Identification of MADS-domain proteins in physic nut
All Arabidopsis MADS-domain protein sequences were used as queries in a BlastP search to identify physic nut proteins. A Hidden Markov Model (HMM) search was also performed against the physic nut protein database using the MADS-domain PF03106. In total, 63 putative MADS-domain proteins were identified in physic nut, with the presence of the MADS-box domain in each of them being confirmed by a SMART website search. These genes were named sequentially from JcMADS01 to JcMADS63 according to their chromosomal locations (Additional File 1). The 63 JcMADS genes ranged in length from 195 (JcMADS35) to 1164 (JcMADS34), thus potentially the proteins encoded would be from 64 to 387 amino acids; their GenBank accession numbers are given in Additional File 1.
Phylogenetic relationships of the JcMADS proteins
To clarify the phylogenetic relationships of the physic nut MADS family proteins with the previously reported members of the family in Arabidopsis and rice, an unrooted tree was constructed by IQ-TREE using the Maximum likelihood method (Fig. 1). On the basis of full length amino acid sequence conserved domain and similarity, we subdivided the 246 typical members of the MADS gene family into 5 groups (designated MIKC, Mα, Mβ, Mγ, MIKC*), according to the previous classification of MADS proteins from Arabidopsis [9]. Of the 63 inferred physic nut JcMADS proteins (Additional File 2), thirty-two were assigned to group MIKC (JcMADS32-JcMADS63), thirteen to group Mα (JcMADS19-JcMADS31), four to group Mβ (JcMADS01-JcMADS04), six to group Mγ (JcMADS05-JcMADS10) and eight to group MIKC* (JcMADS11-JcMADS18). In the phylogenetic tree, some members of the JcMADS gene family formed related sister pairs (Fig. 1): JcMADS02 and 03, 06 and 07, 14 and 15, 19 and 20, 22 and 23, 38 and 39. There were also triplets (JcMADS08, 09 and 10; 28, 30 and 31; 40, 41 and 42). The tree indicated that proteins in group MIKC were the most numerous; it contained 39 AtMADS, 38 OsMADS and 32 JcMADS proteins.
Conserved motifs in JcMADS proteins
The structures of proteins encoded by JcMADS genes were analyzed using the MEME online software tool. Twenty conserved motifs, which we named motifs 1–20, were identified in the 63 JcMADS proteins (Fig. 2 and Additional File 3). As expected, motif 1 and motif 4 corresponded to the typical MADS-box domain, and motif 1 was found in all the JcMADS proteins. Motif 9, specifying the K domain, was found in most MIKC type proteins; the exceptions were JcMADS35, 43, 47, 50, 54 and 62, which have relatively short amino acid sequences. In addition to these known functional motifs, some of unknown function were also detected. Examples included motifs 5, 10 and 20 (detected only in group Mγ), and motif 13 (found only in groups MIKC and Mβ). Motif 12 was found only in group MIKC, while motifs 14 and 19 was detected only in group MIKC*. Additionally, most conserved motifs detected in JcMADS proteins were clade-specifically assigned in different groups, implying similarity of function within a given group.
Chromosomal localization of JcMADS genes
We mapped 62 of the 63 JcMADS genes (all except JcMADS45) to LGs based on a previously published report on the physic nut genome [25]. As shown in Fig. 3, we found that LGs 4 and 7 had more members of the JcMADS gene family than other LGs, with eleven and nine JcMADS genes respectively. They were followed by LGs 2, 3, 5 and 10, each of which had six JcMADS genes. In addition, there was five JcMADS genes on each of LGs 6 and 9, three on LG8, three on LG11 and two on LG1. The results also indicated that most JcMADS genes were on the lower and middle regions of the LGs. Tandem duplications, defined as tandem repeats which are separated by < 4 non-homologous spacers or are genes located within 50 kb of each other [26], were found among these members of the JcMADS gene family. The gene pairs present as tandem repeats (T) were T1 (JcMADS41 and 42), T2 (JcMADS28 and 39), T3 (JcMADS22, 23 and 56), T4 (JcMADS33 and 61), T5 (JcMADS52 and 59), T6 (JcMADS12 and 62), T7 (JcMADS10 and 27), T8 (JcMADS47 and 63), T9 (JcMADS29 and 49), T10 (JcMADS53 and 58), T11 (JcMADS02, 35 and 60) and T12 (JcMADS30, 31 and 50), on LG2, LG2, LG3, LG4, LG5, LG5, LG6, LG7, LG7, LG8, LG9 and LG10 respectively.
Expression profile of JcMADS genes under non-stressed growth condition
To clarify the roles of the JcMADS in regulating physic nut development, we examined the expression profiles of JcMADS genes in roots, stem cortex, leaves, and seeds (S1 and S2) under non-stressed growth conditions based on data from RNA sequencing (RNA-seq) (Additional File 4 and Fig. 4). The result suggested that fifty of the predicted JcMADS genes were expressed in at least one of the organs examined, while thirteen (JcMADS04, 11, 17, 20, 21, 22, 28, 30, 31, 35, 43, 45 and 57) were not expressed in any of these tissues. Of the 50 JcMADS genes for which expression was detected, two (JcMADS03 and 53) were highly expressed across all the tissues sampled, ten (JcMADS09, 10, 16, 19, 23, 40, 41, 42, 56 and 58) were expressed only in seeds, thirteen (JcMADS01, 02, 05, 08, 13, 29, 34, 40, 44, 51, 55, 60, and 63) exhibited highest expression in seeds, four (JcMADS24, 36, 48 and 49) preferred to be expressed in roots, and one (JcMADS32) was most strongly expressed in the stem cortex.
