Proper chromatin condensation and sister chromatid resolution are essential for the maintenance of chromosomal integrity during cell division, and is in part mediated by a conserved multisubunit apparatus termed the condensin complex. The core subunits of the complex are members of the SMC2(Structural Maintenance of Chromosomes) and SMC4 gene families. We have cloned an Arabidopsis gene, AtCAP-E1, which is a functional ortholog of the yeast SMC2gene. A second, highly homologous SMC2 gene, AtCAPE-2, was identified by the Arabidopsis genome project. SMC2 gene expression in Arabidopsis was correlated with the mitotic activity of tissues, with high level expression observed in meristematic cells. The two genes are differentially expressed with AtCAP-E1 accounting for more than 85%of the total SMC2 transcript pool. The titan3 mutant is the result of a T-DNA insertion into AtCAP-E1, but other than subtle endosperm defects, titan3 is viable and fecund. We identified a T-DNA insertion mutant of AtCAP-E2, which showed no obvious mutant phenotype,indicating that the two genes are functionally redundant. Genetic crosses were employed to examine the consequences of reduced SMC2 levels. Both male and female gametogenesis were compromised in double mutant spores. Embryo lethality was observed for both double homozygous and AtCAP-E1-/-, AtCAP-E2+/- plants;arrest occurred at or before the globular stage and was associated with altered planes of cell division in both the suspensor and the embryo. Down regulation of both genes by antisense technology, as well as in AtCAP-E1+/-, AtCAP-E2-/- plants results in meristem disorganization and fasciation. Our data are consistent with the interpretation that threshold levels of SMC2 proteins are required for normal development and that AtCAP-E2 may have a higher affinity for its target than AtCAP-E1.
Chromatin condensation is a fundamental cellular process in which entangled interphase chromatin fibers are actively resolved and packaged into physically discrete and compact chromosomes that undergo segregation during cell division. This complex series of events enables the cell to proceed through mitosis and faithfully distribute genetic information to each daughter cell. The inability to execute this process correctly can result in catastrophic chromosome segregation defects as is observed in mutants defective in topoisomerase II (DiNardo et al.,1984). Failure to disentangle chromatin fibers prevents compaction necessary for mitotic chromosome condensation(Gimenez-Abian et al., 2000; Cuvier and Hirano, 2003). In addition to topoisomerase II, other conserved proteins and histone modifications play roles in chromosome condensation(Hirano, 2002).
The SMC proteins are evolutionarily conserved across phylogenetic kingdoms with representatives in both prokaryotes and eukaryotes. The eukaryotic SMC proteins can be subdivided into six distinct subfamilies (SMC1-6)(Hirano, 2002; Jessberger, 2002), and three functional categories can be defined based on the formation of their heterodimers. The SMC2/SMC4 heterodimers assemble with three other non-SMC proteins to form the `condensin' complex, which is involved in mitotic chromosome condensation and dosage compensation(Hirano and Mitchison, 1994; Lieb et al., 1998; Sutani et al., 1999; Kimura et al., 2001). Members of the SMC1 and SMC3 families form heterodimers to give rise to the `cohesin'complex, which is involved in sister chromatid cohesion(Michaelis et al., 1997; Losada et al., 1998). Lastly,SMC5 and SMC6 members form heterodimers that are involved in DNA recombination/repair (Mengiste et al.,1999; Fousteri and Lehmann,2000). All SMC proteins consist of five distinct domains:nucleotide-binding domains located at the N and C termini lie adjacent to two coiled-coil regions, which are separated by a hinge region(Hirano, 2002; Jessberger, 2002). The catalytic function of the SMC subunits is regulated by non-SMC members of these multisubunit complexes. Mutations in genes encoding components of SMC complexes result in a variety of defects, ranging from lethality to developmental abnormalities (Saka et al.,1994; Strunnikov et al.,1995; Bhat et al.,1996; Lieb et al.,1998; Freeman et al.,2000; Steffensen et al.,2001; Bhalla et al.,2002; Hagstrom et al.,2002).
In both plants and animals, alterations in chromatin structure and dynamics are implicated as factors controlling both cell cycle progression and aspects of development (reviewed by Muller and Leutz, 2001; Reyes et al.,2002; Berger and Gaudin,2003). Whereas in many metazoans mutations in these genes are lethal, orthologous mutants in plants are often tolerated, resulting in developmental and/or physiological defects(Wagner, 2003). For example,plants harboring an antisense construct of histone deacetylase 1(AtHD1) exhibit pleiotropic defects in growth and development, and some of these could be correlated with ectopic expression of the SUPERMAN gene (Tian and Chen,2001). Similarly, disruption of the topoisomerase I gene leads to phyllotaxy defects and bifurcation of lateral shoots(Takahashi et al., 2002). Arabidopsis plants deficient in telomerase function also exhibit developmental abnormalities with grossly enlarged meristems, altered leaf morphology and low viable seed set (Riha et al., 2001). Lastly, mutations in chromatin assembly factor(CAF1) subunits, which assemble nucleosomes onto nascent DNA, underpin the fasciata phenotype (Kaya et al.,2001).
Despite significant progress in our understanding of molecular and biochemical aspects of chromatin remodeling and chromosome condensation in yeast and in animal systems, very little information exists in plants. Arabidopsis affords an excellent system to explore these basic cellular processes as the genome has been sequenced and annotated, genetic manipulations are facile, and there exists a large collection of insertion mutant lines for reverse genetics studies. We therefore sought to identify mutations in Arabidopsis condensin genes. Here we report the characterization of two Arabidopsis SMC2 genes, AtCAP-E1 and AtCAP-E2 (Arabidopsis thaliana Chromosome Associated Protein subunit E, a nomenclature that conforms to the designation given to SMC2 orthologs in higher eukaryotes). Unlike other characterized diploid organisms, Arabidopsis contains two members of the SMC2 family. Recently, a T-DNA insertion into AtCAP-E1 was reported to be responsible for the titan3 mutant, which exhibits enlarged endosperm nuclei and aberrant mitotic figures but otherwise develops normally and is fecund(Liu et al., 2002). Given the essential role for SMC2 proteins in chromosome condensation in other organisms, it could be predicted that disruption of SMC2 function would have profound effects on cell division and plant development. We identified a mutant of AtCAP-E2, which exhibits no obvious developmental defects,but the AtCAP-E1-/-, AtCAP-E2-/-double mutant results in embryo lethality. Transgenic plants expressing an antisense construct exhibited reduced levels of both SMC2 transcripts, and pleiotropic phenotypes including slow growth, disorganized meristems, and fasciation. Our data indicate that proper developmental patterning requires a threshold level of SMC2 proteins to support condensin complex function.
