Ectopic expression of a Brassica rapa AINTEGUMENTA gene (BrANT-1) increases organ size and stomatal density in Arabidopsis

The AINTEGUMENTA-like (AIL) family plays a central role in regulating the growth and development of organs in many plants. However, little is known about the characteristics and functions of the AIL family in Chinese cabbage (Brassica rapa L. ssp. pekinensis). In this study, a genome-wide analysis was performed to identify the members of the AIL family in Chinese cabbage. We identified three ANT genes and six ANT-like genes of Chinese cabbage, most of which were differentially expressed in different organs or tissues. Furthermore, compared with the wild-type line, the size of different organs in the 35S-BrANT-1 line was significantly increased by promoting cell proliferation. Meanwhile, over-expression of BrANT-1 also increases the stomatal number and delays the leaf senescence. Transcriptome analyses revealed that a set of cell proliferation and stoma development genes were up-regulated, while the senescence-associated genes were down-regulated, suggesting these genes may be involved in BrANT-1 regulated processes for controlling organ size, stomatal density and leaf senescence. In summary, this study offers important insights into the characteristics and functions of the ANT genes in Chinese cabbage, and provides a promising strategy to improve yield or head size in Chinese cabbage breeding programs.

in the production of organs with a large size 7,17,18 , while the loss-of-function mutations of the ANT gene have a smaller organ size 4,5,7 . Additionally, many studies revealed that ANT regulates organ size by changing the total cell numbers 7 . Furthermore, cells ectopically expressing ANT in fully differentiated organs exhibit neoplastic activity by producing calli and adventitious roots and shoots 7 . These studies strongly suggest that ANT most likely maintains ongoing cell proliferation coordinately with cell growth 19 . However, to our knowledge, the AIL family from Chinese cabbage has not been characterized. Therefore, identifying and analyzing the AIL family in Chinese cabbage is of great interest.
In this study, we investigated the BrAIL family members in Chinese cabbage through genome-wide bioinformatics analysis, including the identification and characterization of the AIL family members, gene structural analysis, phylogeny and motif analysis. The expression patterns of these genes were characterized in detail in response to the naphthaleneacetic acid (NAA) treatment in different tissues by quantitative real-time PCR (qRT-PCR). Our results show BrANT-1 had the highest sequence similarity to AtANT (AT4G37750), and hence was chosen for further functional analysis. Over-expression of BrANT-1 enhanced organ size by increasing cell number and stomatal density and delaying leaf senescence. Furthermore, by analyzing the potential pathways where BrANT-1 participates, it is possible to understand how the gene regulates Chinese cabbage yield and/or head size.