As shown in Fig. 4, most of the JcMADS genes were expressed more highly in seeds at the S1 stage than at the S2 stage. It was noteworthy that nine genes (JcMADS01, 09, 10, 23, 40, 42, 55, 56, 48 and 58) was detected expression only in seeds at S1 stage. Based on the results of expression pattern analysis, the JcMADS40 gene was chosen for functional research.
Expression profile of JcMADS under abiotic stress conditions
Many studies have suggested that some MADS-box genes encode proteins involved in the regulation of abiotic stresss [21, 27-28]. We therefore further investigated the patterns of expression of JcMADS genes in leaves after 2 d, 4 d and 7 d of drought stress and after 2 h, 2 d and 4 d of salinity stress according to data from RNA-seq. As shown in Fig. 5, the transcript abundances of seven JcMADS genes indicated at least a twofold enhancement or reduction compared with the control in response to at least one stress at one or more time points. Of these seven genes detected as having differential expression, three (JcMADS29, 36 and 53) exhibited significantly induced or inhibited expression in response to drought and salinity stresses, three (JcMADS12, 37 and 54) showed differential expression only in response to drought stress, and JcMADS05 responded solely to salt stress.
JcMADS40 is a nucleus-localized transcriptional activator
To confirm the subcellular localization of the protein encoded by JcMADS40 gene, the 35S:JcMADS40-YFP fusion construct and the 35S:YFP empty vector were introduced into Arabidopsis protoplasts. The fluorescence signals from the protoplasts were then observed immediately by confocal laser-scanning microscopy. As shown in Fig. 6, we observed that the yellow fluorescent signal was distributed throughout the whole of the cell when the 35S:YFP vector was used, whereas in protoplasts harboring the construct 35S:JcMADS40-YFP a strong yellow fluorescent signal was detected in the nuclei. These findings indicate that JcMADS40 gene is located in the nucleus.
A dual-luciferase assay was used to examine the transcription activation activity of JcMADS40 protein. The full-length CDS of JcMADS40 was attached to the vector pBD, then the pBD-JcMADS40 fusion effector vector and the p5×GAL-Reporter vector were transformed into Arabidopsis protoplasts. The results indicated that the LUC/REN ratio was significantly lower in the control protoplasts (pBD) than in the pBD-JcMADS40 group. Our data suggest that the full-length JcMADS40 has transactivation activity (Fig. 7). Based on the above results, we drew the conclusion that JcMADS40 functions as a transcription activator.
Phenotypic analysis of transgenic rice plants expressing JcMADS40
To investigate the role of JcMADS40 gene in regulating plant development, and to assess the feasibility of using JcMADS genes to control seed size in an important crop plant, we overexpressed this gene in rice. Three independent transgenic lines (OE1, OE2 and OE3) were confirmed as expressing JcMADS40 expression using semi-quantitative RT-PCR, and selected for further study. Expression of JcMADS40 were detected in transgenic lines, whereas no expression was found in WT (wild type) plants (Fig. 8C). Phenotypic analysis showed that the growth and flower structure of transgenic plants overexpressing JcMADS40 were similar to those of WT plants (Fig. 8A and B). Statistical analysis indicated that there was no obvious difference in root and shoot lengths in the transgenic plants compared to the WT plants (Fig. 8D and E). Taken together, these results led to the conclusion that JcMADS40 did not have any major effect on the growth of the transgenic plants.
Overexpression of JcMADS40 reduces the grain size in transgenic rice
As described above, JcMADS40 expression was most strongly detected in seed, suggesting that JcMADS40 might have an important role in seed growth and development. To test this, we examined the effects of JcMADS40 overexpression on rice grain size. We found that JcMADS40 transgenic plants produced dramatically smaller seeds than the WT lines (Fig. 9A). The results also showed that JcMADS40 transgenic plants had a significant reduction in both grain length and width compared to the WT plants (Fig. 9B and C). We also detected a significant reduction in 1000-seed weight and yield per plant in JcMADS40 transgenic lines (Fig. 9D and E). Our data suggested that overexpressing JcMADS40 significantly altered seed size in transgenic plants.
To study the molecular mechanism of JcMADS40 gene regulates grain size, we further tested the expression of grain-size-related genes (Fig. 9F). The results showed that expression of some positive regulatory factors, such as GS2, SMG11, was significantly lower in transgenic plants than that in wild type, while expression of some negative regulatory factors, such as OsMKP1, GW2, was obviously higher than that in wild-type. Taken together, these data supported a putative role for JcMADS genes in seed development.