MATERIALS AND METHODS
Identification and cloning of the Arabidopsis SMC2 gene AtCAP-E1
Based on the conserved amino acid sequences, FNAITGLN and PHFLIMQG, which flank the NTP-binding domain of SMC2 proteins, degenerate oligonucleotide primers (5′CCYTGCATRATIARRAARTKWGG3′ and 5′TTYAAYGCIATYACIGGWYTIAAYGG3′) were designed to PCR amplify Arabidopsis genomic DNA. The PCR fragment was used to screen an Arabidopsis genomic library by standard methods, and a lambda genomic clone was isolated. The AtCAP-E1 cDNA was isolated by RT-PCR. A primer corresponding to the last eight amino acids and the stop codon of AtCAP-E1 (5′-TCACTTGGTCTGCTTTGTTACTGTCC-3′) was used for RT with 1 μg of total RNA from young seedlings. PCR was performed using the primer employed for RT and a forward primer(5′-ATGCATATAAAGGAGATATGCTTGG-3′) encoding the first eight amino acids of the AtCAP-E1 protein. The sequence of the 3528 bp cDNA has been deposited in the GenBank database under accession number AF306547.
The AtCAP-E1 cDNA was digested with NotI and SpeI and ligated into the NotI and XbaI sites in pYES2. Transformation of smc2-Δ6 cells was performed according to the method of Gietz and Woods(Gietz and Woods, 1998). Transformants were selected at the restrictive temperature (37°C) on DOB dropout base plus CSM lacking uracil containing 2% D-galactose. For 4′,6-diamidino-2-phenylindole (DAPI) staining, liquid cultures [wild type (YPH499), smc2-Δ6, and transformed cells] were grown overnight at 30°C and then shifted to 37°C for 2 hours. An aliquot of cells was harvested, fixed in 2.5% glutaraldehyde, rinsed, and stained with 2 μg/ml DAPI solution. DAPI staining was monitored using a Zeiss axiophot microscope.
RNA isolation and RT-PCR analysis
Total RNA was prepared from Arabidopsis roots, stems, mature leaves, floral buds and mature green siliques using the TRIZOL (Gibco)reagent. RT-PCR was performed as follows. Forward(5′-GATACATGCAAAGATGAAGGAATG-3′) and reverse(5′-TATCATTCTTCCTATGTTCTGTGTGTG-3′) primers were used to amplify a 762 bp fragment from both AtCAP-E1 and AtCAP-E2. First-strand cDNA synthesis was performed with 1 μg of total RNA and the backward primer in a 25 μl reaction according to the 5′RACE protocol(Gibco). PCR was performed by making tenfold serial dilutions (10-1and 10-2) starting from the same amount of first-strand cDNA. The volume of each tenfold RT-dilution step used as a template was 6% of the final PCR volume. For CAPS analysis (Konieczny and Ausubel, 1993), PCR products were digested with SspI or XbaI and DNA gel blotting was performed with the 762 bp cDNA fragment as the probe.
Construction of AtCAP-E1 promoter::β-glucuronidase gene and plant transformation
A 2.7 kb region upstream of the AtCAP-E1 start codon was amplified by PCR with the primers: 5′-CGGGATCCAGCAATAGCTTTAGCTTTCGCC-3′ and 5′-TCCCCCGGGCTTCCTTCTTCTTCTTCCCC-3′. The amplified fragment was ligated into pBI101.2 (Jefferson et al.,1987) upstream of the β-glucuronidase gene. The construct was mobilized into Agrobacterium (GV3101) and used to transform(Landsberg erecta ecotype) plants by the floral dip method(Clough and Bent, 1998). Transformants were selected on Murashige and Skoog (MS) plates (1× MS salts pH 5.7, 3% sucrose, 0.5 g/l MES, 0.8% phytagar) containing 50 μg/ml kanamycin. The histochemical assay for GUS was performed as described previously (Jefferson et al.,1987).
In situ hybridization
In situ hybridization was performed essentially as described by Jackson(Jackson, 1991). A 196 bp region of the Arabidopsis histone H4 gene was labeled with digoxigenin. In vitro transcription was performed following the manufacturer's(Roche) protocol. Wild-type and antisense seedlings were fixed in FAA (3.7%formaldehyde, 5% acetic acid, 50% ethanol) for 2 hours, dehydrated in an ethanol/Histoclear series, embedded in TissuePrep and sectioned at 10 μm. After the color development, sections were mounted in VectaShield mounting medium (Vector Laboratories, USA) and photographed using a Zeiss axiophot microscope and digital imaging system.
Isolation of an AtCAP-E2-/- mutant
A BLAST search in the SAIL (Syngenta Arabidopsis Insertion Library)database with the AtCAP-E2 genomic sequence (AGI identifier At3g47460) revealed a line (36-C09) with a T-DNA insertion near the 3′end of the gene. Seeds of this line were germinated on MS plates containing glufosinate ammonium (BASTA) to select for plants having a T-DNA insertion. The integration site of the T-DNA was determined by PCR amplification and sequencing of a 1061 bp fragment using a gene-specific primer (5′GATACATGCAAAGATGAGGAATG-3′) and the T-DNA left border primer(5′-TAGCATCTGAATTTCATAACCAATCTCGATACAC-3′). The genotype of the individual plants grown on BASTA was determined by two sets of PCR reactions. The first used gene-specific primers flanking the T-DNA(5′-GATACATGCAAAGATGAAGGAATG-3′;5′-TATCATTCTTCCTATGTTCTGTGTGTG-3′) to identify the presence of a wild-type allele. Plants heterozygous for the T-DNA insertion gave rise to a 1491 bp AtCAP-E2 gene fragment with PCR primers flanking the T-DNA insertion site, while no band corresponding to the AtCAP-E2 locus was amplified from the homozygous plants. The second PCR employed a gene-specific primer and the T-DNA left border primer to confirm the presence of a mutant allele.
