Tangled1: A Microtubule Binding Protein Required for the Spatial Control of Cytokinesis in Maize

Because plant cells are surrounded by walls that immobilize them within tissues, the orientations in which they divide during development constrain patterns of organ growth and are critical for establishing the cellular organization of plant tissues. Plant cells achieve cytokinesis through the action of a phragmoplast, a cytoskeletal structure composed of microtubules (MTs) and F-actin, which directs the formation of a new cell wall between daughter nuclei after mitosis (Staehelin and Hepler 1996). The phragmoplast arises initially in isolation from the parental wall and plasma membrane, and expands centrifugally as cytokinesis proceeds. The location at which the new cell wall will fuse with the parental wall is governed by an interaction between the expanding phragmoplast and a cortical “division site” established before mitosis (Gunning 1982; Smith 2001). This site is marked throughout prophase by a cortical band of MTs and actin filaments, the preprophase band (PPB), which probably plays an essential role in establishing the division site during prophase (Wick 1991; Smith 2001). Very little is known in molecular terms about the nature of the cortical division site or how the phragmoplast interacts with it during cytokinesis.

To better understand the spatial control of cytokinesis in plant cells, we have analyzed the function of a gene in maize, Tangled1 (Tan1), which is required for this process. Throughout the development of tan1 mutant leaves, the majority of cells in all tissue layers divide in abnormal orientations (Smith et al. 1996). Analysis of the cytoskeleton in dividing leaf cells of tan1 mutants showed that all the cytoskeletal structures involved in cell division are formed and appear structurally normal, but are not oriented normally within dividing cells. Abnormally oriented cell divisions can be attributed mainly to the failure of most phragmoplasts to be guided to the former PPB site (Cleary and Smith 1998). Here, we present an initial molecular characterization of the Tan1 gene and its protein product. In combination with our previous analysis of the tan1 mutant phenotype, the results suggest a direct role for TAN1 in orienting cytoskeletal structures during cell division.

tan-Mu1 was isolated from a Mutator mutagenized population (Smith et al. 1996). tan1-py1 was obtained from the Maize Genetics Stock Center. The ethyl methanesulfonate–induced tan-gt1 allele was a gift from Sharon Kessler and Neelima Sinha (University of California at Davis, Davis, CA).

Genomic DNA isolation from leaf tissue and Southern blots was carried out according to standard protocols (Chen and Dellaporta 1994; Ausubel et al. 2000). Blots were hybridized at 65°C in 0.25 M NaPO₄, pH 7.2, with 2% SDS and washed in 0.2× SSC with 0.2% SDS (high stringency) or at 54°C in 0.5 M NaPO₄, pH 7.2, with 7% SDS and washed at 54°C in 100 mM NaPO₄, pH 7.2, with 5% SDS (low stringency). Total RNA was extracted using Trizol reagent (GIBCO BRL) and enriched for poly A⁺ RNA using the PolyATtract mRNA isolation system (Promega). Northern blots were carried out as described by Luehrsen 1994. To demonstrate equal loading, Northern blots were stripped and reprobed with a 700-bp PstI-SacI fragment of the ubiquitin clone pSKUBI (Christensen et al. 1992), a gift from P. Quail (US Department of Agriculture Plant Gene Expression Center, Albany, CA).

The 2.5-kb Mu1-hybridizing SstI fragment cosegregating with the tan1 phenotype was cloned from a size-selected library of SstI-digested genomic DNA from a homozygous mutant constructed in Lambda Zap (Stratagene). Full-length genomic and cDNA clones were isolated using the 600-bp Tan1 fragment (see Fig. 1 A) to screen a B73 genomic DNA library (a gift from Pioneer Hi-Bred, Johnston, IA) and a B73 vegetative shoot tip cDNA library (a gift from B. Veit and S. Hake, US Department of Agriculture Plant Plant Gene Expression Center). The sequences of three different cDNAs and one full-length genomic clone were assembled using MacVector software (v6.5). Sequencing of genomic PCR products amplified from the tan-gt1 allele revealed the presence of a point mutation near the end of exon 2. A combination of PCR and Southern blotting was used to identify a 6-kb insertion of unknown identity in the first intron of the tan-py1 allele.

