Plant Mol Biol (2006) 62:291–304 DOI 10.1007/s11103-006-9021-2 The putative SWI/SNF complex subunit BRAHMA activates flower homeotic genes in Arabidopsis thaliana Lidia Hurtado Æ Sara Farrona Æ Jose C. Reyes Received: 7 March 2006 / Accepted: 16 May 2006 / Published online: 15 July 2006
Springer Science+Business Media B.V. 2006 Abstract Arabidopsis thaliana BRAHMA (BRM, also called AtBRM) is a SNF2 family protein homolog of Brahma, the ATPase of the Drosophila SWI/SNF complex involved in chromatin remodeling during transcription. Here we show that, in contrast to its Drosophila counterpart, BRM is not an essential gene. Thus, homozygous BRM loss of function mutants are viable but exhibit numerous defects including dwarfism, altered leaf and root development and several reproduction defects. The analysis of the progeny of self-fertilized heterozygous brm plants and reciprocal crosses between heterozygous and wild type plants indicated that disruption of BRM reduced both male and female gametophyte transmission. This was consistent with the presence of aborted ovules in the selffertilized heterozygous flowers that contained arrested embryos predominantly at the two terminal cells stage. Furthermore, brm homozygous mutants were completely sterile. Flowers of brm loss-of-function mutants have several developmental abnormalities, including homeotic transformations in the second and third floral whorls. In accordance with these results, brm mutants present reduced levels of APETALA2, APETALA3, PISTILLATA and NAC-LIKE, ACTIVATED BY AP3/PI. We have previously shown that BRM strongly interacts with AtSWI3C. Now we extend our interaction studies demonstrating that BRM interacts weakly Lidia Hurtado and Sara Farrona authors contributed equally to this work L. Hurtado Æ S. Farrona Æ J. C. Reyes (&) Instituto de Bioquı́mica Vegetal y Fotosı́ntesis, Consejo Superior de Investigaciones Cientı́ficas-Universidad de Sevilla, Av. Américo Vespucio 49, E-41092 Sevilla, Spain e-mail: jcreyes@cica.es with AtSWI3B but not with AtSWI3A or AtSWI3D. In agreement with these results, the phenotype described in this study for brm plants is very similar to that previously described for the AtSWI3C mutant plants, suggesting that both proteins participate in the same genetic pathway or form a molecular complex. Keywords Chromatin Æ AtSWI3 Æ SWI/SNF complex Æ Homeotic gene expression Introduction Chromatin structure constrains the interaction of DNA with most nuclear factors involved in transcription, replication, DNA repair, and recombination. ATPdependent chromatin remodeling complexes alter the interactions between histones and DNA in order to create accessible DNA (Cairns 2005). The SWI/SNF complex was the first ATP-dependent chromatin remodeling complex characterized (reviewed in (Smith and Peterson 2005)). swi and snf mutants were first identified in Saccharomyces cerevisiae by their defects in mating type switching (SWI) and/or sucrose fermentation (SNF; sucrose non-fermenting) (Neigeborn and Carlson 1984; Stern et al. 1984). Later analysis demonstrated that lack of both SWI or SNF genes causes pleiotropic phenotypes. Actually, whole-genome expression analysis has shown that expression of about 5% of the yeast genes are affected, positively or negatively, by the loss of SWI and SNF proteins (Sudarsanam et al. 2000). Biochemical analysis demonstrated that yeast SWI and SNF proteins function together as a multi-subunit complex (Cairns et al. 1994; Peterson et al. 1994) able 123 292 to facilitate binding of transcription factors to nucleosomal DNA in an ATP-dependent way (Cote et al. 1994). The purified complex is approximately one megadalton in size and contains eleven subunits: SWI1, SWI2/SNF2, SWI3, SNF5, SWP82, SNF6, SNF11, TFG3, and SWP73, ARP7, ARP9 (Smith et al. 2003). Complexes related to SWI/SNF have also been identified, and characterized in mammals and Drosophila (Wang et al. 1996b; a; Mohrmann et al. 2004). The SWI2/SNF2 protein from yeasts and its counterparts from higher eukaryotes belong to the SNF2 family of DNA dependent ATPases, and many studies have demonstrated that these proteins are the motor subunits responsible for the ATP-dependent nucleosomal remodeling activity of the complexes (Cairns 2005; Smith and Peterson 2005). However, the mechanism by which the interactions between DNA and histones are distorted by the SWI/SNF complexes is still unclear (for review see (Flaus and Owen-Hughes 2004)). Several studies have demonstrated the essential role of the SWI/SNF complexes in animal development and cell differentiation (see for example (Pedersen et al. 2001; Ohkawa et al. 2006). Much less is known about the putative SWI/SNF complex in plants. To date no SWI/SNF-like complex has been purified from plants, although, genome analysis suggests that Arabidopsis thaliana contains several potential SWI/SNF subunits (Plant chromatin database: http://chromdb.org). There are more than 40 ATPases of the SNF2 family in Arabidopsis, but only four belong to the SWI2/SNF2 subfamily based on phylogenetic analysis of the SNF2 ATPase catalytic domains. However, only one, Arabidopsis thaliana BRAHMA (AtBRM) presents all the domains that are characteristic of ATPases of SWI/ SNF complexes: a glutamine rich region at the amino terminus, domains I and II (defined for the Drosophila Brahma protein), the SNF2 ATPase catalytic domain, an AT-hook motif and a bromodomain (Farrona et al. 2004). We have recently renamed the AtBRM gene as BRM following the TAIR and EMBL/GenBank/ DDBJ nomenclature guidelines. The closest homolog of BRM is SPLAYED (SYD), which lacks the glutamine rich region and the bromodomain (Wagner and Meyerowitz 2002). Finally, the related proteins CHR12 and CHR23 also lack domain I. In addition to the ATPase motor subunits, Arabidopsis contains four genes encoding SWI3 homologs named AtSWI3A to AtSWI3D, two genes encoding SWP73 homologs (CHC1 and CHC2) and one gene encoding a SNF5 homolog (BUSHY, BSH). The existence of small gene families for a number of SWI/SNF subunits suggests the presence of several different SWI/SNF complexes in Arabidopsis. 123 Plant Mol Biol (2006) 62:291–304 Increasing evidence suggests that the putative plant SWI/SNF complexes control transcription of genes involved in many developmental processes. The syd mutants were first isolated in a screening to identify phenotypic enhancers of the weak leafy-5 allele (Wagner and Meyerowitz 2002). SYD acts with LEAFY to regulate shoot apical meristem identity. Furthermore, SYD also controls the shoot apical meristem stem cell pool via direct transcriptional control of the WUSCHEL gene (Kwon et al. 2005). Recently it has been shown that mutations of AtSWI3A and AtSWI3B arrest embryo development at the globular stage, while mutations in AtSWI3D and AtSWI3C cause dwarfism, different alterations in the number and development of flower organs and deregulation of flower homeotic genes (Sarnowski et al. 2005). Partial silencing of BRM by RNA interference results in small plants with curled leaves, flowers with small petals and stamens, reduced fertility and early flowering (Farrona et al. 2004). Here we show that brm loss-of-function mutants exhibit a stronger phenotype characterized by curly leaves, defects in