Generation of double mutants and analysis of F2progeny
Double mutant plants were generated by crossing homozygous AtCAP-E1-/- (ttn3 mutant; Wassilewskija ecotype;WS) with AtCAP-E2-/- (Columbia ecotype; Col) plants. F1 seeds were germinated on medium containing both BASTA and kanamycin and 100% resistance to both antibiotics was observed (The T-DNA insertion at the AtCAP-E1 locus confers kanamycin resistance while the T-DNA insertion at the AtCAP-E2 locus confers BASTA resistance). F2 seedlings were selected on MS plates containing both antibiotics, and 180 F2 seedlings were genotyped by employing three sets of primers. Based on the site of insertion of the T-DNAs, the first primer pair (5′-GATACATGCAAAGATGAAGGAATG-3′;5′-TATCATTCTTCCTATGTTCTGTGTGTG-3′) was chosen to flank the insertion sites and amplification of 1.65 kb (amplified from the AtCAP-E1 locus) and 1.49 kb (amplified from the AtCAP-E2locus) fragments are indicative of a wild-type allele. Amplification of a single band (either 1.65 or 1.49) represented heterozygosity at the given locus and homozygosity for a T-DNA insertion in the other gene. The other sets of primer pairs employed an allele-specific primer together with a T-DNA primer to test for the presence of a T-DNA insertion as described in the previous section.
Because the two single mutants are in different genetic backgrounds (WS and Col) we crossed these wild-type plants to determine if any inter-ecotype incompatibilities exist that would compromise development and/or fecundity. We observed no evidence of ovule or embryo abortion, which validates the phenotypic defects observed in the F2 progeny of a cross between the single mutants. Genetic evidence supporting this contention is twofold. In siliques of E1+/-, E2+/- plants, the number of aborted ovules and aborted embryos conformed to the ratios expected based on the parental genotype (data not shown) such as they did in E1+/-,E2-/-. Secondly, the anaphase bridges observed in AtCAP-E1+/-, AtCAP-E2-/- pollen mother cells were not observed in AtCAP-E1+/-,AtCAP-E2+/- pollen mother cells.
Generation of antisense transgenic plants
Arabidopsis plants (Landsberg erecta ecotype) carrying the AtCAP-E1 cDNA in an antisense orientation were generated by inserting a 2.3 kb fragment of the AtCAP-E1 cDNA in the reverse orientation in pBI121, such that transcription of the antisense strand was driven by the CaMV-35S constitutive promoter. Transgenic plants were generated as described above. All plants were grown at 25°C under a 16-hour light/8-hour dark regime.
Siliques were dissected after fixation in an ethanol:acetic acid (3:1)solution, cleared overnight in chloral hydrate:glycerol:water (8:2:1) and imaged by Nomarski optics.
For DAPI staining of pollen mother cells, fresh anthers were dissected on a slide in 40 μl of 10 mM sodium citrate buffer pH 4.0 using a needle under a dissecting misroscope. After releasing the contents of the anthers the sample was mounted in 15 μl of Vectashield containing DAPI. After the addition of a coverslip, the slides were viewed by fluorescence using a Zeiss Axiophot microscope.
For Toluidine Blue and DAPI staining of SAMs, 8-day-old wild-type and 18-day-old antisense seedlings were fixed in FAA, embedded in wax, and 8 μm sections were cut. Images were captured with a Zeiss axiophot microscope equipped with a digital image capture system.
For scanning electron microscopy (SEM), seedlings were fixed in FAA,dehydrated, critical point dried, and coated with gold particles. Specimens were viewed using a Hitachi S-2500 microscope.
Wild-type and antisense seedlings were grown vertically on agar plates for 10 days for confocal laser scanning microscopy (CLSM). For Aniline Blue staining, roots were fixed in a 2.5% NaOH/0.1% sodium dodecyl sulfate solution for 18 hours at room temperature, rinsed in water, and dehydrated by passing through an ethanol series. The tissue was stained for 4 hours with 0.4%Aniline Blue. After staining, tissue was dehydrated to 100% ethanol and cleared by passing through a 3:1, 1:1 and 1:3 series of ethanol:methyl salicylate solution followed by two washes in 100% methyl salicylate. Median longitudinal optical sections of the root tip were obtained using an Olympus Fluoview (FV5-PSU) CLSM.
A 1302 bp DraI to EcoRI fragment of the AtCAP-E1cDNA was cloned between SmaI and EcoRI sites of pGEX2TK(Invitrogen) vector downstream of the glutathione S-transferase (GST) fragment to produce a GST-tagged recombinant protein in bacteria. The recombinant protein was purified and used to immunize rabbits. The affinity-purified polyclonal antibodies were diluted 1:2500 for immunoblotting with total protein extracts prepared from 7-day-old seedlings.
Arabidopsis contains two SMC2 genes
Based on a sequence alignment of SMC2 proteins, we designed degenerate oligonucleotide primers to amplify an Arabidopsis genomic fragment encoding the conserved amino terminal nucleotide-binding domain. Sequencing of this fragment and computer analysis confirmed its identity as an SMC2 gene fragment located on chromosome V, and we used the fragment as a hybridization probe to screen a genomic library for the complete gene, which we termed AtCAP-E1. Based on the genomic sequence information we designed primers corresponding to the ends of the coding region of the gene and employed RT-PCR to amplify the cDNA. Comparison of cDNA and genomic sequences confirmed the intron/exon boundaries predicted by computer annotation. In parallel, the Arabidopsis Genome Initiative (AGI) reported the sequencing of a second SMC2 gene, which resides on chromosome III. We have termed the second gene AtCAP-E2. The gene products of AtCAP-E1 and AtCAP-E2 are 83% identical to each other and likely arose by genome duplication(Simillion et al., 2002).
AtCAP-E1 complements the yeast smc2-Δ6 mutant
To confirm that the AtCAP-E1 gene encodes a functional SMC2 protein, we used it to transform the yeast SMC2 mutant smc2-Δ6, which shows a temperature-sensitive defect in chromosome condensation and segregation(Strunnikov et al., 1995). Mutant cells undergo cell-cycle arrest at the non-permissive temperature(37°C). Some of these cells exhibit the classical `cut' phenotype wherein the cells have a fully elongated mitotic spindle but their chromosomes fail to separate and appear as a stretched mass of DNA between the two daughter cells. This nuclear morphology is indicative of a defect in chromosome condensation.
