A myosin XI adaptor, TAPE, is essential for pollen tube elongation in rice.

Pollen tube (PT) elongation is important for double fertilization in angiosperms and affects the seed-setting rate and, therefore, crop productivity. Compared to Arabidopsis (Arabidopsis thaliana L.), information on PT elongation in rice (Oryza sativa L.) is limited by the difficulty in obtaining homozygous mutants. In a screen of T-DNA insertional mutants, we identified a mutant in the Tethering protein of actomyosin transport in pollen tube elongation (TAPE) gene with an unusual segregation ratio by genotyping analysis. A CRISPR/Cas9 knockout mutant of TAPE that produced a short PT was sterile, and TAPE was expressed specifically in pollen grains. TAPE is a homolog of a myosin XI adaptor in Arabidopsis with three tetratricopeptide repeat and Phox and Bem1 protein domains. TAPE showed latrunculin B-sensitive, actin-dependent localization to the endoplasmic reticulum. Yeast two-hybrid screening and transcriptome analysis revealed that TAPE interacted with pollen-specific LIM protein 2b and elongation factor 1-alpha. Loss of TAPE affected transcription of 1,259 genes, especially genes related to cell organization, which were downregulated. In summary, TAPE encodes a myosin XI adaptor essential for rice PT elongation.

Introduction

Pollen tube (PT) elongation, an essential process for successful sexual reproduction in angiosperms, is a model system for rapid polar growth (Kroeger and Geitmann, 2012). After pollination, pollen grains attach to the stigma, become hydrated (Moon and Jung, 2020), germinate, and elongate the PT into the transmitting tract in the pistil. The immotile male gametes (two sperm cells and a generative cell), initially released into the ovule via a delicately controlled PT, fuse with and fertilize the egg cell and central cell (Johnson et al., 2019).

Several PT elongation genes have been studied in Arabidopsis (Arabidopsis thaliana L.). Two Catharanthus roseus receptor-like kinase (CrRLKs) family proteins, Buddha’s Paper Seal 1 (BUPS1) and BUPS2 have redundant roles in normal PT elongation in the pistil (Zhu et al., 2018). The CrRLKs interact with Rho-of-plants guanine nucleotide exchange factors (RopGEFs), and a mutation of the RopGEFs resulted in depolarized growth of the PT in the pistil. Along with the two CrRLKs, rapid alkalinization factors (RALFs) are also involved in PT elongation. Pollen-expressed peptides, RALF4 and RALF19 interact with leucine-rich repeat extensin to control PT integrity during elongation (Mecchia et al., 2017). In addition, LORELEI-like GPI-anchored protein 2 (LLG2) and LLG3 control PT growth via the interaction of ANXUR and BUPS in a RALF4-dependent manner (Feng et al., 2019). A Rho-type GTPase of plants, ROP1 in Arabidopsis is a key regulator for establishing the polarity in PT growth, which is coupled to Ca2+ influx and actin filament (AF) dynamics (Li et al., 1999; Gu et al., 2003). Additional genes involved in AF organization (Xu and Huang, 2020), cell wall integrity (Chen and Ye, 2007; Vogler et al., 2019), and ion gradient (Michard et al., 2017; Hoffmann et al., 2020) are required for PT elongation.

Actomyosin-dependent transport mediated by the AF-related motor protein myosin is not only in plant organelle transport generating cytoplasmic streaming (Ross et al., 2008; Tominaga and Ito, 2015) but also in PT elongation (Cai and Cresti, 2009; Madison et al., 2015). In Arabidopsis, there are 17 myosins, classified into two subgroups, four myosins in class VIII and 13 myosins in class XI (Avisar et al., 2009). These plant-specific myosins resemble vertebrate myosin V, consisting of a motor domain, six IQ motifs, a coiled-coil domain, and a C-terminal globular tail domain (GTD) (Tominaga and Nakano, 2012). Due to its high velocity, with a 35-nm step size (Tominaga et al., 2003), myosin XI is important for the motility of organelles such as the endoplasmic reticulum (ER), nucleus, Golgi stacks, peroxisomes, and mitochondria (Ueda et al., 2015). For example, myosin XI-K is involved in ER motility, and AF bundle organization depends on interactions with myosin XI-2 (Ueda et al., 2010). Myosin XI-I localizes on the nuclear membrane and moves to the nucleus via interaction with WPP domain-interacting tail-anchored protein 1 (WIP1) and WIP2 (Tamura et al., 2013). In addition, myosin XI-C and myosin XI-E mediate Golgi stacks, peroxisomes, and mitochondrial movement along with secretory vesicle transport in Arabidopsis PT (Wang et al., 2020). Two myosin XI receptors, myosin-binding protein 1 (MyoB1) and MyoB2, identified by a yeast two-hybrid (Y2H) screening assay using the GTD of myosin XI, contain a transmembrane domain and a domain of unknown function 593 (DUF593) (Peremyslov et al., 2013). They are involved in vesicular cargo transport together with myosins XI. In tobacco (Nicotiana tabacum L.), RAC/ROP GTPase5 (RAC5) and the DUF593 containing protein, RAC5 interacting subapical PT protein regulate trafficking of the trans-Golgi network in PT growth (Stephan et al., 2014). Two other adaptor families in Arabidopsis, myosin XI adaptors of family A (MadA) and MadB, have four members in each family (Kurth et al., 2017). Although their role in PT elongation is unknown, mutants with multiple MadB knockout (KO) genes show a progressive reduction in root hair length with increasing numbers of MadB KO genes.

