Auxin Response Factor17 is essential for pollen wall pattern formation in Arabidopsis

: In angiosperms, pollen wall pattern formation is determined by primexine deposition on the microspores. Here, we show that an auxin response factor, ARF17, is essential for primexine formation and pollen development in Arabidopsis thaliana . The arf17 mutant exhibited a male sterile phenotype with normal vegetative growth. ARF17 was expressed in microsporocytes and microgametophytes from meiosis to the bicellular microspore stage. Transmission electron microscopy (TEM) analysis showed that primexine was absent in the arf17 mutant, which leads to pollen wall patterning defects and pollen degradation. Callose deposition was also significantly reduced in the arf17 mutant, and the expression of CALLOSE SYNTHASE5 ( CalS5 ), the major gene for callose biosynthesis, was approximately 10% that of wild type. Chromatin immunoprecipitation (ChIP) and Electrophoretic Mobility Shift Assay (EMSA) showed that ARF17 can directly bind to the CalS5 promoter. As indicated by expression of the DR5::GFP auxin reporter, auxin signaling appeared to be specifically impaired in arf17 anthers. Taken together, our results suggest that ARF17 is essential for pollen wall patterning in Arabidopsis by modulating primexine formation at least partially through direct regulation of CalS5 gene expression.


Introduction
In angiosperms, the pollen wall is the most complex plant cell wall. It consists of the inner wall, the intine, and the outer wall, the exine. The exine is further divided into sexine and nexine layers. The sculptured sexine includes three major parts: baculum, tectum and tryphine (Heslop-Harrison 1971;Piffanelli et al. 1998;Ariizumi and Toriyama, 2011;Fig. 1A). Production of a functional pollen wall requires the precise spatial and temporal cooperation of gametophytic and sporophytic tissues and metabolic events (Blackmore et al., 2007). The intine layer is controlled gametophytically while the exine is regulated sporophytically. The sporophytic tapetum cells provide material for pollen wall formation while primexine determines pollen wall patterning (Heslop-Harrison, 1968).
After meiosis, four microspores were encased in callose to form a tetrad. Subsequently, the primexine develops between the callose layer and the microspore membrane (Fig. 1B) and the microspore plasma membrane becomes undulated (Fig. 1B;Southworth and Jernstedt, 1995;Fitzgerald and Knox, 1995). Sporopollenin precursors then accumulate on peak of the undulated microspore membrane to form the bacula and tectum (Fig. 1B;Fitzgerald and Knox, 1995). After callose degradation, individual microspores are released from the tetrad, and the bacula and tectum continue to grow into exine with further sporopollenin deposition (Fitzgerald and Knox, 1995;Blackmore et al., 2007).
The callose have been reported to affect primexine deposition and pollen wall pattern formation. The peripheral callose layer, secreted by the microsporocyte, act as the mold for primexine (Waterkeyn and Beinfait, 1970;Heslop-Harrison, 1971). Callose  is the major enzyme that is responsible for biosynthesis of the callose peripheral of the tetrad (Dong et al., 2005;Nishikawa et al., 2005). Mutation of Cals5 and abnormal CalS5 pre-mRNA splicing resulted in defective peripheral callose deposition and primexine formation (Dong et al., 2005;Nishikawa et al., 2005;Huang et al., 2013;Sun et al., 2003). Besides CalS5, four membrane associated proteins have also been reported to be involved in primexine formation, including DEFECTIVE EXINE of wild type (Fig. 2E). In contrast, no pollen grains were observed in the arf17 flower ( Fig. 2F). Further Alexander staining of pollen viability (Alexander, 1969) showed that the wild-type anthers contained purple-staining viable pollen grains (Fig. 2G), while only degraded pollen remnants were observed in the arf17 anther (Fig. 2H). These data indicate that knocking out ARF17 leads to aborted pollen development and male sterility.

