The Arabidopsis-mei2-Like Genes Play a Role in Meiosis and Vegetative Growth in Arabidopsis^[W]

The AML Genes Are Widely Expressed in Vegetative and Reproductive Tissues, with Strong Expression in Meristems and Meiocytes

The predicted AML proteins show greatest similarity to Mei2p in the C-terminal portion within a region that encompasses RRM3 (Figure 1A). A phylogenetic analysis of the mei2-related proteins AML1-AML5, At TEL1, At TEL2, OML1-OML5, and Zm TE1 from Arabidopsis, rice (Oryza sativa), and maize along with Mei2p is depicted in Figure 1B. AML2, AML3, and AML5 are clustered in one clade that includes OML2 and OML5, whereas AML1 and AML4, which are closely related, cluster together with OML3 and OML4. At TEL1 and At TEL2 cluster with OML1 and Zm TE1 in a distinct clade. These relationships are in general agreement with those reported by Anderson et al. (2005). Semiquantitative RT-PCR analyses were performed to examine expression of the AML genes in different tissues. Expression of AML1-AML5 was detected in buds, open flowers, green siliques, leaves, and roots, indicating a broad expression pattern of these genes (Figure 1C). To compare these findings with other expression patterns, we queried the Genevestigator microarray expression database https://www.genevestigator.ethz.ch (Zimmermann et al., 2004).

The analysis indicated that all five AML genes are expressed in all organs examined, covering both vegetative and reproductive development. The normalized expression values for the different genes ranged from 823 ± 18 to 1976 ± 66 in adult leaves and from 1911 ± 82 to 2136 ± 162 in the inflorescence. This contrasts, for example, with AP1, whose expression is known to be confined to reproductive tissues (Gustafson-Brown et al., 1994) and for which the normalized expression value for the inflorescence was 20-fold higher than for adult leaves (1879 versus 98).

To obtain specific expression patterns, we examined expression of AML1-AML5 in aerial vegetative and reproductive tissues, using RNA in situ hybridization. The results indicated an overall expression pattern for all five genes that was very similar (Figures 2 and 3). In each case, a basal level of expression was present in most tissues. More concentrated expression was seen in the vegetative shoot apical meristem and the inflorescence meristem (Figure 2) as well as in the axillary buds (data not shown). For all the genes, a strong signal was present in the vegetative apical meristem covering all cell layers and in the emerging leaves (Figure 2A). All the AML genes except AML3 showed strong expression in the inflorescence meristem and floral buds (Figure 2B). In stage 6 flowers, expression was concentrated in the developing anther and pistil primordia and to a lesser extent in the sepals (Figure 2C). In flowers at stage 9, increased expression was observed for all AML genes in developing carpels along the walls of the placenta and ovule primordia (Figure 2D) and in male meiocytes, which showed a strong signal (Figure 3A). After meiosis, reduced expression was observed in developing tetrads (Figure 3B). Ovules showed expression in all cells, including the female meiocyte (Figure 3C), and later in the embryo sac, which showed strong expression (Figure 3D). Therefore, a detailed examination of the expression patterns for AML1-AML5 demonstrated that all genes are expressed at multiple stages of vegetative and reproductive development, including meiosis and gametogenesis, and overlap in their expression patterns. The strong meiotic expression was particularly striking and prompted us to look for a possible function during meiosis as is the case for the related gene mei2 of S. pombe.

RNA Interference with AML5 Causes Sterility and Defects in Gametogenesis

Given the close similarity in expression pattern between all five AML genes, the possibility of functional redundancy between members of the AML family appeared likely. This was in fact borne out in subsequent experiments (see below). We initially adopted an RNA interference (RNAi) approach to address the function of AML genes in meiosis. A 572-bp region of the AML5 cDNA (GenBank entry NM_179396; coordinates 2112 to 2683) encoding the C-terminal part of the protein was selected for RNAi experiments. This contains the region most conserved in all of the AML genes and includes RRM3. To estimate similarity and specificity, we searched the Arabidopsis genome using BLASTN for short nearly exact matches to the 572-bp region. Close similarity was detected only to each of the other four AML genes at E values between 1e⁻⁴⁴ and 1e⁻¹¹. The next level of similarity was at E = 0.034. Hence, the RNAi strategy was designed to specifically target the five AML genes.

