Identification and Characterization of MADS-box Proteins in Setaria italica
A total of 512 genes encoding MADS-box protein were compared in nine representative species(Volvox carteri, Micromonas pusilla CCMP 1545, Arabidopsis thaliana, Ananas comosus, Oryza sativa, Brachypodium distachyon, Sorghum bicolor, Zea mays and Setaria italica) and the number of gene members in different subgroups was also calculated (Online Resource F1 and S1). In Setaria italica, 89 putative MADS-domain proteins were identified by HMMER(Potter et al. 2018). Then, the MADS-domain protein sequences encoded by non-representative transcripts were excluded. The remaining sequences were checked for the existence of complete MADS-domains using CDD, Pfam and SMART. A total of 70 sequences were confirmed as foxtail millet MADS-box proteins and renamed as being SitMADS genes (SitMADS1 to SitMADS70) based on their chromosomal locations (Online Resource S2). The length of SitMADS proteins varied from 61 (SitMADS70) to 477 (SitMADS17) amino acids. The molecular weights ranged from 6.97 KDa (SitMADS70) to 53.18 KDa (SitMADS1). The predicted isoelectric point values of SitMADS proteins were between 4.41 (SitMADS17) to 10.65 (SitMADS70).
Gene structure and conserved motif distribution of SitMADS genes
The exon-intron organizations of all the identified SitMADS genes were examined to gain more insight into the evolution of the MADS family in Setaria italica. The phylogenetic tree and gene structure were integrated together as shown in Online Resource F2, The number of introns varied from 0 to 11. Type-Ⅰ SitMADS genes, with and without one intron accounted for 92%, which showed that the genes within the same subgroup had similar genetic organization.
To study the structural features of the SitMADS proteins, the MEME(Multiple Em for Motif Elicitation) program was used to identify the conserved motifs of SitMADS proteins(Bailey et al. 2009). Meanwhile, the predicted motifs were annotated. A total of four conserved motifs were predicted in 70 SitMADS proteins (Online Resource F2). As expected,typical MADS-box domains were found in all SitMADS proteins. Most type-Ⅱ SitMADS proteins contained a specific conserved K domain which were lacked in all type-I SitMADS genes.
Phylogenetic analysis of SitMADS-box gene family
To analyze the classification and evolutionary relationships among the SitMADS members, we constructed a phylogenetic tree based on the protein sequences in Setaria italica, Oryza sativa and Arabidopsis thaliana (Fig. 1). Phylogenetic analysis of the 70 SitMADSs together with 108 AtMADSs and 72 OsMADSs showed that the SitMADS family could be classified into 12 distinct groups, according to the classification in Arabidopsis thaliana(Parenicova et al. 2003), namely subfamilies MIKCC (39), MIKC∗ (4), Mα (18), Mβ (5) and Mγ (4). Totally 56% of the 70 SitMADS genes belongs to the MIKCC-type genes, which could be further divided into 8 subfamilies, including the SOC1, SEP, AP3/PI, ANR1, A, AGL-12, C/D and TT16 subgroup.
Chromosome distribution and synteny analysis of SitMADS genes
To understand the genomic distribution and gene duplication of SitMADS genes, 70 SitMADS genes were distributed unevenly throughout the nine chromosomes in Setaria italica (Fig. 2A). Among them, the chromosome 5 contained the largest number of SitMADS genes(13), while only two SitMADS genes were localized on chromosome 8. The majority of SitMADS genes were located on the proximate or the distal ends of foxtail millet chromosomes.
There was no positive correlation between the chromosome length and the number of SitMADS genes. Gene replication events, including segmental and tandem duplications, were crucial events for the expansion of the gene family in the process of evolution. The chromosomal regions of two or more genes within the 200kb range were defined as tandem duplications(Holub 2001). Two tandem duplicated clusters(SitMADS19-SitMADS20 and SitMADS53-SitMADS54-SitMADS55) formed two gene clusters that were detected on chromosome 3 and 6, respectively (Fig. 2A). Interestingly, the five tandem duplicated genes all belong to the type-Ⅱ. 12 segmental duplication events with 20 SitMADS genes were indicated in Fig. 2B. A total of 24 (34.3%) SitMADSs were duplicated genes. The result suggested that segmental duplication events may be more important than tandem duplication in the expansion of the SitMADS gene family in Setaria italica.