We transformed smc2-Δ6 cells with the AtCAP-E1 cDNA under the control of the galactose-inducible GAL1 promoter and selected for transformants in the presence of galactose at 37°C. Transformed cells were streaked on plates along with the wild-type and smc2-Δ6 mutant cells in the presence and absence of galactose and incubated at 37°C. Transformed smc2-Δ6 cells grew to nearly wild-type levels at 37°C in the presence of galactose (Fig. 1B), but failed to grow in the absence of galactose(Fig. 1C). Fig. 1D-G shows the morphology of the DAPI-stained chromatin by fluorescence microscopy. The stretched nuclear DNA in the smc2-Δ6 mutant cells and smc2-Δ6 cells transformed with AtCAP-E1 cDNA,in the absence of galactose, can be seen in Fig. 1D and E, respectively. Normal chromosome segregation was observed in wild-type cells(Fig. 1F), and transformed cells (Fig. 1G) when grown in the presence of galactose.
Approximately 500 cells were screened in two independent experiments and the cut phenotype was not observed in transformed cells expressing AtCAP-E1. To determine the growth profiles of the wild type, mutant and transformed cells, exponentially growing cells were shifted to 37°C and growth curves were established. These experiments demonstrated that the mutant cells transformed with the AtCAP-E1 construct had growth properties similar to wild-type cells at 37°C (data not shown). Taken together, these results show that AtCAP-E1 is able to functionally complement the chromosome condensation defect of smc2-Δ6 cells by rescuing the cut phenotype and rendering wild-type levels of growth to mutant cells transformed with the AtCAP-E1 cDNA. These results support an in vivo role of AtCAP-E1 in chromosome condensation in Arabidopsis.
Differential expression of AtCAP-E1 and AtCAP-E2
The Arabidopsis genome contains two SMC2 family members, encoding AtCAP-E1 and AtCAP-E2, which are 90% similar at the protein level. In order to examine the expression profile of both genes, we conducted RT-PCR followed by CAPS analysis (see Materials and Methods). Primers that exactly matched two regions in both genes were used in RT-PCR reactions, and thus the PCR product pool represented transcripts from both genes. The AtCAP-E1 and AtCAP-E2 mRNAs were most abundant in roots and young floral buds. The expression was intermediate in stems and low in fully developed leaves and mature siliques (Fig. 2A).
CAPS analysis was employed to assess the relative contribution of the AtCAP-E1 and AtCAP-E2 genes to the total SMC2 transcript pool. Based on the sequence of AtCAP-E1 and the AGI sequence information of AtCAP-E2, we digested the PCR products with SspI or XbaI, which represent unique restriction sites in AtCAP-E1 and AtCAP-E2 PCR fragments, respectively. Fig. 2B shows that AtCAP-E1 is expressed at a significantly higher level than AtCAP-E2 in all tissues tested. The reciprocal CAPS marker restriction enzyme, which cleaves only the AtCAP-E2 product, gives supporting data (Fig. 2C). Approximately 85% of the transcript pool consisted of AtCAP-E1transcripts.
To study the spatial and temporal expression of the AtCAP-E1 gene,an AtCAP-E1 promoter::β-glucuronidase (GUS) reporter chimeric gene was constructed and used to generate transgenic plants. Several independent transformants were assayed histochemically for GUS activity, and all primary transformants exhibited the same qualitative expression pattern.
In plants, apical meristems and lateral organ primordia are sites of active cell division. Fig. 2E shows GUS staining of the apex of a young seedling, wherein the entire apical dome and the emerging true leaves are stained. The root apical meristem (RAM) and emerging lateral root primordia also stain intensely(Fig. 2F,G). No activity was observed in differentiated root tissue. In mature plants, strong GUS expression was observed in newly formed organs. Uniform GUS expression was detected during early leaf development, which declined in a basipetal gradient starting from the distal region of immature leaves first(Fig. 2H). The onset of mesophyll cell expansion starts at the leaf tip and progresses towards the leaf base forming a gradient of cells at different developmental stages(Pyke et al., 1991). Thus, a decline in GUS activity as development proceeds can be correlated with the decline in mitotic activity. Fig. 2I shows that AtCAP-E1 is strongly expressed in developing floral buds, which are undergoing active cell division. After organ formation, strong GUS activity was observed in all four floral whorls and staining in young anthers and the gynoecium suggested that the SMC2 gene(s)might be active during meiosis (data not shown). Owing to the long half-life of GUS and the brief duration of meiosis in Arabidopsis, it was not possible to be certain if the GUS staining observed in the pollen mother cells(PMCs) was generated during or just prior to meiosis. Therefore, we elected to do RNA in situ hybridization using cross sections of whole inflorescences to precisely define the temporal expression pattern of AtCAP-E1. Expression in stamens was observed in the PMCs and in the surrounding tapetal cells during premeiotic S phase, prophase I, and anaphase I stages of meiosis(data not shown, but see below).
Identification and characterization of SMC2 insertion mutants
We undertook a reverse genetics approach to find mutants in the two Arabidopsis SMC2 genes. The Arabidopsis Biological Resource Center(ABRC) T-DNA lines were screened to identify insertions into AtCAP-E1and we found one pool containing such an insertion. During the course of this work, Liu et al. (Liu et al.,2002) reported that the titan3 mutant was due to an insertion into this gene, and hence we discontinued characterization of the ABRC line, using the available ttn3 mutant for additional studies. A T-DNA insertion mutant in the AtCAP-E2 gene was identified in the Syngenta T-DNA mutant collection (McElver et al., 2001). Sequencing revealed that the insertion occurred at the beginning of the 17th exon. Semi-quantitative RT-PCR revealed no AtCAP-E2 mRNA in 7-day-old AtCAP-E2-/- seedlings(data not shown). To determine the copy number of T-DNA insertions, seeds from the segregating population were grown on BASTA and resistant seedlings were counted (1399 resistant versus 498 sensitive). This ratio closely matched with the expected 3:1 ratio suggesting a single T-DNA insertion. Southern blotting also confirmed that AtCAP-E2-/- plants contained a single T-DNA insertion (data not shown).
We could not identify any obvious phenotype associated with the AtCAP-E2 mutation when the plants were grown under standard conditions.