Results
Identification of AIL family members in Chinese cabbage. A total of nine BrAIL family members were identified in the Chinese cabbage genome, according to the taxonomy of the AP2 superfamily (Supplementary Tables S1 and S2). These included three ANT genes and six AIL genes ( Table 1). The sequences of each member were downloaded from the Brassica database (http://brassicadb.org/brad/) 20 and the two AP2 domains were confirmed according to the SMART database (http://smart.embl-heidelberg.de/). All nine BrAIL family members were named by the A. thaliana orthologs based on the sequence similarity of the protein sequences. The detailed description of sequence similarity between Arabidopsis thaliana and Brassica rapa is shown in Supplementary  Table S3. When more than two Chinese cabbage genes were mapped to one homologous gene in A. thaliana, one additional number was added to the end of the gene name 21 . For example, three ANT genes, Bra017852, Bra011782 and Bra010610, were homologs of AtANT (AT4G37750). Accordingly, they were named BrANT-1, BrANT-2 and BrANT-3, respectively. The BrAIL family members were randomly mapped to different chromosomes of B. rapa (chromosome number: 01, 02, 03, 06, 08 and 10). The chromosome 02 and 03 contained three and two BrAIL genes, respectively, while the chromosome 01, 06, 08 and 10 contained only one BrAIL gene. Subsequent sequence analysis showed that the coding sequences of these nine BrAIL genes had a length of 1323 to 1725 bp, which encode a peptide of 440 to 574 aa. The encoded proteins had a predicted molecular weight varying from 49.6 to 64.1 kDa and a theoretical isoelectric point (pI) varying from 6.09 to 7.81.
Phylogenetic relationships and gene structure of the AIL protein in Arabidopsis, rice and Chinese cabbage. In order to understand the classification of the BrAIL genes in Chinese cabbage, the AIL genes from two other model plants were selected for comparative analyses, including a model C3 monocotyledon plant (rice) and a model eudicots plant (Arabidopsis). A phylogenetic tree was constructed based on the full-length protein sequences of Chinese cabbage, rice and Arabidopsis, by using the bootstrap-neighbor-joining method. The AIL genes of the other two species were obtained from previous reports on Arabidopsis 22 and rice 23 . According to the phylogenetic analysis, the 21 members were divided into seven groups (Class A-Class G) (Fig. 1A). Low bootstrap values were obtained because the difference of the AP2 domain sequences among these three species was small, indicating that BrAILs have high similarity to AtAILs. On the contrary, BrAIL proteins were only remotely related to the OsAIL proteins. To further explore the evolutionary relationships of the coding sequences, the structural analyses of intron/exon of the three species were performed by an online tool GSDS (http://gsds.cbi.pku.edu.cn/). It was found that most of the AIL family members in the three species had at least three introns, except OsAP2/EREBP#058, while the number of introns in the BrAIL genes ranged from six to eight (Fig. 1B). Furthermore, most of the genes (7 out of 9) had six introns, except BrAIL1 and BrAIL5, which contained seven and eight introns, respectively. Syntenic analysis of the AIL family members in Chinese cabbage. To understand the evolution mechanism of the BrAIL genes in Chinese cabbage, the syntenic genes were analyzed between A. thaliana and B. rapa with the BRAD program (http://brassicadb.org/brad/searchSynteny.php) 24 . The results showed that nine BrAIL genes were from five blocks of four translocation Proto-Calepineae Karyotype (tPCK) chromosomes of the ancestor, respectively. All genes were uniformly distributed on three subgenomes, namely, less fractionized (LF), more fractionized 1 (MF1), and more fractionized 2 (MF2). Additionally, two sets of triplicated genes were found in the Chinese cabbage genomes: BrANT-1, BrANT-2 and BrANT-3 in the U block; BrAIL6-1, BrAIL6-2 and BrAIL6-3 in the R block ( Table 2). The other genes (BrAIL1, BrAIL5 and BrAIL7) were singletons. Interestingly, the tandemly duplicated gene was not found in the BrAIL family.  Supplementary Table S4. Additionally, the length of the linker region was conservative (30 aa). The motif of the BrAIL members was analyzed using an online tool MEME (meme.nbcr.net/meme/intro.html), and ten motifs were identified, including motif 1, 2, 3, 4 and 10 in nine BrAIL proteins, motif 5 and 6 in eight BrAIL proteins, motif 8 in seven BrAIL proteins, and motif 7 and 9 in three BrAIL proteins, respectively ( Supplementary Fig. S2).

Expression patterns of the BrAIL family members in various organs.
Accumulating experiments have shown that the AIL genes were expressed in multiple tissues and involved in regulating the development process of different tissues, such as root 25 , shoot 7 , floral organ [26][27][28] , leaf 29 and seed 18 . To explore the potential roles of the BrAIL genes in regulating the growth and development of Chinese cabbage, the expression patterns of the BrAIL genes was investigated in root (R), dwarf stem (DS), old leaf (OL), young leaf (YL), blooming flower (BFL) of Chinese cabbage by qRT-PCR. Eight genes were detected in different tissues, except BrAIL1, which was undetectable in all tissues (Fig. 2). For example, four BrAIL family members, including BrANT-2, BrANT-3, BrAIL6-1 and BrAIL6-2, showed higher expression levels in R than other tissues; three BrAIL genes (BrANT-1, BrAIL6-3   and BrAIL7) were mainly expressed in DS; BrAIL5 was mainly expressed in YL. Moreover, none of the genes were detected in old leaves.