Expression Patterns of ARF17
ARF17 is widely expressed in Arabidopsis tissues including the roots, leaves, stems, siliques, seedlings and buds (Mallory et al., 2005). To analyze the spatio-temporal expression of ARF17 during anther development, we performed an RNA in situ hybridization assay with an ARF17-specific anti-sense probe. Anther development can be divided into 14 stages (Sanders et al., 1999). At Stage 5, no obvious hybridization signal was detected (Fig. 3A). Apparent hybridization signals were observed in microsporocytes at Stage 6 ( Fig. 3B). Expression of ARF17 was reduced in tetrads at Stage 7 (Fig. 3C), and remained detectable in the microspores until Stage 9 (Fig. 3, D and E). In mature pollen, no ARF17-specific hybridization signal was detected (Fig. 3, F and G). As a control, the ARF17 sense-strand probe did not yield any hybridization signal (Fig. 3H).
We further investigated the distribution of the ARF17 protein. The genomic fragment of ARF17 was fused with the gene for green fluorescent protein (GFP; Supplemental Fig. S1B) and this construct was introduced into the arf17 mutant (Supplemental Fig. S1C).
The ARF17-GFP fusion protein was shown to be functional because it was able to rescue the male sterile phenotype in the arf17 mutant (Supplemental Fig. S1E). Confocal microscopy of GFP allows detection of the respective ARF17-GFP fusion proteins. At stage 5, no GFP fluorescence was observed (Fig. 3I). At stage 6, significant GFP signals were observed in microsporocytes (Fig. 3, J and K). At stage 7, ARF17-GFP fluorescence could be detected within microspores in tetrads (Fig. 3, L and M). GFP expression was also found in the microspores at uninucleate pollen stage (Fig. 3,N and O). In bicellular pollen, GFP signals were present in both the germ cell and the vegetative nuclei ( Fig. 3 (Fig. 3, R-U), which is consistent with previous male gametophyte transcriptome (Honys and twell, 2004). When ARF17-GFP pollen grains were germinated in vitro, GFP could also be detected in the imbibed pollen grain (Fig. 3, V and W) and pollen tube ( Fig. 3, X-Z); however, the expression level was lower than that observed from stage 6 to the bicellular microspore stage.

Primexine Is Absent in the arf17 Mutant
In order to understand the defects in arf17 in detail, we analyzed cross-sections prepared from wild type (Col-0) and mutant inflorescences. Up to stage 6, no detectable differences could be observed between wild type and the arf17 mutant ( Fig. 4, A, B, E and F). DAPI (4, 6-diamidino-2-phenylindole) staining showed that meiosis is normal in arf17 (Supplemental Fig. S2). At stage 7, meiosis is completed and tetrads have formed in the wild type (Fig. 4C). In contrast, the shape of arf17 tetrads appeared to be distorted ( Fig. 4G). At stage 8, individual microspores could be observed in both wild type and arf17 anthers, indicating the normal release of free microspores ( Fig. 4D and H). From stage 9 to 11, wild-type microspores developed an obvious exine layer ( Fig. 4I and J). In arf17, however, microspores failed to form exine layer and became vacuolated and degraded ( Fig. 4L and M). Eventually, the pollen wall developed successfully and mature pollen grains were formed in wild type (Fig. 4K), whereas all arf17 microspores degenerated completely, leaving only aborted pollen debris (Fig. 4N).
To further investigate the detailed abnormalities in arf17 exine development, transmission electron microscopy (TEM) observation was employed. In the wild-type locules at Stage 6, callose was deposited around microsporocytes with medium electron-dense staining (Fig. 5A). The callose in arf17 was much reduced compared to wild type (Fig. 5B). At Stage 7, the wild-type tetrad developed primexine between microspore membrane and the callose layer; microspore membrane became undulated ( Fig. 5C). In contrast, the callose in the arf17 mutant seemed thinner; the primexine and membrane undulation was not observed (Fig. 5D). At the uninucleate and the bicellular microspore stage, sporopollenin continued to accumulate onto the surface of the microspore to form the bacula，tectum and exine in wild type (Fig. 5E and G). However, in arf17, the sporopollenin randomly accumulated around the microspore surface and no exine could be observed for those shriveled pollen grains ( Fig. 5F and H). The TEM observations revealed that arf17 pollen had an absence of primexine, leading to subsequent aborted exine pattern and microspore degradation.

