA family. OsIAA26 was shown to be important for rice root elongation and the protein abundance is indirectly regulated by TIR SCF1 -IAA signaling 21 . There is no data directly deciphering the function of A8 groups IAAs. However, studies suggested that the rice A8 groups IAA genes OsIAA14 and OsIAA24 are regulated by root development regulators, Crown Rootless1 (CRL1), MADS-box transcription factor (OsMADS25), and Nonexpressor of Pathogenesis-Related Genes1 (OsNPR1), indicating that this group of IAAs may function specifically in root development [65][66][67] . Likewise, there is no direct data to infer the function of B3 IAAs. Interestingly, the subgroup B5 only contains barley and Arabidopsis IAAs. It is intriguing to ask whether this group of IAAs is related to the xeromorph of barley and Arabidopsis. All the three Arabidopsis members, IAA32, IAA33and IAA34 were shown to act on high auxin level 19,20 . chromosomal locations of HvIAA genes. The physical location of 35 HvIAAs determines the location on the chromosome. 35 HvIAAs (97.2%, 35/36) locate unevenly on 7 chromosomes, HvIAA36 was not able to be mapped to chromosome (Supplementary Table S1). ChrUn is composed of sequence fragments originating from BAC (bacterial artificial chromosome) overlap clusters not placed in the Hi-C (high-throughput/resolution chromosome conformation capture) map-, or gene-bearing fragments of BAC sequences and Morex WGS (whole genome shotgun) contigs selected in addition to the non-redundant sequence 68 . The nine HvIAAs are located on chromosome 5, two on chromosome 2 and 4, five are located on chromosome 1, eight on chromosome 3, three on chromosome 6, and six on chromosome 7, respectively ( Supplementary Fig. S1).
Phylogenetic analyses can often be used to uncover the duplication events that led to the generation of large sets of tandemly duplicated genes 69 . In this study, 36 HvIAAs only formed 8 sister pairs (Supplementary Fig. S2) with strong bootstrap support (>97%), at least 11 paralogs of HvIAAs might have undergone gene duplication, which may be caused by segment replication and tandem replication events ( Fig. 3; Supplementary Fig. S2) 70 . Segment duplication leads to many homologies of HvIAAs between the chromosomes. www.nature.com/scientificreports www.nature.com/scientificreports/ In order to explore which type of selective pressure determines the divergence process of HvIAA after replication. Ka/Ks is the ratio of nonsynonymous substitution (Ka) to synonymous substitution (Ks), the Ka/Ks substitution ratio is used to assess the coding sequences of 11 pairs of HvIAAs. A Ka/Ks ration greater than 1 represents positive selection, a ratio of 1 represents neutral evolution and a ratio less than1 represents purifying selection 71 . We found that nine out of the eleven Aux/IAA paralog pairs had the ratio of Ka/Ks among 11 pairs genes was less than 1 (Supplementary Table S2), suggesting a purified selection that favors synonymous substitutions than nonsynonymous substitutions to prevents the change of an amino acid residues had occurred following the duplications.
Gene structure and conserved motif analysis of the barley Aux/IAA genes. In the previous research, gene structure diversity provides the main impetus for the evolution of polygenic families [72][73][74] . Therefore, the gene structural diversity of HvIAAs is further studied with exon/intron analysis. The result shows that the sequence length and the number of introns (exons) of 36 HvIAAs vary greatly. The number of introns ranges from 0 to 13, of which HvIAA25 has a complete lack of introns, and HvIAA35 has most introns, with 13 introns. Additionally, some HvIAAs in the same phylogenetic subgroup had the same number of exons such as HvIAA2 and HvIAA9, HvIAA34 and HvIAA27. Interestingly, these two pairs of HvIAAs have similar gene sequence length (Fig. 4).