Polyclonal rabbit antibodies were raised against a COOH-terminal TAN1 peptide (CGLKQRPGYSLTVRTVSSKISSR) coupled to keyhole limpet hemocyanin at Covance Research Products (Denver, PA) using their standard protocols. Antibodies were affinity-purified on peptide-coupled SulfoLink beads (Pierce Chemical Co.) as described by Harlow and Lane 1988. For the peptide competition experiments (see Fig. 5O), 1.5 μg of affinity-purified COOH-terminal peptide antibody in 200 μl of PBS with 1 mg/ml BSA was absorbed with beads coupled to ∼33 μg of peptide and used without further dilution for cell labeling experiments. mAbs were raised against the portion of TAN1 encoded by exons 1 and 2 (expressed as a glutathione S-transferase [GST] fusion protein from the pGEX-4T-2 vector [Amersham Pharmacia Biotech] and cleaved from glutathione S-transferase with thrombin protease) at Covance Research Products according to their standard protocols. The His-TAN1 fusion protein used for MT overlay assays was constructed by cloning a full-length Tan1 cDNA in frame with the histidine tag of pQE-30 (QIAGEN).

Plant extracts were prepared by homogenizing fresh tissue with an Omni TH homogenizer on ice in an extraction buffer of 100 mM Tris, pH 7.4, 10% sucrose, 5 mM EGTA, 5 mM EDTA, and a cocktail of protease inhibitors from Roche Molecular Biochemicals. Crude extracts were centrifuged at various speeds for 10 min in an Eppendorf microcentrifuge at 4°C. Proteins were separated by SDS-PAGE as described by Ausubel et al. 2000. Western blots were carried out as described by Harlow and Lane 1988, using the affinity-purified COOH-terminal peptide antibody at 5–10 μg/ml or hybridoma TAN75 culture supernatant at 1:20, and alkaline phosphatase–conjugated secondary antibodies (Promega). To confirm equal loading of proteins (see Fig. 3C and Fig. D), blots were reprobed with antiactin mAb C4 (ICN Biomedicals) and detected with alkaline phosphatase–conjugated anti–mouse Ig as described above.

The MT overlay assay was based on the method of Saunders et al. 1997. Proteins were separated on SDS-PAGE gels and transferred to polyvinyldifluoride membranes (Millipore). Membranes were blocked in TBST (50 mM Tris, pH 7.5, 150 mM NaCl, 0.05% Tween 20) containing 5% BSA for 1 h and washed three times in PEM buffer (80 mM Pipes, 1 mM EGTA, 0.5 mM MgCl₂, 10% glycerol, 5 mM DTT) containing 0.5% BSA and protease inhibitors. MTs were polymerized from bovine brain tubulin (Cytoskeleton) in PEM buffer, diluted to 10 μg/ml in PEM buffer containing protease inhibitors, 0.5% BSA and 1 μm taxol, and incubated with filters for 2 h at room temperature After washing three times in TBST with 0.5% BSA (TBST/BSA), filters were incubated with primary antibodies (TAN75 or anti–β-tubulin mAb DM1B [Amersham Pharmacia Biotech] diluted in TBST/BSA). Subsequent washes and detection were carried out as described above for Western blots.

Leaf primordia ≤1 cm in length were removed from 2–3-wk-old seedlings, fixed as described in Cleary and Smith 1998, and squashed onto coverslips essentially as described in Wick et al. 1981. Alternatively, tissue slices were plunge frozen in liquid nitrogen–cooled propane, freeze substituted, embedded in butylmethylmethacrylate resin, sectioned at 3 μm, and incubated in acetone to remove the resin as described by Baskin et al. 1996. These methods gave very similar results. Coverslips/slides were blocked in PBS with 5% normal goat serum (PBS/NGS), incubated with primary antibodies (anti-COOH peptide at 5–10 μg/ml, TAN75 supernatant at 1:50 in PBS/NGS) for 2 h at room temperature, washed in PBS with 0.2% Triton X-100, incubated with appropriate FITC-conjugated secondary antibodies (Jackson ImmunoResearch Laboratories), and washed again as above. For double labeling experiments, coverslips/slides were then blocked in PBS with 5% normal mouse serum (PBS/NMS) and incubated for 2 h in Cy3-conjugated anti–β-tubulin mAb Tub 2.1 (Sigma-Aldrich) diluted into PBS/NMS (incubations were always done in this order to ensure that FITC labeling coinciding with MTs could not result from cross-reaction of the FITC secondary antibodies with the tubulin antibody). After washing as described above, coverslips/slides were mounted in Vectashield (Vector Laboratories) and viewed on a E-600 microscope using EF-4 FITC HYQ and G-2A filter sets and a CFI Plan Apo 60× 1,4 NA objective (Nikon). Images were captured on an MDS-120 digital camera (Eastman Kodak Co.) and processed using Adobe^® Photoshop™ v4.0.