Expression profiles of the BrAIL family members under NAA treatments. Previous studies have
shown that auxin plays an important role in plant growth and developmental processes 30 . In addition, the auxin signal may regulate organ growth by modulating ANT expression 10 . To understand the expression responses of the BrAIL members to auxin, we assessed the transcript levels of the BrAIL genes upon NAA treatments by qRT-PCR. As shown in Fig. 3, the transcriptional levels of most genes were induced by the NAA treatment, except BrAIL1. The expression level of BrAIL5 was only up-regulated at 3 h. Additionally, six genes were increased at 1 h and 3 h compared with the untreated control, including BrANT-1, BrANT-3, BrAIL6-1, BrAIL6-2, BrAIL6-3 and BrAIL7. However, the expression level of BrANT-2 showed a trend of decrease at 1 h, followed by an increase at 3 h.

Over-expression of the BrANT-1 gene caused a variety of phenotypic changes in Arabidopsis.
Compared with other BrAIL family members, BrANT-1 had the highest sequence similarity to Arabidopsis thaliana ANT (AT4G37750), while BrANT-1 was different from other AtAIL genes (Supplementary Table S5). Moreover, BrANT-1 was expressed in different tissues, suggesting that BrANT-1 might play an important role in growth and development of Chinese cabbage. To explore the potential function of BrANT-1, Arabidopsis plants were transformed with BrANT-1 under the control of the CaMV 35 S promoter. A series of pleiotropic effects were significantly distinguishable from the wild-type (WT) at 7 days after sowing, with the 35S-BrANT-1 transgenic seedling exhibiting a longer root (Fig. 4A). Additionally, the hypocotyl length of the transgenic seedling also showed a >24% increase compared with that of the WT (Fig. 4C). At 15 days after acclimating to the nutrient soil conditions, the seedling size of the transgenic plants was bigger than that of the WT (Fig. 4E). At 40 days after sowing, the leaf dimension of the transgenic plants was significantly increased compared with that of the WT (Fig. 5). Rachis length, flower dimension, seed size and silique were also enlarged ( Fig. 6), while the seed number per silique of the 35S-BrANT-1 transgenic line was slightly more than that in the WT (Fig. 6D). Furthermore, SEM was used to compare the epidermal cells in the adaxial of mature leaves of 35S-BrANT-1 and WT. The results showed the numbers of the adaxial epidermal cells of the 35S-BrANT-1 transgenic line had a >50% increase while the cell size of the 35S-BrANT-1 transgenic line was modestly decreased (<6%) compared with the WT (Fig. 5E). However, the slightly smaller cell size could not account for the big leaf size. Therefore, we conclude that the enlarged leaf area was caused by the increase of the cell number (Fig. 5B). All phenotypic analysis data were shown in Supplementary Table S6.
Additionally, at 40 days after sowing, the transgenic lines had a delayed leaf senescence phenotype, compared with the WT, in that the fifth rosette leaf (numbered from the bottom) became senescent, while the second rosette leaf was still green (Fig. 5A).
In addition, other interesting phenotypes were observed. For example, the stomatal density of the 35S-BrANT-1 line was significantly increased in the mature leaves compared with that of the WT (Fig. 5B, detailed in Supplementary Fig. S3). Consistently, the stomatal conductance (Gs) and transpiration rate (Tr) of the 40-day 35S-BrANT-1 plants were significantly increased compared with those of the WT, while the intracellular  Table S7). Additionally, ten DEGs were randomly selected for further verification by qRT-PCR; all genes exhibited the same expression tendency as shown in the original data (Fig. 8). According to the Arabidopsis genome sequence and the DEGs could be assigned to different families, such as MADS-box, TCP, extension, expansin, early auxin-responsive genes (small auxin-up RNA, SAUR), VQ, STOMAGEN, SAGs and transcription factors. The DEGs that were related to plant growth and development were selected and characterized (Table 3). Further analysis indicated many genes were up-regulated in 35S-BrANT-1, which were mainly involved in cell proliferation (MADS-box protein, AT1G59920; TCP21, AT5G08330) and stoma development (STOMAGEN, AT4G12970). On the other hand, the genes inhibiting plant growth (SAUR33, AT3G61900; VQ22, AT3G22160; VQ3, AT1G21326 and ANAC036, AT2G17040) and promoting leaf senescence (AtNAC2, AT5G39610; SUAR36, AT2G45210; SAG13, AT2G29350) were down-regulated in the 35S-BrANT-1 line.