Double mutants of AtCAP-E1 and AtCAP-E2 result in embryo lethality
Since ttn3 (AtCAP-E1-/-) plants exhibited a subtle endosperm phenotype with viable embryos and no post-embryonic phenotype, and the AtCAP-E2-/- plants demonstrated no obvious defects, we performed genetic crosses between the two homozygous single mutant plants to generate double mutants. The expected relative abundance of each F2 genotype is given in Table 1. Both mutants are the result of T-DNA insertions: E1 (referring to insertion in AtCAP-E1)is marked by an insertion of a kanamycin resistance gene while E2 (referring to insertion in AtCAP-E2) contains a BASTA resistance gene. Thus,E1- alleles and E2- alleles would confer resistance to kanamycin and BASTA, respectively (Table 1, column 2). To enrich for double mutants, F2 seeds were germinated on media containing both antibiotics.
|Genotype* .||Antibiotic resistance† .||Relative frequency‡ .||Observed/expected % ratio of genotypes§ .|
|Genotype* .||Antibiotic resistance† .||Relative frequency‡ .||Observed/expected % ratio of genotypes§ .|
All potential F2 genotypes of AtCAP-E1-/-× AtCAP-E2-/- are presented
E1 mutants contain a T-DNA conferring kanamycin resistance while E2 mutants contain a T-DNA conferring BASTA resistance
Genotypic frequency relative to the 16 groups in a dihybrid cross
Of the F2 seedlings that germinated on kanamycin+BASTA, the observed/expected ratios of each genotype as confirmed by PCR; n=180 plants
PCR analysis was performed to genotype 180 of the kanamycin and BASTA resistant F2 plants. If all kanamycin and BASTA resistant genotypes were equally viable, then 1 in 9 of the resistant plants would be expected to have the double mutant genotype (Table 1). However, we did not find any double homozygous mutant plants among the 180 F2 plants, which is a strong indication that this genotype might be embryo lethal. PCR analysis also revealed another very interesting result. Both the E1+/-E2-/- and E1-/-E2+/- genotypes were expected to each represent 2/9 of the double antibiotic resistant F2 progeny(Table 1, column 3); thus in 180 plants 40 plants of each genotype would be expected. However, no E1-/-E2+/- plants were observed, and only 24 E1+/-E2-/- plants were recovered. Therefore, analysis of the F2 population strongly suggests that E1-/-E2-/- and E1-/-E2+/-individuals are genotypes that result in lethality early in development.
The E1-/-E2-/-, E1-/-E2+/-, and E1+/-E2-/- diploid genotypes all require that at least one of the gametes has the haploid genome E1-E2-. Since the three aforementioned genotypes were either missing or under-represented in the F2 generation, it raised the possibility that the haploid spores with the E1-E2- genotype could not successfully complete gametogenesis. To test this hypothesis we performed reciprocal crosses of wild type with double heterozygotes,E1+/-E2+/-, and scored the F1 seeds for their ability to grow on the kanamycin/BASTA selection media. The types of haploid genotypes that can be produced by the double heterozygote are E1+E2+, E1+E2-,E1-E2+, and E1-E2-. Since the wild-type parent only produces the E1+E2+ (which is sensitive to both kanamycin and BASTA), only the F1 offspring that receive the E1-E2- genotype from the double heterozygous parent would be able to grow on the selection medium. Thus only 25% of the F1 plants would be expected to be resistant to both kanamycin and BASTA. When the E1+/-E2+/- plants were used as the male parent only 3.2% of the F1 seedlings (n=248) were resistant to both antibiotics. Thus, the greatly reduced number of E1+/-E2+/- individuals indicates that only 13% (3.2%observed versus 25% expected) of the expected number of E1-E2- pollen grains are able to successfully complete male gametogenesis. Similarly, in the reciprocal cross in which the E1+/-E2+/- plants were used as the female parent, only 12.5% of the F1 seedlings (n=202) were resistant to both antibiotics. This indicates that there is a 50% reduction in the ability of the megaspore to complete female gametogenesis. Thus, failure during gametogenesis in E1-E2- gametes is a major, but not exclusive source of lethality.
To further investigate the issue of viability we examined the contents of the siliques. An indicator of early lethality in Arabidopsis is a paucity of normal seeds in developing siliques. Wild-type and the single mutant plants yielded normal siliques filled with full, green seeds(Fig. 3A-C). However, severe seed defects were observed in the mature siliques of the self-crossed E1+/-E2-/- individuals(Fig. 3D). Among the 26 green siliques that we examined, we observed 43.43% normal green seeds, 30.86%discolored and mis-shapened seeds, and 25.7% unfertilized ovules, which appeared as empty spaces (aborted ovule development presumably associated with failure to complete female gametogenesis) in the silique. We used differential interference contrast microscopy (DIC) to determine the developmental stage at which the embryos in these abnormal seeds were arrested. The normal-appearing green seeds contained normal embryos at the globular to torpedo stages(Fig. 3E). Embryo development clearly differed in the abnormal sibling seeds with embryos arrested at the globular stage of embryogenesis (Fig. 3F). The seeds harboring the mutant embryos were usually smaller. Occasionally, we observed mutant embryos at the late globular stage with distinct defects in cell patterning of the suspensor as well as the embryo proper. The suspensors of these embryos consisted of randomly arranged multi-cell layers with both transverse and longitudinal cell division planes,whereas the suspensors of wild-type embryos consist of single files of six to 11 cells resulting from a series of transverse divisions. Similarly, the abnormal embryos exhibited an altered morphology probably due to atypical cell division planes, resulting in an arrested globular-like embryo(Fig. 3G,H). Thus, our silique analyses show that self-crosses of E1+/-E2-/-individuals often produce many abnormal seeds containing aborted embryos and some seeds fail to form at all.
In our earlier PCR analysis of F2 progeny of the self-cross of the dihybrid, in addition to not recovering any double homozygous mutants, we also did not recover any plants of the E1-/-E2+/-genotype. To establish that this genotype also is associated with lethality during early development we crossed E1+/-E2-/- (female parent) with E1-/-E2+/+ (male parent). In this cross 50%of the progeny are expected to have the E1+/-E2+/-genotype and the other 50% are expected to have the E1-/-E2+/- genotype, but 25% of this latter group are expected to fail because of female gametic inviability (for E1-E2-). Thus we expected to find 25% nonviable seeds. We dissected 8 developing siliques from this cross and observed the embryo lethal phenotype in 27.28% of the seeds (n=164). We genotyped 32 progeny seedlings of this cross, and all of the plants were identified as E1+/-E2+/- genotype. Thus, in addition to the double null plants, E1-/- mutants which are heterozygous at the E2 locus,also cannot successfully complete embryogenesis.
