2.1 Plant materials
The seeds of C.pilosula No. 1 ('We1') and C.pilosula No. 3 ('We3') grown in Weiyuan County, Dingxi City were germinated in laboratory and transplanted to the Experimental Park of Northwest Normal University. ' We3 ' was produced by ' We1 ' radiation mutagenesis breeding (Su 2012; Wang and Song 2015). An annual C.pilosula var. Modesta ('Wn') growing in Wenxian County of Longnan City was transplanted into the Experimental Park of Northwest Normal University. Female, male, petal and leave of 'We1-1', 'We1-2', 'We1-6', 'Wn2', 'Wn3', 'Wn5' plants were collected at bud stage, which was quickly frozen in liquid nitrogen, and then stored with dry ice and sent to Beijing Nohe Zhiyuan Technology Co., Ltd. for RNA extraction and related sequencing.
2.2 Natural pollination
During the flowering period, 20 flowers with the same growth condition were selected and marked with labels, and the seed setting rate was counted after the seeds were mature.
2.3 Stigma dehiscent stage pollination
Six flowers were chosen from a specific section of the designated plant for the purpose of being emasculated and bagged prior to flowering. Following stigma dehiscence, the flowers were subsequently pollinated. Pollen was obtained from either the same or different flowers that had released pollen on the day of flowering.
2.4 Pollination during different flowering cycles
A total of 50 flowers were selected from the same part of the plant two days before flowering, and 5 flowers were de-pollinated every day. After pollination, they were bagged for self-pollination, and the pollen was all selected from the pollen released by the flowers of the same plant on the day of flowering.
2.5 Observation of single flowering stage of Codonopsis pilosula
Three flowers of 'We3' plants two days before flowering were selected for hanging and marking, and photos were taken from 8:00 to 9:00 every morning to record the status of stigma and anther until the 7th day after flowering.
2.6 Determination of pollen viability
Acetic magenta method (Yu et al. 2013).
2.7 Determination of pollen germination rate
Anthers were collected from 2 days before flowering to 7 days after flowering (-2 to 5) at the full flowering stage (July to August) and stored in a 1.5 mL centrifuge tube in a seed storage closet at 4℃. The sucrose solution was set at 5 concentrations of 2%, 5%, 8%, 10% and 15%, and counted once every 1 h for 6 h to investigate the optimal sucrose solution and germination time of Codonopsis pilosula pollen.
2.8 Determination of stigma acceptability
Stigma receptivity was determined using the benzidine-hydrogen peroxide method (Duan et al. 2019; Gong et al. 2021). The flowers were collected from 2 days before flowering to 7 days after flowering, and the stigma was removed and immersed in a concave slide containing the reaction liquid (benzidine: hydrogen peroxide: water = 4:11:12). If the stigma were permeable, the surrounding stigma would appear blue and a large number of bubbles would appear. The receptivity of the stigma is judged by the depth of dyeing, the speed of bubble formation and the number of bubbles.
2.9 Microscopic observation of pollen tube growth
12 flowers of 'We' plant with the same growth state were selected, castrated and bagged. Pollination was pollinated on the day of complete stigma dehiscence (the third day after flowering), with 6 outcrosses and 6 selfcrosses. 24 hours after pollination, the stigma and style of the flowers were completely removed and placed in the Carnot fixative. The style fixed for 24 hours and rehydrated with 80%, 50%, 30% ethanol and ddH2O. The samples were softened in 8M NaOH solution for 24 h and rinsed 3 times with ddH2O (Sogo and Tobe 2006). The samples were transferred to aniline blue dye solution for dark staining overnight (Regan and Moffatt 1990). After complete staining, the style and stigma were observed under a fluorescence microscope.