Discussion
AIL genes duplication in Chinese cabbage. The genome of the mesopolyploid crop species Brassica rapa has undergone whole genome triplication (WGT) since divergence from Arabidopsis thaliana, resulting in a significant increase of the numbers of the duplicated genes 31 . However, only nine AIL family members were identified in Chinese cabbage, including three BrANTs and six BrAILs, which were only 1.5-fold more than those in Arabidopsis, suggesting many genes were lost during genome duplication 32 . Similar results were also found in other gene families, such as BrGRF and BrVQ genes, which were about 1.89-and 1.9-fold more than those in Arabidopsis, respectively 13,33 . Additionally, the expansion of the BrAIL family members mainly depended on the segmental duplication, because no tandem duplicated pairs were found. Many studies on genome duplications have shown that the genes involved in transcription, protein binding, response to biotic stimuli and signal transduction path are preferentially retained by segmental duplication [34][35][36] . Duplication events within a genome can result in paralogs, and these genes may have different expression patterns following duplication indicative of subfunctionalization 37 . For example, two duplicated genes, BrVQ22-1 and BrVQ22-2, are differentially expressed in different tissues 33 . In this study, the triplicated genes of BrAIL6-1/-2/-3 and BrANT-1/2/3 also exhibited a different expression pattern in different organs and in response to the auxin treatment. Additionally, similar cases have also been reported in the AP2/ERF family in Chinese cabbage 38,39 . BrAIL family members were expressed in various tissues. By quantitative real time PCR, we have shown that most of the BrAIL family members were differentially expressed in both vegetative and reproductive tissues (Fig. 2). Similar to the AtAIL genes, the BrAIL genes had higher expression in young tissues (seedling   and roots) and lower expression or absent in mature leaves 40 . Multiple BrAIL genes were expressed in different tissues, suggesting that the BrAIL genes play different roles in different organs, which has been confirmed in other plants. For example, ANT and AIL6 are involved in the regulation of flower or seed development in Arabidopsis or Medicago truncatula 9,17,18 and the VviANT1 gene plays an important role in the regulation of berry size 28 . Additionally, we observed the BrAIL genes were lowly expressed in blooming flowers. A previous study also found that the AIL genes are mainly expressed in the advanced stage B (B2), flowers from inflorescences at stage G (G), and the early stage H (H1) in grapevine, while the expression of the AIL genes is low in blooming flowers 28 . Besides, AtANT, AtAIL5, AtAIL6, and AtAIL7 also exhibit the same expression tendency during the flower development 40 .

Transgenic lines over-expressing BrANT-1 had enlarged organs by increasing the cell number.
Generally, the AIL genes play an important role in regulating organ growth through increasing cell number 7 or cell size 29 . In the present study, the cell number in the leaf was increased in the 35S-BrANT-1 transgenic lines, implying that BrANT-1 positively regulates cell proliferation. This result was consistent with the function of the AtANT gene in the leaf or floral organ 7 . However, it is different from the roles of PnANTL1 and PnANTL2 genes, which increase leaf length through increasing cell size in tobacco 29 .
Recently, Liu et al. 41 found that the OsMADS1 gene can positively regulate cell proliferation in rice. Therefore, the most up-regulated MADS-box gene (AT1G59920) among the DEGs may play critical roles in increasing the cell number in the 35S-BrANT-1 line. Additionally, previous studies have shown that the TCP gene can be divided into two classes (I and II). The class I genes like TCP20 function as positive regulators of cell growth 42 , while the class II genes like the Antirrhinum genes CINCINNATA (CIN) function as negative regulators of cell growth 43 . Here, AtTCP21 (AT5G08330), a member of the class I genes 44 was found to be up-regulated, suggesting that it may partially promote cell proliferation in the 35S-BrANT-1 line. Apart from the up-regulated genes, some down-regulated genes, such as SAUR36 (AT2G45210) 45 , VQ22 (AT3G22160) 46 and ANAC036 (AT2G17040) 47 , might also play an important role in increasing organ size in the 35S-BrANT-1 line, as indicated by previous studies, which have shown that some SAUR, VQ and NAC genes are involved in the negative control of plant growth [46][47][48][49][50] . Interestingly, some genes, such as LRX5 (AT4G18670) 51 and expansinB1 (AT2G20750) 52 , positively controlling cell size were also identified from the up-regulated DEGs, as indicated in our histological results which showed a slight reduce in cell size between the 35S-BrANT-1 and WT plants. The result was consistent with our previous studies on the BrARGOS gene, which regulates the ANT gene 10 and may also promote the transcription of AtEXP10 in transgenic Arabidopsis 53 . This is probably because meristematic competence is disrupted locally, cell division gradually ceases and differentiation begins with the expansion of the postmitotic cells 19 . In addition, cell proliferation is also coupled with a limited amount of cell expansion during the proliferation phase 54 . Therefore, we speculate that the extensin and expansin proteins promote cell expansion following a significant increase of cell proliferation. The modest reduction in cell size may compensate the increase in cell number 19 .
BrANT-1 might regulate leaf senescence. Leaf senescence constitutes the final phase of leaf development and is a highly complex but genetically programmed process involving the expression of many senescence-associated genes (SAGs) 55,56 . Our study shows that, compared with the wild type, leaf senescence in the 35S-BrANT-1 transgenic line was delayed, which is consistent with that in Arabidopsis over-expressing the AtANT gene 57 . Additionally, a number of SAGs such as AtNAC2 (AT5G39610) 58 , AtSAUR36(AT2G45210) 45 , and AtSAG13(AT2G29350) 59 were prominently down-regulated in the 35S-BrANT-1 line. AtSAUR36 (or SAG201), a member of the early auxin-responsive gene family, was remarkably up-regulated during leaf senescence. In Arabidopsis, a saur36 knockout line shows a delayed leaf senescence phenotype, but the transgenic plant overexpressing SAUR36 displays an opposite phenotype 45 . In addition, SAG13 as an early senescence marker is strongly induced before visible yellowing 60 . Accordingly, we speculate that BrANT-1, similar to the AtANT gene, is one of the negative factors that prevent premature senescence.