Meiotic segregation defects in AtCAP-E1+/-, AtCAP-E2-/- plants
The E1+/-E2-/- genotype was the only genotype we recovered that carries only one functional SMC2 allele. Plants of this genotype often exhibit vegetative abnormalities (see the next section), and we examined meiosis in these plants to assess chromosome condensation/segregation defects. In plants, meiosis punctuates the sporophytic and gametophytic phases of the life cycle. Male meiosis (microsporogenesis) occurs in the pollen mother cells (PMCs) that give rise to four haploid microspores, each of which undergoes mitosis and differentiates into a mature pollen grain. Our GUS reporter gene studies as well as the in situ hybridization experiments indicated that the SMC2 genes are expressed in PMCs, and this fact prompted us to ascertain whether there exists an aberrant chromatin phenotype during meiosis in E1+/-E2+/- and E1+/-E2-/- plants. During meiosis I, chromatin condensation reaches a maximal level at metaphase I/anaphase I and hence we examined PMCs undergoing these two stages during meiosis I. In wild-type PMCs,congression of discrete bivalents was observed, and later, at the onset of anaphase, the chromosomes segregated properly(Fig. 4A,D). Cytological analysis of DAPI-stained PMCs of E1+/-E2+/- plants showed no aberrant chromatin behavior during meiosis I, and normal homologous chromosome segregation was observed (data not shown). In contrast, PMCs of E1+/-E2-/- plants often contained diffuse and non-discrete bivalents at metaphase I (Fig. 4C), and exhibited regions of stretched chromatin bridges during anaphase I with less condensed chromosomes(Fig. 4F). Nonetheless, these chromosomes appeared more condensed than the interphase chromatin as observed by DAPI staining (data not shown). Thus, this meiosis I defect mimics those observed in other condensin mutants during mitosis in yeast(Saka et al., 1994; Strunnikov et al., 1995; Bhalla et al., 2002) and Drosophila (Bhat et al.,1996; Steffensen et al.,2001) and during mitosis and meiosis II in C. elegans(Hagstrom et al., 2002).
Developmental defects of condensin antisense plants
To examine the effects of reduced AtCAP-E1 and AtCAP-E2activity during Arabidopsis development, we generated transgenic plants carrying an antisense fragment of the AtCAP-E1 cDNA under the control of the constitutive CaMV-35S promoter. Ten independent antisense lines were characterized, all of which exhibited varying degrees of similar phenotypic abnormalities. All T1 plants were selfed and seeds of the T2 generation were selected on kanamycin. Three independent T2 lines with most severe defects were selected for further analyses.
Although the antisense seedlings germinated with no obvious embryonic defects, pleiotropic phenotypes affecting various stages of post-embryonic development were observed. The transcript levels of both AtCAP-E1 and AtCAP-E2 in these seedlings were dramatically reduced(Fig. 5A-C), confirming a strong antisense suppression of the in vivo AtCAP-E1 and AtCAP-E2 transcript pool. In the most severe cases, the shoot apical meristem (SAM) developed into an enlarged dysfunctional mass(Fig. 5E,F) and asymmetric leaf-like structures often formed from the primary SAM(Fig. 5G). However, these plants usually produced axillary buds, which ultimately produced flowering shoots that were developmentally delayed(Fig. 6A,B). Typically, the antisense plants produced shoots with varying degrees of stem and inflorescence fasciation (Fig. 5H/I). Fasciation represents a disruption in the pattern of organogenesis and is associated with breakdown or disruption in SAM structure(Leyser and Furner, 1992),resulting in altered leaf and floral phyllotaxy, and broadening, flattening,and in extreme cases, bifurcation of the stem. Interestingly, stem and floral fasciation was also observed in some E1+/-E2-/- plants(Fig. 5J,K).
DAPI-stained sections of the SAMs of wild-type and antisense plants with the terminal dysfunctional meristem revealed a clear distinction in nuclear morphology (Fig. 7A-D). In wild-type cells (Fig. 7A,B),uniform and centrally localized nuclear staining typical of interphase nuclei was observed. In contrast, the chromatin of the enlarged meristem of the antisense plants was stretched and localized at the periphery of the cell(Fig. 7C,D). The intensity of DAPI staining in these cells is indicative of partially condensed chromatin,which in addition is displaced to one side of the cell probably by a large vacuole in an irregularly shaped cell (Fig. 7D). The presence of a large vacuole suggests cell differentiation, as the vacuoles of wild-type SAM cells cannot be distinguished by light microscopy. The normal tunica/corpus cellular arrangement of a wild-type SAM into distinctive histological zones consisting of L1, L2 and L3 layers (Fig. 7A,E) was disorganized in the antisense SAMs(Fig. 7C,F). Scanning electron micrographs of the apex revealed an enlarged SAM bearing morphologically altered leaf primordia. Cells of irregular shapes and sizes were observed both in the meristem and in the developing leaf primordia(Fig. 7G-I and data not shown). We used in situ hybridization with an S-phase-specific marker, histone H4, to assess the levels of cell division activity in the SAMs of wild-type and antisense plants. Compared to the high mitotic activity in the wild-type SAMs(Fig. 7J), very few dividing cells were observed in the antisense SAMs(Fig. 7K,L). However, axillary meristems of these plants were relatively normal in appearance(Fig. 7G) and contained an abundance of cycling cells (Fig. 7K).
Morphological abnormalities were also evident in roots of the antisense plants. Root development was delayed with as much as a tenfold reduction in root length 1 week after germination (Fig. 6C). Arabidopsis roots can be divided into three zones along the apical-basal axis: the cell differentiation zone, the cell elongation zone, and the meristematic zone (RAM). The cell elongation zone,located just above the RAM, consists of rapidly proliferating and expanding cells with a dense cytoplasm. Cells in this zone are responsible for root growth (Dolan et al., 1993). The cell differentiation zone lies above the elongation zone and consists of fully differentiated cells. Antisense seedlings had a reduced elongation zone,and root hairs emanated from the epidermal cells adjacent to the RAM(Fig. 8A,B). This is in contrast with the wild-type roots where root hairs are elaborated in the differentiation zone. Ten days after germination, antisense roots contained fewer densely cytoplasmic cells in the elongation zone(Fig. 8C,D). The epidermal and cortical cells were enlarged and vacuolated(Fig. 8D,F), a phenotype similar to that observed in the SAM. In addition, the arrangement of the four types of initials in the RAM around the quiescent center was perturbed(Fig. 8E,F).