root growth, flowers with homeotic transformations, complete sterility, and misregulation of homeotic genes. We also show that BRM is required for normal gametophyte development. Materials and methods Plant material and growth conditions Wild-type Arabidopsis thaliana (Columbia) and T-DNA mutant (Columbia) plants were grown either in pots containing a mixture of substrate and vermiculite (3:1) or aseptically in Petri dishes containing Murashige and Skoog media supplemented with 1% (wt/v) of sucrose and 0.37% (wt/v) of Phytagel (Sigma). Plants were grown in a cabinet under long-day (16 h light/8 h dark) or short-day (10 h light/14 h dark) photoregimes at 22C (day)/20C (night), 70% relative humidity, and a light intensity of 130 lE m–2 s–1 supplied by white fluorescent lamps. The SIGnal database (http://signal.salk.edu) was used to select lines containing T-DNA insertions. Lines from the Salk (SALK_038610, SALK_030046, SALK_002500) (Alonso et al. 2003) and Syngenta Arabidopsis Insertion Library (SAIL_224_B10) collections were obtained from NASC (Nottingham University, UK). T-DNA mutant GABI-854D01 was generated in the context of the GABI-kat program and provided by Bernd Weisshaar (MPI for Plant Breeding Research, Cologne, Germany) (Rosso et al. 2003). Plant Mol Biol (2006) 62:291–304 Analysis of T-DNA insertion lines The mutant lines were backcrossed twice with wild type before phenotypic analysis. Plant DNA was extracted from one to two young leaves using a previously described miniprep procedure (Guidet et al., 1991). The locations of the T-DNA insertion in the mutant lines was verified by PCR amplification of genomic DNA with the T-DNA left border LBa1 (5¢-GCGTGGACCGCTTGCTGCAACT-3¢) and BRM or AtSWI3C specific primers. The amplified fragments were then sequenced to identify the exact location of the inserts within the genes. To identify lines that were homozygous for the insertion, a second PCR was performed using two gene-specific primers. Only the wildtype allele was amplified by this PCR, whereas no amplification product was detectable from plants homozygous for the insertion. Segregation analysis was carried out by PCR-analysis of the genotype of 30–40 offspring obtained after self-pollination of heterozygous plants. BRM mutants are described in the Results section. The AtSWI3C mutant line (SAIL_224_B10) contains a single T-DNA (data not shown) inserted in exon 2 at nucleotide 211 of the ORF. RT-PCR analysis confirmed the lack of wild type AtSWI3C mRNA in the homozygous line. Given the fact that atswi3c-1 and atswi3c-2 mutant alleles have been already described by (Sarnowski et al. 2005) this new mutant was named atswi3c-3. BRM antibodies and immunoblot analysis Two different anti-BRM rabbit polyclonal antibodies were used. The a-AtBRMb antibody described in (Farrona et al. 2004) was raised against amino acids 2047–2187 of BRM and therefore recognizes the C-terminal part of the protein. For clarity, in this paper this antibody will be designed a-BRM-C. A second antibody (a-BRM-N) against a peptide encompassing amino acids 307–424 of BRM has been developed in order to recognize the possible truncated protein generated in the mutant strains. To produce the BRM-N antigen, a fragment of the BRM cDNA encompassing nucleotides 919–1273 was inserted into the pGEX-4T-2 plasmid in-frame with the GST. Purified GST-BRM307–424 was used to raise polyclonal antiserum in rabbit. Immunoblot analysis was carried out as described in (Farrona et al. 2004). Whole-mount preparations for microscopy Fruits of different developmental stages were prepared as described (Kohler et al. 2003). Briefly, the tissue was 293 fixed in ethanol-acetic acid (9:1) overnight at 4C. The preparation was then rehydrated and cleared in chloral hydrate solution (chloral hydrate:water:glycerol (8:2:1)) overnight at 4C. Ovules were dissected and observed using DIC optics. Anthers were stained with DAPI (4¢,6-diamidino-2-phenylindole) as follows. Anthers were teased apart and incubated for 1–24 h in coloration buffer. Coloration buffer contains equal volumes of extraction buffer (0.1% Nonidet P40, 10% DMSO, 5 mM ethylene glycol bis(2-aminoethyl ether)N,N,N¢N¢-tetraacetic acid (EGTA) (pH 7.5) and 50 mM piperazine-1,4-bis(2-ethanesulfonic acid) (PIPES) pH 6.9) and DAPI solution (1 M DMSO, 2.86 mM DAPI). Yeast two-hybrid analysis Yeast two-hybrid analysis was performed with the PROQUEST two-hybrid system (Invitrogen). Constructs that express protein fusions to the GAL4 DNA-binding domain (GDB) or transcriptional-activation domain (GAD) were performed using the vectors pDBleu and the pPC86, respectively. Full length cDNAs for AtSWI3A (clone number U16949) and AtSWI3D (clone number U19881) were obtained from the Arabidopsis Biological Resource Center (ABRC). Full length cDNA for AtSWI3B (accession number BX820245) were obtained from the National Center for Plant Genomic Resources (INRACNRGV). Full length cDNAs for AtSWI3C (accession number AV524064) was obtained from the Kazusa DNA Research Institute. cDNA fragments were generated by standard PCR techniques and cloned into the appropriate plasmids. All the clones were validated by sequence analysis. Details about the constructs will be provided upon request. Interaction experiments were carried out in the yeast strain MaV203. Expression of GAD and GDB fusion proteins was verified by protein gel blot using antibodies anti-GAL4-AD (G9293, Sigma) and anti-GAL4 DNA-BD (G3042, sigma). Total extracts of yeast were carried out as described by (Hoffman et al. 2002). Interaction was tested by two methods: Growth of transformed yeast in selective minimal medium SC medium-His-Leu-Trp (Bio101 Systems) supplemented with different concentrations of 3-amino-1,2,4-triazole (3-AT) and b-galactosidase activity, determined using the colony lift assay and the liquid assay as recommended by the manufacturer instruction manual. All the constructs were tested for autoactivation in the presence of the partner plasmid (pDBleu or pPC86) without cloned insert. 123 294 Plant Mol Biol (2006) 62:291–304 Recombinant protein expression and purification RT-PCR analysis Glutathione-S-transferase (GST) fusion recombinant proteins were used for pull-down experiments. PCRgenerated DNA fragments were cloned into the appropriate restriction sites of plasmids pGEX-4T in-frame with the GST gene. Details about the constructs and oligonucleotides used will be provided upon request. Fusion proteins were expressed in Escherichia coli BL21 cells. Two hundred ml to 1l of culture was grown in Luria broth to an optical density at 600 nm of 0.6, induced with 1 mM isopropyl-ß-D-thiogalactopyranoside for 2.5 h, harvested by centrifugation, and resuspended in 5–8 ml of PBS buffer (150 mM NaCl, 16 mM Na2HPO4, 4 mM NaH2PO4), supplemented with 4 mM phenylmethylsulfonyl fluoride and 1% Triton X-100. Cells were broken by sonication on ice, and insoluble debris was pelleted by centrifugation. Extracts were mixed with 0.5–1 ml of Glutathione-Sepharose 4B (Amersham) and incubated for 2 h at 4C with gentle agitation. Thereafter, beads were transferred to a column and washed extensively with PBS buffer until no more protein was eluted from the column. GST or GST fusion proteins were eluted with 3 ml of 50 mM