PECTATE LYASE-LIKE 9 from Brassica campestris is associated with intine formation

Plant Science 229 (2014) 66–75

Contents lists available at ScienceDirect

Plant Science journal homepage: www.elsevier.com/locate/plantsci

PECTATE LYASE-LIKE 9 from Brassica campestris is associated with intine formation Jingjing Jiang a,b , Lina Yao a , Youjian Yu a , Ying Liang a , Jianxia Jiang a , Nenghui Ye b , Ying Miao c , Jiashu Cao a,∗ a

Laboratory of Cell & Molecular Biology, Institute of Vegetable Science, Zhejiang University, Hangzhou 310058, China State Key Lab of Agrobiotechnology Shenzhen Base, Shenzhen Research Institute, The Chinese University of Hong Kong, Shenzhen 518057, China c The Center of Molecular Cell and Systems Biology, College of Life Science, Fujian Agriculture and Forestry University, Fuzhou 350002, China b

a r t i c l e

i n f o

Article history: Received 29 April 2014 Received in revised form 13 August 2014 Accepted 15 August 2014 Available online 23 August 2014 Keywords: Brassica campestris Chinese cabbage Intine Pectate lyase PLL Pollen development

a b s t r a c t Brassica campestris pectate lyase-like 9 (BcPLL9) was previously identified as a differentially expressed gene both in buds during late pollen developmental stage and in pistils during fertilization in Chinese cabbage. To characterize the gene’s function, antisense-RNA lines of BcPLL9 (bcpll9) were constructed in Chinese cabbage. Self- and cross-fertilization experiments harvested half seed yields when bcpll9 lines were used as pollen donors. In vivo and in vitro pollen germination assays showed that nearly half of the pollen tubes in bcpll9 were irregular with shorter length and uneven surface. Aniline blue staining identified abnormal accumulation of a specific bright blue unknown material in the bcpll9 pollen portion. Scanning electron microscopy observation verified the abnormal outthrust material to be near the pollen germinal furrows. Transmission electron microscopy observation revealed the internal endintine layer was overdeveloped and predominantly occupied the intine. This abnormally formed intine likely induced the wavy structure and growth arrest of the pollen tube in half of the bcpll9 pollen grains, which resulted in less seed yields. Collectively, this study presented a novel PLL gene that has an important function in B. campestris intine formation. © 2014 Elsevier Ireland Ltd. All rights reserved.

1. Introduction In flowering plants, microspores undergo meiosis to produce a tetrad of four haploid microspores, each of which will develop into a pollen grain. During pollen development, several layers of pollen walls are synthesized and degraded in a specific order and in turn produce a mature bi-layered structure pollen wall that includes the intine and exine. This pattern of pollen development has been clearly identified along with the identification of various important genes involving in male gametogenesis [1–3] and pollen wall development processes [4]. Pectin is a class of polysaccharide polymer typically characterized by the linear backbone of the ␣-1,4-linked galacturonic acid (GalA) residues, which include homogalacturonan (HG), rhamnogalacturonan-I (RG-I) and rhamnogalacturonan-II (RG-II)

∗ Corresponding author. Tel.: +86 571 88982188; fax: +86 571 88982188;. E-mail addresses: [email protected] (J. Jiang), [email protected] (L. Yao), [email protected] (Y. Yu), [email protected] (Y. Liang), [email protected] (J. Jiang), [email protected] (N. Ye), [email protected] (Y. Miao), [email protected] (J. Cao). http://dx.doi.org/10.1016/j.plantsci.2014.08.008 0168-9452/© 2014 Elsevier Ireland Ltd. All rights reserved.

[5]. Pectinases target methyl-esterified HG to yield substrates for polygalacturonases (PGs) and pectate lyases (PLs, or pectate transeliminases) to cleave GalA backbones. PLs cleave ␣-1,4-glycosidic linkages in HG to produce unsaturated oligosaccharides [6–8]. PLs have been extensively studied in several plant pathogenic microorganisms. Erwinia chrysanthemi, for example, is an extracellular causal agent in soft root disease that affects a wide range of plant species [9]. In plants, pectate lyase-like (PLL) genes Late Anther Tomato 56 (LAT56) and LAT59 that present strong sequence homology with the PelC isoform of bacterial PLs were first isolated to be expressed at maximal levels in mature anthers in tomatoes (Solanum lycopersicum) [10]. Since then, several other PLL genes were identified as expressed in pistils, tracheary elements, ripening fruits and latex [11–19]. PLL genes were also expressed in response to plant hormones, environmental stress, cell separation and pathogen infections [20–22]. PLL genes are abundant in plants, including 26 PLL homologous genes in Arabidopsis and 12 in rice (Oryza sativa) [20]. These genes are also abundant in tomato, tobacco, alfalfa and Chinese cabbage (Brassica campestris ssp. chinensis) [10,12,14,23]. Out of the 26 Arabidopsis PLL genes, 14 are expressed in the pollen [20]. Our previous study also revealed that two Chinese cabbage PLL genes

J. Jiang et al. / Plant Science 229 (2014) 66–75

were exclusively expressed in the pollen and pistil [4]. Researchers have proposed that the tomato pollen expressed genes, LAT56 and LAT59, are related to a requirement for pectin degradation during pollen tube growth [12,14]. In addition, numerous promoter activities of the PLL genes are similar to those exhibited by multiple PGs, which suggests a close functional association among PLLs and PGs, particularly in the digestion of the pollen grain cell wall prior to germination and during pollen tube growth [24–26]. However, to our knowledge, no direct experimental evidence has been presented regarding the involvement of PLL genes in pollen development or in remodelling the pollen tube wall. Little is known about the spatial and temporal regulation of PLL genes and their functions during reproductive growth in plants. In this study, we provide evidences that a novel PLL gene, BcPLL9, has an important function in pollen wall development. Ectopic expression of BcPLL9 antisense gene in B. campestris ssp. chinensis resulted in abnormal intine formation during pollen wall development as well as pollen tube growth retardation, partial male sterility and reduced seed set.

2. Materials and methods 2.1. Plant material and sampling A Chinese cabbage pak-choi ‘Aijiaohuang’ (B. campestris ssp. chinensis syn. B. rapa ssp. chinensis cv. Aijiaohuang) GMS AB line named ‘Bajh97-01A/B’ was used in this study. ‘Bcajh97-01A’ completely failed in pollen formation but several of its tissues were the same as those of ‘Bcajh97-01B’ [27–29]. ‘Bajh97-01A/B’ was developed through continuous backcrossing within the population for more than ten generations and was the sibling (sister) line segregated with homozygous male sterile plants (Bcajh 97-01A, genotype: msms) and heterozygous male fertile plants (Bcajh 97-01B, genotype: Msms) in a 1:1 ratio. The plants were grown at natural light conditions (12 h light/12 h dark) at 14–25 ◦ C during early April in Hangzhou, Zhejiang, China.