The motif discovery (classic mode) of the MEME online tool 49  Analysis of cis-elements in HvIAA genes. In order to identify the potential regulatory elements response to auxin. We used the Plant Care database to identify the cis-acting elements existing in the promoter region of HvIAAs. The detailed information of cis-acting elements at the promoter regions was listed in (Supplementary  Table S3), except for one gene (HvIAA6, no clear promoter sequence). According to the previous study, we classify them into six major functional categories; development/tissue specificity, promoter/enhancer element, light response, circadian control, stress and hormone response 75 . The result reveals that two types of cis-acting elements are involved in hormone response and related to auxin response, including AuxRRs and TGA-element (Marked in red color in Supplementary Table S3). It can be inferred that HvIAAs can response to exogenous auxin 76 . Moreover, CAAT-box and TATA-box are ubiquitously presented in the promoter/enhancer element of HvIAAs ( Supplementary Fig. S4), it is one of the binding sites for RNA polymerase and it also determines the  Table S4). Hierarchical clustering was used to cluster the genes with similar expression patterns, which divided the HvIAA into three clusters. The HvIAAs in Cluster I (including eleven members) display low expression levels in most tissues, whereas HvIAA16 and HvIAA17 have high expression in the specific tissues compared with other genes in Cluster I. In Cluster II (including three members) genes are general highly expressed in different tissues, for example, HvIAA20 shows a strong expression level in developing tillers (NOD) compared with other 14 tissues. HvIAA transcripts are less abundant in developing grain (CAR 15) and senescing leaves (SEN). In Cluster III (including twelve members), genes are expressed at a moderate level.
HvIAAs in Cluster I is merely mildly expressed, suggesting that these genes might be less required in those tissues. Remarkably, HvIAA16 is only expressed in developing grain (CAR 5). The expression of genes in Cluster II, in etiolated seedling (ETI), inflorescences, rachis (RAC) and developing tillers (NOD) was higher than other genes in this group. The expression of genes in 15 different tissues of Cluster III shows a fluctuating trend. For instance, the expression of HvIAA19 is highest in developing tillers (NOD), but it is the lowest in developing grain (CAR 15). HvIAA19 may play a key role in development tillers (NOD) in the barley development. These results suggest that these genes have specific functions in different tissue and development processes. expression of HvIAAs genes in the spike development and response to NAA treatment. To investigate whether HvIAAs respond to auxin, the expression levels of HvIAAs are evaluated in the barley one-week seedling by qRT-PCR under NAA treatment. According to expression pattern after NAA treatment, then clustered using hierarchical clustering algorithms, these genes are divided into C1-C8. The data showed that the transcript levels of group C2, 4, 5, and 6 of the HvIAAs are upregulated at 4 h after NAA treatment. Especially, HvIAA30 is upregulated over ten-fold at 4 h, HvIAA2 and HvIAA30 are upregulated over two-fold at 2 and 4 h (HvIAA4, HvIAA16 and HvIAA21 are not detectable). Interestingly, HvIAA33 is decreased at 0.5 h after NAA treatment (Fig. 7).
To investigate the expression of HvIAAs associated with barley spike development, the transcript levels of each HvIAA is monitored in the spike development and their expression patterns are analyzed using qRT-PCR. The result shows that most of the HvIAAs are detected in the barley spike development compare with two-week-old seedling (SD) (HvIAA1, HvIAA4, HvIAA16 and HvIAA30 are not detect data), such as double ridge stage (DR), lemma primordium stage (LP), stamen primordium stage (SP), awn primordium stage (AP) and white anther www.nature.com/scientificreports www.nature.com/scientificreports/ stage (WA) (Supplementary Fig. S5). However, some HvIAAs genes display a high expression during barley spike development. For example, HvIAA3, HvIAA7, HvIAA8 and HvIAA18 are highly expressed in the DR, LP, SP, AP and WA. Interestingly, HvIAA24 shows the highest expression level in WA. HvIAA34 displays a higher expression level in the DR, LP and SP than in the other stages (Fig. 8).

Discussion
Here, we identified 36 Aux/IAAs from the barley genome, each of them has at least one conserved Aux/IAA domain. Intriguingly, barley (5.1 GB) has a large difference in genome size compare to rice (430 MB) and Arabidopsis (135 MB), but the numbers of Aux/IAA members are very close 13,58 . In contrast to the higher plants, Marchantia and Physcomitrella have only one and three Aux/IAAs, respectively [79][80][81] .