In summary, partial depletion of AtCAP-E1 and AtCAP-E2 proteins compromises primary meristem structure and function, consistent with a requirement for SMC2 proteins in actively dividing cells during embryonic as well as post-embryonic development.
Dosage effect of SMC2 proteins
We used immunoblotting to assess the SMC2 protein levels in wild type and in different mutant backgrounds, using a polyclonal antiserum raised against a fragment of AtCAP-E1. The antiserum detects a single protein of approximately 130 KDa in a total protein extract from 7-day old wild-type seedlings(Fig. 9, lane1). Based on conceptualized translation, this corresponds well with the size expected for AtCAP-E1 and AtCAP-E2. In antisense seedlings, the signal is reduced dramatically compared to wild type (lanes 2-4). In E1-/-E2+/+ plants (titan 3), no band was detected confirming that these plants did not contain detectable levels of AtCAP-E1 protein (lane 5). Although AtCAP-E2 should be present in this extract, its levels are too low to be detected by immunoblotting. A faint band was detected from these extracts when six- to eightfold more protein extract was used for immunoblotting (data not shown), suggesting that the sensitivity of immunoblotting is most likely the limiting factor in detecting AtCAP-E2. The protein band detected in the E1+/+E2-/- extract is similar in strength to the wild type (lane 6). Extracts of E1+/-E2-/- plants gave rise to a signal, which was reduced relative to wild type, but significantly more than that of antisense plants (lane 7), pointing to a significant knock-down of both AtCAP-E1 and AtCAP-E2 gene products in the antisense plants. Taken together, our data suggest that Arabidopsis plants can use both proteins redundantly, that a reduction in SMC2 levels gives rise to fasciated plants, and that a threshold level of SMC2 molecules is required for normal embryogenesis and development.
Expression of SMC2 genes
Our RT-PCR analyses demonstrated that AtCAP-E1 is expressed at much higher levels than AtCAP-E2, but both genes are expressed in a pattern that parallels that of the mitotic activity of the tissues. Meristems and developing tissues of transgenic plants harboring an AtCAP-E1::GUS construct stained intensely, whereas more mature tissues showed little or no expression. Similar expression patterns have been observed for other Arabidopsis cell cycle genes [Cdc2a(Hemerly et al., 1993), Cyc1At (Ferreira et al.,1994; Donnelly et al.,1999) and PROLIFERA(Springer et al., 2000)] where gene activity was observed only in proliferating tissues. Gene expression of AtCAP-E1 and AtCAP-E2 may be regulated by cell cycle-specific factors. Sequence analysis of the promoters of these genes revealed a putative binding site for the E2F transcription factor in each promoter. E2F transcription factors regulate cell cycle-specific gene expression at the G1/S transition (Olivier and Nicholas, 2002). Given our expression data and the presence of putative E2F sites in the promoters, it is likely that these genes are transcriptionally regulated during the cell cycle.
It is also possible that protein activity is controlled by post-translational modifications. In S. pombe, the condensin complex is translocated to the nucleus after the phosphorylation of the SMC4 ortholog by the cdc2 mitotic kinase(Sutani et al., 1999). In Xenopus and humans, phosphorylation of a non-SMC component of the complex by cdc2 has been implicated in supercoiling and condensation activities (Kimura et al.,1998; Kimura et al.,2001), and in C. elegans, AIR-2/AuroraB kinase-dependent phosphorylation is required for the cell cycle-dependent localization of the condensin complex (Hagstrom et al.,2002). We are currently employing immunocytochemistry to address questions of cell cycle-dependent expression and localization of SMC2 proteins using synchronized suspension cultured cells.
How do reduced levels of SMC2 influence viability?
In several systems, in vitro experiments have demonstrated that the condensin complex possesses a DNA-dependent ATPase activity that induces positive supercoiling, which probably results in a reorganization of the topology of DNA loop domains (Kimura et al., 1999; Kimura et al.,2001). In Drosophila, mitotic defects in embryos and formation of chromatin bridges are caused by mutations in the BARRENgene, which encode a non-SMC condensin subunit(Bhat et al., 1996). Similarly,a complete or partial depletion of SMC2/SMC4 transcripts by RNAi treatment in C. elegans, results either in embryo lethality or in mature adults with a variety of developmental defects, respectively(Hagstrom et al., 2002). Analysis of the mutant progeny revealed chromosome segregation defects in both somatic and germline cells, and as we observed in Arabidopsis,affected cells were abnormally shaped and fecundity was compromised. Thus, in the absence of a functional condensin complex, chromatin may remain relatively decondensed and tangles may remain unresolved, leading to chromosome breakage and abnormal chromosome segregation.
We observed that the highest levels of Arabidopsis SMC2 transcripts occur in tissues with high mitotic indices. We were interested in determining if reduced levels of SMC2 protein also affected meiotic chromosome behavior. During meiosis, two distinct nuclear divisions occur to reduce the chromosome number; in the first, homologous chromosomes segregate, while at meiosis II, sister chromatids segregate.
Our GUS and in situ hybridization studies of wild-type plants demonstrated that SMC2 genes are expressed in meiotic cells. But from these studies it was not possible to determine if the SMC2 protein plays a role in the first or second meiotic division, or both. We were not able to recover either E1-/-E2-/- or E1-/-E2+/- mutants due to lethality early in development. However, we did recover a limited number of E1+/-E2-/- plants and examined their chromosome behavior in the first division. These plants showed a reduction in condensation at metaphase and anaphase I, and a mild cut-type phenotype at anaphase I. In order for homologous chromosomes to segregate from each other they must resolve the chiasmata, which hold them together. Chiasmata resolution requires that sister chromatids decatenate from each other along the chromosome length. Thus the cut-phenotype at meiosis I is probably due to failure to completely decatenate sister chromatids. We did not examine the second meiotic division, but given that the chromosomes did not properly segregate at the first meiotic division, it would be difficult to distinguish affects specific to the second division from secondary effects due to failure at the first division. Our study is the first report demonstrating a role for SMC2 proteins in the first meiotic division. The only other report on the role of SMC2 proteins in meiosis focused on RNAi depletion of the C. elegans SMC2 gene [Mix-1(Hagstrom et al., 2002)]. These authors reported that depletion of Mix-1 in meiotic cells did not affect first division segregation, but did affect meiosis II. Compared with our observations, this contrasting result may be due to the fact that the Mix-1/SMC4 condensin complex, unlike the condensin complexes of other organisms, is not associated along the central axis of the chromosome, but instead outlines condensing prophase chromosomes, then localizes specifically to the centromere region.