Rice (Oryza sativa L.) is an important crop plant that supplies calories for almost half of mankind. Unlike Arabidopsis, rice has an ovule in a carpel; therefore, there has been little characterization of rice genes related to pollen germination and PT elongation (Kim et al., 2019b). Using genome editing to generate homozygous (HM) KO mutants at the T0 generation, several genes involved in rice PT elongation have been identified. Ruptured PT and Male-gene transfer defective 2, members of the CrRLK family in rice, maintain PT integrity by interacting with K+ transporters and by balancing reactive oxygen species, respectively (Liu et al., 2016; Kim et al., 2021b). In addition, Germinating modulator of rice pollen and its downstream interactor, an AP180 N-terminal homology (ANTH) domain-containing protein, OsANTH3 affect PT growth in rice by regulating clathrin-mediated endocytosis (Kim et al., 2021a; Lee et al., 2021). Although actomyosin transport is important in the PT elongation process and 14 class XI myosins were identified in rice (Jiang and Ramachandran, 2004), the mechanism of PT elongation is unknown.

Here, we report a gene, PT (TAPE), which encodes a myosin XI adaptor essential for PT elongation in rice. We present a regulatory model of TAPE in PT elongation that is supported by various experimental evidence and characterize a myosin XI adaptor in PT elongation. This information could be useful in understanding the actomyosin transport mechanism for PT elongation in plants and be used for applications to maintain crop yield.

Results

Identification of TAPE and its impact on rice male gamete transmission

From a large-scale T-DNA mutant screening for 627 late pollen-preferred genes (Moon et al., 2018), we identified genes related to gamete gene transfer in rice. A T-DNA insertion line in TAPE (2B-00091) had a 1:1 wild-type (WT) to heterozygote (TAPE/tape) segregation ratio without HM progeny. We confirmed by genotyping that a T-DNA was located in the intron of TAPE on the backside of three tetratricopeptide repeat (TPR) and Phox and Bem1 (PB1) domains (Figure 1A). To determine whether the mutation caused a defect in male or female gamete transfer, we performed reciprocal crosses. The TAPE/tape plant as a paternal line (male donor) crossed with the WT produced only TAPE/TAPE progeny. In contrast, when the TAPE/tape plant was the maternal line (female donor), there was almost a 1:1 ratio of TAPE/TAPE to TAPE/tape progeny (Table 1). Thus, TAPE is an essential gene for male gamete transmission.

Figure 1

Gene structure of TAPE and the phenotype of the tape-1 mutant. A, Gene structure of TAPE and the T-DNA insertion and CRISPR/Cas9 mutants (tape-1 and tape-2). The black and white boxes indicate exon and untranslated regions, respectively. The gray box represents the T-DNA fragment inserted in TAPE. The P-value from the χ2 test result compared to the predicted 1:2:1 ratio for the WT:HZ:HM is shown. Red letters indicate insertions or deletions in the tape-1 and tape-2 mutants with altered amino acids. B and C, Phenotypes of the WT and tape-1 plants showing sterility of tape-1 plants without vegetative growth defects. Bar = 10 cm (B), 2 cm (C). D, Morphology of the spikelet and floret of WT and tape-1 plants. Bar = 2 mm. E–F, pollen stained with I2-KI (E), auramine O, or calcofluor white solutions (F) for starch content, and the exine and intine layers of the pollen, respectively. Bar = 200 μm (E), 100 μm (F).

Table 1

Reciprocal crossing assay of TAPE

Genotype of Parents		Genotype of Progeny			Expected Ratio	Observed Ratio	χ 2 test (1:1:0)
♀	♂	TAPE/TAPE (WT)	TAPE/tape (+/−)	tape/tape (−/−)	WT:(+/−):(−/−)	WT:(+/−):(−/−)	P-value
TAPE/TAPE	TAPE/tape	19	0	0	1:1:0	19:0:0	1.307E-0.5
♀	♂	TAPE/TAPE (WT)	TAPE/tape (+/−)	tape/tape (−/−)	WT:(+/−):(−/−)	WT:(+/−):(−/−)	P-value
TAPE/tape	TAPE/TAPE	28	42	0	1:1:0	2:3:0	0.09

tape-1 exhibits short PT growth without germination defects

Since we could not obtain the HM mutant of TAPE by traditional methods, we generated two independent HM mutants (tape-1 and -2) using the CRISPR/Cas9 system and found a sterile phenotype in two independent gene-edited lines (Figure 1, A–C; Supplemental Figure S1, A and B). To determine whether sterility resulted from developmental defects of the anther or pollen, we analyzed the morphology of the anther and pollen in the maturation stage (Figure 1, D–F; Supplemental Figure S1, C–E). The anthers in the WT and two HM strains (tape-1 and -2) both developed normally (Figure 1D; Supplemental Figure S1C). Starch accumulation, seen by staining with I2-KI, and wall formation in pollen grains, seen by staining with calcofluor white and auramine O for intine and exine layers, respectively, were normal (Figure 1, E and F; Supplemental Figure S1, D and E). Therefore, the sterile phenotypes in the tape-1 and tape-2 mutants were not caused by developmental defects of the anther and pollen at the maturation stage.