BrANT-1 positively increased the stomatal number in Arabidopsis.
Stomata is an important part of the epidermal tissues of leaves in plants, which controls gas exchange by paired subsidiary cells, and participates in the global carbon cycle 61 . Stomatal development in leaf is positively regulated by signaling factor STOMAGEN (At4g12970) through interacting with cell-surface receptor TOO MANY MOUTHS (TMM) 62 . In this study, it was found that the number of stoma significantly increased in the 35S-BrANT-1 line compared with the WT. Besides, the expression level of STOMAGEN (At4g12970) was significantly up-regulated in the transgenic line. The overexpression of STOMAGEN was found on many agminate stomata in mature leaves of Arabidopsis 62 . However, the homologous gene of AtANT regulating the stomatal development has not been reported. Additionally, previous studies have shown that the stomatal density in Arabidopsis influences the leaf photosynthetic capacity through regulating gas diffusion 63 . In this study, we assessed the leaf photosynthetic capacity of the BrANT-1-overexpressing transgenic and wild type lines. The result indicated that, with the increase of the stomatal number, the stomatal conductance (Gs) and transpiration rate (Tr) of mature leaves were significantly increased in the 35S-BrANT-1 transgenic plants. However, the net photosynthetic rate (Pn) was only slightly increased in the 35S-BrANT-1 plants, probably due to the low intracellular CO 2 concentration (Ci). Severe stomatal patchiness results in the underestimation of Pn due to the lack of uniform Ci in the leaf 64 . Taken together, these results indicated that BrANT-1 might be involved in regulation of the development of stomata by enhancing the expression level of STOMAGEN (At4g12970).
In summary, we identified three BrANT and six BrAIL proteins in Chinese cabbage, which can be classified into four subgroups. Phylogenetic analysis showed that the AIL genes of Chinese cabbage had high sequence similarity with those of Arabidopsis. Furthermore, multiple sequence alignment suggests that the nine genes belonged to the euANT subgroup according to the conserved motifs in Arabidopsis. Finally, ectopic expression of BrANT-1 in Arabidopsis controlled the organ size by regulating the cell number. BrANT-1 regulated the stomatal density and leaf senescence by increasing the expression of STOMAGEN and reducing the expression of SAGs. Taken together, these results not only enhance the understanding of the role of the AIL genes in controlling the organ size and other tissue growth, but also provide a promising tactic for Chinese cabbage molecular breeding program.