2.2. Cloning and analysis of Chinese cabbage BcPLL9 gene DNA was isolated from the leaves of Bajh97-01B plants via the CTAB method. Total RNA was extracted from inflorescence of Bajh97-01B by using Trizol reagent (Invitrogene, USA) treated with DNase (TaKaRa, Japan) and reverse-transcribed into first strand complementary DNA by using the PrimeScript 1st strand cDNA synthesis kit (TaKaRa, Japan) from 1 ␮g of total RNA. The DNA and cDNA sequences of At2g02720 (AtPLL9) were downloaded from the Arabidopsis website (http://www.arabidopsis.org/) and used for basic local alignment search in the Brassica database (http://brassicadb.org/brad/) with the basic local alignment search tool (BLAST) with default settings. The BAC clone AC189452 showed the highest identity. Second, the full length of the cDNA sequence, including 5 -untranslated region (UTR) and 3 UTR, was amplified with the forward primer, P9-fwd, and the reverse primer, P9-rev. The DNA sequence of the gene was cloned with the forward primer, PD9-fwd, and the reverse primer, PD9-rev. Gene-specific primers were designed using the Primer Premier 5.0 software and are listed in Table S1. The PCR products were sub-cloned into a pGEM-T Easy vector (Promega, USA) and sequenced. The DNAstar software was used to analyse the gene structure, including the largest open reading frame (ORF) and the character of the deduced amino acids. The secondary structure prediction of the protein was performed in the NPSA-MLRC website (http://npsa-pbil.ibcp.fr). Motif prediction was carried out in the SignalP-4.0 (http://www.cbs.dtu.dk/services/SignalP/)

67

and Pfam (http://pfam.sanger.ac.uk/) websites. Default parameters were used. 2.3. Quantitative RT-PCR (qRT-PCR) analysis First-strand cDNA was reverse-transcribed using the PrimeScript 1st strand cDNA synthesis kit (TaKaRa, Japan) as described above. A sample of cDNA (100 ng) was subjected to qRT-PCR in a final volume of 20 ␮L by using a SYBR Premix Ex Taq Kit (TaKaRa, Japan) with CFX96 Real Time System (Bio-Rad, France). Data was normalized to the expression levels of the internal control gene, BcUBC10 [30]. In addition, the end-time PCR products were separated on agarose gel and sequenced to verify the amplification product. Three biological and three technical replicates for each sample were performed. Relative expression levels were calibrated via the 2–Ct method [31]. The primers (P9q-fwd and P9q-rev) are listed in Table S1. Error bars represent ±SE, **P < 0.01, by 2 test. 2.4. In situ hybridization Flower buds at uninucleate and mature pollen stages, as well as pollinated pistils at 24 h after pollination (hap), were fixed in 4% formaldehyde–PBS solution, dehydrated with 30% sucrose solution, transited with 30% sucrose:optimal cutting temperature (OCT) compound (1:1), and then embedded in pure OCT compound. The embedded samples were sectioned at 14–16 ␮m in thickness with Thermo Shandon Finesse 32 freezing microtome (Thermo Scientific, USA) and were hybridized to digoxigenin (DIG)-labelled RNA probes. A 190-bp fragment was amplified with a primer pair (P9ifwd and P9i-rev, listed in Table S1) specific to BcPLL9. Sense and antisense probes were synthesized and labelled using a DIG RNA Labelling Kit (Roche Diagnostics, New Jersey, USA). Flower buds at uninucleate and mature pollen stages were used in the present experiment [28]. The male sterile line was used as the female parent in conducting hand-pollination experiments to prevent extraneous pollen contamination. Unpollinated pistils were collected from inflorescences next to pollinated inflorescences and used as negative control samples. The flower buds at mature stages were collected from the bagged shoots. 2.5. Antisense-RNA construction and plant transformation A 190-bp fragment within BcPLL9 gene (from 287 bp to 476 bp) was amplified with primers (P9i-fwd and P9i-rev) containing the restricted enzyme sites of Xba I and BamH I, respectively. The PCR product was cloned into pGEM-T Easy vector and then sub-cloned into the binary vector, pBI121, which contained a CaMV35S promoter. The construct was confirmed via PCR assay and restriction enzyme digestion, sequenced, and named as pBI121-35S::aBcPLL9. This antisense-RNA construct was then transformed into flowering Chinese cabbage (B. campestris ssp. chinensis var. parachinensis) by using the Agrobacterium tumefaciens EHA105-mediated plant transformation system [32]. Empty vector pBI121 was transformed using the same method as negative control sample. 2.6. Molecular confirmation of the transgenic plants To confirm that the T-DNA fragment in pBI121-35S::aBcPLL9 construct was migrated into the plant genome, PCR and Southern blot were performed. Genomic DNA was extracted from young leaves of the putative antisense-RNA transformed plants and the control plants. PCR and Southern blot were performed to detect a 371-bp fragment within the GUS gene in the pBI121 vector (primer: GUS-fwd and GUS-rev, listed in Table S1). For PCR assay, 35 cycles of amplification were performed using 20 ng of each gDNA sample. For

68

J. Jiang et al. / Plant Science 229 (2014) 66–75

Southern blot assay, 30 ␮g of gDNA were digested with EcoR I, fractionated on 1% agarose gel, transferred to a Hybond-N membrane (Amersham, UK), and hybridized to digoxigenin (DIG)-labelled DNA probes. The probes were prepared by labelling the fragment by using a DIG DNA Labelling Kit (Roche, Germany) according to the manufacturer’s instructions. The plant tissues used for the qRT-PCR assay were the inflorescence collected from bagged shoots carrying the gene-specific primers (P9q-fwd and P9q-rev, Table S1). The calculation method was the same as aforementioned. 2.7. Pollen germination, microscopy and pollen analysis Aniline blue staining of in vivo pollen tube growth was performed according to Ishiguro et al. [33]. The pistils were fixed in Carnoy’s solution overnight, washed three times with distilled water, treated in 16 M NaOH solution for 24 h, rinsed three times in distilled water and stained with aniline blue solution (0.1% aniline blue in 0.1 M K2 HPO4 –KOH buffer, pH 11) for 4 h in the dark. The pistils were then observed via Leica DMLB fluorescence microscopy (Leica, Germany), and images were obtained. Twenty pistils were observed for each sample. During the in vitro pollen germination experiment, pollen grains were incubated in the culture medium [15% sucrose (w/v); 0.4 Mm HBO3 ; 0.4 Mm Ca(NO3 )2 ; 0.1% of agar (w/v)] at 20–25 ◦ C with 100% relative humidity. The pollen germinating ratio was counted, and pollen morphology was observed under the microscope after 6 h of incubation. Three independent transgenic lines and the corresponding control lines were examined. To calculate the germination rate, six flowers from each plant were mixed. For each sample, ten visual fields were observed, and within each visual field, more than 50 pollen grains were examined. Measurement of the average pollen tube length was based on 30 pollen tubes. Error bars represent ±SE, **P < 0.01, by 2 test. To analyse pollen viability, the pollen grains were dissected from anthers and stained with different dyes. Alexander staining assay was performed to identify the cytoplasm and pollen wall [34]. 4 ,6diamidino-2-phenylindole (DAPI; Roche, Switzerland) was used to analyse the nuclei. For scanning electron microscopy (SEM) assays, the pollen grains from transgenic plants and control plants were mounted on SEM (HITACHI TM-1000, Japan) stubs, coated with palladium–gold and vacuum desiccated. For each line, pollen from three plants were collected and observed. Digital images were then obtained and modified using Adobe Photoshop CS3 software. For transmission electron microscopy (TEM), the anthers were fixed with 2.5% glutaraldehyde (containing 0.01% Tween-20) overnight, rinsed with 0.1 M phosphate buffer (PB), transferred to 1% osmic acid for 1 h, washed again in PB, and dehydrated through an ethanol series. The samples were then embedded in Spurr’s resin with the ultrathin sections stained with uranyl acetate and lead citrate before viewing in a JEM-1230 electron microscope.