Callose Deposition around the Tetrad Is Significantly Reduced in arf17
The above observations indicated that callose deposition was decreased in the arf17 mutant. To confirm this, we stained tetrads from both arf17 and wild-type anthers with toluidine blue. Under light microscopy, both wild type and ARF17-GFP complemented plants showed apparent callose layer in tetrads (Fig. 6, A and C; n >150). By contrast, the callose layer of arf17 tetrads was ambiguous (n=295; Fig. 6B). We then analyzed callose fluorescence by aniline blue staining. Both aniline blue staining and fluorescence analysis showed that the fluorescence of the callose in arf17 was obviously weaker than that in wild type and ARF17-GFP complemented plants (Fig. 6, D-I). These results confirmed that arf17 tetrads had a reduced callose deposition.
CalS5 is the major enzyme responsible for peripheral callose synthesis in the tetrad in Arabidopsis (Dong et al., 2005). We analyzed the expression of CalS5 in arf17.
Real-time PCR revealed that the expression of CalS5 was dramatically decreased in the arf17 mutant compared to wild type (Fig. 6J). To further verify these results at the cellular level, we performed RNA in situ hybridization using a CalS5-specific probe. The CalS5 signal reached its peak in microsporocytes and tapetum cells at Stage 6 ( Fig. 6, K-M), consistent with a previous report (Nishikawa et al., 2005). In the arf17 mutant, CalS5 transcript abundance was much less than in the wild type (Fig. 6

ARF17 Directly Regulates CalS5 Expression
ARF17 is an ARF transcription factor, and would be expected to regulate gene expression by binding to the AuxRE of target gene promoters. An AuxRE (5'-GAGACA-3') was found between -863 and -858 in the CalS5 promoter (Fig. 7A). To corroborate the interaction of the ARF17 protein with the CalS5 promoter in vivo, we performed a chromatin immunoprecipitation (ChIP) assay using buds from a transgenic Arabidopsis line that expressed the functional GFP-tagged ARF17 (Fig. 3). Real-time PCR results indicated that one DNA fragment, pCalS5-2, containing the predicted ARF protein binding site (GAGACA) was enriched when the GFP antibodies were used (Fig. 7B).
However, when three other fragments of the CalS5 promoter were amplified, no obvious enrichment was observed (Fig. 7B).
In order to confirm the physical interaction between the ARF17 protein and the CalS5 promoter region, we performed an electrophoretic mobility shift assay (EMSA) using recombinant ARF17 proteins fused to glutathione S-transferase (GST-ARF17) and a 64-bp CalS5 promoter fragment containing an AuxRE motif. The EMSA result showed that the recombinant ARF17 protein is able to bind this DNA fragment (Fig. 7C). The binding specificity was further confirmed by using unlabeled promoter fragment as a competitor. When the unlabeled DNA fragment was added, the excess CalS5 competitor probe reduced the signal in a concentration-dependent manner (Fig. 7C). These results indicate that ARF17 directly binds to the CalS5 promoter to regulate its expression.