Tris–HCl (pH 8.0) containing 10 mM reduced glutathione, and the eluates were dialyzed against the appropriate buffer. RNA was isolated by using the RNeasy Mini Kit (Qiagen). For semi-quantitative RT-PCR, 5 lg of total RNA were used to generate the first-strand cDNA with the SuperScript First-Strand Synthesis System for RT-PCR kit (Invitrogen). PCR amplification was performed using 2 ll of 20 ll of RT reaction and specific primers for each analyzed gene. About 15–25-amplification PCR cycles were employed and DNA products were detected by Southern blot hybridization. The number of PCR cycles chosen for each gene was shown to be in the linear range of the reaction in a separate experiment. Primer sequences and details about the probes used for DNA gel blot experiments are available upon request. Pull-down experiments A 1-lg portion of each of the purified GST-BRM691– 942, and GST proteins was dialyzed against interaction buffer (buffer I) (10 mM Tris–HCl [pH 8.0], 350 mM NaCl, 0.3% NP-40, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 0.25% bovine serum albumin [BSA]). These proteins were bound to glutathione-agarose resin equilibrated in buffer I (20 ll of 50% slurry). GST-BRM691–942 and GST containing beads were incubated for 2 h at 4C with in vitro transcription–translation reaction mixtures containing [35S]methionine-labeled AtSWI3C385–807 protein, under gentle stirring. After the supernatant containing unbound proteins was removed, the beads were washed three times with 1 ml of buffer I and once with BSAfree buffer I. The beads were boiled in 1· Laemmli buffer, and the samples were analyzed by SDS-PAGE using a 10% polyacrylamide gel. The gels were stained with Coomassie Brilliant blue and then transferred to Whatman 3 MM paper, dried and subjected to autoradiography. 123 Results Identification of T-DNA insertion mutants of BRM We have previously reported the phenotype of transgenic plants with reduced levels of expression of BRM by RNA interference (Farrona et al. 2004). These plants presented non-detectable levels of BRM protein in the aerial part of the plant and about 5% of the BRM mRNA, detectable by RT-PCR experiments but not by RNA gel blot. Surprisingly, levels of the BRM transcript and protein product were not altered in the roots of the transgenic plants (our unpublished observation). To better characterize the role of BRM in Arabidopsis development we decided to look for brm null mutants. For that, a search of T-DNA insertion lines in different T-DNA collections was carried out. Four different T-DNA lines were originally identified, three from the SALK collection (Alonso et al. 2003) and one from the GABI collection (Rosso et al. 2003) (Fig. 1A). The insertion site in the line SALK_038610 was mapped 339 bp upstream of the first translated nucleotide. Analysis of BRM mRNA levels by RTPCR in plants homozygous for this insertion demonstrated that this line presented normal levels of this transcript. Consistently, homozygous plants exhibited a wild-type phenotype. Line SALK_002500 contains a T-DNA in the last exon of the BRM gene, 39 bp upstream of the stop codon. RT-PCR experiments using oligonucleotides 1 and 2 (see Fig. 1A) demonstrated that homozygous plants of this line also presented normal levels of BRM mRNA. These plants also exhibited a wild type phenotype, suggesting that the last 13 amino acids of BRM are not required for its normal function. Line SALK_030046 contains a Plant Mol Biol (2006) 62:291–304 295 Fig. 1 Molecular characterization of BRM mutant alleles. (A) Molecular structure of the BRM locus and site of T-DNA integration. Colors represent the location of DNA regions encoding the different protein domains of BRM. Black, glutamine rich region; yellow, domain I; blue, domain II; red, ATPase domain; green, AT-hook motif; pink, bromodomain. Oligonucleotides used for RT-PCR analysis are numbered and depicted by arrows. (B) RT-PCR analysis of the amount of transcript found in wild-type (WT) and brm-1 plants. Oligonucleotides 1 and 2 shown in A were used for 35 cycles PCR amplification. A reaction without reverse transcriptase using brm-1 plants RNA was carried out as control (-RT). Amplification of the Ubiquitine (Ub) mRNA was used as control. (C) RT- PCR analysis of the amount of transcript found in wild-type (WT) and brm-2 plants. PCR reactions were carried out with the indicated oligonucleotide pairs. Reactions without reverse transcriptase were carried out as control (-RT). (D) Immunoblot analysis of the level of BRM protein. Nuclear extracts from WT, brm-1 and brm-2 inflorescences were subjected to immunoblotting using antibodies against the C-terminal (a-BRM-C) or the N-terminal (a-BRM-N) parts of BRM. Antibodies against the nuclear protein PICKLE were used as control (a-PKL). (E) Whole 25-day-old plants of WT (Columbia), brm-2, and brm-1 homozygous plants grown under long day conditions T-DNA in the first translated exon of BRM, 20 bp downstream of the first translated nucleotide. Although BRM transcripts were not detectable by RNA gel blot (data not shown), a small amount of mRNA was detected by RT-PCR experiments using oligonucleotides 1 and 2 (Fig. 1A, B) and 35 PCR cycles. To determine the amount of BRM protein in the homozygous plants, immunoblot analysis of nuclear proteins isolated from inflorescences was carried out with two different antibodies against BRM. In contrast to wild type plants, BRM was not detected with these antibodies in SALK_030046 homozygous (Fig. 1D). Despite these results, we cannot exclude the possibility that a very small amount of BRM protein, lacking the first amino acids, is expressed in this mutant. We have designated the BRM mutant allele present in this line as brm-1. Line GABI-854D01 contains a T-DNA in the seventh intron. RT-PCR experiments using different sets of oligonucleotides indicated that a small amount of truncated transcripts that contain exons upstream and downstream of the T-DNA insertion point are formed in the homozygous plants (Fig. 1C). However, no cDNA was amplified by RT-PCR using oligonucleotides 1 and 3 (Fig. 1A), indicating that correct splicing of the seventh intron was not possible in plants carrying the T-DNA insertion. Furthermore, no BRM protein was detected in immunoblotting experiments (Fig. 1D). Consequently, our data indicate that the BRM mutant allele present in this line, which we named brm-2, is a null allele. Characterization of BRM mutants brm-1 and brm-2 homozygous plants presented identical phenotypes and will be described simultaneously. The progeny of self-fertilized heterozygous BRM/brm plants followed a non-Mendelian segregation: 41.1% BRM/BRM, 48.6% BRM/brm-1, and 10.3% brm-1/ brm-1 (n = 107) for the brm-1 allele; 44.8% BRM/ BRM, 48.3% BRM/brm-2 and 6.9% brm-2/brm-2 123 296 Plant Mol Biol (2006) 62:291–304 Table 1 Genotype of the progeny obtained after self-pollination or backcrosses between different brm mutants and wild-type plants Female · Male brm-1/BRM · brm-1/BRM brm-2/BRM · brm-2/BRM brm-1/BRM · BRM/BRM BRM/BRM · brm-1/BRM brm-1/brm-1 · BRM/BRM BRM/BRM · brm-1/brm-1 brm-2/brm-2 · BRM/BRM BRM/BRM · brm-2/brm-2 Genotype of F1 plants BRM/ BRM brm/ BRM brm/ brm n 41.1% 44.8% 65.0% 71.5% –c – – – 48.6% 48.3% 35.0% 28.5% – – – – 10.3% 