When we compared in vitro pollen germination between the WT and tape-1 strains, we observed short PTs in the tape-1 strain compared to the WT (Figure 2, A and B) without substantial differences in the pollen germination rate (Supplemental Figure S2). The short PTs in the tape-1 strain were also seen in the in vivo germination assay using the pistils after 2 h of pollination (Figure 2, C and D). The PTs in the tape-1 strain were attached and germinated on the stigma but, unlike the WT, did not reach the transmitting track on the style. In the magnified PTs of WT and tape-1, the PTs of the tape-1 strain had a short but intact PT (Figure 2, E and F), which was reduced about 30% in length in tape-1 (Figure 2G) leading to the male gamete defect. The short PT phenotype was also observed in the tape-2 allele in both in vitro and in vivo germination assays (Supplemental Figure S1, F and G).

Figure 2

In vitro and in vivo germination of WT and tape-1 pollen. A–B, In vitro germination of WT and tape-1 pollen showing the short PTs of tape-1. Bar = 200 μm. C–D, In vivo germination 2 h after pollination. The tape-1 pollen did not elongate sufficiently to penetrate the stigma. Bar = 200 μm. E–F, Magnified photograph of fully elongated PTs from the WT compared to tape-1. Bar = 20 μm. G, Box plot of PT length in WT and tape-1. It indicates upper and lower quartiles, median, and outlier of measured PT length. At least 150 PTs were counted. *** indicates P-value < 0.001 calculated from the Student’s t test.

TAPE is expressed specifically in the pollen grain and PT

We measured gene expression by reverse transcription-quantitative PCR (RT-qPCR) and tissue and developmental expression patterns using a histochemical β-glucuronidase (GUS) assay (Figure 3, A and B). RT-qPCR analysis of nine samples including vegetative tissues, panicle, developing anthers, mature pollen, and developing seeds showed that TAPE was expressed specifically in mature pollen at levels 4 times higher than a reference gene, OsUbi5 (Figure 3A). For plant expression patterns of TAPE, we generated a native TAPE promoter-driven GUS transgenic plant (pTAPE::TAPE-GUS) and performed histochemical analysis during late anther developmental stages (Figure 3B; Zhang et al., 2011). This revealed that TAPE was expressed specifically in mature pollen but not in other floret tissues. The expression began after the bicellular pollen stage as predicted by the rice male gamete expression database (Chandran et al., 2020). TAPE-GUS proteins were localized on PTs in the germinated pollen of the pTAPE::TAPE-GUS line.

Figure 3

Expression and histochemical assays of TAPE. A, Relative TAPE expression in various rice tissues and tetrad, microspore, and vacuolated stages of the anther as determined by RT-qPCR. Values are presented as the mean ± standard deviation for three biological replicates, normalized to the reference gene OsUbi5 (Os01g22490) using the 2−ΔΔCt method. B, Florets from the pTAPE::TAPE-GUS transgenic plants were stained for GUS expression. Pollen was released from the anther by tweezing with forceps. Bar = 2 mm. Stages above the floret images indicate anther development stages (Zhang et al., 2011). YM, young microspore; VP, vacuolated pollen; BP, bicellular pollen; TP, tricellular pollen; MP, mature pollen.

TAPE protein is mainly localized to the ER

To determine the subcellular localization of the TAPE protein, we constructed a C-terminal green fluorescent protein (GFP)-fused TAPE (p35s::TAPE-GFP) in a plasmid for a tobacco (Nicotiana benthamiana L.) infiltration assay. In N. benthamiana epidermal cells, we observed that localization of the TAPE-GFP protein differed from GFP localized in the nucleus and cytosol (Figure 4A). TAPE-GFP was localized in a mesh structure all over the cells (Supplemental Movie S1), similar to the reported pattern of the ER in N. benthamiana cells (Cao et al., 2016). Using co-localization of TAPE-GFP with ER or cis-Golgi marker proteins, OsSPP1-mCherry-HDEL and OsManI-mCherry, respectively (Dangol et al., 2017), we found co-localization of TAPE-GFP with the ER marker protein but not with a cis-Golgi marker (Figure 4, B and C; Supplemental Movies S2 and S3).

Figure 4

Subcellular localization of TAPE-GFP in N. benthamiana epidermal cells and rice PT. A, Subcellular localization of GFP and TAPE-GFP in N. benthamiana epidermal cell. The images were processed with the orthogonal projection method to check the in-depth subcellular localization of the TAPE-GFP. Bar = 20 μm. B–C, Co-localization of TAPE-GFP with the ER (B) and cis-Golgi (C) in N. benthamiana cells. The signal peptides from OsSPP1 and OsManI were fused with mCherry and used as ER and cis-Golgi marker proteins, respectively. The images were processed by orthogonal projection. Bar = 20 μm (B) and 10 μm (C). D–E, Subcellular localization of TAPE in PTs of the pTAPE::TAPE-GFP transgenic line (TAPE-GFP line). A representative pollen grain was photographed before (D) and after pollen germination (E) by confocal microscopy. Bar = 5 μm (D and E). F–G, Co-localization of the PM and ER in PTs of TAPE-GFP lines. The PTs were probed with FM4-64 (F) and ER-Tracker dyes (G). Bars = 2 μm and 5 μm, respectively.