The arf17 Mutation Affects Pollen Tube Growth
In the F 2 population, only about 12% (62 of 493) of the progeny displayed the male sterile phenotype. This ratio was stable over five consecutive generations. The segregation ratio deviated significantly from the expected 25% (1:3; Table 1), suggesting that the gametophyte function may also be affected in arf17/ARF17 (arf17/+) plants. We pollinated arf17/+ stigmas with wild-type pollen. The segregation ratio of the T-DNA in the F 1 progeny was approximately 1:1 (111:114; Table1), indicating that the transmission of the T-DNA through the female gametophyte is not affected and that the arf17 ovule is fertile. When wild-type stigmas were pollinated with arf17/+ pollen, the T-DNA ratio of the F 1 progeny was 24.9% (n = 289), instead of the expected 50% (1:1; Table 1), indicating that male gametophyte transmission is reduced in the arf17/+ mutant.
To determine whether the reduced transmission was derived from impaired pollen development, pollen germination, or pollen tube growth, homozygous arf17 plants were pollinated with qrt1 pollen (Preuss et al., 1994) to generate arf17/+qrt1/qrt1 plants. In qrt1 plant, the four microspores remain associated as tetrads (Preuss et al., 1994;Francis et al., 2006). DAPI and Alexander staining showed that arf17/+ pollen grains developed normally (n = 300; Supplemental Fig. S4). We next carried out an in vitro pollen germination assay to test whether the pollen tube performance of arf17/+ was affected.
After 6 h, 78% (n = 760) of pollen grains from arf17/+ plants germinated (Fig. 8A and C), in contrast to 82% (n = 560) of wild-type pollen grains. This indicated that pollen grain germination in arf17/+ may not be affected (P>0.05). However, the average pollen tube length from arf17/+ pollen grains was dramatically reduced, and more short pollen tubes were observed compared with wild type (Fig. 8A, B and D). Statistical analysis showed that 36% of arf17/+ pollen tubes were in the range of 10-200 μ m, whereas only 9% of wild-type pollen tubes were in this range; 19% of arf17/+ pollen tubes were >400 μ m long, while 46% of wild-type pollen tubes were in this range (Fig. 8E). This indicates that the arf17 mutation affected pollen tube growth.
The pollen tube growth defect was further validated in vivo. We pollinated wild-type stigmas in emasculated flowers with arf17/+ pollen grains and examined the distribution of wild-type and mutant seeds in the upper and lower parts of crossed siliques using a Basta herbicide screen. The ratio of mutant seeds from the upper part was obviously higher than that from the lower part ( showed that arf17 pollen tubes grow slower than wild type.

Expression of an Auxin Reporter Gene Is Impaired in arf17 Anthers
The ARF family proteins are involved in auxin signaling and can directly regulate the transcription of auxin response genes. To investigate the role of ARF17 in the auxin response, we introduced the synthetic auxin response promoter DR5 (Ulmasov et al., 1997) driving expression of the GFP gene (DR5::GFP) into the arf17 mutant by crossing. In five-week-old wild-type plants, GFP fluorescence could be detected in the root tip (Fig. 9A), leaf tips ( Fig. 9C) and anthers at later stages when mature pollen is formed inside (Fig. 9E). In arf17, GFP signals could also be observed in the root tip and leaf tips (Fig. 9, B and D; n = 24). However, expression of DR5::GFP was barely detected in arf17 late stage anthers ( Fig. 9F; n = 24). Since ARF17 functions at the tetrad stage, we chose to observe the DR5::GFP signal at this stage. In the wild type, apparent GFP signals could be observed at the anther apex and the joint between the anther and filament ( Fig. 9G; n = 30). However, GFP signal was not observed in tetrads, which is consistent with a previous report (Cecchetti et al., 2008). In arf17, a similar pattern of GFP expression was detected to that in wild type ( Fig. 9H;