service was used to find the sequences sharing the highest identity with At2g02720 gene in B. campestris database, which identified a BAC clone, AC189452. Homologous cloning was performed, and fragments of 1473 and 1952 bp were amplified from cDNA pools of flower buds and DNA pools of leaves, respectively, in B. campestris ssp. chinensis (Fig. S1). The gene contained a 62-bp 5 untranslated region (UTR), a 31-bp 3 UTR, a 1380-bp largest ORF and three introns based on the results from the DNAstar software. In addition, the gene encoded a 459-amino acid (aa) putative protein with a calculated molecular mass at 51.37 kDa and a theoretical pI at 9.56. The aliphatic index of the putative protein was 75.01, and the grand average of hydropathicity was −0.436. The protein sub-cellular location prediction programme, TargetP 1.1 server, predicted that the putative protein probably contains a secretory signal peptide. Meanwhile, the SignalP 4.1 server confidently predicted a signal sequence at the N-terminus between aa residue 29 and 30 (IQA-HV), a 56 aa of N-terminal conserved domain and a 186 aa of C-terminal Pec Lyase C domain, which is the typical pectate lyase characteristic. In addition, the nucleotide sequence and the amino acids of the gene showed 81% and 82% identities with Arabidopsis PLL9 gene, respectively. Thus, this gene was named BcPLL9 (GeneBank accession number KM114227). The secondary structure prediction of the BcPLL9 protein in NPSA-MLRC website indicated the composition to have a unique ␣ super helix, a random coli, and an extension strand (Fig. S2), with each accounting for 26.14, 50.54 and 23.31% of the protein, respectively. The expression pattern of BcPLL9 had been analysed via semiquantitative RT-PCR in our previous study [23]. BcPLL9 was specifically expressed in mature pollen and pistil but not in pollen of the other developmental stages, including pollen mother cell stage, meiosis stage, tetrad stage, or uninucleate stage [23]. In the present study, in situ hybridization was carried out to identify the precise localization of the BcPLL9’s mRNA (Fig. 1). Cross-sections of anthers at the uninucleate stage and the mature pollen stage of the wild-type plants, as well as pistils at 24 hap, were prepared. In accordance with the qRT-PCR results, the hybridization signal was not observed in the anther at the uninucleate stage (Fig. 1A) and was specifically observed in microspores in the anther at mature pollen stages (Fig. 1B). The signal was also detected in stigma (Fig. 1C) and the upper half part of the style at 24 hap (Fig. 1D) and was not observed in the lower half part of the style at 24 hap (Fig. 1E), in the ovaries (Fig. 1F), or in anthers at earlier developmental stages. Sections hybridized with sense probes did not exhibit any detectable signal, as shown by the anthers at mature pollen stage (Fig. 1G) and the stigma at 24 hap (Fig. 1H). Collectively, these data demonstrated that BcPLL9 was expressed in the microspore at the mature pollen stage and in pollinated pistils after different developmental periods. In the pistil, the BcPLL9 was highly expressed in the stigma and in the upper half part of the style at 24 hap. Thus, BcPLL9 is probably correlated with pollen development and subsequent fertilization processes.

3. Results 3.1.1. 3.1 Characterization and expression pattern of BcPLL9

3.2. Expression inhibition of BcPLL9 results in less seed production in B. campestris

In our previous study, the Arabidopsis ATH1 gene chip was used to screen differentially expressed genes in anther during pollen development as well as up-regulated genes expressed in pistils after pollination in B. campestris ssp. chinensis. Two PLL genes (homologous genes of At2g02720 and At3g01270) and eight other genes were identified [4]. RT-PCR and in situ hybridization assays demonstrated that the homologous gene of At3g01270 was specifically expressed in microspores and pollinated pistils in B. campestris ssp. chinensis [4]. The present study focused on the homologous gene of At2g02720 in B. campestris. First, BLAST

To analyse the function of BcPLL9 during pollen development, as well as in the fertilization process, an antisense construct (pBI12135S::aBcPLL9) of BcPLL9 was prepared (Fig. S3) and transformed into the flowering Chinese cabbage (B. campestris ssp. chinensis var. parachinensis) via the tissue culture method (Fig. S4). Out of 11 kanamycin-resistant lines, seven putative independent lines (termed bcpll9-1 to bcpll9-7, respectively) were screened via PCR analysis (Fig. 2A) and verified via Southern blot hybridization (Fig. 2B) and qRT-PCR (Fig. 2C). In Southern blot analysis, bcpll92 and bcpll9-6 were verified as negative. In qRT-PCR analysis, the

J. Jiang et al. / Plant Science 229 (2014) 66–75

69

Fig. 1. Expression pattern analyses of BcPLL9 during pollen development and fertilization process in Brassica campestris ssp. chinensis using in situ hybridization. Cross-sections are hybridized with a BcPLL9 antisense probe (A–F) and with a BcPLL9 sense probe (G–H). Specific hybridization signal can be apparently observed in the microspores in the mature pollen (B), in the stigma (C) and the upper style (D). No signals were observed in the anthers at the uninucleate stage (A), and in the lower style (E) and the ovary (F). Cross-sections of anthers at mature pollen stage (G) and stigmas at 24 h after pollination (H) are hybridized with a BcPLL9 sense probe and were used as negative control samples. No signal was detected. Scale bars are 20 ␮m in A, B and G and are 100 ␮m in C to F and H. Arrows are pointed towards the examples of hybridization signals.