Supplementary Materials including details of the relationship between TAN1 and vertebrate adenomatous polyposis coli (APC) proteins, and additional biochemical characterization of TAN proteins, can be found.

The tan-Mu1 mutation arose in a stock with active Mutator transposons. A Mu1-hybridizing 2.5-kb SstI fragment that cosegregated with the mutant phenotype was cloned and used as a probe to isolate three cDNAs and a full-length genomic clone. Sequencing of these clones revealed the gene structure illustrated in Fig. 1 A. The Mu1 insertion in tan-Mu1 coincides approximately with the presumed start site of transcription, as estimated from the size of Tan1 mRNA (∼1.5 kb, see below). Southern blot analysis and sequencing of genomic PCR products showed that the tan1-py1 mutant allele contains an ∼6-kb insertion of unknown identity in the first intron of the gene, and the ethyl methanesulfonate–induced tan-gt1 allele contains a premature stop codon near the end of the second exon, which would truncate the TAN1 protein to <10 kD.

The 600-bp fragment shown in Fig. 1 A hybridizes to the 2.5-kb SstI fragment in genomic DNA of homozygous mutant individuals, a 1.2-kb fragment in their homozygous wild-type siblings, and to both fragments in heterozygous siblings (Fig. 1 B). This is the expected pattern of hybridization for a fragment of the Tan1 gene, since the wild-type allele should be smaller than the tan1-Mu1 allele by the size of Mu1 (1.3 kb). Proof that the 600-bp fragment is a part of the Tan1 gene came from the analysis of somatic revertant sectors. Because tan1 mutant leaves have a rough crepe papery texture, large wild-type sectors in mutant leaves, such as that illustrated in Fig. 1 D, are readily visible. Four such sectors showed the result illustrated in Fig. 1 C: the 600-bp fragment hybridizes to a 2.5-kb SstI fragment in DNA from mutant tissue (Fig. 1 C, lane 1) and to both 2.5 and 1.2-kb fragments in DNA from the adjacent wild-type sector (Fig. 1 C, lane 2). Thus, appearance of the wild-type phenotype in this sector correlates with excision of Mu1 from one of the two mutant alleles. At high stringency, the 600-bp Tan1 probe hybridizes to a single fragment of maize genomic DNA digested with various restriction enzymes (Fig. 1 E, asterisks). At low stringency, this probe hybridizes to the same fragment (Fig. 1 F, asterisks) and to one additional fragment (Fig. 1 F, arrows).

Northern blot analysis using the 600-bp Tan1 probe illustrated in Fig. 1 A was carried out to examine the pattern of Tan1 gene expression. As illustrated in Fig. 2 A, this probe hybridizes to a single mRNA of ∼1.5 kb in wild-type vegetative shoot tips consisting of leaf primordia and the bases of immature leaves, the shoot apical meristem, and unexpanded stem tissue (highly enriched in actively dividing cells). This mRNA is greatly reduced in tan1-Mu1 mutants and to a lesser extent in tan1-py1 mutants. In addition, a smaller transcript is observed in tan-py1 mutants (and their wild-type siblings, mostly tan-py1 heterozygotes), presumably a product of aberrant RNA splicing (Fig. 2 A, asterisks). Tan1 mRNA accumulation is only slightly reduced in shoot tips of homozygous tan-gt1 mutants, consistent with the presence of a premature stop codon in this allele. As shown in Fig. 2 B, compared with the 0–2-cm shoot segment enriched in dividing cells (Fig. 2 B, lane 1), Tan1 mRNA is vastly reduced in the 3–5-cm shoot segment composed mainly of postmitotic expanding leaf cells (Fig. 2 B, lane 2), and in the 6–8-cm segment composed mainly of fully expanded leaf cells that are differentiating (Fig. 2 B, lane 3), as well as in mature (fully expanded and differentiated) leaf tissue (Fig. 2 B, lane 4). In addition, Tan1 is strongly expressed in other tissues enriched in dividing cells: ear primordia (Fig. 2 B, lane 5), and embryos (Fig. 2, lane 6). Thus, Tan1 gene expression correlates with cell division in the shoot.