DISCUSSION
In this work, we showed that in the arf17 mutant, primexine deposition within tetrads is completely absent. Sporopollenin seems to be synthesized in arf17 (Fig. 5). Exine and pollen wall patterning fail to form; pollen grain degradation and male infertility were observed at later stages (Fig. 4). Because primexine has a central role in exine formation and pollen wall patterning, it was suggested that the male sterile phenotype of the arf17 mutant is caused by absent primexine. ARF17 is expressed in microsporocytes and tetrads, supporting its role for primexine formation. Currently, several genes have been ARF17 belongs to the ARF family of transcription factors, which are considered to be a component of the auxin signaling pathway. Therefore, our results provide a new mechanism that determines primexine formation, as well as subsequent pollen wall formation.
ARF proteins can regulate the expression of genes with an AuxRE in their promoter (Ulmasov et al., 1999). The CalS5 promoter contains one AuxRE (-863 to -858).
Both ChIP and EMSA assays showed that ARF17 can directly bind to the promoter region of CalS5. The expression of ARF17 and CalS5 in microsporocytes and tetrads support thebinding of ARF17 to the CalS5 promoter. Expression of CalS5 in the arf17 mutant was approximately 10% of the level in the wild type (Fig. 6). In arf17, the peripheral callose layer of the tetrad was detected, although the amount is significantly reduced compared with wild type (Fig. 6). Therefore, these results showed that ARF17 directly regulates CalS5 expression in callose synthesis. Recently, Huang et al (2013) showed that CDKG1 can associated with spliceosome to regulate CalS5 splicing and pollen wall formation. In addition, CalS5 expression is also decreased in rpg1 and npu mutants (Sun et al., 2013;Chang et al., 2012), indicating CalS5 expression is regulated at multiple levels.
The cals5 knockout mutants can produce a reduced amount of primexine and are able to generate several seeds (Dong et al., 2005;Nishikawa et al., 2005), whereas arf17 is completely sterile and shows no primexine deposition (Fig. 4). This suggests that there are additional genes that can be regulated by ARF17 in primexine formation. RPG1 is down-regulated in arf17 (Fig. 6). However, no AuxRE is present in the RPG1 promoter; it is likely that ARF17 regulates RPG1 expression through an unknown intermediate transcription factor. Therefore, reduced expression of both CalS5 and RPG1 should partially contribute to the primexine defect in arf17. Expression of DEX1, NEF1 and NPU was not altered in the arf17 mutant (Fig. 6), indicating that ARF17 does not regulate their expression at the transcription level. We propose that there are some novel genes required for primexine formation that can also be regulated by ARF17. Further identification of direct targets of ARF17 could help us to understand the regulatory role of ARF17 and the underlying mechanism controlling primexine formation and pollen wall development.
The expression of ARF17 in germinated pollen grains and pollen tubes, and the reduced male gametophyte transmission efficiency growth in heterozygous plants indicated that ARF17 plays a role in pollen tube growth. It was reported that IAA could stimulate pollen tube growth in Torenia fournieri (Wu et al., 2008). ARF17 may act as an important component in auxin signaling to regulate pollen tube growth in Arabidopsis. ARF17 is essential for pollen wall formation (Fig. 5). However, the genetic analysis showed that arf17 pollen can grow and transmit, although the pollen tube growth is slower than wild type (Table1; Fig. 8). This suggests that other ARFs may also involve in pollen tube growth.
Evidence for the localization, biosynthesis and transport of auxin indicates that auxin regulates dehiscence, pollen maturation, and filament elongation during late anther development (Cecchetti et al., 2008). Expression of an auxin-sensitive reporter marker gene (DR5::GUS/GFP) in anthers also supports the involvement of auxin in these biological processes (Aloni et al., 2006;Cecchetti et al., 2008;Feng et al., 2006). In anther, high-level expression of DR5::GUS/GFP could be detected after the release of the microspores from the tetrads, at which time ARF17 was also expressed. The reduced expression of DR5::GFP in arf17 anthers indicated that ARF17 is important for anther auxin signaling and expression of auxin response genes. DR5::GFP expression in wild-type tetrads was not observed (Fig. 9), which is consistent with a previous report (Cecchetti et al., 2008). However, ARF17 is present at this stage. The failure to detect DR5::GFP in the tetrad could be due to its low level of expression or a lack of necessary co-factors at this stage (Nakamura et al., 2003).

Plant Materials and Growth Conditions
Arabidopsis thaliana wild-type and mutant plants in this study were in the Columbia (Col-0) ecotype background. The arf17 mutant was isolated from the pSKI15 activation tagging T-DNA mutant pools (Qin et al., 2003). Seeds were sown on vermiculite and allowed to imbibe for 3 days at 4°C. Plants were grown under long-day conditions (16-h light/8-h dark) in a growth room at approximately 22°C.
Anthers at the tetrad stage were squeezed onto the slide and stained with toluidine blue (100μg/L in 10 mM KPO4 buffer, pH 7.2) or aniline blue solution (0.1g/L in 50 mM KPO4 buffer, pH 7.5). After covering with cover glasses, the slides stained by toluidine  Table S1) and arbitrary degenerate primers. The TAIL-PCR procedure and arbitrary degenerate primers used were as described (Liu et al., 1995). Linkage analysis of the T-DNA insertion with the mutant phenotype was analyzed with primer AtLB3 and the ARF17-specific primers LP and RP (Supplemental Table S1).
The LP/RP primers were also used for detecting ARF17 gene expression.