Methods
Identification and analysis of the AIL genes in Chinese cabbage. The gene and amino acid sequences of the AIL family members were confirmed according to the genome of the B. rapa line Chiifu (http:// brassicadb.org). The difference of the AIL genes was analyzed using DNAMAN 6.0.40 (Lynnon Biosoft, USA). Gene Structure Display Server (GSDS) (http://gsds.cbi.pku.edu.cn/) was used to perform the intron/exon structure analysis. Phylogenetic trees were constructed based on the amino acid sequences using the neighbor-joining method by MEGA 5.0 software 65 . The physicochemical properties of the AIL proteins were calculated by using the ProtParam tool (http://web.expasy.org/protparam/). MEME (meme.nbcr.net/meme/intro.html) was used to analyze the conserved motifs of the AIL proteins in Chinese cabbage 66 . The AP2 domain sequences of the AIL proteins from Chinese cabbage were identified with SMART (http://smart.embl-heidelberg.de/).

Plant materials, growth conditions and plant phytohormone treatments. The wild-type and
35S-BrANT-1 transgenic Arabidopsis plants were pre-treated for 3 days at 4 °C under dark condition before transferring to pots with nutrient media (Sheng Xiang Agricultural Science and Technology Co, China; peat: vermiculite: pearlite = 1: 1: 1). The plants were cultivated in a growth room with a continuous artificial light period of 16 h and a dark period of 8 h, and a constant temperature between 19-23 °C. The Chinese cabbage line "Guangdongzao" was used for all stress experiments, and the growth condition was the same as that for Arabidopsis.
Chinese cabbage young seedlings at the four-leafed stage (21 days after sowing) were used for the phytohormone treatments, during which, the plant leaves were treated with 100 μM naphthaleneacetic acid or distilled water (DW), respectively. Plant leaves were harvested after 0, 1 and 3 h of the phytohormone treatment.  promoter was used instead of the intrinsic promoter. The product of polymerase chain reaction (PCR) was cloned into the binary vector pCAMBIA2300-35SOCS using the XbalI and SalI restriction sites. The constructed vector was used to transform Arabidopsis plants using the planta Agrobacterium-mediated method 67  Sequencing and data processing. The matured leaves of the WT and 35S-BrANT-1 transgenic Arabidopsis plants (40 days after sowing) were used for RNA-seq. Each line was biologically repeated three times. The sequencing of six cDNA libraries was performed at Beijing Genomics Institute (BGI, Shenzhen, China) using an Illumina HiSeqTM 2000 sequencing platform (Illumina Inc., San Diego, CA, USA). Clean reads were obtained from the raw reads that were clipped by abandoning the adaptor sequences and low-quality reads (reads with >10% ambiguous "N" bases or reads in which >50% of the bases had a Quality-score ≤ 5). The clean reads were mapped to the reference genome using a rapid short-read mapping program, namely SOAP aligner/soap2 68 . More than two mismatches were abandoned in the sequence alignment. The quality of sequencing was controlled by quality assessment of reads, statistics of alignment, sequencing saturation analysis and randomness assessments.

Identification and functional annotation of differentially expressed genes. The reads per kb per
Million reads (RPKM) method 69 was used to calculate the expression level of each unigene. Therefore, the RPKM values can be directly used for comparing the difference of gene expression among the WT and 35S-BrANT-1 lines. The DEGs from these two lines (six cDNA libraries) were enriched for further analysis according to the standard with false discovery rate (FDR) ≤0.01, and the absolute value of log2 ratio ≥ 1.

Quantitative RT-PCR (qRT-PCR) analysis.
To confirm the quality of the sequencing data and distinguish the expression level, the Chinese cabbage AIL genes were subjected to qRT-PCR analysis. qRT-PCR was performed under the following conditions: 94 °C for 2 min, followed by 45 cycles of reaction (94 °C for 20 s, followed by 60 °C for 34 s). The actin gene was used as a constitutive expression control. qRT-PCR was performed on an IQ5 Real-Time PCR System (BIO-RAD, Hercules, CA, USA). The specific primers for qRT-PCR (Supplementary  Table S8 Data availability. The RNA-seq raw data are deposited in the Sequence Read Archive (SRA) under the number "SPR136061". Phenotype datasets are available in this article and its supplementary files.