How do reduced levels of SMC2 affect development?
Development of the aerial parts of plants relies on the activity of meristems, which continuously initiate new organs. Functional meristems are maintained by a coordinated balance between stem cell divisions at the summit of the meristem and loss of cells by differentiation to form primordia at its flanks. The meristem phenotypes we observed could be due to two separate, but not exclusive, mechanisms. First, the morphogenetic consequences could be a direct effect of an altered cell cycle. If chromatin remains relatively decondensed and catenated, the propensity for DNA topological stress is elevated in actively proliferating cells, and this may activate a checkpoint control to result in cell cycle arrest. Indeed, very few cells of the enlarged meristems in antisense plants contain histone H4 mRNA, indicating that most cells have ceased cycling. Other studies involving the perturbation of cell cycle regulatory factors have shown that these also can give rise to meristem defects. Arabidopsis plants expressing a dominant-negative form of cdc2a can exhibit defects in apical/basal patterning of the embryo,and in some instances meristems elaborate deformed leaves in an incorrect phyllotactic pattern (Hemerly et al.,2000). Similarly, the overexpression of cyclin D3 results in a variety of developmental defects, including meristem disorganization(Riou-Khamlichi et al., 1999). In root initials, alterations in the activity of a CDK-activating kinase led to differentiation of the initials, followed by cessation of cell division(Umeda et al., 2000).
The second potential mechanism involves epigenetic control of morphogenesis. Evidence is accumulating that condensins and other factors involved in DNA/chromatin dynamics epigenetically alter gene expression programs. One of the regulatory proteins of the Drosophila condensin subunit interacts with polycomb group proteins to maintain transcriptional silencing of homeotic genes (Lupo et al.,2001). This indicates that condensin might act in the epigenetic control of gene expression. A similar mechanism has been suggested in the fasciata CAF-1 mutant that ectopically expresses the WUSCHELand SCARECROW genes (Kaya et al.,2001). Molecular and cytological studies in Arabidopsishave included analyses of centromeres and associated heterochromatin, and surprisingly, these regions contain expressed genes, including some encoding putative transcription factors and other signaling molecules(Copenhaver et al., 1999). If the Arabidopsis condensin preferentially associates with heterochromatin, it may dictate how and when genes located in heterochromatin are expressed, and in antisense or mutant plants having reduced levels of SMC2, this may lead to misregulation of gene expression. Whether the developmental defects we observed are a direct consequence of chromosomal damage or an indirect effect mediated by an epigenetic mechanism needs to be addressed.
In antisense plants exhibiting severe phenotypes, the primary SAM enlarges into a callus-like mass of tissue and aborts, but functional, albeit fasciated, axillary meristems arise. The difference in sensitivity of the primary and axillary meristems to impaired chromosome condensation in our antisense plants could be due to low expression of the antisense gene from the 35S promoter in axillary meristems or a differential ability of the cells to employ a checkpoint delay in response to defective chromosome segregation. It is possible that the duration of the cell cycle is longer in axillary meristems, which could allow more time for the cells to resolve the chromosomes, allowing them to circumvent damage and proceed through an arrest checkpoint. Although Laufs et al. (Laufs et al., 1998) compared the mitotic indices and cell cycle dynamics in inflorescence and floral meristems in Arabidopsis and found them to be similar, to our knowledge no comparative information exists on cell cycle parameters of primary versus axillary meristems.
Threshold levels of SMC2 are required for proper development
Amongst organisms in which SMC2 genes have been characterized, Arabidopsis is unique in that it harbors two closely related genes instead of one. Our experiments show that AtCAP-E1 and AtCAP-E2 can act redundantly, as indicated by mild or no defects in the single mutants(compared to the catastrophic consequences of the double null mutant), and the similar expression patterns as gauged by RT-PCR. Although the qualitative patterns of expression are similar, the two genes are differentially expressed, with AtCAP-E1 transcripts accounting for most of the SMC2 transcript pool. Differential regulation of duplicated genes has been observed in a number of species and probably drives functional divergence(Pickett and Meeks-Wagner,1995). Such differential regulation might ensure normal development despite physiological or genetic perturbations. It is not known if the duplication of SMC genes in Arabidopsis represents a unique situation, or if plants in general have evolved this mechanism to avoid problems associated with improper chromosome condensation.
Although both single mutants are relatively normal, heterozygosity for insertion into one gene in a null background for the second gene confers two distinct phenotypes: lethality during embryogenesis(E1-/-E2+/-) or fasciation of relatively normal plants(E1+/-E2-/-). This suggests that the effect of the SMC2 proteins is dose dependent. Reducing the AtCAP-E2 level to one half of its level in E1-/-E2+/+ plants, which develop normally,results in death during embryogenesis, and suggests that AtCAP-E2, despite its low levels in the cell, has a higher specificity for its target. In the antisense lines, where the expression of both genes is reduced, and in E1+/-E2-/- plants, fasciation occurs, indicating that a critical threshold level of SMC2 proteins are required to sponsor events of chromosome condensation and support normal morphogenesis. The use of SMC2 antibodies in chromatin immunoprecipitation assays may provide some clues on binding specificity and identify target sequences important for condensin complex function.
The authors wish to thank Dr Douglas Koshland for kindly providing the yeast smc2-Δ6 and wild-type strains, Dr Nicole Chaubet-Gigot for providing the histone H4 clone, and the Arabidopsis Biological Resource Center (Ohio State University) for providing seeds of titan3. We also thank Drs Nancy Dengler and Michele Heath for sharing equipment, Sasha Singh and Danielle Vidaurre for assistance with microscopy,David Douda for assistance with genotyping, and Dr Nancy Dengler and Scott Douglas for critical review of the manuscript. We appreciate the helpful advice of three anonymous reviewers. N.U.S. was supported by an Ontario Graduate Scholarship and a University of Toronto Research Fellowship. P.E.S. was supported by a post-graduate fellowship from the Natural Sciences and Engineering Research Council of Canada (NSERC). This research was supported by NSERC research grants to C.A.H. and C.D.R.