The Tan1Polakis 1997), suggesting that TAN1 might perhaps be an MT binding protein. cDNA sequence (sequence data are available from GenBank/EMBL/DDBJ under accession no. AF305892) contains an open reading frame encoding a predicted protein of ∼41 kD that is highly basic with a pI of 12.6. Analysis of this protein sequence reveals no strong homology with other proteins or motifs of known function, but a distant similarity was observed between 85% of the TAN1 protein and the basic regions of vertebrate APC proteins (see supplemental material available). The basic region of APC has been shown to bind to tubulin in vitro and associate with MTs in vivo (

Antibodies were raised against various fragments of TAN1 to provide tools for analysis of this protein. Fig. 3 A shows a Western blot of proteins extracted from wild-type vegetative shoot tips probed with an affinity-purified antibody raised against a TAN1 COOH-terminal peptide. In Fig. 3 B, an equivalent blot is probed with an mAb raised against the NH₂-terminal portion of TAN1 encoded by exons 1 and 2 (TAN75). Both antibodies predominantly label a single protein band at ∼43 kD, close to the predicted molecular mass of TAN1 (41 kD). Sequential centrifugation of crude cell extracts at gradually increasing speeds of 3,000–15,000 g shows that the majority of this protein sediments at 6,000–9,000 gTan1 mRNA, this protein is not detected in 3,000–10,000 g pellets from extracts of shoot segments composed of expanding (EXP) or differentiating (DIF) leaf tissue (Fig. 3 A)., indicating that it is associated with a relatively large or dense structure in cell extracts. However, since this protein cannot be solubilized with detergents, it does not appear to be associated with a membrane-bound organelle (see supplemental material available). Parallel to results from the Northern blot analysis of

As shown in Fig. 3C and Fig. D, the 43-kD protein identified by both antibodies is reduced, but not eliminated, in mutants homozygous for each of the three tan1 mutant alleles. Notably, the truncated form of TAN1 protein encoded by the tan-gt1 allele (see Fig. 1 A) would not be recognized by the COOH-terminal peptide antibody and would not comigrate with full-length TAN1. Thus, it appears that the 43-kD protein recognized by anti-TAN1 antibodies in tan-gt1 shoot tip extracts is the product of another gene, which is similar enough to TAN1 to be recognized by antibodies raised against nonoverlapping NH₂- and COOH-terminal fragments of TAN1 as well as three other polyclonal antibodies raised against other fragments of TAN1 covering its entire length (data not shown).

An overlay assay was used to show that TAN1 protein can bind directly to MTs. Fig. 4 A shows Coomassie blue–stained Escherichia coli extracts separated by SDS-PAGE before or after induction of a His-tagged TAN1 protein (arrow). After transfer to membranes, E. coli extracts were incubated with or without MTs polymerized from bovine brain tubulin. Subsequent detection with an anti–β-tubulin antibody showed that MTs bound to the His-tagged TAN1 protein (Fig. 4 A). A similar experiment with maize vegetative shoot tip extracts is shown in Fig. 4 B. Anti–β-tubulin detects endogenous maize tubulin with or without prior incubation of the blots with polymerized MTs. In addition, this antibody detects MTs bound to a 43-kD protein that comigrates with TAN1 and is reduced in abundance in the tan-gt1 mutant extract. The residual MT-binding protein of the same molecular mass as TAN1 seen in the tan-gt1 mutant extract presumably corresponds to the 43-kD protein in this extract recognized by anti-TAN1 antibodies (see Fig. 3C and Fig. D), suggested earlier to be the product of a Tan1-related gene.