The 572-bp region was cloned in both sense and antisense orientations in the vector pKANNIBAL (Wesley et al., 2001). Lines carrying the AML5 RNAi construct expressed under control of the 35S promoter were generated and examined for phenotypes. The main phenotypes we observed in T1 plants were sterility coupled with increased branching, suggesting a loss of apical dominance (Figure 4). Out of 70 T1 plants, 11 displayed >90% sterility, whereas none of the control transformants (0/80) showed sterility. To investigate the basis of sterility, we examined embryo sac and pollen development in T1 and T2 plants from lines exhibiting strong sterility (>90%). In ovules, the phenotype ranged from arrest of embryo sac development at early stages to degeneration of the embryo sac. At low frequency (∼5% of ovules), we also observed two enlarged cells along with degenerating material from the nonfunctional megaspores (Figure 4L). The two cells could represent the products of an abnormal division of the functional megaspore, or they could possibly represent two surviving megaspores. Although the frequency of occurrence of such two-celled structures was low, they were observed in several independent RNAi plants as well as in antisense transformant plants (J. Kaur, unpublished data) and in plants carrying triple mutant combinations of aml alleles (described below). We did not observe such a phenotype in wild-type or control transformants. Defects in male gametophyte development were also observed. The pollen was shrunken and showed reduced viability as revealed by Alexander staining. In addition to defects in embryo sac and pollen development, we also observed defects in seedling development in several T2 and T3 generation RNAi lines (described below).

To test if the defects in gametogenesis correlated with a reduction in the mRNA levels of the AML genes, we quantitated RNA levels for each of the five AML genes in the inflorescence using semiquantitative RT-PCR (Figure 5). Lines showing strong sterility, such as R12, R45, and R47, exhibited a twofold or greater reduction in level for at least three of the AML genes, whereas line R25, which showed moderate sterility, exhibited less reduction in RNA levels. RNAi line R20 that did not show sterility failed to show such a reduction. Therefore, the phenotype correlates with reduction in RNA levels. The analysis suggested that AML1 and AML4 are more important, as their expression was substantially reduced in all the strong RNAi lines. A subset of RNAi lines exhibiting sterility was also crossed to the wild type, and in each case, the phenotype was found to be heritable and dominant in F1 (data not shown), consistent with it being due to RNAi.

Combinations of aml Insertion Alleles Cause Defects in Gametogenesis and Vegetative Growth

We obtained one insertion line for each of the five AML genes from the SALK T-DNA collection (Alonso et al., 2003). The presence of the T-DNA insertion in each of the lines was confirmed by PCR using T-DNA border primers in combination with gene-specific primers on both sides of the insertion for each of the genes so as to also establish the presence and orientation of T-DNA border sequences at both ends of the insertion (Figure 6; data not shown). The aml1, aml2, and aml4 insertions are in exons, whereas the aml3 and aml5 insertions are in introns. We used gene-specific primers within the coding region and downstream from the point of insertion to examine transcript levels by RT-PCR. For aml2, aml3, and aml5, no product was detected, indicating that they are probably null alleles. For aml1, we failed to detect a transcript using a T-DNA primer and a gene-specific primer downstream of the insertion, although we could detect a transcript using primers that were upstream of the insertion. These results suggest that the transcript is truncated. The aml1 insertion site is 12 bases after the predicted end of RRM3 and, hence, in a region that is likely to be important for function. In the case of aml4, which carries an insertion in exon 6, we detected the presence of transcripts downstream of the insertion site. However, we could also detect the presence in transcripts of T-DNA, which interrupted the coding region. Our analysis therefore indicated that for each of the insertion lines, the insertion is likely to cause a severe reduction in function of the respective AML gene.