This implies that Aux/IAA gene families are expanded when plants start to conquer the land during evolution and the current number for higher plants is essential for their well-being. Nevertheless, only a few of the Aux/ IAAs loss-of-function mutants in Arabidopsis have only mild phenotypes 82 . Investigation of additional species outside of Arabidopsis is recommended for a better understand of the true biological functions of Aux/IAAs 82 . Consistent with the finding in Arabidopsis, there is no premature stop-codon found in barley Aux/IAA genes, suggested that they are not likely pseudogenes 57 .
The 36 barley, 29 Arabidopsis, and 31 rice Aux/IAA proteins sequences were used to construct evolutionary tree by using Neighbor-Joining method.The 36 HvIAAs can be classified into two groups, Group A (17 members) and B (19 members), which is consistent with the classification in rice 13 and Arabidopsis 57 . We found that most of the HvIAAs are located at the distal ends of the chromosome. Similar findings on other barley gene families were also reported 53,83,84 . This may due to the fact that there are overall more abundant genes at both ends than in the middle of the chromosome in barley 36 . It is the meiotic homologous chromosome recombination mainly confined to the distal regions of all chromosomes that renders the uneven density of genes on cereal chromosomes and most of the genes concentrated in the distal regions 85 . Gene duplication events accelerate the rapid expansion and evolution of gene families. In this study, at least 11 pairs of HvIAAs undergone gene duplication ( Fig. 3; Supplementary Fig. S2), including segment duplication and tandem duplication events. For instance HvIAA2 and HvIAA9 may be the products of genomic segment replication. HvIAA24 and HvIAA25 are the products of genomic tandem replication. Generally, this kind of genes arrangement is difficult to segregate via hybridization www.nature.com/scientificreports www.nature.com/scientificreports/ in breeding or research. These members belonging to the same clades usually share similar conserved motif and exon/intron, and they may have similar functions. Nine out of eleven pairs of the paralogous genes are Ka/Ks less than 1, indicating they have undergone purifying selection.
The domain I containing the LxLxLx motif (L, leucine, x represents any amino acid residue) is important for the IAAs as active transcriptional repressors. It couples the TOPLESS protein scaffold (TPL) and thereafter recruiting histone deacetylase to the ARF transcription factor. The Lego bricks structure transforms the ARF to a transcriptional repressor 7,86,87 . Nevertheless, there are abundant numbers of IAA lacking domain I in various species and they are also able to suppress the expression of auxin response genes. IAAs lacking domain I oligomerize with ARF to prevent the activation activity of ARF 87 . In this case, Aux/IAA is an indirect transcription activator acting through modulating the free ARFs level. Besides, IAAs were shown to repress ARF activity without dimerizing with ARF 88 . In our study, we identified HvIAA25 from A5 subgroup, HvIAA21 from A7 subgroup, HvIAA4 from A8 subgroup, HvIAA13 from B2 subgroup, HvIAA16 from B3 subgroup and HvIAA6, HvIAA7, HvIAA11, HvIAA15, HvIAA18, HvIAA27, HvIAA34 and HvIAA35 from B5 subgroup are lacking domain I ( Fig. 2 and  Fig. 5). However, how different in the transcription repression activity of the barely IAAs with/without domain I need to be further checked.
The core sequence of Domain II with GWPPV/I is the target site for ubiquitination degradation of Aux/IAA protein by interaction with TIR1 of SCF TIR1 complex. The absence of domain II in AtIAA20 and AtIAA30 leads to the longer half-life of the proteins comparing to the canonical Aux/IAA proteins, and overexpression of either of them disturbs auxin physiology and causes auxin-related aberrant phenotypes 89 . Previous studies showed non canonical IAA proteins lacking domain II are not subject to the degradation by the SCF TIR1 protein complex and stabilized by the phosphorylation of IAAs on high auxin level 19,20 . In our study, we identified HvIAA21 from A7 subgroup, HvIAA4 from A8 subgroup, HvIAA6, HvIAA7, HvIAA11, HvIAA15, HvIAA18, HvIAA27, HvIAA31, HvIAA34 and HvIAA35 from B5 subgroup, and HvIAA1 from B6 subgroup and HvIAA16 from B3 subgroup are lacking domain II (Fig. 2 and Fig. 5). They may share the common function mechanism as these reported non canonical IAAs. Similar with lacking domain II, mutations in the conserved motif GWPPV/I motif in the domain II, such as AtIAA31 in Arabidopsis and OsIAA4 and OsIAA10 in rice, and the Arabidopsis mutants shy2-2 and shy2-3 (IAA3), iaa18-1 (IAA18), arx2-1 (IAA7), arx3-1 and arx3-3 (IAA17), also stabilize the protein and the overexpression of these genes exhibits the typical auxin-related aberrant phenotypes such as dwarfism, increased tiller angles, reduced gravity response [89][90][91][92] . In our study, similar dominant mutation in IAA domain II if it has domain II was not found in the barley genome ( Supplementary Fig. S3).