To verify and further characterize TAPE localization, we studied the subcellular localization of a C-terminal, GFP-fused transgenic line (pTAPE::TAPE-GFP; Figure 4, D–G). TAPE-GFP was localized evenly throughout the pollen grain (Figure 4D) and at growing PTs during germination (Figure 4E). In elongated PT, the TAPE-GFP did not colocalize with the plasma membrane (PM; Figure 4F) but colocalized with the ER at the subapical region of the PT (Figure 4G). These results confirmed that TAPE-GFP localized in the ER in both N. benthamiana epidermal cells and in the growing PT in rice, and it moved toward the tip of the PT when the pollen germinated.

TAPE encodes a myosin XI adaptor protein in rice

To obtain clues to the function of TAPE, we compared the amino acid sequence to its homologs in the Phytozome database (Supplemental Figure S3). We used homologs of TAPE with three TPR domains and a PB1 domain from 29 dicot and 13 monocot plants. Interestingly, these homologs were conserved in angiosperms, and they clustered according to the monocot and dicot classification. In this analysis, we found that TAPE is a putative ortholog of a myosin XI adaptor coding gene MadB1 in Arabidopsis (Kurth et al., 2017). In addition, one of the maize (Zea mays L.) homologs, ZmHIP (GRMZM2G023275), interacted with OPAQUE1, a myosin XI in maize required for ER motility in the endosperm (Wang et al., 2012). These results suggested that TAPE might function as a myosin XI adaptor in rice. To confirm our hypothesis, we tested the binding between TAPE and the GTD of a myosin XI (LOC4330906) that is closely related to the maize OPAQUE1 in rice (Figure 5, A–D; Supplemental Figure S4A). Using a bait constructed with the C-terminal region of TAPE, we performed a Y2H assay (Figure 5A). The bait did not self-activate (Figure 5B) and interacted with a GTD of myosin XI upon selection (Figure 5C). To confirm this interaction, we performed bimolecular fluorescence complementation (BiFC) and co-immunoprecipitation (co-IP) assays (Supplemental Figure S4A; Figure 5D). Both experiments confirmed that TAPE was bound to myosin XI which is expressed in rice pollen (Supplemental Figure S4B).

Figure 5

Myosin XI binding and AF-dependent localization of TAPE. A, Gene structure of TAPE and the bait region used for myosin XI binding assays. The orange dotted box indicates the bait region, corresponding to a homolog in maize (Wang et al., 2012). B, Y2H self-activation test for the TAPE bait encoding the GAL4 DNA binding domain fused to the C-terminal of TAPE. The bait vector contained murine p53 and prey vector with SV40 large T-antigen as a positive control. Empty vectors were the negative control. C, TAPE bound to the GTD of the myosin XI (LOC4330906) in the Y2H assay. D, co-IP of TAPE and myosin XI GTD. Ehd1-HA and Myc were the negative and positive controls, respectively. E–F, LatB treatment for AF-dependent localization of TAPE in N. benthamiana epidermal cells (E) and rice PT (F). Nicotiana benthamiana and pollen were treated with LatB at 10 nM and 2 nM, respectively, for an hour. The red asterisk in F indicates abnormally localized TAPE-GFP in the PT. Bar = 5 μm.

In addition, we determined whether localization of TAPE-GFP depended on AF for TAPE to function as a myosin XI adaptor (Figure 5, E and F). Initially, we infiltrated the N. benthamiana leaf with TAPE-GFP and an AF marker, Lifeact-mCherry (Riedl et al., 2008), and treated the leaf with 10 nM of latrunculin B, an AF inhibitor (LatB). When LatB was added, the TAPE-GFP localization changed forming a small lump compared to the MOCK treatment (Figure 5E), and the dynamic movement of the TAPE with ER marker stopped with LatB (Supplemental Movie S4). TAPE localized differently in pollen grains and PT when it was treated with 2-nM LatB forming lumps (Figure 5F; Supplemental Figure S5). These results confirmed that TAPE is a gene encoding a myosin XI adaptor.

TAPE interacts with OsPLIM2b and EF1a

TPR and PB1 are well-known domains that mediate various protein-protein interactions (D’Andrea and Regan, 2003; Sumimoto et al., 2007). Therefore, we searched for interacting proteins of TAPE by Y2H screening using a yeast prey library with rice anther cDNAs (Figure 6; Supplemental Figure S6). When we constructed baits A and B, the GAL4 DNA binding domain fused to the TPR and PB1 domain, respectively, and Y2H screening (Figure 6A; Supplemental Figure S6A) identified five candidate proteins. We found one candidate protein for bait A (Os04g54010 encoding LIM-domain containing protein, PLIM2b) and two proteins for bait B (Os03g08050 encoding elongation factor 1 alpha, EF1a, and Os10g40090 encoding expansinB9 [ExpB9]; Figure 6B; Supplemental Figure S6B). To confirm these interactors, we performed subcellular localization and BiFC assays (Supplemental Figure S7). OsPLIM2b and EF1a both showed co-localization and interaction with TAPE inside the N. benthamiana epidermal cells (Supplemental Figure S7, A and B), and those two interactions were confirmed by co-IP analysis indicating that TAPE interacts with OsPLIM2b and EF1a (Figure 6C). The RT-qPCR results of these two interactors support that these interactions could occur in rice pollen (Supplemental Figure S7C).