We identified plants homozygous for the insertion in each case. None of the single mutant lines displayed a visible phenotype; therefore, double mutants were generated in all possible combinations between aml1, aml2, aml3, aml4, and aml5. In the aml1 aml4 double mutant, we observed a partially penetrant phenotype at the seedling stage, wherein approximately one-quarter (20/86) of the seedlings were arrested at cotyledon expansion and failed to grow new leaves (Figure 7). A similar phenotype was observed in the aml1 aml2 aml4 triple mutant (obtained by crossing aml1 aml2 and aml1 aml4) as well as in T2 and T3 generation RNAi lines, and some of these seedlings also showed defects in root development (Figure 7B). Scanning electron microscopy of arrested seedlings (Figures 7D to 7H) showed that the shoot apical meristem did initiate leaf primordia; however, these were much slower in growth and expansion when compared with the wild type, suggesting a defect in meristem activity. The AML1 and AML4 genes show the highest level of sequence similarity (75% amino acid sequence identity) among the five AML genes and are hence likely to share a greater degree of functional redundancy. No phenotype was observed for any of the other double mutant combinations.

Triple mutant combinations were generated by crossing together homozygous double mutants. Sterility and defects in male and female gametogenesis were observed in 8/8 of the heterozygous aml1 aml2/+ aml4/+ F1 plants. Homozygous triple mutant plants were identified and analyzed in the F2 generation. The 5/5 aml1 aml2 aml4 triple mutant plants that were identified showed sterility and gametogenesis defects (Figure 8) but not aml1 aml2 aml3 and aml2 aml3 aml4 combinations. The sterility was ∼30 to 60% and correlated with defects in female gametogenesis (Tables 1 and 2). Male gametogenesis was defective as indicated by the presence of ∼30 to 40% shrunken pollen. The female phenotypes representing arrest at early stages or degenerating embryo sacs were similar to those observed for the RNAi lines.

AML Genes Play a Role in Meiosis

The defects in male and female gametophyte development described above could arise from a defective meiosis. We therefore examined chromosome spreads from various stages of meiosis in meioctyes from T1 and T2 generation RNAi plants and the aml1 aml2 aml4 triple mutant combination. In observations on several T1 and T2 RNAi lines, an overall 115/457 (25%) of the male meioses showed abnormalities (Table 3), whereas for the wild type, 0/500 meioses were abnormal. The abnormalities included pairing defects ranging from partial desynapsis to the formation of univalents, fragmentation and appearance of acentric pieces, and clumping of chromosomes (Figure 9). Chromosome bridges were observed resulting from exchange between what appeared to be nonhomologous chromosomes (Figure 9L). Similar defects were also observed in female meiosis (4/11 at diplotene) (Figure 9XX). We observed multiple instances of the appearance of an acentric fragment prior to anaphase I, in what was an otherwise normal-looking meiocyte (Figure 9N). At prometaphase I to metaphase I, there was clumping of chromosomes (Figure 9M). Defects were also observed during meiosis in the triple mutant combination of aml1 aml2 aml4 and aml1 aml2/+ aml4/+ plants. The 181/738 meioses covering diplotene to tetrad stages in aml1 aml2 aml4 plants were abnormal. Abnormalities included desynapsis and the formation of univalents (Figures 9P and 9W), chromosome bridges (Figures 9R and 9U), presence of an acentric fragment (Figure 9T), and clumping (Figures 9S and 9X). In addition, 5/25 prometaphase II meiocytes lacked an organelle band at the center (Figure 9Y), whereas all (26/26) wild-type prometaphase II stages examined showed an organelle band. The range of meiotic abnormalities is therefore similar between the RNAi lines and the triple mutant allelic combinations examined here. Together, these data provide evidence for a role for the AML genes in meiosis in plants.

Evidence for a Gametophytic Function for AML Genes

The postmeiotic defects observed during gametophyte development could arise from a defective meiosis or additionally from a gametophytic requirement for AML function after meiosis. Evidence in support of the latter came from the finding that transmission of the aml1 aml4 double mutant haplotype is defective in a parental background of aml1 aml4/+, where there was no evidence for sterility. Genotyping of all seeds (51 total) from a single silique (100% seed set and germination efficiency) obtained from an aml1 aml4/+ plant indicated recovery of the aml4, aml4/+, and +/+ genotypes at frequencies (5/51, 17/51, and 29/51, respectively) that were significantly different from those predicted by randomness [χ² = 28.3; P(χ² > 13.8) = 0.001 for 2 df]. A likely explanation is that the aml1 aml4 pollen haplotype shows reduced fitness and is outcompeted by the aml1 AML4 haplotype during pollination.