70

J. Jiang et al. / Plant Science 229 (2014) 66–75 Table 1 Seed number in the genetic analysis for the BcPLL9 antisense-RNA transgenic plants. Types

Total number of harvested siliques

Average seed number per silique (pellet)*

bcpll9 self-pollination CK × bcpll9 bcpll9 × CK CK self-pollination

30 28 28 25

9.5 ± 0.7A 10.4 ± 0.6A 16.6 ± 0.9B 17.5 ± 0.5B

*Duncan’s test. The capital letter means significantly different at 0.01 level.

and pollinated pistils, more attention was paid to pollen development and fertilization processes. No apparent differences were detected in the sepal, petal, pistil, nectary, anther, or other floral organs between the three lines (Fig. 3a–c) and the control plants (Fig. 3A–C). However, cross-pollination experiments identified different results (Table 1). When the pollen from the bcpll9 lines was used as donors, considerably lower seed production was observed. The average numbers of seeds per silique were 9.5 ± 0.7 and 10.4 ± 0.6 in self-pollinated bcpll9 plant and control plants pollinated by bcpll9 pollen, respectively (Table 1). Meanwhile, the seed number was 17.5 ± 0.5 in control plants of the self-pollination group (Table 1). Pollinating bcpll9 plants with the control pollen could produce a fruit pod with 16.6 ± 0.9 seeds, which was not significantly different from the results from the self-pollination group of the control plants. However, a statistically significant difference was found between the first two groups with the latter two groups (Duncan’s test, P value < 0.01) (Table 1). These data demonstrated the normal function of female organs, such that the reduction of the seed set should be attributed by the defective male organs in bcpll9. 3.3. Half of the pollen tubes exhibited arrested development in pistils and showed deformity in their outer surface in bcpll9 Fig. 2. Molecular detection of BcPLL9 antisense-RNA transgenic plants. (A) PCR amplification of the GUS gene within the T-DNA fragment was performed to screen the BcPLL9 putative antisense RNA transgenic plants. Lane M indicates the DNA molecular makers. The templates used in PCR for each lane are as follows: lane 1, pBI121; lane 2, ddH2 O; lane 3, genome DNA of wild-type plants; lanes 4 to 6, genome DNA of empty vector control transgenic plants (CK); lanes 7 to 13, genome DNA of putative BcPLL9 antisense-RNA transgenic plants. (B) Southern blot of the GUS gene within the T-DNA fragment is performed to verify the putative BcPLL9 antisense RNA transgenic plants. The templates used in PCR for each lane are as follows: lane 1, pBI121; lane 2, genome DNA of wild-type plant; lanes 3 to 5, genome DNA of empty vector control transgenic plants (CK); lanes 6 to 12, genome DNA of putative BcPLL9 antisense-RNA transgenic plants. (C) Expression levels of BcPLL9 gene in transgenic plants were detected via qRT-PCR analysis. Forty cycles were performed for amplification, and the relative amount of each gene was computed via the 2–Ct method.

mRNA expression level of BcPLL9 was extremely low in bcpll9-1, bcpll9-3 and bcpll9-7 lines and was found to decrease in bcpll92, bcpll9-4, bcpll9-5 and bcpll9-6 lines, while not being affected in the control lines (indicated as CK in Fig. 2C). In summary, three lines, namely, bcpll9-1, bcpll9-3 and bcpll9-7, which were verified as positive antisense RNA transgenic lines in all three experiments, including PCR, Southern bolt and qRT-PCR, were selected for further analysis. The phenotypes of these three lines were observed separately and analysed statistically. If the lines displayed similar characteristics, those of bcpll9-1 would be used as an example. Otherwise, the lines would be respectively displayed in the following text. Transgenic lines were observed every other day before flowering. No morphological differences could be observed between the three lines and the control plants (data not shown). The plants were monitored every day once they started flowering. Considering that the mRNA levels of BcPLL9 were abundant in anthers

Given that less seed production may be caused by poor pollen viability as well as unsuccessful fertilization process, we tested pollen viability by using several staining dyes and we performed in vivo (Fig. 4C) and in vitro pollen germination experiments (Fig. 4A and B, and Table 2). First, in vivo pollen germination assays were obtained. We conducted cross-pollination as previously described and collected pistils at 16 hap because pollen tubes reached the bottom region of ovary between 4 and 16 hap in B. campestris [4]. Aniline blue staining was used to monitor the growth status of pollen tubes in the transmitting tracks of pistils (Fig. 4C). Ten pistils from each line were analysed. One hundred per cent of the pollen tubes reached the bottom of the pistil when wild-type pollen was used as the donor to pollinate the pistils of bcpll9 and wildtype plants. Meanwhile, when the bcpll9 pollen was used as the donor, about half of the pollen tubes reached the bottom, with the remaining half mostly arresting halfway of the transmitting tracks (Fig. 4C). Apparently, the fertilization process was defective when pollen from bcpll9 lines was used as the donor (Fig. 4C). Again, in vitro Table 2 Pollen germination assay in the BcPLL9 antisense-RNA transgenic plants. Types

Total number of tested pollen (pellet)

Number of abnormal pollen (pellet)

Ratio of abnormality (%)*

Average length of pollen tube (␮m)*

bcpll9-1 bcpll9-3 bcpll9-7 CK

515 585 605 541

238 252 275 39

46.21 ± 1.88A 43.08 ± 1.59A 45.12 ± 1.23A 7.21 ± 0.50B

30.50 ± 0.55A 32.50 ± 1.05A 31.20 ± 1.12A 95.50 ± 4.05B

*Duncan’s test. The capital letter means significantly different at 0.01 level.

J. Jiang et al. / Plant Science 229 (2014) 66–75

71

pollen germination assay was performed, in which the pollen was incubated at 23 ◦ C for 6 h (Table 2). In average, 46.21, 43.08 and 45.12% of abnormal pollen were identified in bcpll9-1, bcpll9-3 and bcpll9-7, respectively, but only 7.21% was observed in the control pollen (Table 2, Fig. 4B). The abnormal pollen ratios from transgenic plants all showed statistically significant difference compared with that of the control group (Duncan’s test, P value < 0.01). In addition to germination ratio, the length of pollen tubes was considerably shorter in bcpll9 plants than that in the control plants (Table 2). The average length of the germinated pollen tubes was approximately 31 ␮m in bcpll9 plants and 95 ␮m in control plants, thereby showing statistically significant difference (Duncan’s test, P value < 0.01). In addition, the pollen tubes of bcpll9 plants were not as smooth as those of the control plants (Fig. 4A). Almost half of the pollen from bcpll9 plants showed a jagged outer surface (indicated in the columnar graph in Fig. 4B). 3.4. Intine development was impaired in bcpll9 pollen

Fig. 3. Morphological and cytological observation of the bcpll9 antisense RNA transgenic plants. No distinct differences were observed during the vegetative growth phases between transgenic plants bcpll9 (a) and the control plants (A). No differences were also identified in the flower organs (B and b), including the anthers (C and c). (D and d) Alexander staining was used to detect the pollen wall and cytoplasm of bcpll9 (d) and the control (D) pollen. Several bcpll9 pollen grains were not as round as the control pollen, as pointed out by the black arrow (d). (E and e) DAPI staining detection was used to detect the two vegetative nucleus and the sperm cells of the pollen grains. However, no apparent difference was observed between the bcpll9 (e) and the control (E) pollen. (F–G and f–g) Aniline blue staining. Unknown bright blue materials were observed in the bcpll9 pollen grains (g). F and f, Pollen under white light field; and G and g, pollen under UV light field. The white arrow indicates the bright blue materials. (H–I and h–i) Scanning electron microscope was used to analyse the pollen surface of bcpll9 plants. Compared with the fine reticulated exine pattern in control pollen