Domain III and IV are conserved between the Aux/IAA and ARF protein families, they are mainly responsible for the homo-and hetero-dimerization among the Aux/IAA proteins and ARFs to inhibit the transcription of auxin-responsive genes 11 . Additionally, the βαα motif in domain III composed of a β-sheet and two α-helices (α 1 and α 2) is critical for the dimerization of Aux/IAA protein 7 . Most of the HvIAA proteins have two putative nuclear localization signals (NLS), the first one is composed of a bipartite structure, conserved double KR between domain I and domain II and basic amino acids in domain II, while the second one includes SV40-type www.nature.com/scientificreports www.nature.com/scientificreports/ NLS located in domain III 93 , assuming that HvIAA proteins may play a role in the nucleus. Furthermore, multiple phosphorylation sites are discovered in the HvIAA proteins, these different phosphorylation sites may regulate nuclear transportation ( Supplementary Fig. S3) 94 . Domain III and domain IV in the C-terminal region, has well-known Phox and Bem 1 (PB1) protein-protein interaction domain that mediates homo-and heterodimerization 95 . In our study, we identified HvIAA7, HvIAA18, HvIAA27, HvIAA31, HvIAA34 and HvIAA35 from subgroup B5, are lacking domain III. HvIAA14 and HvIAA36 from subgroup A1, HvIAA25 from subgroup A5, HvIAA28 and HvIAA30 from subgroup B4, and HvIAA1 from subgroup B6, which are lacking domain IV. Genetic and biochemical data on their function are needed to further elucidate these non-canonical IAAs.
Gene expression gives hints for the function of the gene. In order to characterize the expression regulation of HvIAAs, we check cis-elements in the promoter regions of the HvIAAs. We found that the promoter regions of HvIAAs contain cis-elements related to development/tissue specificity, promoter/enhancer element, light response, circadian control rhythm, stress and hormone response (Supplementary Table S3), indicating that expression of HvIAAs is tailored to adapt the multiple functions in various biological processes. The result shows that 18 HvIAAs contains one or two auxin response elements, which may be regulated by auxin. HvIAA8, HvIAA15, HvIAA20, HvIAA27 have two auxin response elements, and the other 18 HvIAAs do not have any auxin response elements (Supplementary Table S3). The presence of these cis-acting elements suggests that the HvIAAs play important roles in the early response of auxin in barley. Previous work shows that auxin and light, brassinolide and abiotic stress signal are mutually regulated [96][97][98] . OsIAA1 and HvIAA9, OsIAA6 and HvIAA13 are in the same clades, and cis-acting elements analysis suggesting that canonical HvIAA9 and HvIAA13 contain TGA-element in their promoters, it can respond to exogenous auxin. Undoubtedly, several of cis-acting elements in the promoter regions of HvIAAs coordinate the regulate expression of the HvIAAs to facilitate their functions in barley development. It is helpful to validate the expressional pattern of HvIAAs experimentally to reveal the gene function of HvIAAs.