Plant Material and Growth Conditions

The Columbia ecotype of Arabidopsis thaliana was used for generating transgenic lines expressing the AML5 RNAi construct. Lines carrying a T-DNA insertion in each of the AML1-AML5 genes were obtained from the ABRC (SALK_015088, SALK_029713, SALK_006041, SALK_019467, and SALK_061664). Plants were grown at 21°C under a 16-h-light and 8-h-dark cycle as described previously (Siddiqi et al., 2000).

Phylogenetic Analysis

Sequences were aligned using ClustalW with default parameters (Higgins et al., 1994). The Phylip 3.6a3 package (Felsenstein, 1989; http://bioweb.pasteur.fr) was used for construction of the phylogenetic tree using 100 replicates with neighbor joining as the set criterion and Mei2 as the outgroup in the consensus tree.

Generation of AML5 RNAi Lines

A 572-bp fragment corresponding to the C-terminal region of AML5 was amplified using primer pair AML2F1 (5′-GACTCGAGTCTAGAGTCCGGGATTCTCGGACTACC-3′) and AML2R1 (5′-TAGAATTCGGATCCTGCATGCATCACTTAACTCTG-3′), which engineered XhoI-XbaI and EcoRI-BamHI enzyme sites at the 5′ and 3′ ends, respectively. The 572-bp fragment was cloned in the sense orientation as a XhoI-EcoRI fragment and in the antisense orientation as a BamHI-XbaI fragment in the pKANNIBAL vector (Wesley et al., 2001). This AML5 RNAi construct was further subcloned into the binary vector pART27 and transformed into Agrobacterium tumefaciens strain AGL1 for in planta transformations. The pKANNIBAL vector without the insert was used as a control. In planta transformation was performed as described by Bechtold and Pelletier (1998). The AML5-RNAi transformants were selected on Murashige and Skoog plates containing 2% sucrose and kanamycin at 50 μg/mL and transferred to soil. The sterility was scored as the mean percentage reduction of seeds in approximately five siliques, assuming a wild-type seed set of 50 per silique. Plants scored as >90% sterile gave less than three seeds per silique. Wild-type plants set 52 ± 6 seeds per silique.

Expression Analysis by Semiquantitative RT-PCR

Plant tissue was frozen in liquid nitrogen and stored at −80°C. Total RNA was isolated from the inflorescence using Trizol (Invitrogen) according to the manufacturer's protocol and treated with RNase-free DNase1 (RQ1; Promega) before proceeding for reverse transcription. Equal amounts of RNA (∼1 μg) from different samples was reverse transcribed using MMLV RT (Gibco) according to the manufacturer's instructions. cDNA product was diluted 10 to 50 times, and 2 μL of these dilutions was used to determine the linear range for AML1-AML5. GAPC expression was measured using GAPC1 and GAPC2 primers after determining the linear amplification range and used for normalization. PCR was performed after equalizing the reverse transcription products for each sample with respect to GAPC. Primers AML1up and AML1sac were used for detecting expression of AML1, 140jUp and 140jD for AML2, Chr4F and Chr4R for AML3, AML5F and AML5R2 for AML4, and 337c and 40BD for AML5. Sequence information for all the primers used in this study is given in Supplemental Table 1 online. RT-PCR products were separated on a 1% agarose gel, blotted onto Hybond N⁺, and hybridized with the respective α³²P-labeled probes for AML1-AML5 and GAPC. Hybridization signals were detected using a Fuji FLA-3000 phosphor imager and quantified using Image Gauge software.

Microscopy

Developmental analysis of whole-mount anthers and ovules was done after fixing and clearing the inflorescence in methyl benzoate as described previously (Siddiqi et al., 2000). The slides were observed on a Zeiss Axioplan 2 imaging microscope under differential interference contrast optics using a ×40 oil immersion objective. Pollen viability was examined using the method of Alexander staining (Alexander, 1969). Images were captured on an Axioplan CCD camera using Axiovision (version 3.1) and processed using Adobe Photoshop 6.0.