To identify the reason for pollen tube growth arrest and pollen tube deformity, several staining dyes were used to detect the different parts of mature pollen grains (Fig. 3D–I, and d–i). First, Alexander’s stain was used to assess pollen viability in bcpll9 (Fig. 3D and d) (Alexander, 1969). The results indicated that most of the pollen contained round full cytoplasm and light blue pollen wall stained with the Alexander dye (Fig. 3D and d), but a portion (10.1%) of bcpll9 pollen (Fig. 3d) appeared shrivelled in comparison with 3.0% in the control plants (Fig. 3D). Second, DAPI was used to ensure the status of the pollen nucleus (Fig. 3E and e). The sperm nucleus and the vegetative nuclei of the bcpll9 pollen (Fig. 3e) did not show apparent difference with those of the control pollen (Fig. 3E). Thirdly, a difference was suddenly observed when bright blue materials appeared to accumulate in the aniline blue staining assay (Fig. 3F and G and f and g) in nearly half (46.2%) of the bcpll9 pollen (Fig. 3g), with none appearing in the control pollen (Fig. 3G). This bright blue material may be attributed to the deposition of callose, which can be easily detected using the aniline blue dye during pollen development. SEM was used to characterize the surface morphology of the pollen (Fig. 3H and I and h and i). Unknown materials were observed to be accumulated on the surface of the bcpll9 pollen (Fig. 3h and i) at the germinal furrow, which were absent in the control pollen (Fig. 3H and I). In addition, the bcpll9 pollen were also even in size and had clear reticulate patterns and regular germinal furrows as that of control plants. The unknown materials were observed in all three transgenic lines, which indicate that these materials originated from some common changes and were probably the identified potential callose aforementioned among the transgenic lines. The question is how callose accumulated. We conducted extrathin section assay of the anther via TEM (Fig. 5). Buds of different sizes representing the five pollen developmental stages were fixed and processed to prepare similar to the sections as described in Section 2 (data prior to the binucleate stage were not shown). The cytoplasm of the bcpll9 pollen (Fig. 5a–l) and that of the control plants did not show any obvious differences during the entire pollen developmental process (Fig. 5A–L). Clear exine structures consisting of foot layer, bacula, tectum and pollen coat were synthesized around the cytoplasm in both lines. No abnormalities were displayed in both layers of the pollen wall in bcpll9 plants compared with that of the control plants. Differences were first observed in

((H and I), the pollen grains of bcpll9 plants (h and i) were observed to present specific overly protruding material in the germinal furrow regions, which are pointed out by arrow heads in the graph.

72

J. Jiang et al. / Plant Science 229 (2014) 66–75

Fig. 4. In vitro and in vivo pollen germination experiments on bcpll9 plants. (A) In vitro pollen germination assay. The left picture shows normal pollen germination and pollen tube growth in control plants (CK), and the right one shows the abnormal grown pollen tube from bcpll9. The pollen tube wall was wavy and not as smooth as that of the control ones, as indicated by the black arrow. (B) Comparison of the pollen tube morphological abnormality ratios between bcpll9 and control plants. The data of the three bcpll9 lines were calculated separately. All three lines show more abnormal pollen tubes than the control. Error bars represent ±SE, **P < 0.01, by 2 test. C, cross- and self-pollination analyses of bcpll9 plants. The growth of pollen tubes was arrested in the middle of the styles when the bcpll9 pollen was used as the donor, whereas the control pollen grew to the bottom of the pistils. The white arrows indicate the destination of the pollen tubes.

the mature pollen, at which time the lipid materials of pollen coat accumulated into the cavities among the tectum (Fig. 5D and d). The sites of the germinal furrows appeared to protrude in bcpll9 pollen (Fig. 5d) compared with normal ones (Fig. 5D), which indicates that bcpll9 pollen tubes immediately started to germinate after their maturation. To determine the reason for abnormality by zooming in the pollen wall at both germinal and non-germinal furrow regions during the different developmental stages, differences were observed in the double-layered structure of the intine. In binucleate pollen, the outer and inner layers of intine were formed in the control pollen (Fig. 5A), where the intine within the germinal furrows (Fig. 5I) was considerably thicker than that in other regions

(Fig. 5E). At this stage, intine formation in the bcpll9 was just beginning (Fig. 5a), and the intine within the germinal furrows (Fig. 5i) and other regions (Fig. 5e) was both thinner than that in the control pollen, specifically in the germinal furrow regions. Later, at the early mature pollen stage, intine underwent further thickening in both bcpll9 and control pollen. However, the thickening become more apparent in the bcpll9 pollen (Fig. 5j) compared with that of the control pollen (Fig. 5J). Notably, the relative proportions of exintine and endintine between the bcpll9 pollen (Fig. 5b–d, f–h and j–l) and the control pollen (Fig. 5B–D, F–H and J–L) were different. Outside the germinal furrow regions, the bcpll9 pollen had no obvious demarcation between the exintine and endintine layers, with the microfibrillar endintine facing the interior and predominantly occupying the intine (Fig. 5f–h). In the same region of the control pollen, although the demarcation of the exintine and endintine layers was not clearly delineated, the intine appeared to be occupied by both layers (Fig. 5F–H). The demarcation, which was a multilamellar intine within the germinal furrows in the control pollen with a granular exintine towards the exterior and a microfibrillar endintine towards the interior and found beneath the nexine (Fig. 5J and K), was not clear inside the germinal furrows in bcpll9 pollen (Fig. 5j and k). However, in the bcpll9 pollen, the exintine overdeveloped into a thicker layer and predominantly occupied the intine layer, and the internal microfibrillar layer occupied only an extremely small proportion of the total (Fig. 5j and k). Moreover, this abnormal proportion of the two layers in intine may induce an early germination-like characteristic in the late stage of mature pollen, which shows abnormal protrusions at the three germinal furrow regions (Fig. 5d and l). This observation was not seen in the control pollen as mentioned above (. 5D and L). The exine of bcpll9 normally developed, including integrative foot layer, bacula, tectum and pollen coat, similar to that of the control pollen. From the above results, we can conclude that the deformity of pollen grains detected by aniline blue staining assay and SEM observation was probably caused by the abnormal pollen wall formation. The abnormal component ratio of intine in the germinal furrows may result in earlier pollen germination. This ratio may also be due to imbalanced force deduced from the abnormal endintine and exintine proportion upon pollen wall development, which can cause the twisted shape of the pollen tube in bcpll9 and arrested growing in style. As such, fertilization failure and less seed formation occur.