Genetic Complementation
For complementation, a 4.9 kb ARF17 genomic fragment, which included 1615-bp of upstream and 0.5 kb of downstream sequence, was amplified from wild-type genomic DNA, using the primer pair CMP-F and CMP-R. For GFP fusion, the ARF17 genomic fragment without a stop codon was cloned into the modified GFP-pCAMBIA1300 vector using primer pair CMP-F and CMP-F/FU (Supplemental Table S1). All amplified DNA fragments were cloned into pMD18-T (Takara) for sequencing and sub-cloning. After verification by sequencing, the fragments were released by restriction enzyme digestion and sub-cloned into the binary vector pCAMBIA1300 (CAMBIA). The resulting constructs were introduced into heterozygous arf17 plants using the flower dip method (Clough and Bent, 1998) with Agrobacterium tumefaciens strain GV3101 (pMP90).
Hygromycin-resistant transformants were indentified in the arf17 background using primer pairs MC-F/RP and MC-F/AtLB3 (Supplemental Table S1).

RT-PCR and Real-time PCR Analysis
Total RNA was isolated from floral buds using the TRIzol kit (Invitrogen). After treatment with DNase (Promega), first-strand cDNA was synthesized from 2  Table S1.

In Situ Hybridization
The probe fragments were amplified from wild type Col-0 cDNA using gene-specific primers (CalS5in-F and CalS5in-R; ARF17in-F and ARF17in-R). After sequencing, the PCR products were cloned into the pBluescript SK vector. Plasmid DNA was completely digested with EcoRIor BamHI. The digestion products were used as templates for transcription into sense and anti-sense probes by T3 or T7 RNA polymerases, respectively (Roche). Images were taken using the Olympus BX-51 microscope.

Chromatin Immunoprecipitation (ChIP)
The ChIP procedure was performed on 35-day-old flower buds as described by   Supplemental Table S1.

Electrophoretic Mobility Shift Assays (EMSA)
To obtain purified ARF17 protein for the EMSA experiments, the full-length fragment of the ARF17 gene was amplified using the primer pairs ARF17GST-F and ARF17GST-R and cloned into the pGEX-4T vector (GE Healthcare, http://www.gehealthcare.com) to produce the GST-ARF17 construct. Expression and purification of the fusion protein were performed according to the manufacturer's instructions. The DNA fragment containing the AuxRE (GAGACA) in the ARF17 regulatory region was generated by PCR amplification with the following pecific primers (EMSAbio-F/EMSAbio-R and EMSAcomp-F/EMSAcomp-R) that were used to generate a biotin-labeled and competitor probe respectively. The EMSA was performed with a LightShift Chemiluminescent EMSA Kit (Thermo Scientific, http://www.thermoscientific.com). The bingding reactions containing binding buffer [10 mM Tris-HCl, pH 7.5, 50 mM NaCl, 1 mM EDTA, 5% glycerol, 0.05 mg mL -1 poly(dI-dC)], 0.5μg GST-ARF17 recombinant fusion protein and 10 fmol biotin-labeled DNA were performed at room temperature for 20 min. The subsequent processes were performed according to the manufacturer's instructions.       Sporopollenin is randomly deposited around the arf17 microspore. G and H, bicellular stage. The exine layer is established in the WT (G). Microspore exine is absent and the microspore is degenerate in the arf17 mutant (H). Ba, bacula; Cl, callose; dMsp, degenerate microspore; Ex, exine; Ms, microsporocyte; Msp, microspore; Pr, primexine; Sp, sporopollenin; Tc, tectum; Tds, tetrads. Bar = 2μm.  Enrichment was only observed when the fragment containing the AuxRE was amplified when affinity-purified GFP antibody was used. C, EMSA assay for ARF17 binding to the CalS5 gene promoter. GST tagged ARF17 protein was mixed with a biotin-labeled 64-bp probe containing the AuxRE and 5-or 25-fold unlabeled probe as competitor. The arrowhead indicates shift band. FP, free probe.