6.9% 0% 0% – – – – 107a 60a 40a 35a 10b 10b 10b 10b Genotype was determined by PCR using allel-specific primers as described in Material and methods a Number of plants genotyped b Number of crosses analyzed c – No seeds were found (n = 60) for the brm-2 allele (Table 1). The small amount of viable homozygous might initially suggest an incomplete penetrant zygotic embryo lethality. However, the data approached to a 1:1 segregation ratio rather than 2:1 as expected for a zygotic embryo lethal mutation, suggesting that absence of BRM causes reduced gametophytic transmission. Transmission of the brm-1 allele was studied by reciprocal backcrosses of heterozygous plants with WT Columbia plants (Table 1). Male and female transmission efficiencies (as defined by (Howden et al. 1998)) were 39.9% and 53.8%, respectively, indicating incomplete penetrant male and female gametophyte alterations. Analysis of BRM/brm-1 and BRM/brm-2 heterozygous siliques showed that 21.3% (n = 1478) and 37.5% (n = 421), respectively, of the seeds were arrested at an early developmental stage, while only 2.1% (n = 548) were arrested in wild-type siliques (Fig. 2A). The 21.3% aborted seeds observed in the BRM/brm-1 heterozygous fruits approach to the 23% calculated based on transmission data (100/2 – 0.538 · 100/ 2 = 23.1%). Microscopic analysis of cleared developing seeds, 2 days post anthesis evidenced that 100% of the ovules of WT plants contained embryos in early globular stage (Fig. 2B). However, heterozygous BRM/ brm-2 and BRM/brm-1 siliques contained a percentage of ovules with enlarged integument cells and without visible embryos or with embryos arrested at a very early stage (one or two terminal cells stage) (Fig. 2C– E). Therefore, our results suggest that a percentage of the zygotes containing one or two brm mutant alleles were arrested early after fertilization by a gametophytic parental effect not fully penetrant. Taken 123 Fig. 2 Seeds and embryo development in BRM/brm heterozygous plants. (A) Open siliques of self-fertilized WT and BRM/ brm-1 and BRM/brm-2 plants. Arrowheads indicate aborted ovules. (B–E) Developing seeds three days after flowering on WT (B) and BRM/brm-2 (C–E) plants. Arrowheads indicate the position of the embryo. Scale bars, 20 lm together, our data indicate that BRM is required for normal gametophyte development. Besides the segregation defects, heterozygous brm plants did not show other alterations indicating that brm-1, -2 are recessive mutations. brm-1 and brm-2 homozygous plants exhibited slow growth, delayed development and a dramatic reduction of plant size (Figs. 1E, 3A). Rosette and cauline leaves were small and curled downwards as well as coiled in the proximo-distal axes, both under short day (SD) and long day (LD) conditions (Fig. 3B–E). Leaf developmental phases were not altered. Thus, under LD conditions, the first four leaves of brm-1 and brm-2 plants presented features of juvenile leaves such as lack of trichomes in the abaxial surface, and displayed a small size reduction. However, adult leaves exhibited a strong size reduction (Fig. 3B). Similar abnormalities, although less severe, were previously described in BRM silenced plants (Farrona et al. 2004). Under SD conditions filamentouslike structures rolled in spiral were sometime observed in the position of cauline leaves (Fig. 3F). These structures were never observed in the BRM silenced plants. As discussed above, levels of the BRM transcript in the Plant Mol Biol (2006) 62:291–304 297 Fig. 3 Phenotypic characterization of brm mutants. (A) 50-daysold adult WT, brm-2 and brm-1 plants grown in soil under LD conditions. (B) Rosette leaves of WT and brm-2 plants grown under LD conditions, 25 days after sowing. Scale bar: 1 cm. (C) Curled rosette leaf of brm-1 plants grown under LD. Scale bar: 2 mm. (D) Curled rosette leaf of brm-2 plants grown under LD. Scale bar: 2 mm. (E) Curled cauline leaf of brm-2 plants grown under SD. Scale bar: 2 mm. (F) Filamentous structure in the position of cauline leaf of brm-2 plants grown under SD. Scale bar: 2 mm. (G) Roots of WT (2, 4, and 5), heterozygous (3,6,7, and 8) and homozygous (1 and 9) brm-1 seedlings grown on a vertical plate for 10 days roots of previously described BRM silenced plants were not reduced (data not shown). Consistently, root development in these plants was unaffected (Farrona et al. 2004). However, root development was strongly impaired in brm-1 and brm-2 plants (Fig. 3G). The phenotypes observed in the plants homozygous for brm-1 and brm-2 indicate an essential role for BRM in the reproductive phase of development including control of flowering time and flower development. We have shown that BRM silenced plants present a precocious transition from the vegetative to the reproductive phases. Thus, BRM silenced plants flowered earlier and with less leaves than WT Columbia plants (Farrona et al. 2004). In contrast, brm-1 and brm-2 plants flowered with less leaves but later than WT plants both under SD and under LD conditions (Table 2). We also observed that approximately a 20% of the mutant plants never flower in SD. brm-1 and brm-2 primary inflorescences gave rise to a reduced number of flowers (10–20). Primary inflo- rescence meristem stopped growing and flowers remained in an immature state. Then, secondary inflorescences arose and grew above the primary inflorescence but shortly after that, secondary meristems also stopped growing, suggesting that BRM is required for the maintenance of the inflorescence shoot apical meristem. As shown in Fig. 4 and Table 3, brm-1 and brm-2 flowers displayed several developmental abnormalities. brm-1 and brm-2 flowers were smaller than WT flowers and remained closed (Fig. 4A). About 50% of the brm-1 and brm-2 flowers displayed fused sepals and asymmetric position of these organs (Fig. 4D). Petals were shorter than sepals and deformed (Fig. 4B, C). Stamens were short and all flowers presented altered number of these organs (from 2 to 5) (Fig. 4B). About 40% of the flowers displayed fused stamen filaments (Fig. 4E). Anthers development was severely retarded and pollen grains were rarely found (Fig. 4F–H). The gynoecia of all the brm1 and brm-2 flowers were curved or aberrantly shaped (Fig. 4I). Stigmatic papillae and style were not correctly formed. A small number of mutant flowers (about 15%) presented an open gynoecium with ectopic growth of carpelloid tissue (Fig. 4J, K). About 50% of the flowers exhibited defects in second- and/or thirdwhorl organ identity. For instance, petals with patches of sepaloid green tissue, third-whorl organs with carpelloid tissue and ovule-like structures in the edges or staminoid filaments with anthers replaced by stigmatic papillae (Fig. 4L–N). brm-1 and brm-2 homozygotes were both male and female sterile, the gynoecium did not elongate and seeds were not developed, even when homozygous brm-1 and brm-2 flowers were manually Table 2 Flowering time of WT and brm mutants Line LD (n > 15)a no. of leavesb SD(n > 15) no. of daysc no. of leaves no. of days Columbia 10.68 ± 1.42 22.00 ± 1.35 52.67 ± 5.28 64.67 ± 8.29 brm-1 8.08 ± 1.32 27.08 ± 1.80 22.0 ± 4.16 68.5 ± 11.32 brm-2 7.89 ± 1.50 26.43 ± 2.30 n.d. n.d. a n, number of plants analyzed b Number of rosette leaves