Leaf primordium cells labeled with mAb TAN75 are shown in Fig. 5A–G. This antibody labels cells at all stages of the cell cycle in a punctate manner. In interphase cells, antibody labeling is distributed uniformly throughout the cytoplasm but is excluded from the nucleus (Fig. 5 A). Examination of thin sections confirmed that TAN75 labeling is not restricted to the cortex as interphase MTs are but is evenly distributed throughout the cytoplasm. In mitotic cells, TAN75 labeling is also observed throughout the cytoplasm but is preferentially associated with the cytoskeleton. A prophase cell labeled with TAN75 (Fig. 5 B) and anti–β-tubulin (Fig. 5 C) illustrates the coincidence of TAN75 labeling with the PPB. Similarly, mitotic cells labeled with TAN75 (Fig. 5D and Fig. F) and anti–β-tubulin (Fig. 5E and Fig. G) show that proteins recognized by TAN75 are predominantly associated with the spindle and phragmoplast. Fig. 5H–N, shows that labeling of leaf primordium cells with the affinity-purified COOH-terminal peptide antibody produces very similar results.

Consistent with the results of Western blot experiments suggesting that these antibodies recognize a TAN1-related protein in tan1 mutants, we found that cell labeling with TAN75 and the COOH-terminal peptide antibody is not eliminated or qualitatively altered in any of the tan1 mutants, including tan-gt1 mutants. However, peptide competition experiments confirm that the COOH-terminal peptide antibody specifically labels proteins sharing epitopes with TAN1: preincubation of the antibody with COOH-terminal TAN1 peptide–coupled beads reduced labeling to background levels (Fig. 5 P), but preincubation with beads coupled to an unrelated peptide had no effect (Fig. 5 O). From these results, in combination with the Western blot data and the observation that antibodies directed against nonoverlapping parts of TAN1 show equivalent cell labeling patterns, we conclude that these patterns reflect the intracellular distribution of a family of two or more related proteins that includes TAN1.

Previous work showed that Tan1 is required for spatial control of cytokinesis during maize leaf development. In tan1 mutant leaf cells, cytoskeletal structures involved in establishing planes of cell division (PPBs) and forming new cell walls (phragmoplasts) appear structurally normal but are frequently misoriented (Cleary and Smith 1998). Here, we report the cloning of Tan1 and initial molecular analysis of its protein product.

The Tan1 gene is expressed in regions of active cell division and encodes a highly basic protein distantly related to the basic MT-binding domain of vertebrate APC proteins. Antibodies raised against various fragments of the TAN1 protein specifically recognize a 43-kD protein band detectable only in extracts from regions of active cell division, which apparently includes at least one TAN-related protein in addition to TAN1 itself. A genomic DNA fragment shown to hybridize with the Tan1 probe at low but not high stringency could correspond to a gene 80–90% identical to Tan1 at the nucleotide and protein levels, sharing most if not all epitopes recognized by the anti-TAN1 antibodies employed in this study. Similarly, antibodies raised against individual maize profilins cross-react with other family members that are 85% identical at the amino acid and nucleotide levels (Kovar et al. 2000). Efforts are currently underway to identify TAN1-like proteins for direct comparison with TAN1.

TAN1 protein produced in E. coli and extracted from maize shoot tips can bind directly to MTs in an MT overlay assay. Furthermore, proteins recognized by anti-TAN1 antibodies are preferentially associated with the MT-containing structures in dividing cells that are misoriented in tan1 mutants (PPBs, spindles, and phragmoplasts). The apparent lack of association of these proteins with MTs in interphase cells suggests that their interaction with the cytoskeleton is regulated in a cell cycle–dependent manner. Together, our results suggest that TAN1 protein participates in the orientation of cytoskeletal structures in dividing cells through an association with MTs. An interesting possibility is that TAN1 mediates interactions between these structures and the cell cortex that are necessary for their proper orientation, such as those that guide phragmoplasts to cortical division sites previously occupied by PPBs.

We thank Scott Hammer and Andy Teng for their help with analysis of tan1 mutant alleles; Sharon Kessler, Neelima Sinha, and the Maize Genetics Stock Center for seeds; Tama Hasson for advice on preparation of protein extracts; and Bob Schmidt, Raffi Aroian, and members of the Smith lab for valuable input. Finally, special thanks to Sarah Hake for financial support and mentorship during the early stages of this work.

This work was supported by National Institutes of Health grant R01 GM53137 to L.G. Smith.