4. Discussion 4.1. BcPLL9 is a novel PLL gene preferentially expressed in mature pollen Pectate lyases catalyse the eliminative cleavage of de-esterified homogalacturonan in pectin, which disintegrates the cell wall structure [7]. PLL genes are expressed in different organs at different stages of seedling development and in response to various hormones and stresses. As was proven in Arabidopsis, all AtPLLs are expressed in flowers, and several of them are expressed highly in the pollen [20,22]. In addition, specific PLL genes are expressed in the pollen at late stages of anther development and in the pistil [4,14,22]. This batch of genes is considered to be necessary for pollen maturation and pollen tube elongation. During pollen maturation, intine is secreted by the microspore during the ringvacuolated microspore stage [35]. Intine is the innermost layer located adjacent to the pollen plasma membrane, which is a highly ordered complex of polysaccharides and structural proteins similar to other plant cell walls [35]. During pollen germination, hydrated pollen grains release numerous proteins anchored into the pollen wall, including PME and PG, to disintegrate the style cell wall,

J. Jiang et al. / Plant Science 229 (2014) 66–75

73

Fig. 5. TEM observation of pollen development in BcPLL9 antisense RNA transgenic plants. Figs. a–l indicate the pollen wall of bcpll9 plants, and A–L indicate the pollen wall of control plants (CK). In the binucleate pollen stage, the two layers of intine formed in the control pollen (A). The intine within the germinal furrows (I) was much thicker than that of the other regions (E), whereas the intine formation was just beginning in bcpll9 (a), with the intine within the germinal furrows (j) and other regions (e) were both thinner. Later, at the early mature pollen stage, intine within the germinal furrow of both bcpll9 and control pollen underwent further thickening, but the thickening became more apparent in the germinal furrow regions in the bcpll9 pollen (j) compared with that of the control (J). In addition, the relative proportion of exintine and endintine was different between the two pollens. In the non-germinal furrow regions, bcpll9 pollen had no obvious demarcation between the exintine and endintine layers, and the endintine predominantly occupied the intine (f–h). In the control pollen, the intine was occupied by both layers (F–H). In the germinal furrow region of the bcpll9 pollen, the exintine overdeveloped into a thicker layer and predominantly occupied the intine layer, whereas the endintine layer occupied only an extremely small proportion (j and k). By contrast, the endintine and the exintine each occupied a relative appropriate proportion in control pollen (J–L). The white arrows indicate the germinal furrows. The black asterisks indicate the protruding parts of the bcpll9 pollen. SN, VN, Endi, Exin, Try, Tec, Bac, Fl and CP indicate the sperm nucleate, the vegetative nucleate, the endintine, the exintine, the tryphine, the tectum, the bacula, the foot layer and the cytoplasm, respectively.

74

J. Jiang et al. / Plant Science 229 (2014) 66–75

thereby allowing pollen tubes to pass through [36–38]. However, to date, little direct evidence regarding the function of PLL genes in the pollen maturation and pollen tube growth have been identified, apart from information on the spatial and temporal expression patterns of PLL genes and their promoters [20,22]. PLL genes were first identified as dominantly expressed in tomato pollen. LAT56 and LAT59 are needed for pollen germination and pollen tube growth [10]. The promoter sequence of the LAT59 gene had also been identified as pollen specific [39]. In addition, homologous genes of LAT59 were again identified in tobacco and Arabidopsis as Nt59 and At59 [14]. In the present study, the BcPLL9 gene was isolated from the inflorescence of B. campestris. The gene presents typical pectate lyase characteristics and a 186 aa Pec Lyase C domain [7,10,20,40]. Expression pattern analysis showed that BcPLL9 was expressed in mature pollen and stigma, as well as the upper part of the style (Fig. 1). The expression of BcPLL9 genes in late developmental pollen may suggest its relevance to other pollen-specific, pectin-related genes because pollen PME genes and PG genes are considered to be necessary for pollen maturation or pollen tube growth [41,42]. In addition, the expression of BcPLL9 in the stigma indicates its possible association with pollen germination. 4.2. BcPLL9 takes part in intine formation To study the function of BcPLL9, an antisense RNA construct under the drive of a CaMV 35S promoter was transformed into a fertile flowering Chinese cabbage plant. Self- and cross-fertilization revealed that the transgenic bcpll9 plants with inhibited BcPLL9 expression displayed substantially reduced male fertility but female fertility remained normal. Seed yields were apparently lower in the bcpll9 lines compared with those of the control plants (Table 1). In vivo and in vitro pollen germination assay identified shorter pollen tubes (Table 2, Fig. 4) and abnormal pollen tube wall. Further observation with aniline blue staining revealed that nearly half of bcpll9 pollen showed putative abnormal callose accumulation (Fig. 3g), which was verified via SEM (Fig. 3h and i) and TEM analysis (Fig. 5). In the germinal furrow region, exintine overdeveloped to a thicker layer and predominantly occupied the intine layer, whereas the internal microfibrillar endintine layer occupied only an extremely small proportion in bcpll9 pollen (Fig. 5j and k). By contrast, outside the furrow region, endintine developed into a thicker layer and occupied the whole intine. This disordered proportion in the two intine layers probably induced the early germinationlike characteristic in the late-stage mature pollen, which showed abnormal protrusions at the three germinal furrow regions (Fig. 5d and l). Disordered intine formation was also found in Arabidopsis ms33, usp, fla3 and tek mutants [43–46], as well as in Chinese cabbage antisense lines of a PG gene, BcMF2, and a BcPLL9’s homologous gene, BcPLL10, in our previous study [47,48]. In Chinese cabbage PG antisense line, bcmf2, the abnormal thickening in the intine region was due to the overdevelopment of exintine instead of the endintine both inside and outside of the germinal furrow [47]. Exine was normal in bcmf2, which is the same as bcpll9. Compared with the delay of intine formation in bcpll9, the intine is precociously formed in Arabidopsis ms33 pollen with thicker endintine and thinner exintine, which resulted in abnormalities in pollen grain desiccation and pollen viability [43]. Thinner intine was also found in fla3 mutant at the bicellular stage. However, whether this fasciclin-like arabinogalactan proteins 3 (FLA3) affected exintine or endintine formation, or both [45], was not determined. In usp mutants, the whole intine was absent in mature pollen [44]. Meanwhile, whether AtUSP (UDP-sugar pyrophosphorylase) prevented intine formation at the first time is not apparent, or if intine was abnormally degraded after biosynthesis. In tek mutants, the knockout of TRANSPOSABLE ELEMENT SILENCING VIA AT-HOOK (TEK) gene

caused failure in the development of both intine and nexine, which resulted in microspore abortion [46]. In addition, in bcpll10 pollen, which is relatively similar to bcpll9 pollen, and within the furrow region, the exintine overdeveloped to a thicker layer and predominantly occupied the intine layer, with the endintine layer occupying only an extremely small proportion. Outside the germinal furrow region, the bcpll10 pollen had no evident demarcation between exintine and endintine layers, and endintine was considerably thicker [47]. However, bcpll10 did not have protruded materials in germinal furrow regions and early pollen germination compared with that of bcpll9. Instead, exine structure defect was observed in bcpll10 with pollen grains covered by over biosynthesized and accumulated tryphine, which indicated disordered lipid metabolism and transportation. In bcpll9, all exines were well developed. This difference between bcpll9 and bcpll10 also indicated the partially redundant function between these two homologous genes during pollen wall formation. Therefore, the abnormal phenotypes observed in ms33, usp, fla3, tek, bcmf2, bcpll10 and bcpll9 plants are all evidently associated with intine formation but were not exactly the same. These genes may perform partially redundant functions in pollen wall formation. Further experiments are needed to prove and clarify their functions and connections.