Figure 6

Identification of interaction partners of the TAPE protein. A, Diagrams of two bait vectors for Y2H screening. BD, DNA binding domain of yeast GAL4 transcription factor. B, Partners interacting with TAPE by Y2H. Empty, an empty bait vector used as a negative control. Hits, the number of times the interaction occurred in the Y2H screening. C, co-IP of TAPE with proteins OsPLIM2b and EF1a.

TAPE affects the transcription of genes related to cell organization, especially kinesin motors

Using anthers from the WT and tape-1 strains we performed RNA-Seq analysis to understand the transcriptional regulation of TAPE. We identified 1,259 differentially expressed genes (DEGs), in the tape-1 anther (799 were upregulated and 460 were downregulated) which were analyzed using the cell function overview in MapMan (Figure 7A). We found that genes for regulation, protein modification or degradation, and transport were differentially expressed in the tape-1 anther. Notably, the cell organization gene category was substantially downregulated in the mutant anther. A Gene Ontology (GO) enrichment with the downregulated DEGs (Figure 7B) identified microtubule (MT)-based movement as the most enriched GO term. The GO term contains seven kinesin motor genes showing significant transcript reduction in the tape-1 anther. We confirmed the reduction in expression for three kinesin motor-coding genes in the tape-1 anther by RT-qPCR analysis (Figure 7C). These data suggest that TAPE is involved in the transcription of MT-related genes, especially kinesin motors.

Figure 7

Transcriptome analysis of the tape-1 anther. A–B, Functional classification of DEGs in the tape-1 anther compared to the WT anther. A, Cell function overview in the MapMan program. The red and blue squares indicate upregulated and downregulated DEGs in tape-1, respectively. The cell organization category is highlighted with red letters. B, A Gene Ontology (GO) enrichment analysis of the downregulated DEGs to identify GO terms related to cell organization. Only GO terms with more than five query numbers were used. The size and color of the dots indicate the statistical significance and enrichment value for each GO term. C, Expression of genes for MT motors in tape-1 anther. Relative expression was determined by RT-qPCR and calculated as 2−ΔΔCt with normalization to OsUbi5. Bars indicate the mean values of the three replicates with their standard deviations. P-values from the Student’s t test are indicated as follow: *P < 0.05, **P < 0.01, and ***P < 0.001.

Discussion

TAPE is an essential myosin XI adaptor for normal PT elongation in rice

PT elongation in rice is important for successful male gamete delivery and seed setting, which affects yield. In addition, the genes for PT elongation are a valuable genetic resource for the two-line hybrid system in the seed industry. Despite their importance, few genes have been characterized (Kim et al., 2019b). Here, we identified and characterized TAPE, which encodes a protein with three TPRs and a PB1 domain (Figure 1A) and is expressed in pollen after the bicellular pollen stage (Figure 3). We demonstrated by segregation experiments and reciprocal crossing using T-DNA lines that TAPE is a key player in male gamete transfer (Figure 1A; Table 1). HM KO mutants tape-1 and tape-2, generated by CRISPR/Cas9, were sterile (Figure 1, B and C; Supplemental Figure S1, A and B). The tape-1 strain had PTs that were short but intact and ruptured normally after elongation (Figure 2) and had normal germination compared to the WT (Supplemental Figure S2). Thus, TAPE is responsible for PT elongation in rice. Comparison of TAPE with other proteins revealed it to be a myosin XI adaptor protein (Supplemental Figure S3) which was confirmed by its interaction with the GTD of pollen-expressed myosin XI in protein-binding assays (Figure 5, C and D; Supplemental Figure S4). Together, our data indicate that TAPE is a key myosin XI adaptor that is required for normal PT elongation in rice.

TAPE is a possible mediator of actomyosin-driven transport of organelles and proteins in rice PT elongation

TAPE was mainly localized to the ER in both N. benthamiana epidermal cells and rice PT (Figure 4, B and G), consistent with a previous report for ZmHIP, a TAPE-homolog in maize (Wang et al., 2012). ZmHIP binds to the myosin XI OPAQUE1 and is involved in ER motility in the maize endosperm. By Y2H screening we identified interactions between TAPE and OsPLIM2b and EF1a that are expressed in rice pollen (Figure 6; Supplemental Figure S7). LatB treatment showing AF-dependent movement of TAPE (Figure 5, E and F; Supplemental Figure S5; Supplemental Movies S2 and S4) and co-localization assays of TAPE with its interactors and an ER marker (Supplemental Movies S5 and S6) indicate that TAPE might mediate the actomyosin-dependent transport of the ER and the interactors in rice PT elongation.

