Transcriptome profiling provides new insights into florets number difference of inflorescence in lavandula angustifolia

Background Lavender flowers essential oil had been for a variety of therapeutic and cosmetic purposes, and had been popular for centuries. The previous studies of lavender mainly focused on essential oil composition and extraction methods, ignoring the factors which affected the production of essential oils, such as the floret number. This study aims to get a deeper insight into florets number difference mechanism. Results Hormone profile showed positive correlation between ABA content and the number of florets while IAA was negatively correlated. RNA-Seq results showed that 2848 differentially expressed genes screened by comparing different florets samples in one plant. By analyzing dynamic changes of differentially expressed genes, many potentially interesting genes were identified that encoded putative regulators or key components of ABA metabolism and signaling transduction, such as NCED , PYL , PP2C , SnRK2 . These genes were highlighted to reveal their importance in regulation of florets numbers. concentrations difference


Conclusions 1. The different ABA concentrations lead to florets difference in the
Lavandula angustifolia "JX-2" clusters; 2. ABA may affect the florets number by regulating IAA transport and accumulation.
The results will be useful for a better understanding of the molecular mechanism on florets number difference that could be laid the foundation for molecular breeding of muti-flortes varieties.

Background
Lavenders belonging to the family labiatae (Lamiacae) has been used as dried flowers and extracting essential oil which is being used for a variety of therapeutic and cosmetic purposes for so long time [1]. This essential oil continued to be as popular till today as it had been down the centuries past. Lavender essential oil is extracted from both the flowers and foliage by steam distillation with varying chemical composition but most of the aromatic oil is being derived from the flowers [2]. The previous studies mainly focused on lavender essential oil composition and extraction methods ignoring factors which affected the production of essential oils, such as the floret number.
The type of Lavandula angustigolia inflorescence is verticillaster where numbers of floret form a false whorl at the node. Some of whorls are borne on main floral axis forming a spike. One whorl as example, the main floral axis gives rise to two lateral branches and the succeeding branches bear only one branch each on alternate sides. Practice had shown that the number of floret and branch was positively correlated. The axillary meristems had the same developmental potential as the primary inflorescence primordium playing an important role in regulating floret number, but cannot go on infinitely meristem [3,4]. Axillary meristems develop into flower buds or form a new meristem at their base [3]. The flexibility in axillary meristem activity made possible substantial variation in inflorescence architecture, allowing the Lavandula angustifolia to adapt to the prevailing environmental conditions.
The inflorescence pattern of lavender resembled with the developmental process of plant stems, all them relied on constant branching. The previous studies showed that the shoot architecture was regulated by light, nutrition and density [5,6].
These environmental signals were likely to be relayed through the action of plant hormones, which of particular importance were auxin, cytokinin, strigolactone, ABA [4,7]. Auxin was the first hormone to be linked to regulation of shoot branching and it had been in the spotlight for more than 100 years. Cytokinin played a fundamental role in regulation of apical meristem size [8,9,10]. Strigolactones blocked the transport of IAA through the Pin protein to regulate branching [11]. ABA likewise inhibited the expression of genes associated with the bud autonomous auxin pathway and hindered the accumulation of IAA in the bud [12].
The biggest difference between the formation of inflorescences and stems was the termination signals; stems relied on dormant buds while inflorescences were dependent on flower. The difference of the termination signal may be caused by the difference of regulation mechanism.
In Lavandula angustifolia 'Jx-2' cluster, all inflorescences had five whole flowers but the floret number on different inflorescence was from 30 to 70 (about 2.5-fold difference) which directly affected the production of essential oils. However, the reason for florets number difference was poorly understood. In this study, we aimed to understand florets number difference in "Lavandula angustifolia 'Jx-2' by addressing the following issues; 1. Which hormones regulate the number of florets? 2. What are the key genes in hormonal metabolic pathways; 3. Possible regulatory mechanism of the difference in the number of florets.

Inflorescence characteristics of lavender
The botanical characteristics of Lavender were dwarf shrubs, branch with long inflorescence and short leafy shoots. One three-year-old plant had about 100 flowering branches (Fig. 2a). Each flowering shoots borne five whorl flower (Fig. 2c).
Take one side of one round flower for example, the 2 nd flower had symmetry but the flowers begun to be inward hyperplasia from the third flower (Fig.2 b). The inflorescence of lavender "JX-2" was 30-70 flowers per shoots, 6-14 florets per round, (Fig. 2c).