4.3. Disturbed intine formation resulted in early pollen germination in bcpll9 The pollen tube is the extension of the intine and germinates through germinal furrow to deliver sperms [4,49,50]. The exintine is composed of pectin and various proteins, and the endintine consists of cellulose [28]. Pectin biosynthesis and degradation is a complex process involving hundreds of genes such as structure protein-encoding genes, regulatory protein-coding genes and regulatory sequences [4]. The disturbance of pectin biosynthesis not only affects intine formation but also the metabolism of other materials associated with pollen wall development, such as cellulose, callose and lipids [4,46,51]. The down-regulation of BcPLL9 caused by disproportionally developed endintine and exintine indicates unbalanced metabolism of pectin and cellulose (Fig. 5). The accumulation of callose is probably caused by early pollen germination because callose plug invariably forms soon after pollen germination and along with pollen tube elongation [2]. Studies have also suggested that the disorder in intine formation would result in pollen tube burst [28]. In the present study, the disproportionality between exintine and endintine of intine possibly changed the pressure of pollen wall, thereby leading to early pollen germination, as indicated by the accumulation of callose at the germinal furrow regions (Fig. 4C). These early germinated pollen tubes experienced arrested growth halfway through delivering the sperms and resulted in failure of fertilization. Therefore, we deduced that BcPLL9 is probably associated with pectin metabolism during intine formation and may be related to the degradation of the cell wall in the stigma to allow pollen tube to penetrate through the style. This initial analysis indicated a function of BcPLL9 during pollen development and germination. However, the molecular mechanism underlying BcPLL9 gene involvement in pectin metabolism cannot be clearly explained. Moreover, we cannot explain how the pressure for the pollen germination formed and changed and how it influences the growth state of pollen tube wall. Further studies will be needed to explore the interaction proteins of BcPLL9, its connection to BcPLL10, as well as their substrates and associated working model during pollen wall formation and pollen tube growth, which affect the fertilization process and results in partial male sterility.

J. Jiang et al. / Plant Science 229 (2014) 66–75

Acknowledgements This work was supported by the National Program on Key Basic Research Projects (No. 2012CB113900), Natural Science Foundation of China (Nos. 31071805, 31301790) and the Key Sci-Technology Project of Zhejiang Province (No. 2010C12004), Guangdong Natural Science Foundation (S2013040016220), China Postdoctoral Science Foundation (2013M530375, 2014T70827). Appendix A. Supplementary data

[25]

[26] [27]

[28]

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.plantsci. 2014.08.008.

[29]

References

[31]

[1] S. McCormick, Control of male gametophyte development, Plant Cell 16 (2004) s142–s153. [2] S. Blackmore, A.H. Wortley, J.J. Skvarla, J.R. Rowley, Pollen wall development in flowering plants, New Phytol. 174 (2007) 483–498. [3] M. Borg, L. Brownfield, D. Twell, Male gametophyte development: a molecular perspective, J. Exp. Bot. 60 (2009) 1465–1478. [4] J. Jiang, Z. Zhang, J. Cao, Pollen wall development: the associated enzymes and metabolic pathways, Plant Biol. (Stuttg) 15 (2013) 249–263. [5] K.H. Caffall, D. Mohnen, The structure, function, and biosynthesis of plant cell wall pectic polysaccharides, Carbohydr. Res. 344 (2009) 1879–1900. [6] A. Collmer, N.T. Keen, The role of pectic enzymes in plant pathogenesis, Ann. Rev. Phytopathol. 24 (1986) 383–409. [7] M.C. Marin-Rodriguez, J. Orchard, G.B. Seymour, Pectate lyases, cell wall degradation and fruit softening, J. Exp. Bot. 53 (2002) 2115–2119. [8] S. Yadav, P.K. Yadav, D. Yadav, K.D.S. Yadav, Pectin lyase: a review, Process Biochem. 44 (2009) 1–10. [9] F. Barras, F. van Gijsegem, A.K. Chatterjee, Extracellular enzymes and pathogenesis of soft-rot Erwinia, Ann. Rev. Phytol. 32 (1994) 201–234. [10] R.A. Wing, J. Yamaguchi, S.K. Larabell, V.M. Ursin, S. McCormic, Molecular and genetic characterization of two pollen-expressed genes that have sequence similarity to pectate lyases of the plant pathogen Erwinia, Plant Mol. Biol. 14 (1989) 17–28. [11] G.B. Seymour, K.C. Gross, Cell wall disassembly and fruit softening, Postharvest News Infor. 7 (1996) 45–52. [12] Y. Wu, X. Qiu, S. Du, L. Erickson, PO149, a new member of pollen pectate lyaselike gene family from alfalfa, Plant Mol. Biol. 32 (1996) 1037–1042. [13] E. Dominguez-Puigjaner, I. Llop, M. Vendrell, S. Prat, A cDNA clone highly expressed in ripe banana fruit shows homology to pectate lyases, Plant Physiol. 114 (1997) 1071–1076. [14] R. Kulikauskas, S. McCormick, Identification of the tobacco and Arabidopsis homologues of the pollen-expressed LAT59 gene of tomato, Plant Mol. Biol. 34 (1997) 809–814. [15] N. Medina-Escobar, J. Cardenas, E. Moyano, J.L. Caballero, J. Munoz-Blanco, Cloning molecular characterisation and expression pattern of a strawberry ripening-specific cDNA with sequence homology to pectate lyase from higher plants, Plant Mol. Biol. 34 (1997) 867–877. [16] R. Medina-Suarez, K. Manning, J. Fletcher, J. Aked, C.R. Bird, G.B. Seymour, Gene expression in the pulp of ripening bananas, Plant Physiol. 115 (1997) 453–461. [17] C. Domingo, K. Roberts, N.J. Stacey, I. Connerton, F. Ruiz-Teran, M.C. McCann, A pectate lyase from Zinnia elegans is auxin inducible, Plant J. 13 (1998) 17–28. [18] D. Milioni, P.E. Sado, N.J. Stacey, C. Domingo, K. Roberts, M.C. McCann, Differential expression of cell-wall-related genes during the formation of tracheary elements in the Zinnia mesophyll cell system, Plant Mol. Biol. 47 (2001) 221–238. [19] I. Pilatzke-Wunderlich, C.L. Nessler, Expression and activity of cell-walldegrading enzymes in the latex of opium poppy, Papaver somniferum L, Plant Mol. Biol. 45 (2001) 567–576. [20] S.G. Palusa, M. Golovkin, S.B. Shin, D.N. Richardson, A.S. Reddy, Organ-specific, developmental, hormonal and stress regulation of expression of putative pectate lyase genes in Arabidopsis, New Phytol. 174 (2007) 537–550. [21] J.P. Vogel, T. Raab, C. Somerville, PMR6 a pectate lyase-like gene required for powdery mildew susceptibility in Arabidopsis, Plant Cell 14 (2002) 2095–2106. [22] L. Sun, S. van Nocker, Analysis of promoter activity of members of the PECTATE LYASE-LIKE (PLL) gene family in cell separation in Arabidopsis, BMC Plant Biol. 10 (2010) 152. [23] J. Jiang, L. Yao, Y. Miao, J. Cao, Genome-wide characterization of the pectate lyase-like (PLL) gene in Brassica rapa, Mol. Genet. Genomics 288 (2013) 601–614. [24] S. Dai, T. Chen, K. Chong, Y. Xue, S. Liu, T. Wang, Proteomics identification of differentially expressed proteins associated with pollen germination and