The interaction of plant organelles with membrane contact sites is important in organelle motility (Perico and Sparkes, 2018). Consistently, we observed similar dynamics in N. benthamiana cells between TAPE and the cis-Golgi marker in our co-localization assay (Figure 4C; Supplemental Movie S3). In Arabidopsis, MadB1, also known as AtPhox1 (Prasad et al., 2010), and a putative ortholog of TAPE (Supplemental Figure S3), encodes a carboxylate clamp-type TPR protein and a co-chaperone of heat shock protein 70 (Hsp70) and Hsp90. MadB2, which is a paralog of MadB1 in Arabidopsis, is also known as CLUMPED CHLOROPLAST1 (CLMP1), and a KO clmp1-1 mutant showed abnormal plastid distribution in various organs (Yang et al., 2011). How TAPE can localize to the ER without a transmembrane domain remains to be determined; however, these results suggest that TAPE might mediate the actomyosin-dependent transport of various organelles and proteins during PT elongation in rice.

TAPE is involved in the transcriptional regulatory network in rice PT elongation

Transcriptome analysis of tape-1 versus WT anthers revealed that TAPE affects transcriptional regulation for various biological functions, especially cell organization (Figure 7A). Paradoxically, the absence of TAPE affected the transcription of MT-related motors, such as kinesins, even though TAPE is a myosin XI adaptor for the AF-related motor (Figure 7, B and C). Further research will be needed to determine whether there is a transcriptional connection between the two cytoskeletons, AF and MT, in addition to the physical connection previously reported (Schneider and Persson, 2015).

Conclusion

In this study, we report that TAPE encodes an essential myosin XI adaptor associated with the actomyosin-dependent movement in rice PT elongation, and we provide a model for the role of TAPE in the rice PT elongation process (Figure 8). The functional characterization of a myosin XI adaptor, TAPE will improve our understanding of myosin XI-mediated transport in plants and the PT elongation process in rice.

Figure 8

Model for TAPE functions in rice PT elongation. TAPE transcriptionally regulates 1,259 DEGs, especially MT-movement associate motors, kinesins. TAPE is a myosin XI adaptor protein that is localized to the ER where it mediates actomyosin-based transport by binding with interactors such as OsPLIM2b and EF1a in rice PT elongation. In the tape KO mutant, actomyosin movement of the ER and interactors is disrupted so that the PT cannot elongate sufficiently for fertilization to occur.

Materials and methods

Plant materials and growth conditions

WT rice (O.sativa spp. Japonica cv. Dongin and Hwayoung) and transgenic lines were grown in a growth chamber under the following conditions: 28°C/25°C (day/night), 16/8 h (light/dark), and 80% humidity for 2 weeks and transferred to the paddy field located at Kyung Hee University in Yongin, Republic of Korea, from May 2018 to October 2020. The T-DNA mutant of TAPE (2B-00091; cv. Hwayoung), was obtained from the seed bank in the Kyung Hee University (Yi and An, 2013). We used the heterozygous (HZ) mutant for the segregation analysis and reciprocal crossing. The seeds were grown on Murashige and Skoog media for 2 weeks in a CO2 incubator and transferred to the growth chamber. Nicotiana benthamiana plants for the agro-infiltration assay were grown in a growth chamber at 25°C, in 12/12 h (light/dark) and 60% humidity for 3 weeks.

Vector cloning and tissue culture

Guide RNAs (gRNAs) to generate two KO mutants (tape-1 and tape-2) of TAPE were designed using CRISPRdirect and CAFRI-Rice web software (Naito et al., 2015; Hong et al., 2020). Oligomers corresponding to the gRNAs were annealed and ligated into the BsaI-treated pRGEB32 binary vector (Addgene plasmid number: 63142).

To generate the TAPE promoter-driven GFP (pTAPE::TAPE-GFP), and GUS fusion transgenic lines (pTAPE::TAPE-GUS), a 1,233-bp promoter with the 2,523-bp coding sequence (CDS), except for stop codon, of TAPE was amplified. The amplicon was ligated into the BamHI-treated pGA3519 binary vector and the BamHI and HindIII-treated pGA3427 binary vector for C-terminal (C-term) GUS and GFP fusions, respectively, using the In-Fusion HD Cloning Kit (639648; Takara, Shiga, Japan). The constructs were transformed into Agrobacterium tumefaciens LBA4404 for stable transformation by Agrobacterium-mediated co-cultivation (Lee et al., 1999).

The CDSs of TAPE and its interactors (OsPLIM2b, Os04g45010; EF1a, Os03g08050; OsExpB9, and Os10g40090) were amplified to generate p35S driven C-term GFP (p35S::TAPE-GFP) and mRFP-fused constructs (p35S::interactors-mRFP). The amplicons were cloned into HindIII-digested vectors. For the BiFC assay, the amplified CDSs were ligated into the HindIII-digested p35S driven N- (NV) or C-term Venus (CV) binary vectors. The primers used in cloning are listed in Supplemental Table S1.

Plant phenotyping and pollen germination analysis

The WT and tape-1 plants were photographed before harvesting using a COOLPIX P900s camera (Nikon, Tokyo, Japan), and the spikelets from those plants at the anthesis stage were collected and imaged using a SZX61 microscope (Olympus, Tokyo, Japan).

To observe pollen morphology, the pollen grains were collected by tweezing the anther and were stained for starch content, and inner and outer cell wall layers. Briefly, a 1% w/v I2-KI solution was used to stain starch and 0.1% calcofluor white and 0.001% auramine O solutions were used to stain intine and exine layers, respectively. The stained pollens were observed using a BX61 microscope (Olympus, Tokyo, Japan) with brightfield, ultra-violet (UV), and fluorescein isothiocyanate channels.