The Ka, Ks, and Ka/Ks values of the homologs were calculated using a KaKs_calculator (Online Resource S3). Three pairs of segmental duplication genes were found to have been subject to purifying selection pressure during evolution and negative purifying selection was present.
For better understanding the evolution of SitMADS gene family, we constructed three comparative syntenic maps of foxtail millet associated with three representative species, including Arabidopsis thaliana, Oryza sativa and Zea mays. The numbers of orthologous gene pairs between Arabidopsis, rice and maize, were 4, 37 and 23(Fig. 3, Online Resource S4). Between foxtail millet and rice genes, SitMADS21, SitMADS35, SitMADS41, SitMADS46, SitMADS53 and SitMADS59 were found to be associated with two syntenic gene pairs, which implied that these genes may have played an important role of MADS gene family during evolution.
Expression Profiles of SitMADS Genes in Different Tissues
To predict the possible functions of SitMADS genes, we further detected the expression profiles of 10 selected SitMADS genes in seedlings, roots, leaves, stems and panicles by RT-qPCR(Real-time quantitative PCR) including SitMADS6 (ANR1), SitMADS9 (ANR1), SitMADS23 (Mα), SitMADS42 (Mβ), SitMADS46 (AP3/PI), SitMADS51 (ANR1), SitMADS52 (MIKC*), SitMADS58 (ANR1), SitMADS63 (SEP) and SitMADS64 (A). The results revealed that ten SitMADS genes were highly expressed in roots relative to seedling tissues (Fig. 4). In addition, compared to gene expression in seedling, SitMADS42, SitMADS46 and SitMADS64 also had relatively high transcriptional activity in leaf tissue, SitMADS46, SitMADS63 and SitMADS64 had relatively high transcript abundance in the panicle (Fig. 4). These results implied the differential roles of ten genes in tissue development.
Cis-acting regulatory elements in SitMADS promoters
For further analyze the role of SitMADS in plant resistance to stress, the region about 2000 bp upstream of the start codon (ATG) was used here to analyze cis-acting regulatory elements. The cis-acting element in the SitMADS gene promoter could be divided into three categories according to their functional annotations including light-response, biotic and abiotic stress-response and plant hormones-response (Online Resource F3 and S5). The results showed that many abiotic stress (WUN-motif, GC-motif, ARE, TC-rich repeats, MBS) related elements participate in regulating gene expression, such as SitMADS6, SitMADS9, SitMADS23, SitMADS42, SitMADS46, SitMADS51, SitMADS52, SitMADS58, SitMADS63 and SitMADS64). In addition, ABRE (abscisic acid response element), which is involved in the response to ABA (Abscisic Acid), were detected in SitMADS51, SitMADS63 and SitMADS64. These elements were thought to play important roles in responses to drought stress.
Expression patterns of SitMADS genes in response to different treatments
To investigate whether the expression of SitMADS genes were influenced by drought stresses and related signaling at the transcriptional level, 10 SitMADS genes (SitMADS6, SitMADS9, SitMADS23, SitMADS42, SitMADS46, SitMADS51, SitMADS52, SitMADS58, SitMADS63 and SitMADS64) were selected for further expression profiling after PEG, drought and ABA treatments. Overall, except SitMADS9 up-regulated in 3 h, nine SitMADS genes (SitMADS6, SitMADS23, SitMADS42, SitMADS46, SitMADS51, SitMADS52, SitMADS58, SitMADS63 and SitMADS64) were significantly induced and up-regulated after 1 h, whereas down regulated after 3 h under PEG-6000 treatment (Fig. 5). To further confirm whether the expression of SitMADS genes was influenced by drought stresses, we treated the ten genes with real drought treatment. Ten SitMADS genes (SitMADS6, SitMADS9, SitMADS23, SitMADS42, SitMADS46, SitMADS51, SitMADS52, SitMADS58, SitMADS63 and SitMADS64) significantly responded to drought stress. The results were consistent with the trend of PEG simulated drought experiment. In addition, the expression pattern of ten SitMADSs under 100 μM ABA treatment were also analyzed. SitMADS51, SitMADS63 and SitMADS64 had an increase in transcript levels at 3 h by ABA treatment, while SitMADS52 showed a trend of first decreased and then slightly increased (Fig. 5).