2.10 Sequencing and Raw Data Processing
Using the Illumina high-throughput sequencing of cDNA library sequencing platform get a lot of high quality reads, referred to as the original data (raw data) (Luff et al. 2019). The original data obtained were evaluated and cut by fastqc and trimmomatic, and the splice, sample identification sequence, low-quality Reads and Reads with more N bases were removed to obtain high-quality Clean reads (Zeng et al. 2019). Software Trinity was used to assemble the obtained Clean reads (Grabherr et al. 2011). After assembly, the script perl 5.22 was used to deredundant and cluster the assembled transcripts, and the unigene database of this species was obtained. The transcriptome sequencing library was further evaluated for quality (Gu et al. 2017). After qualified evaluation, a series of analysis were carried out for expression level and gene structure.
2.11 Gene expression analysis
Reads sequenced were compared with Unigene library using Bowtie1.3.0 (Langmead et al. 2009). According to the comparison results, expression level was estimated using RSEM 1.2.19 software (Li and Dewey 2011), and the FPKM value was used to represent the expression abundance of corresponding Unigene.
2.12 HMMER search
Hmmer 3.0 (http://www.hmmer.org/) is an application package that uses a statistical Model or Hidden Markov Model (HMM) to search for gene sequences (Eddy 2008). HMM search was conducted on the protein data in the ' We 'transcriptome, and the expected value (E value) was set to be less than 10 − 5 to obtain the candidate sporulation self-incompatibility factors SRK and SCR in ' We ', Self-incompatibility factors pS and sS of papiferae type and self-incompatibility factors T2 nuclease and SLF/SLFL/FBK of S-RNase type.
2.13 Blast search
The candidate T2 nucleases and SLF/SLFL/FBK obtained by HMMER search were used as queries to perform TBLASTN search on the Unigene of ' Wn ' and the genome of Radix Codonopsis Lanceolatae and Platycodon grandiflorus., and then get their candidate S-RNase and SLF/SLFL/FBK fragment. The Fgenesh tool (http://linux1.softberry.com/berry.phtml?topic= fgenesh&group = programs&subgroup = gfind) in Softberry website was further used to predict genes.
2.14 Phylogenetic tree construction, protein structure and cell localization analysis
The candidate style S factor and pollen S factor were compared with the identified style S factor and pollen S factor in Solanaceae, Rosaceae and Scrophulariaceae by muscle (Edgar 2004) of MEGA7.0 (https://www.megasoftware.net/) (Kumar et al. 2016). The phylogenetic tree was constructed by Neighbor-Joining method and the Bootstrap value was set to 1000 (Stamatakis 2014).
The number, molecular weight, theoretical isoelectric point, hydrophobicity and other physicochemical properties of amino acids were predicted and analyzed by using Protparam (https://web.expasy.org/protparam/), an online tool provided by ExPaSy (Chen and Liu 2013). The online tool PSORT (https://psort.hgc.jp/) was used for subcellular localization of pollen S factor and style S factor (Meinken et al. 2014). The secondary structure of pollen S factor and style S factor was predicted by online SOPMA program (Chen and Liu 2013).
2.15 RT-PCR and sequencing
To ensure the accuracy of transcriptome sequencing, the total RNA extracted from the samples was reverse transcribed using TaKaRa reverse transcription Kit PrimeScriptTM II 1st Strand cDNA Synthesis Kit, followed by cDNA as a template. The style S gene and pollen S gene were screened by reverse transcription PCR in 6 tissue 'We1-1', 'We1-2', 'We1-6', 'We3-1', 'We3-2', 'We3-5', Cpactin gene was obtained by blast search of Cpactin gene in Arabidopsis Thaliana (Wang et al. 2019).
Table 1
Primer sequence of RT-PCR
Gene | Primer | Bases 5’→3’ | Bases |
We66712 | 66712_F | CCAAAAATGTCGGCATGGGG | 20 |
66712_R | ACATGTGAGTGAGAGAGCGAG | 21 |
We57191 | 57191_F | CCTTCATGAAAGATGCCACTCC | 22 |
57191_R | TCAGCCCCAGAGCTCAGATT | 20 |
Cpactin | ACT_F | GTGGWCGTACMACHGGTATTGTG | 23 |
ACT_R | AGACCTAACCTMACCKCCAAG | 21 |