For pollen germination analysis, released pollen grains from the dehiscent anther were collected on a solid or liquid pollen germination medium as previously described (Kim et al., 2020). PT length and pollen germination ratio were measured for at least 150 pollen grains using the ImageJ software. To observe PT elongation in the pistil, we stained the pistil with aniline blue solution as described by Liu et al., (2016) with some modifications. The pistils were collected 2 h after pollination and incubated in Carnoy’s solution overnight. The fixed pistils were washed 5 times with distilled water (DW), incubated in 1 M NaOH solution for 6 h until they became transparent, washed twice with DW, and stained with 0.05% aniline blue (in 0.1 M K2HPO4, pH 8.5 solution) for 1 h in the dark. The stained pistils were observed under the UV channel.

Nucleic acid extraction and RT-qPCR analysis

Genomic DNA, extracted from the leaf sample with cetyltrimethylammonium bromide, was used for genotyping T-DNA mutants and sequencing CRISPR/cas9 mutants. For RNA extraction, various tissue samples as described previously (Moon et al., 2019) were immediately frozen with liquid N2. The pollen grains were stabilized with RNAlater (Invitrogen, Waltham, MA, USA) after collecting fresh samples from dehiscent anthers in the paddy field. Total RNA was extracted according to the manufacturer’s instructions for RNeasy Plant Mini Kit (Qiagen, Hilden, Germany). Complementary DNA (cDNA) was synthesized from RNA using a SuPrimeScript RT premix kit (GeNet Bio, Daejeon, South Korea) with a 1-h incubation at 50°C.

RT-qPCR analysis was performed with the Roter-Gene Q system (Qiagen, Hilden, Germany) and 2X Prime Q-Mastermix (GeNet Bio). Expression was normalized to the rice ubiquitin 5 gene (OsUbi5, Os01g22490) as previously described (Moon et al., 2020) and the 2−ΔΔCt method was used to calculate relative expression (Schmittgen and Livak, 2008). The primers for RT-qPCR analysis are listed in Supplemental Table S1.

Histochemical GUS assay

Spikelets of the GUS fusion transgenic line (pTAPE::TAPE-GUS) at different developmental stages and pollen grains were stained with GUS solution for 2 h at 37°C. The stained spikelet and pollen grains were imaged using the SZX61 microscope.

Subcellular localization analysis

For subcellular localization analysis in N. benthamiana epidermal cells, the cloned p35S-driven constructs were transformed into A. tumefaciens GV3101 using the freeze-thaw method and transformed GV3101 were infiltrated into N. benthamiana leaves using a N. benthamiana infiltration assay as previously reported (Sparkes et al., 2006). After 72 h of infiltration, leaves were observed using an LSM 510 META confocal scanning laser microscope (CSLM; Carl Zeiss, Jena, Germany). GFP and RFP fluorescence was observed under 488/505–530 and 543/560–615 nm filters (excitation/emission wavelength), respectively. Images were analyzed and exported using Zeiss ZEN lite software.

The subcellular localization of TAPE in the PT of the pTAPE::TAPE-GFP plant was also imaged using CSLM. To visualize the PM and ER, the elongated PTs were stained with 10 µΜ FM4-64 (T13320; Invitrogen, Waltham, MA, USA) and 1 µM ER-Tracker Red dye (E34250; Invitrogen, Waltham, MA, USA). To determine the AF-dependent movement of TAPE, 2 nM or 10 nM LatB was added to the infiltrated leaves and PT, respectively.

Phylogenetic analysis

TAPE was compared to the protein sequences of its homologs in angiosperm for 29 dicots and 13 monocots that were retrieved from the Phytozome database (Goodstein et al., 2012). The most similar protein homologs with a similarity score of over 1,000 were selected for each species. The sequences were aligned with the MUSCLE program and a phylogenetic tree based on a neighbor-joining method with 1,000 bootstrapping was created using MEGA X software (Kumar et al., 2018). The information of the homologs is listed in Supplemental Table S2.

Y2H assays

The 4–561-bp, 1,252–1,488-bp, and 1,726–2,619-bp fragments of the TAPE CDS were amplified and cloned into the EcoRI/BamHI-digested pGBKT7 vector to generate GAL4 DNA binding domain fused bait (bait A, B, and C respectively). The constructs were transformed into the yeast AH109 for the self-activation test. These yeast transformants did not grow on a minimal synthetic defined (SD) medium and showed no transcriptional activation. Bait A and B were used for the Y2H screening assay, performed by Panbionet Corp (Pohang, Republic of Korea; www.panbionet.com) with the same rice anther cDNA library used in our previous study (Kim et al., 2021a).