Hormone contents of different florets samples
Hormone content of different parts in LaF6, LaF10 and LaF14 samples were measured by ELISA (Table 1). Hormones content of florets and flower axis were no significantly difference. In leaves and shoots, GA and ZR change trends was similar, while ABA and IAA content of leaves in three samples had the biggest difference (Table 1). ABA content was the lowest in LaF6 sample, followed by that in the LaF10, and highest in the LaF14. However, as far as IAA is concerned, leaves from LaF6 sample had highest content followed by that in the LaF10 and lowest in the LaF14.
Further evidence by GC-MS supported this result that ABA content of leaves from sample with more floret numbers was higher, but IAA content of leaves from sample with more floret numbers was lower (Fig. 3).   (Table 3). E-value indicated the extent of sequence homology and E-value is smaller, the sequence homology is higher. For E-value distribution of unigenes blastx hits in the Nr database which had the largest number of annotated unigenes, 37% homolog sequences ranged between 1E -5 to 1E -60 , while 63% sequence had a threshold E-value less than 1E -60 that showed strong homology (Fig.4a). Species distribution of the first blastx hits of each unigene in the Nr database indicated that Sesamum indicum provided the best blastx matches with 56.8% unigenes in lavender, and Erythranthe guttata was the second closest species, which had 13.8% homology with lavender ( Fig. 4b).
To gain insight into functions of the annotated genes from the macro level, GO functional classification was performed. A total of 55 GO terms were categorized into three domains: biological process, cellular component and molecular function

Identification and KEGG enrichment analysis of DEGs
With qvalue < 0.005 and an absolute log 2 fold change > 1, a total of 2848 genes were found to be differently expressed among the groups LaF6 vs LaF10, LaF6 vs LaF14 and LaF10 vs LaF14. In the group LaF6 vs LaF10, 1089 genes were found, and 522 genes were down-regulated and 567 genes were up-regulated. In the LaF6 vs LaF14 group, 1344 genes were found, of which 741 genes were up-regulated and 603 genes were down-regulated. In LaF10 vs LaF14, 320 genes were up-regulated and 299 genes were down-regulated (Fig. 5a). These results indicated that both upand down-regulation of gene expression occurred, and the transcript abundance of genes changed dynamically among the difference florets samples.
To validate DEGs expression profiling obtained by RNA-seq, 7 DEGs in LaF10 vs LaF14 group, 17 DEGs in LaF6 vs LaF14 group and 12 DEGs in LaF6 vs LaF10 group with higher or lower expression levels were selected for qRT-PCR analysis (Fig. 5bd).
The results obtained from the qRT-PCR and RNA-Seq indicated that their correlation coefficients (R 2 ) ranged from 0.94 to 0.97, with a very significant level (P<0.01).
Thus, Data from RNA-Seq in this study were available.

DEGs associated with ABA metabolism and signal transduction
Hormone content showed that the difference in the number of florets was related to ABA anabolism in the flowering shoots and leaf Table 1. Fig. 3). Here, we also found that carotenoids biosynthesis pathway including ABA metabolism was enriched among the difference florets samples (Table 4). Thus, the DEGs involved in the entire ABA metabolism and signaling transduction were focused.
The expression of BCH gene (CL-10130. 19560) in LaF6 sample was lower than that of LaF10, and that in LaF14 sample was the highest (Fig. 6c) 6c). Above all, the expression levels of five ABA biosynthetic genes except one were higher in sample with higher ABA content.
In the ABA signaling transduction (Fig. 6b), one PYL gene, nine PP2C genes, three SnRK2 genes and two ABF genes were found in response to florets difference (Fig.   6c). The expression level of PYL gene declined with florets number increasing. The nine PP2C genes all had a similar expression profile that had highest expression in LaF14, followed by that in LaF10 and lowest expression level in LaF6. However, there were complicated expression patterns for SnRK2 and ABF genes. Some of genes encoding homologous products were upregulated, but others were downregulated in the same sample (Fig. 6c).