The diversity of the expression profiles of HvIAAs in tissues and developmental stages indicates that HvIAAs regulate multiple developmental processes. The expression pattern of HvIAA provides hints to further investigation of their biological function. For example, the high expression of HvIAA20 in inflorescence (LOD) implies that HvIAA20 may regulate inflorescence development. HvIAA16 displays expression in developing grain (CAR5), which suggests that HvIAA16 may be related to grain development. Furthermore, HvIAA9, HvIAA19, HvIAA20 and HvIAA36 have a high expression level in developing tillers (NOD), indicating that these four genes may together involve in developing tillers development (Fig. 6). In previous studies, OsIAA9 promotes lateral www.nature.com/scientificreports www.nature.com/scientificreports/ root formation in rice, and mediates geotropic of deletion is associated with starch granule synthesis in root tips 99 . According to expression analysis and phylogenetic relation of OsIAA9 with HvIAA28 and HvIAA30, HvIAA28 and HvIAA30 may influence the root growth. OsIAA11 inhibits the development of lateral roots and influences inflorescence genes in rice 100 . The expression profile and phylogenetic tree indicate that the subcluster including OsIAA11, HvIAA20 and HvIAA22 may probably participate in the root and inflorescence growth. The mutation of AtIAA12 in Arabidopsis specifically affects embryonic development 101 .
Although Aux/IAA transcription was initially thought to be primary auxin responsive, different expression patterns in response to auxin were found among the gene family in various species 13,102,103 . In C 2, 4, 5 and 6 group, the expression levels are significantly upregulated at 4 h after NAA treatment. Especially, HvIAA30 is upregulated over ten-fold at 4 h, HvIAA2 and HvIAA30 are upregulated over two-fold at 2 and 4 h (Fig. 7). Our promotor analysis identified two auxin signaling transduction-related cis-elements presenting in the promoter regions of the 18 HvIAAs (Supplementary Table S3). The diversity of the numbers and locations of these cis-elements may partly explain the different expression patterns of HvIAAs under NAA treatment. However, HvIAA2, HvIAA5, HvIAA10, HvIAA14, HvIAA19 and HvIAA28 are not found the auxin-responsive elements in the promoter regions, the relative mRNA levels of these 6 genes increased at 4 h after the NAA treatment (Fig. 7). Similar results have been reported in previous research 104 . The transcript levels of the OsIAA9, OsIAA19, OsIAA20 and OsIAA31 are prominent upregulated under auxin treatment 13 , HvIAA28, HvIAA5, HvIAA30 and HvIAA19 are homologs of these four OsIAA genes, respectively (Fig. 2), and detect the expression of upregulated after NAA treatment. www.nature.com/scientificreports www.nature.com/scientificreports/ Traits of barley spikelet largely contribute to the yield. Therefore, the investigation of genes regulating spikelet development may serve for the barley breeding aimed for improving yield. There are accumulating evidence that Aux/IAA genes involve in regulating inflorescence structure and thus spikelet in various species 34,35,105 . In the previous study, overexpression of IAA1 in Arabidopsis significantly reduced cell length and cell number in inflorescences and leaves, and affected cell shape 105 . BIF1 (BARREN INFLORESCENCE1) and BIF4 (BARREN INFLORESCENCE4) encoded the Aux/IAA protein regulate the early steps required for inflorescence formation 34 . In our study, the expression of HvIAA3, HvIAA7, HvIAA8, HvIAA18, HvIAA15 and HvIAA34 are increased during the barley spike developmental process, suggesting that these 6 genes might play a key role in the development of barley spike. Interestingly, HvIAA24 displays the highest expression level in the WA (floral organ differentiation was completed), implying that this gene may act in awn development. HvIAA34 exhibits higher expression in the DR, LP and SP than in the other tissues, indicating that this gene may function specifically in DR, LP and SP stages during barley spike development (Fig. 8). Again, there is not likely IAA pseudogenes in barley as all of them are found expressed at least in one condition.
As in other plants, the Aux/IAAs could play important roles in the growth and development of barley. However, the biological function of the barley Aux/IAAs remains to be elucidated in detail. Our data on the phylogeny, gene structure, conserved motifs, chromosome locations, cis-acting elements and expression profiles provide essential clues for exploring the biological functions of the 36 Aux/IAAs in barley.

Conclusion
In this study, we identified 36 Aux/IAAs in barley, similar number as that in rice and Arabidopsis. Half of them are canonical Aux/IAAs containing the typical four conserved IAA domains and the rest are non-canonical Aux/ IAAs. We identified 14 expressional auxins responding HvIAAs and numbers of HvIAAs that may regulate spike development. Our findings would facilitate the functional study of Aux/IAA genes and molecular breeding of barley.