ABAP1 is a novel plant Armadillo BTB protein involved in DNA replication and transcription

In an attempt to identify protein complexes formed with Arabidopsis pre‐RC members, a search for pre‐RC‐interacting proteins was performed. Because AtORC1a harbours two domains involved in transcription regulation—BAH and PHD domains (Masuda et al, 2004), its complete open‐reading frame was assayed in a two‐hybrid screen against an Arabidopsis cDNA two‐hybrid library. Among the interacting proteins, a novel 737‐amino‐acid protein (At5g13060) of approximately 81 kDa was identified. It harbours eight predicted repeats of the β‐catenin‐type Armadillo domain (ARM repeats) in its N‐terminal region (amino acids 111–162, 164–212, 213–254, 256–296, 298–338, 340–380, 381–421 and 496–536) as well as a BTB/POZ domain (acronym for Broad complex/Tram‐Track/Bric‐a‐brac/Poxyvirus and zinc‐finger), located in the C terminus (amino acids 568–665) (Figure 1A). Therefore, the gene was designated ABAP1. Yeast two‐hybrid assays revealed that ABAP1 interaction with AtORC1a was mediated by the ARM repeats and by AtORC1a N terminus (amino acids 1–341) that contained the BAH and PHD domains (Supplementary Figure 1A). β‐Catenin ARM repeats are approximately 40‐amino‐acid‐long tandemly repeated sequence motifs and occur in various families of eukaryotic proteins involved in different cellular processes, such as cell–cell adhesion, nuclear import, cell signalling and transcriptional regulation (Coates, 2003; Hatzfeld, 2005). The BTB/POZ domain is usually found in leucine zipper proteins involved in transcriptional repression with a possible role as E3 ligase (Perez‐Torrado et al, 2006).

In agreement with the presence of a putative nuclear localization signal in ABAP1 N terminus (Figure 1A), confocal and fluorescent microscopy of Arabidopsis plants producing GFP::ABAP1 indicated that ABAP1 was exclusively located in the nucleus, homogeneously distributed or enriched in nuclear domains, and GFP::ABAP1 was absent in mitotic cells (Figure 1B). Analyses of endogenous levels of ABAP1 in synchronized Arabidopsis cell cultures indicated that it accumulated during G1 and/or early S phase (Supplementary Figure 2B–E). This result was corroborated by the studies on BY2 tobacco cells producing GFP::ABAP1 after treatment with inhibitors of cell cycle progression (Supplementary Figure 2A and Supplementary Video 1).

A genomic search revealed the existence of at least 108 ARM domain and 80 BTB domain proteins in the Arabidopsis genome (Mudgil et al, 2004; Gingerich et al, 2005). ABAP1 has the unique feature to harbour both Armadillo and BTB domains in its primary sequence. Arabidopsis has an ABAP1 homologue (At5g19330), characterized earlier as ARIA (ARM repeat protein interacting with ABF2) that shares 59% identity of the deduced amino‐acid sequence with ABAP1 and is involved in abscisic acid response through interaction and regulation of the transcription factor ABF2 (Kim et al, 2004). Remarkably, no ABAP1 homologue was found in eukaryotes other than plants.

In a yeast two‐hybrid assay against all other pre‐RC proteins, except MCM2–7, ABAP1 associated directly with AtORC1b, AtCDT1a and AtCDT1b (Supplementary Figure 1B). The interactions between ABAP1 and the pre‐RC components were confirmed in GST‐pulldown experiments with ABAP1–GST and radiolabelled AtORC1a^35S, AtORC1b^35S, AtCDT1a^35S and AtCDT1b^35S. The pre‐RC components bound to ABAP1 but not to GST alone (Figure 1C and D). The specificity of the interactions was demonstrated in an assay where ABAP1–GST did not bind to radiolabelled AtORC4^35S and bound to AtORC1a N terminus^35S (Figure 1E). The formation of a triple complex between AtORC1a, ABAP1 and AtCDT1b was shown in a GST‐pulldown experiment with AtORC1a–GST, ABAP1^35S and AtCDT1b^35S (Figure 1F). As AtCDT1b and AtORC1a did not bind directly, the presence of two radioactive bands indicated that ABAP1 can associate with pre‐RC subcomplexes or eventually with the full complex.

The association between ABAP1 and pre‐RC was further confirmed in vivo by immunoprecipitations with the polyclonal anti‐ABAP1 antibody (Supplementary Figure 3) and protein extracts of Arabidopsis cell suspension LMM‐1 protoplasts that produced ABAP1, AtORC1a–GST, AtORC3–HA and AtCDT1a–FLAG (Figure 1G). In two‐hybrid assays with pre‐RC proteins, AtORC3 did not interact either with ABAP1 (Supplementary Figure 1B) or with the AtCDT1 homologues, but with all AtORCs, except AtORC1a and itself, possibly having a central function in maintaining ORC associations (Masuda et al, 2004). Similar interaction profile of AtORC3 with the other ORC proteins was reported by GST‐pulldown experiments (Diaz‐Trivino et al, 2005). In immunoprecipitation experiments with anti‐ABAP1, ABAP1 interacted in vivo with AtORC1a–GST, AtCDT1a–FLAG and also with AtORC3–HA (Figure 1G), suggesting that ABAP1 might associate with pre‐RC subcomplexes or eventually with the full complex in plant cells. Possibly, ABAP1 can form different complexes with different affinities to anti‐ABAP1, as the immunoprecipitations depleted the interacting AtORC3 and AtCDT1a, but not AtORC1a and ABAP1 itself.