Discussion
Increasing the florets number is a desired target of lavender breeding. To explore the differences in the number of florets intrinsic mechanism or influencing factors were the basis for cultivating multi-flowering varieties. 2848 DEGs were revealed that 10 metabolic pathways and signal transduction pathways were significantly enriched (Table 4). According to these results, we tried to understand Previous studies showed that axillary buds outgrowth responded to different light quality, ABA regulated stem branching by affecting the accumulation of IAA in the buds [21,22]. In our experiment, the ABA contents gradually increased with the florets number increases (Table.1, Fig. 3), which could be caused by expression of genes involved in the ABA biosynthesis pathway. Four ABA biosynthetic genes expressed increasingly with the florets number increases. The NCED3 gene was a key gene in ABA metabolism process, and high expression of AtNCED3 or high homology BgNCED1 in Arabidopsis can significantly increase the level of endogenous ABA [23,24,25,26]. Our result also indicated that the expression level of three NCED3 or NCED1 genes was accord with endogenous ABA content. It was speculated that ABA was involved in florets number regulation in Lavandula angustifolia.
Like ABA metabolic pathways, some components of ABA signaling transduction were also different among the different florets samples, such as the protein PYR/PYL, PP2C, SnRK2 and ABF. These core component proteins constituted a dual negative regulatory system that regulates ABA signaling [27,28,29]. The previous studies showed that seed germination, vegetative growth, gene expression and stomatal movement were more sensitive to ABA in the plant in which PYL9/RCAR1,PYL5/RCAR8 or PYL8/RCAR3 genes over expressed [28,29]. In the experiment, PYR/PYL, PP2C, SnRK2 and ABF proteins aggregated differential genes ( Fig. 6). Molecular mechanisms further showed that ABA is involved in the regulation of florets number in Lavandula angustifolia inflorescence.
Possible mechanism of ABA regulating the number of florets ABA has been known as the plant stress hormone; however, it also plays a role in non-stress-related plant development and growth processes. ABA may turn out to be one of the earliest upstream factors which shoot branching response to environment signal [30]. ABA was one of the first upstream factors regulating apical dominance responses to the R: FR ratio [22,31]. Previous studies also showed that ABA and auxin are interaction each other in many developmental processes, especially, in part of regulating root growth [32,33,34]. ABA antagonizes to inhibit primary root growth by promoting production of reactive oxygen species [33,35]. ABA also regulate lateral root formation and elongation by suppressed the expression of PIN-FORMED1 through ABA-insensitive (ABI4) [34]. Xing et al. [36] showed that PYL9 may regulate auxin-responsive genes in vivo through direct interaction with MYB77 and MYB44. The prolonged quiescent phase of the pyl8-pyl9 double mutant was reversed by exogenous IAA.
In the study, there are different genes in the protein PYL, and the gene expression is down regulation in the LaF6 compared to LaF14 florets sample. The IAA contents change trends in the LaF6, LaF10, LaF14 samples were gradient decrease, while the ABA contents were gradient increase. The IAA and ABA contents in the LaF6 were 369.2, 452.6 ng/g, and in the LaF14 sample was 90.3, 4534.5 ng/g (Fig. 3).
The characteristics of verticillaster inflorescence indicated that the key factors affecting the number of "JX-2" florets were the axillary meristems and their breaching (Fig. 2). The studies of root growth had confirmed that ABA and IAA interaction regulated the formation and growth of lateral root [37]. Comprehensive analysis of previous research and our experimental results, we speculated that ABA may affect the IAA transport or accumulation which leads to axillary meristems increase and the number florets difference.

Study site
The inflorescence samples of experiment were collected from Qing Shui He town and long-term annual precipitation is 460 mm. Lavender essential oil production in this area account of 95 % of the national total, and known as the hometown of lavender in China [13].

Plant materials
In the lavender planting area of 65 th group farm in Xin Jiang, 667m 2 was selected as were measured. Firstly, hormones were extracted and purified according to modified method described by [14]. Secondly, the extracted hormone elution was analyzed according the ELISA method described by [15]. LaF6, LaF10 and LaF14 samples were measured more accurately using modified GS-MS according to the previous described method [16].

De novo assembly and functional annotation
Adaptor sequences, reads with unknown sequences and low quality reads were removed and the clean reads were then assembled using Trinity software [17]. Data obtained for each sample were separately assembled and the assembly sequences were called unigenes. The unigenes from all samples were further subjected to sequence splicing and redundancy removal with sequence clustering software to acquire non-redundant unigenes as long as possible. These unigene sequences were aligned and annotated to protein databases like Nr, Swiss-Prot, PFAM, COG [18], GO [19], KO [20], and nucleotide database Nt with a threshold of E < 10 -5 . The best aligning results were used to decide sequence direction of unigenes. If results from different databases conflicted with each other, a priority order of Nr, Swiss-Prot, KEGG and COG should be followed when deciding sequence direction of unigenes.

Analysis of differentially expressed genes (DEGs)
Expression of the unigenes was calculated using the FPKM (Fragments Per Kb per Million reads) method which eliminated the influence of lengths and sequencing discrepancies of different genes on the gene expression calculatons. DEGs were selected on condition of qvalue < 0.005 and an absolute value of log 2 foldchange > 1.

Quantitative real-time PCR (qRT-PCR) validation
Total RNA was separately extracted from leaves for LaF6, LaF10 and LaF14 samples as described earlier. First-strand cDNA was synthesized using the PrimeScript® RT reagent kit (Takara) according to the manufacturer's instructions. The primers for the selected genes were designed using Primer Premier 5.0 software and listed in

Ethics approval and consent to participate
Plant samples used in the study were from field cultivation and not collected from national park or natural reserve. According to national and local legislation, no specific permission was required for collecting plant materials.

Consent for publication
Not applicable

Availability of data and materials
The datasets used during the current study are available from the corresponding author on reasonable request.    Table.S1 .xlsx