at bolting ± standard deviation c Number of days from sowing to bolting ± standard deviation n.d. Not determined 123 298 Plant Mol Biol (2006) 62:291–304 Fig. 4 Flower development of brm mutant plants. Plants were grown under long day conditions (16 h light/8 h dark). brm-1 and brm-2 presented identical phenotypes. Representative pictures are shown. (A) Side view of WT, and brm-2 flowers. Scale bar: 1 mm. (B) Short petals and stamens in brm-2 flowers. Scale bar: 1 mm. (C) Short and deformed petals in brm-2 flowers. Scale bar: 0.5 mm. (D) Fused sepals in brm-2 flowers. Scale bar: 100 lm. (E) Fused stamen filaments in brm-1 flowers. Scale bar: 100 lm. (F) DAPI staining of WT anther. Scale bar: 100 lm. (G and H), DAPI staining of brm-1 anthers. Scale bar: 100 lm. (I) Deformed gynoecium in brm-2 (right) flowers. Left, WT flowers gynoecium. Scale bar: 0.5 mm. (J and K) Ectopic growth of internal carpelloid tissue in brm-1 flowers. (J) Scale bar: 1 mm. (K) Scale bar: 100 lm. (L) Petals with a patch of green sepaloid tissue (arrowhead) in brm-1 flowers. Scale bar: 0.5 mm. (M) Second-whorl organ with carpelloid tissue including an ovulelike structure in brm-1 flowers. Scale bar: 100 lm. (N) Staminoid filaments with stigmatic papillae at the tip in brm-2 flowers. Scale bar: 100 lm pollinated with pollen from wild-type plants (Table 1), suggesting that sterility is not only the consequence of the reduced amount of pollen in mutant anthers. The phenotype of the BRM mutants shows important similarities to that previously described of the AtSWI3C mutants (Sarnowski et al. 2005), suggesting that both proteins work in the same genetic pathway or form a molecular complex. Table 3 Phenotype of brm-1 and brm-2 homozygous flowers Variable WT brm-1 (n = 10) (n = 20) brm-2 (n = 20) Flowers with fused sepals Flowers with petals shorter than sepals Number of petals per flower Number of stamens per flower Flowers with fused stamens filaments Flowers with deformed and immature gynoecium Flowers containing petals with sepalloid tissue Flowers containing thirdwhorl organs with carpelloid tissue 0% 0% 55% 100% 50% 95% 4.0 ± 0.0 3.85 ± 0.36 3.75 ± 0.43 6.0 ± 0.0 4.05 ± 0.58 3.85 ± 0.57 0% 45% 40% 0% 100% 100% 0% 30% 30% 0% 25% 35% Plants were grown under LD conditions. Numbers are percentages or means ± standard deviation 123 Expression levels of flower homeotic genes are reduced in brm mutants The finding that flower organ identity is affected in brm-1 and brm-2 mutant plants suggests that, as its Drosophila counterpart, BRM may regulate homeotic gene expression. Furthermore, we have previously shown that BRM is expressed in floral buds (Farrona et al. 2004). In Arabidopsis, flower organ identity is determined by three classes of homeotic genes: the class A genes APETALA1 (AP1) and APETALA2 (AP2), the class B genes PISTILLATA (PI) and APETALA3 (AP3) and the class C gene AGAMOUS (AG), together with the SEPALLATA genes (SEP1, 2 and 3). In order to evaluate whether mutations in BRM cause deregulation of homeotic genes, we have Plant Mol Biol (2006) 62:291–304 299 determined mRNA levels of some of these genes by semi-quantitative RT-PCR experiments using total RNA isolated from inflorescences of WT and homozygous brm-1 mutants. Sarnowski et al (2005), have showed that atswi3c mutants present low levels of AP2, AP3 and PI transcripts. We have also analyzed the expression of these genes in atswi3c mutant plants, grown under our conditions, as control for our experiments. As shown in Fig. 5, AP2, AP3, and PI transcript levels are down-regulated between 2- and 4-fold in brm-1 plants with respect to WT plants. Expression of these genes was also reduced about 2-fold in the atswi3c mutants. In contrast, the level of AG transcript was not altered in mutant plants. The most dramatic floral phenotypes of brm-1 and brm-2 plants affect the second and third whorls. Consistently, RT-PCR experiments showed that expression of the class B homeotic gene PI is strongly altered in the brm-1 plants (about four-fold downregulated). The best characterized immediate target of the AP3/PI heterodimers is the NAP gene (NACLIKE, ACTIVATED BY AP3/PI) (Sablowski and Meyerowitz 1998). Interestingly, both sense and antisense 35S::NAP plants show flowers with small petals and stamens, provoked essentially by a defect in cell elongation, which is strikingly similar to the phenotype of brm flowers. Therefore, we tested whether NAP expression is altered in the brm mutants. RTPCR experiments demonstrated that NAP is strongly down-regulated in brm-1 plants (Fig. 5). Similar results were obtained in atswi3c mutants. These results suggest that BRM is an activator of flower homeotic genes. Molecular interaction between BRM and AtSWI3 proteins Fig. 5 Semi-quantitative RT-PCR analysis of AP2, AP3, PI, AG and NAP transcript levels in WT, brm-1, and atswi3c-3 flowers. Total RNA was isolated from inflorescence apices of plants grown under long day conditions. RT-PCR was performed as indicated in Materials and methods. DNA products were transferred to nylon filters and hybridized with radiolabeled probes for each gene. In order to allow comparison among independent samples, the expression level of a given gene in each cDNA sample, was normalized to the GAPC gene transcript level. The level of cDNA amplified in WT plants was considered 100%. Values are the average of three independent experiments. Bars indicate standard error of the mean The existence of small gene families for several of the putative SWI/SNF subunits rises the question of the specificity of the interactions among the paralogous proteins within the complex. We have previously shown that the N-terminal part of BRM (amino acids 16–952) interacts in the yeast two-hybrid system with AtSWI3C (At1g21700), an Arabidopsis homolog of SWI3 (Farrona et al. 2004). Arabidopsis contains four genes encoding SWI3 homologs named AtSWI3A to AtSWI3D. We decided to investigate whether BRM is able to interact with the other AtSWI3 proteins by yeast two-hybrid. For that, full-length AtSWI3A, AtSWI3B and AtSWI3D proteins were expressed as ‘‘prey’’ fusions with the GAL4 activation domain (GAD). GADAtSWI3C fusion was also included as control. The BRM16–952 fragment fused to the GAL4 DNA binding domain (GDB) was used as ‘‘bait’’. As shown in Fig. 6A, only yeast strains co-expressing the GADAtSWI3B or GAD-AtSWI3C and GBD-BRM16–952 were able to grow in selective medium without histidine, indicating the activation of the GAL1::HIS3 reporter gene. The size of the colonies of the strain expressing GAD-AtSWI3B and GBD-BRM16–952 was smaller than that of yeast expressing GAD-AtSWI3C and GBD-BRM16–952. Consistently, the GAL1::LacZ reporter, assayed as ß-galactosidase activity, was induced six-fold more in cells co-expressing GADAtSWI3C and GBD-BRM16–952 as compared to cells co-expressing GAD-AtSWI3B and GBD-BRM16–952. All fusion proteins were expressed at similar levels as 123 300 Fig. 6 Interaction between BRM and AtSWI3 proteins. For yeast two-hybrid experiments (A–C) GAL4-binding domain (GBD) and GAL4-activation domain (GAD) fusion proteins were co-expressed in the yeast strain MaV203. Activation of the GAL1::HIS3 reporter gene was tested by a nutritional assay. Growth on SC medium -L-T (-Leu, -Trp) selects for markers carried on the bait and prey plasmids, whereas growth on -L-TH + 3AT (-Leu, -Trp, -His plus 50 mM of 3-amino-1,2,4-triazole) also indicates activity of the HIS3 reporter gene. ß-galactosidase activity (GAL1::LacZ reporter) was determined by a liquid quantitative assay. Each value is an average from three independent determinations; the standard error of the mean is also indicated. (A) Interaction of BRM with the four AtSWI3 proteins. (B) Analysis of the region of BRM involved in the interaction with AtSWI3C. A schematic representation of the Nterminal part of BRM is shown. pQ, Glutamine rich region 123 Plant Mol Biol (2006) 62:291–304 (black box); DI, domain I as defined by (Tamkun et al. 1992) (yellow box); DII, domain II as defined by (Tamkun et al. 1992) (blue box). (C) Analysis of the region of AtSWI3C involved in the interaction with BRM. A schematic representation of AtSWI3C is shown. SWIRM, SWIRM domain (blue box); ZF, zinc finger (red box); SANT, SANT domain (pink box); LZ, leucine-zipper (green box); and PQ, proline-glutamine rich region (grey box). (D) In vitro interaction between the Cterminal region of AtSWI3C and a GST fusion protein containing the domain II of BRM. One lg of GST-BRM691–942 or GST proteins bound to glutathione-agarose beads were incubated with in vitro translated 35S-labelled SWI3C385–807. Bound and 20% of unbound proteins were subjected to SDSPAGE. The gels were transferred to Whatman 3 MM paper, dried, and subjected to autoradiography Plant Mol Biol (2006) 62:291–304 verified by protein gel blot using antibodies against the GAD and GDB (data not shown). Therefore, the data indicate that BRM interacts with AtSWI3C and also with AtSWI3B albeit less strongly, but not with AtSWI3A or AtSWI3D. To map the domains involved in the AtSWI3C-BRM interaction, different deletions of AtSWI3C and BRM16–952 were used as prey and bait, respectively, in the same two-hybrid system. Figure 6B shows that domain II (DII) of BRM is necessary and sufficient for the interaction between AtSWI3C and BRM. However, cells expressing the complete BRM Nterminal region (BRM16–952) grew better than those expressing the domain II alone (BRM691–952), indicating that additional points of contact may exist between AtSWI3C and other domains of the N-terminus of BRM. Figure 6C shows that the C-terminal half of AtSWI3C is necessary and sufficient to interact with BRM. Three domains of this region seem to be required for the correct interaction: the SANT (SWI-SNF, ADA, N-CoR, TFIIIB) domain, a leucine-zipper (LZ) and a proline-glutamine (PQ) rich region. Deletion of the LZ and PQ domains completely abolished the interaction. However, single deletion of the LZ, PQ or SANT regions decreased the strength of the interaction although did not suppress the interaction. Immunoblot analysis confirmed that different GAD-AtSWI3C deletions were expressed at comparable levels (data not shown). The existence of more than one point of contact between the yeast RSC complex subunits RSC8/ SWH3 (a SWI3 paralog) and STH1 (a SWI2/SNF2 paralog) has also been suggested (Treich and Carlson 1997). To confirm that BRM and AtSWI3C interact directly we performed in vitro pull-down experiments. Figure 6D shows that a GST-BRM691–942 recombinant protein interacts with a truncated AtSWI3C protein encompassing amino acids 385–807 (AtSWI3C385–807), confirming the yeast two-hybrid data. Therefore, our results indicate an elevated level of specificity in the interactions between the different paralogous subunits, which should result in a functional specificity. Discussion We have previously reported that plants with reduced levels of BRM by RNA interference present a pleiotropic phenotype including defects in leaf morphology, flowering time and flower developments (Farrona et al. 2004). Here we describe the identification and characterization of loss of function mutants of BRM. In contrast to the Drosophila brahma mutant brm plants are viable, probably due to the existence in Arabidopsis of other ATPases of the SNF2 subfamily such as SYD 301 (Wagner and Meyerowitz 2002). The phenotype observed in the brm-1 and brm-2 mutants was more dramatic than the previously described for silenced plants. For example, BRM silenced plants presented small flowers with short petals, and short stamens with immature anthers mostly under SD, whereas brm-1 and brm-2 displayed these abnormalities both under SD and LD conditions. While BRM silenced plants presented a reduced fertility, homozygous brm-1 and brm-2 plants are completely sterile. Partial homeotic transformations of second whorl organs into sepals were also restricted to SD in BRM silenced plants, but generally observed both under SD and LD in the null mutants. In addition, fused sepals, open gynoecium and third-whorl organs with carpelloid features observed in brm-1 and brm-2 mutants were never found in BRM silenced lines. Furthermore, levels of homeotic genes mRNA was not altered in the BRM knockdown plants, however, expression of AP2, AP3 and PI was downregulated in the null mutants (see below). These data suggest that the RNAi silenced plants synthesize a small amount of BRM, and therefore they behave as hypomorphic mutants. Two important phenotypic differences were observed between silenced and null mutant plants. First, while root development was completely normal in BRM silenced plants, brm mutants showed a dramatic reduction in root growth. As previously commented this is consistent with the absence of BRM silencing in the roots of the transgenic plants. The reason for this tissue-specific silencing defect is unknown. Second, brm mutant plants flowered later than the BRM silenced lines but with a similar number of leaves, confirming the strong growth retardation observed in the null mutants. The different root-to-shoot ratio between the silenced strains and the null mutants may be responsible of this strong difference in time to flowering. Gametophyte phenotypes were not investigated in the BRM silenced plants. Now we show that brm mutations affect female and male gametophyte development and function. Consistent with these results, Affymetrix GeneChip analysis (Genevestigator microarray database) revealed high BRM expression in ovules and uninucleate microspores but not in mature pollen (Honys and Twell 2004). Interestingly, the BRM paralogous protein SYD is also strongly expressed in ovules and during microgametogenesis. A partial overlapping of functions of both ATPases in these tissues may explain the partial penetrance of the gametophyte defects in brm mutants. High expression of SWI/SNF ATPases in unfertilized egg cells has been observed in other organisms. Thus, brahma is strongly expressed in Drosophila egg cells (Tamkun et al. 1992) and mBRM and BRG1, the two ATPases of the 123 302 mammalian SWI/SNF complexes, are also strongly expressed in mouse oocytes (LeGouy et al. 1998). Composition of putative Arabidopsis SWI/SNF complexes associated to BRM The existence of a small gene family of SWI3-type genes in Arabidopsis raises the question of which of the AtSWI3 proteins interact with BRM in the putative SWI/SNF complex, and how is this specificity, if any, controlled. One possibility is that all AtSWI3 proteins display