[30]

[32]

[33]

[34] [35]

[36] [37] [38]

[39]

[40]

[41]

[42] [43]

[44]

[45]

[46]

[47]

[48]

[49]

[50] [51]

75

tube growth reveals characteristics of germinated Oryza sativa pollen, Mol. Cell Proteomics 6 (2007) 207–230. Y. Yu, Y. Liang, M. Lv, J. Wu, G. Lu, J. Cao, Genome-wide identification and characterization of polygalactronase genes in Cucumis sativus and Citrullus lanatus, Plant Physiol. Biochem. 74 (2014) 263–275. J.C. Mollet, C. Leroux, F. Dardelle, A. Lehner, Cell wall composition, biosynthesis and remodelling during pollen tube growth, Plant 2 (2013) 107–147. L. Huang, J. Cao, W. Ye, T. Liu, L. Jiang, Transcriptional differences between the male-sterile mutant bcms and wild-type Brassica campestris ssp. chinensis reveal genes related to pollen development, Plant Biol. (Stuttg) 10 (2008) 342–355. L. Huang, W. Ye, T. Liu, J. Cao, Characterization of the male-sterile line Bcajh9701A/B and identification of candidate genes for genic male sterility in Chinese cabbage-pak-choi, J. Am. Hort. Sci. 134 (2009) 632–640. L. Huang, X. Zhao, T. Liu, H. Dong, J. Cao, Developmental characteristics of floral organs and pollen of Chinese cabbage (Brassica campestris L. ssp. chinensis), Plant Syst. Evol. 286 (2010) 103–115. T. Czechowski, M. Stitt, T. Altmann, M.K. Udvardi, W.R. Scheible, Genome-wide identification and testing of superior reference genes for transcript normalization in Arabidopsis, Plant Physiol. 139 (2005) 5–17. K.J. Livak, T.D. Schmittgen, Analysis of relative gene expression data using realtime quantitative PCR and the 2(-Delta Delta CT) method, Methods 25 (2001) 402–408. X. Yu, J. Cao, W. Ye, Y. Wang, Construction of an antisense CYP86MF gene plasmid vector and production of a male-sterile Chinese cabbage transformant by the pollen-tube method, J. Hort. Biotech. 79 (2004) 833–839. S. Ishiguro, A. Kawai-Oda, J. Ueda, I. Nishida, K. Okada, The DEFECTIVE IN ANTHER DEHISCENCE1 gene encodes a novel phospholipase A1 catalyzing the initial step of jasmonic acid biosynthesis, which synchronizes pollen maturation, anther dehiscence, and flower opening in Arabidopsis, Plant Cell 13 (2001) 2191–2209. R. Peterson, J. Slovin, C. Chen, A simplified method for differential staining of aborted and non-aborted pollen grains, Intr. J. Plant Biol. 1 (2010) e13. H.A. Owen, C.A. Markaroff, Ultrastructural of microsporogenesis and microgametogenesis in Arabidopsis thaliana (L.) Heynh. ecotype Wassilewskija (Brassicaceae), Protoplasma 185 (1995) 7–21. M. Bosch, A.Y. Cheung, P.K. Hepler, Pectin methylesterase, a regulator of pollen tube growth, Plant Physiol. 138 (2005) 1334–1346. M. Bosch, P.K. Hepler, Pectin methylesterases and pecin dynamics in pollen tubes, Plant Cell 17 (2005) 3219–3226. L. Jiang, S.L. Yang, L.F. Xie, C.S. Puah, W.Q. Zhang, W.C. Yang, V. Sundaresan, D. Ye, VANGUARD1 encodes a pectin methylesterase that enhances pollen tube growth in the Arabidopsis style and transmitting tract, Plant Cell 17 (2005) 584–596. C. Curie, S. McCormick, A strong inhibitor of gene expression in the 5 untranslated region of the pollen-specific LAT59 of tomato, Plant Cell 9 (1997) 2025–2036. S.J. Tamaki, S. Gold, M. Robeson, S. Manulis, N.T. Keen, Structure and organization of the pel genes from Erwinia chrysanthemi EC16, J. Bacteriol. 170 (1988) 3468–3478. R.J. Allen, D.M. Lonsdale, Molecular characterization of one of the maize polygalacturonase gene family members which are expressed during late pollen development, Plant J. 3 (1993) 261–271. D. Honys, D. Twell, Comparative analysis of the Arabidopsis pollen transcriptome, Plant Physiol. 132 (2003) 640–652. H. Fei, V.K. Sawhney, Ultrastructural characterization of male sterile33 (ms33) mutant in Arabidopsis affected in pollen desiccation and maturation, Can. J. Bot. 79 (2001) 118–129. J.A. Schnurr, K.K. Storey, H.J. Jung, D.A. Somers, J.W. Gronwald, UDP-sugar pyrophosphorylase is essential for pollen development in Arabidopsis, Planta 224 (2006) 520–532. J. Li, M. Yu, L.L. Geng, J. Zhao, The fasciclin-like arabinogalactan protein gene, FLA3, is involved in microspore development of Arabidopsis, Plant J. 64 (2010) 482–497. Y. Lou, X.F. Xu, J. Zhu, J.N. Gu, S. Blackmore, Z.N. Yang, The tapetal AHL family protein TEK determines nexine formation in the pollen wall, Nature Comm. 5 (2014) 3855. L. Huang, J. Cao, A. Zhang, Y. Ye, Y. Zhang, T. Liu, The polygalacturonase gene BcMF2 from Brassica campestris is associated with intine development, J. Exp. Bot. 60 (2009) 301–313. J. Jiang, L. Yao, Y. Yu, M. Lv, Y. Miao, J. Cao, Pectate lyase-like 10 is associated with pollen wall development in Brassica campestris, J. Integr. Plant Biol. (2014), http://dx.doi.org/10.1111/jipb.12209. A.A. Dobritsa, S. Nishikawa, D. Preuss, E. Urbanczyk-Wochniak, L.W. Sumner, A. Hammond, A.L. Carlson, R.J. Swanson, LAP3, a novel plant protein required for pollen development, is essential for proper exine formation, Sex Plant Reprod. 22 (2009) 167–177. K.J. McNeil, A.G. Smith, A glycine-rich protein that facilitates exine formation during tomato pollen development, Planta 231 (2010) 793–808. T. Ariizumi, K. Toriyama, Genetic regulation of sporopollenin synthesis and pollen exine development, Ann. Rev. Plant Biol. 62 (2011) 437–460.