To determine whether TAPE bound myosin XI, a Y2H assay was performed using bait C. The GTD of myosin XI (LOC4330906) was fused with the GAL4 activation domain. SD medium lacking Leu, Trp, and His was the selection media with 5 mM 3-amino-1,2,4-triazole (3-AT).

co-IP assays

For constructing the Myc or HA fusions, TAPE, myosin XI, OsPLIM2b, and EF1a full-length cDNAs without the stop codons were inserted into pGA3817 with the 6X-Myc coding region or into pGA3818 with 3X-HA. The fusion molecules were transformed into Oc cell protoplasts by polyethylene glycol transformation. After incubation for 12–14 h at 28°C, the protoplasts were isolated and used for co-IP assays, as previously described (Cho et al., 2016; Yoon et al., 2017). Before the protoplasts were harvested, anti-HA mouse monoclonal antibodies (12CA5; Roche, Mannheim, Germany, http://www.roche.com) were incubated for 3 h with Protein A/G agarose beads (Millipore, Billerica, MA, USA; http://www.emdmillipore.com) in a binding buffer [75 mM NaCl, 50 mM Tris–HCl (pH 7.5), 5 mM EDTA, 1 mM dithiothreitol, 1 mM phenylmethanesulfonylfluoride, and 1% Triton X-100]. Then the protoplasts were isolated and resuspended in an IP buffer [75 mM NaCl, 50 mM Tris–HCl (pH 7.5), 5 mM EDTA, 1% Triton X-100, 1 mM dithiothreitol, 1 mM phenylmethanesulfonylfluoride, 2 mM NaF, 100 µM MG132, and protease inhibitor cocktail (Roche, Basel, Switzerland)]. Pre-cleared samples were incubated for 6 h with the anti-HA mouse monoclonal antibodies that were bound with the Protein A/G-conjugated agarose beads. After washing with a buffer [75 mM KCl, 50 mM Tris–HCl (pH 7.5), 5 mM EDTA, 0.1% Triton X-100, 1 mM dithiothreitol, and 1 mM phenylmethanesulfonylfluoride], the bound proteins were eluted with IP buffer and a protein loading dye. The eluted proteins were resolved by 10% SDS-PAGE and transferred to a polyvinylidene difluoride membrane (Millipore, Burlington, MA, USA). The blotting membrane was incubated in a blocking solution with HRP-conjugated anti-Myc monoclonal antibodies (#2040; Cell Signaling, Beverly, MA, USA). Proteins bound to the membrane were detected using an enhanced chemiluminescence prime western blotting detection reagent (WSE-7120S; ATTO, Tokyo, Japan; https://www.atto.co.jp/) with the ChemiDoc MP Imaging System (17001402; Biorad, Hercules, USA). Primers for full-length amplifications are listed in Supplemental Table S1.

Transcriptome analysis

Mature pollen grains containing anthers at stages 13–14 (Zhang et al., 2011) were used for RNA-Seq analysis. After the sample quality control process, a sequencing library was constructed using TruSeq Stranded mRNA LT Sample Prep Kit (Illumina, San Diego, CA, USA) following the manufacturer’s instruction (Part #15031047 Rev. E). The library was sequenced using the Illumina NovaSeq 6000 platform at MacroGen Inc with FASTQ formatted raw data files.

The raw files were trimmed using the Cutadapt program (Martin, 2011) with parameters: -a AGATCGGAAGAGC -A AGATCGGAAGAGC -q 30 -m 20 and the reads were mapped to a rice reference genome from the rice genome annotation project (RGAP; http://rice.plantbiology.msu.edu/) using HISAT2 aligner with default parameters (Kim et al., 2019a). The raw reads were counted using the featureCounts program (Liao et al., 2014) and the DEG analysis was performed with the DESeq2 R package (Love et al., 2014). The DEGs were selected based on these criteria: p and adjusted P-value < 0.05, and | log2(fold change) | ≥ 1. GO enrichment and MapMan analyses of the DEGs were performed as previously described (Shin et al., 2020). Among enriched GO terms, we adopted the terms with a query number over five for reliable terms.

Statistical analyses

For statistical analysis on PT length and pollen germination rate, >150 pollen grains of the WT and tape-1 plants were counted. Student’s t test was performed using T.TEST() function embedded in EXCEL software.

Accession numbers

The locus and sequence information for TAPE were retrieved from the RGAP database (TAPE: Os09g03890). The raw data were deposited at the ArrayExpress (https://www.ebi.ac.uk/arrayexpress/) with the accession number E-MTAB-10439.

Supplemental data

The following materials are available in the online version of this article.

Phenotype of the tape-2 mutant.

Germination ratio of the WT plant and tape-1.

Comparative analysis of TAPE and its homologs in angiosperms.

BiFC analysis of TAPE with the C-terminal GTD of myosin XI and expression analysis of the myosin XI gene.

A 2 nM LatB treatment disturbs the subcellular localization of TAPE in the pollen grain and PT.

Self-activation test of the baits used in Y2H assay.

Co-localization and BiFC assays of TAPE with its interactors and expression analysis of genes encoding the interactors.

Primers used in this study.

Information on the TAPE homologs used in comparative studies.

Z-stack images of TAPE-GFP in N. benthamiana epidermal cells.

Time-series images of TAPE-GFP and OsSPP1-mCherry-HDEL in N. benthamiana epidermal cells.

Time-series images of TAPE-GFP and OsManI-mCherry in N. benthamiana epidermal cells.

Time-series images of TAPE-GFP and OsSPP1-mCherry-HDEL in N. benthamiana epidermal cells with 10 nM LatB treatment.

Time-series images of TAPE-NV and PLIM2b-CV with OsSPP1-mCherry-HDEL in N. benthamiana epidermal cells.

Time-series images of TAPE-NV and EF1a-CV with OsSPP1-mCherry-HDEL in N. benthamiana epidermal cells.

Click here for additional data file.