similar affinity for BRM but specificity is controlled by gene expression. This implies that interaction only occurs between subunits that are simultaneously expressed in the same cell type. All AtSWI3 proteins seem to be expressed in the same organs as BRM (Zhou et al. 2003). However, a detailed temporal and spatial analysis of the expression of AtSWI3 genes has not been reported. Another obvious possibility is that BRM displays different affinities for the different AtSWI3 proteins. Our yeast two-hybrid data support this second possibility since BRM interacts more efficiently with AtSWI3C than with AtSWI3B, and does not interact with AtSWI3A and AtSWI3D. Studies from yeast demonstrate the existence of two SWI3 proteins per complex (Smith et al. 2003). In human SWI/SNF complexes, BAF and PBAF, two different SWI3 homologs (BAF170 and BAF155) are present (Wang et al. 1996a). Yeast two-hybrid experiments also demonstrate that Arabidopsis AtSWI3 proteins form homo- or heterodimers (Sarnowski et al. 2002, 2005). These results indicate that six dimer combinations are possible (AtSWI3A: AtSWI3A; B:B; A:B; B:D; A:C; B:C). In addition, BSH, another subunit of the core SWI/SNF complex (SNF5 homolog), interacts with AtSWI3A and AtSWI3B but not with AtSWI3C and AtSWI3D ((Sarnowski et al. 2005) and our unpublished observation), further supporting that either AtSWI3A or AtSWI3B has to be present in all complexes. It is unknown whether one or both SWI3 homologs of the dimer interact with the ATPase. Thus, according to our results and those reported by Sarnowski et al. BRM might form a complex with the following AtSWI3 dimers: A:C and B:C and less efficiently with B:B, A:B and B:D. brm-1 and brm-2 mutants display abnormalities similar to those of atswi3c mutants during vegetative and reproductive developmental phases (our data and (Sarnowski et al. 2005)), supporting the presence of BRM in complexes containing AtSWI3C. In addition, these phenotypes are different from those observed in other atswi3 or syd mutants, indicating a high degree of functional speci- 123 Plant Mol Biol (2006) 62:291–304 ficity for the putative complex containing BRM and AtSWI3C. However, certain differences exist between the brm and the atswi3c mutants phenotype. While brm-1 and brm-2 plants show a not fully penetrant gametophytic defect, atswi3c heterozygous plants display a normal Mendelian segregation, suggesting no deficiencies in gametophyte or embryo development. atswi3a and atswi3b mutants are lethal during embryogenesis, and atswi3b mutations result in lethality of about half of both macro and microspores (Sarnowski et al. 2005). These data are consistent with a role of BRM during gametophyte development in complexes containing AtSWI3B:AtSWI3B, or AtSWI3A:AtSWI3B dimers. Our data indicate that the BRM paralog protein SYD interacts with AtSWI3A and AtSWI3B (R. March, F.J. Florencio and J.C.R., unpublished observation), which may contribute to the incomplete penetrance of the gametophyte phenotype in brm mutants. We speculate that BRM and SYD have overlapping and specific functions. Functions of the putative BRM-AtSWI3C complex would be specific. However, BRM and SYD may share functions when they are associated to AtSWI3B:AtSWI3B, or AtSWI3A:AtSWI3B containing complexes. In summary, we are starting to glimpse the existence of a large number of SWI/SNF-like complexes in Arabidopsis with different combinations among four AtSWI3, two SNF2-like proteins (BRM and SYD) and probably other components, with diverse functions. Another two SNF2 subfamily proteins can be found in the Arabidopsis genome, CHR12 and CHR23. However, our preliminary results indicate that these proteins do not interact with AtSWI3 proteins in yeast two-hybrid assays (R. March, F.J. Florencio and J.C.R., unpublished observation). The putative Arabidopsis SWI/SNF complex controls homeotic gene expression In the Drosophila embryo the pattern of expression of homeotic genes is established by transiently expressed regulators and maintained by the trithorax group (trxG) and Polycomb group (PcG) proteins (Ringrose and Paro 2004). PcG products maintain memory of the repressed state. The trxG was defined as the group of genes able to suppress the phenotype of mutations in PcG genes. Among the trxG genes, three different genes encoding SWI/SNF subunits were originally identified: brahma (SNF2 type), moira (SWI3 type), and osa (SWI1 type) (Kennison and Tamkun 1988). Another member of the trxG is trithorax, the gene that gives name to the group. Trithorax encodes a histone H3 methyltransferase in lysine 4. Importantly, Arabidopsis Plant Mol Biol (2006) 62:291–304 polycomb-like proteins have been identified as repressors of homeotic genes (reviewed in (Goodrich and Tweedie 2002; Reyes and Grossniklaus 2003)). Here we show that BRM activates the expression of a subset of Arabidopsis homeotic genes, also activated by AtSWI3C (Sarnowski et al. 2005), suggesting that one of the plant SWI/SNF complexes activates flower homeotic genes. Interestingly AP1, AP2 and PI are not deregulated in syd mutants (Wagner and Meyerowitz 2002) further demonstrating specificity of functions between BRM and SYD. In further analogy with the Drosophila system, ATX-1, an Arabidopsis homolog of Trithorax, also activates homeotic genes. Thus, lack of ATX-1 causes AP1, AP2, and PI, but not AP3 or SEP3 downregulation. In fact, atx1-1 homozygotes display similar floral phenotypes as do brm plants, including closed flowers, short petals and stamens, staminoid petals or third-whorl organs with stigmatic papillae (AlvarezVenegas et al. 2003). These results suggest that ATX-1 and BRM might act in the same genetic pathways. Genetic and physical interactions have been demonstrated between the Drosophila and human Trithorax protein and subunits of the SWI/SNF complex (Dingwall et al. 1995; Rozenblatt-Rosen et al. 1998). Moreover, both Trithorax and Brahma proteins are recruited to the Fab-7 regulatory region of the Drosophila homeotic gene Abdominal-B (Dejardin and Cavalli 2004). These results suggest an overall conservation of the Polycomb-Trithorax system for homeotic genes control between animals and plants. How the interplay between these different machineries works and how they are recruited to plant homeotic gene regulatory regions will be of great interest in the future. Acknowledgments We thank the Arabidopsis Biological Resource Center (ABRC), Kazusa DNA Research Institute, National Center for Plant Genomic Resources (INRA-CNRGV), The European Arabidopsis Stock Centre (NASC, Nottingham University, UK) and Bernd Weisshaar from the MPI for Plant Breeding Research (Cologne, Germany) for providing cDNA clones or T-DNA insertion mutants used in this study. We thank Frederic Berger for his comments and advice about the analysis of the gametophyte phenotype. We thank José L. Crespo, Marika Lindahl, Javier Florencio and Rosana March for critical reading of the manuscript. L.H. and S.F. are recipients of fellowships from the Spanish Ministerio de Ciencia y Tecnologı́a. This work was supported by Ministerio de Ciencia y Tecnologı́a (grant BMC2002-03198 and BFU2005-01047) and by Junta de Andalucı́a (group CV1-284). References Alonso JM, Stepanova AN, Leisse TJ, Kim