Meiotic chromosome spreads of male meiocytes were prepared and analyzed according to the protocol of Ross et al. (1996), with minor modifications as described in Agashe et al. (2002). Chromosomes were stained with 4′,6-diamidino-2-phenylindole (1 μg/mL) and observed on a Zeiss Axioplan 2 imaging microscope using a 365-nm excitation and 420-nm long-pass emission filter and a ×100 oil objective. The images were captured on an Axioplan CCD camera using Axiovision (version 3.2) and processed using Adobe Photoshop 6.0.

For scanning electron microscopy analysis, the plant material was fixed in formaldehyde-acetic acid-alcohol containing 1% Triton X-100 at 4°C overnight and dehydrated through an acetone series followed by critical point drying in liquid CO₂. The dried samples were mounted on stubs and sputter coated prior to examination using a Leica 240 scanning electron microscope.

Analysis of T-DNA Mutant Lines and Multiple Mutant Combinations

Lines carrying T-DNA insertions in AML1-AML5 were obtained from the ABRC. In each case, the presence of the T-DNA insert was confirmed using a left border outwardly directed primer, LB1, in combination with a gene-specific primer flanking the site of insertion. Both junction fragments were amplified for aml1 to aml5 using gene-specific primers on each side of the insertion in combination with a T-DNA left border primer. In the case of aml1, the two junction fragments were amplified using LB1-K22R and K22F-InvF (a left border inwardly directed primer), respectively. For aml2, the junction fragments were amplified using the gene-specific primers F7F1-LB1 and F7R1 in combination with internal left border primers Tail4 and Tail3 for the other end. This structure is consistent with a head-to-head insertion of the T-DNA. The aml3, aml4, and aml5 insertions were also head-to-head insertions, and both the junction fragments were identified using the same internal left border primers in combination with the gene-specific primers F15J5F1 and F15J5R1 for aml3, T28F1 and T28R1 for aml4, and F15D2F1 and F15D2R1 for aml5. For genotyping of plants, primers flanking the insertion were used to identify the wild-type alleles, and the mutant alleles were identified using LB1 in combination with K22R1, F7F1, F15J5F1, T28F1, and F15D2F1 for aml1 to aml5, respectively.

For generating double mutant combinations, homozygous insertion lines were crossed, and the F2 population was genotyped to identify plants carrying both the mutations in homozygous condition. For generating triple mutants, homozygous double mutants were used as parents. The two parental lines carried one common mutation in the homozygous condition. Homozygous triple mutants were identified and analyzed in the F2 generation.

In Situ Hybridizations

In situ hybridizations were performed as described previously (Siddiqi et al., 2000). Antisense RNAs specific to each of the AML genes were used as probes along with sense controls. The regions used for strand-specific cDNA probes were as follows: AML1 (NM_125589), 1583 to 2120; AML2 (NM_201945), 1080 to 1853; AML3 (NM_117922), 923 to 1761; AML4 (NM_120811), 1590 to 2623; AML5 (AY091327), 946 to 1612. Specificity of the probe was validated in cross-hybridization experiments using radioactively labeled probe under conditions of stringency that were the same as those used for in situ hybridizations. For each gene, the probe used showed >12-fold specificity over hybridization to the related genes. Floral stages are according to Smyth et al. (1990).

Accession Numbers

Sequence data from this article can be found in the GenBank/EMBL data libraries under accession numbers X07180 (Mei2p), At5g61960 (AML1), At2g42890 (AML2), At4g18120 (AML3), At5g07290 (AML4), At1g29400 (AML5), At3g26120 (At TEL1), At1g67770 (At TEL2), BAB92568 (OML1), BAD12869 (OML2), AAW56930 (OML3), BAD46727 (OML4), BAD28947 (OML5), and AF047852 (Zm TE1).

Supplemental Data

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

• Supplemental Figure 1. ClustalW Multiple Sequence Alignment.

• Supplemental Table 1. List of Primers Used in This Study.