CJ, Chen H, Shinn P, Stevenson DK, Zimmerman J, Barajas P, Cheuk R, Gadrinab C, Heller C, Jeske A, Koesema E, Meyers CC, Parker 303 H, Prednis L, Ansari Y, Choy N, Deen H, Geralt M, Hazari N, Hom E, Karnes M, Mulholland C, Ndubaku R, Schmidt I, Guzman P, Aguilar-Henonin L, Schmid M, Weigel D, Carter DE, Marchand T, Risseeuw E, Brogden D, Zeko A, Crosby WL, Berry CC, Ecker JR (2003) Genome-wide insertional mutagenesis of Arabidopsis thaliana. Science 301:653–657 Alvarez-Venegas R, Pien S, Sadder M, Witmer X, Grossniklaus U, Avramova Z (2003) ATX-1, an Arabidopsis homolog of trithorax, activates flower homeotic genes. Curr Biol 13:627– 637 Cairns BR (2005) Chromatin remodeling complexes: strength in diversity, precision through specialization. Curr Opin Genet Dev 15:185–190 Cairns BR, Kim YJ, Sayre MH, Laurent BC, Kornberg RD (1994) A multisubunit complex containing the SWI1/ADR6, SWI2/SNF2, SWI3, SNF5, and SNF6 gene products isolated from yeast. Proc Natl Acad Sci USA 91:1950–1954 Cote J, Quinn J, Workman JL, Peterson CL (1994) Stimulation of GAL4 derivative binding to nucleosomal DNA by the yeast SWI/SNF complex. Science 265:53–60 Dejardin J, Cavalli G (2004) Chromatin inheritance upon Zestemediated Brahma recruitment at a minimal cellular memory module. Embo J 23:857–868 Dingwall AK, Beek SJ, McCallum CM, Tamkun JW, Kalpana GV, Goff SP, Scott MP (1995) The Drosophila snr1 and brm proteins are related to yeast SWI/SNF proteins and are components of a large protein complex. Mol Biol Cell 6:777–791 Farrona S, Hurtado L, Bowman JL, Reyes JC (2004) The Arabidopsis thaliana SNF2 homolog AtBRM controls shoot development and flowering. Development 131:4965–4975 Flaus A, Owen-Hughes T (2004) Mechanisms for ATP-dependent chromatin remodelling: farewell to the tuna-can octamer? Curr Opin Genet Dev 14:165–173 Goodrich J, Tweedie S (2002) Remembrance of things past: chromatin remodeling in plant development. Annu Rev Cell Dev Biol 18:707–746 Guidet F, Rogowsky R, Taylor C, Song W, Langridge P (1991) Cloning and characterization of a new rye-specific repeat sequence. Genome 34:81–87 Hoffman GA, Garrison TR, Dohlman HG (2002) Analysis of RGS proteins in Saccharomyces cerevisiae. Methods Enzymol 344:617–631 Honys D, Twell D (2004) Transcriptome analysis of haploid male gametophyte development in Arabidopsis. Genome Biol 5:R85 Howden R, Park SK, Moore JM, Orme J, Grossniklaus U, Twell D (1998) Selection of T-DNA-tagged male and female gametophytic mutants by segregation distortion in Arabidopsis. Genetics 149:621–631 Kennison JA, Tamkun JW (1988) Dosage-dependent modifiers of polycomb and antennapedia mutations in Drosophila. Proc Natl Acad Sci USA 85:8136–8140 Kohler C, Hennig L, Bouveret R, Gheyselinck J, Grossniklaus U, Gruissem W (2003) Arabidopsis MSI1 is a component of the MEA/FIE Polycomb group complex and required for seed development. Embo J 22:4804–4814 Kwon CS, Chen C, Wagner D (2005) WUSCHEL is a primary target for transcriptional regulation by SPLAYED in dynamic control of stem cell fate in Arabidopsis. Genes Dev 19:992–1003 LeGouy E, Thompson EM, Muchardt C, Renard JP (1998) Differential preimplantation regulation of two mouse homologues of the yeast SWI2 protein. Dev Dyn 212:38–48 123 304 Mohrmann L, Langenberg K, Krijgsveld J, Kal AJ, Heck AJ, Verrijzer CP (2004) Differential targeting of two distinct SWI/SNF-related Drosophila chromatin-remodeling complexes. Mol Cell Biol 24:3077–3088 Neigeborn L, Carlson M (1984) Genes affecting the regulation of SUC2 gene expression by glucose repression in Saccharomyces cerevisiae. Genetics 108:845–858 Ohkawa Y, Marfella CG, Imbalzano AN (2006) Skeletal muscle specification by myogenin and Mef2D via the SWI/SNF ATPase Brg1. Embo J Pedersen TA, Kowenz-Leutz E, Leutz A, Nerlov C (2001) Cooperation between C/EBPalpha TBP/TFIIB and SWI/ SNF recruiting domains is required for adipocyte differentiation. Genes Dev 15:3208–3216 Peterson CL, Dingwall A, Scott MP (1994) Five SWI/SNF gene products are components of a large multisubunit complex required for transcriptional enhancement. Proc Natl Acad Sci USA 91:2905–2908 Reyes JC, Grossniklaus U (2003) Diverse functions of Polycomb group proteins during plant development. Semin Cell Dev Biol 14:77–84 Ringrose L, Paro R (2004) Epigenetic regulation of cellular memory by the Polycomb and Trithorax group proteins. Annu Rev Genet 38:413–443 Rosso MG, Li Y, Strizhov N, Reiss B, Dekker K, Weisshaar B (2003) An Arabidopsis thaliana T-DNA mutagenized population (GABI-Kat) for flanking sequence tag-based reverse genetics. Plant Mol Biol 53:247–259 Rozenblatt-Rosen O, Rozovskaia T, Burakov D, Sedkov Y, Tillib S, Blechman J, Nakamura T, Croce CM, Mazo A, Canaani E (1998) The C-terminal SET domains of ALL-1 and TRITHORAX interact with the INI1 and SNR1 proteins, components of the SWI/SNF complex. Proc Natl Acad Sci USA 95:4152–4157 Sablowski RW, Meyerowitz EM (1998) A homolog of NO APICAL MERISTEM is an immediate target of the floral homeotic genes APETALA3/PISTILLATA. Cell 92:93–103 Sarnowski TJ, Swiezewski S, Pawlikowska K, Kaczanowski S, Jerzmanowski A (2002) AtSWI3B, an Arabidopsis homolog of SWI3, a core subunit of yeast Swi/Snf chromatin 123 View publication stats Plant Mol Biol (2006) 62:291–304 remodeling complex, interacts with FCA, a regulator of flowering time. Nucleic Acids Res 30:3412–3421 Sarnowski TJ, Rios G, Jasik J, Swiezewski S, Kaczanowski S, Li Y, Kwiatkowska A, Pawlikowska K, Kozbial M, Kozbial P, Koncz C, Jerzmanowski A (2005) SWI3 subunits of putative SWI/SNF chromatin-remodeling complexes play distinct roles during Arabidopsis development. Plant Cell 17:2454–2472 Smith CL, Peterson CL (2005) ATP-dependent chromatin remodeling. Curr Top Dev Biol 65:115–148 Smith CL, Horowitz-Scherer R, Flanagan JF, Woodcock CL, Peterson CL (2003) Structural analysis of the yeast SWI/ SNF chromatin remodeling complex. Nat Struct Biol 10:141–145 Stern M, Jensen R, Herskowitz I (1984) Five SWI genes are required for expression of the HO gene in yeast. J Mol Biol 178:853–868 Sudarsanam P, Iyer VR, Brown PO, Winston F (2000) Wholegenome expression analysis of snf/swi mutants of Saccharomyces cerevisiae. Proc Natl Acad Sci USA 97:3364–3369 Tamkun JW, Deuring R, Scott MP, Kissinger M, Pattatucci AM, Kaufman TC, Kennison JA (1992) brahma: a regulator of Drosophila homeotic genes structurally related to the yeast transcriptional activator SNF2/SWI2. Cell 68:561–572 Treich I, Carlson M (1997) Interaction of a Swi3 homolog with Sth1 provides evidence for a Swi/Snf- related complex with an essential function in Saccharomyces cerevisiae. Mol Cell Biol 17:1768–1775 Wagner D, Meyerowitz EM (2002) SPLAYED, a Novel SWI/ SNF ATPase Homolog, Controls Reproductive Development in Arabidopsis. Curr Biol 12:85–94 Wang W, Xue Y, Zhou S, Kuo A, Cairns BR, Crabtree GR. 1996a. Diversity and specialization of mammalian SWI/SNF complexes. Genes Dev 10:2117–2130 Wang W, Cote J, Xue Y, Zhou S, Khavari PA, Biggar SR, Muchardt C, Kalpana GV, Goff SP, Yaniv M, Workman JL, Crabtree GR. 1996b. Purification and biochemical heterogeneity of the mammalian SWI-SNF complex. EMBO J 15:5370–5382 Zhou C, Miki B, Wu K (2003) CHB2, a member of the SWI3 gene family, is a global regulator in Arabidopsis. Plant Mol Biol 52:1125–1134