Studies on localization of ABAP1 expression in plant tissues showed that ABAP1 was weakly expressed in the emerging lateral roots (Figure 2A; Supplementary Figure 2B), being mainly expressed in the shoot apex, young leaves (Figure 2B; Supplementary Figure 2C) and flower buds (Supplementary Figure 4A and D), suggesting a possible function in these organs. Strong GUS expression was detected in developing leaves of 7‐ and 9‐day‐old seedlings, as observed for other pre‐RC components, such as PROLIFERA (AtMCM7), AtCDC6, AtCDT1 homologues and AtORC1b (Springer et al, 2000; Castellano et al, 2001; Castellano et al, 2004; Diaz‐Trivino et al, 2005). As the leaf grew, GUS expression decreased in a tip‐to‐base gradient (Figure 2C and D). In agreement, ABAP1 mRNA levels were higher in developing leaves during proliferation stage (9‐day‐old plants) and rapidly declined as observed in leaves of 13 till 21‐day‐old plants (Figure 2E). GUS expression in the meristemoid and young stomatal cells illustrates ABAP1 expression during stomatal cell differentiation (Supplementary Figure 4E).

To assess the function of ABAP1 during development, plants with increased or reduced expression levels of ABAP1 were characterized. Seventeen lines of Arabidopsis plants expressing higher levels of ABAP1 mRNA and protein under the control of the cauliflower mosaic virus (CaMV) 35S constitutive promoter were generated (here denoted as ABAP1^OE). The increase in mRNA and protein levels varied among the different overexpressing lines and the developmental stage, ranging from 5‐ to 17‐fold and 2‐ to 5‐fold, respectively (Supplementary Figure 5A and B). A heterozygous enhancer trap line (ET13614, here denoted as ABAP1^ET) with interference T‐DNA inserted into the first exon of ABAP1 gene (10 base pairs after the initial ATG) was analysed (Supplementary Figure 6). The plants showed an average of two‐ and five‐fold reduction in ABAP1 mRNA and protein levels, respectively (Supplementary Figure 6A and B). Homozygous plants could not be rescued in ABAP1^ET, suggesting that the absence of ABAP1 causes a lethal phenotype (Supplementary Figure 6C).

Comparative phenotype analyses between ABAP1 overexpressor lines (ABAP1^OE) and control lines showed that plants with higher levels of ABAP1 developed normally, except that rosette area and leaf growth were moderately diminished during development (Figure 2F–H). Quite the opposite, reduction of ABAP1 levels in ABAP1^ET plants caused an increase in the rosette and leaf growth (Figure 2I–K). Kinematics studies of developing leaves showed that cell division rates were higher in ABAP1^ET plants and lower in ABAP1^OE plants during early leaf development when compared with wild‐type controls (Figure 3A and C). Though leaf cell organization (data not shown) and cell sizes were similar to those of control plants, cell numbers were significantly reduced in ABAP1^OE and increased in ABAP1^ET plants (Figure 3E). Our data suggested that ABAP1 negatively regulated overall cell divisions in leaves, differing from the pre‐RC subunits AtCDT1 and AtCDC6 that could specifically affect stomata formation (Castellano et al, 2004). An effect of ABAP1 levels in early leaf development is consistent with its high expression levels in shoot meristem and young leaves.

As ABAP1 interacted with members of the pre‐RC and accumulated at G1 and early S phase of the cell cycle, the cell proliferation phenotypes observed in ABAP1^OE and ABAP1^ET leaves could be at least partly triggered by a misregulation of the DNA replication machinery. Cell proliferation assays with [³H]thymidine in ABAP1^OE and ABAP1^ET seedlings showed that the DNA replication levels were significantly lower in ABAP1^OE plants than those in the wild‐type plants (Figure 3B). In contrast, ABAP1^ET seedlings showed higher levels of DNA replication than wild type (Figure 3D). A direct role of ABAP1 in pre‐RC assembly and association with chromatin was investigated by measuring MCM loading onto chromatin, the hallmark step in the formation of the pre‐RC. Cells were separated into cytoplasmic, nucleus‐soluble and chromatin‐enriched fractions and levels of MCM7 (PROLIFERA) and ABAP1 were assayed by immunoblot analysis (Figure 3F). PROLIFERA association with chromatin was lower in 6‐day‐old ABAP1^OE plants than in wild‐type plants, and the excess of ABAP1 was mostly found as a soluble nuclear protein. When ABAP1 levels were reduced in ABAP1^ET, higher levels of PROLIFERA were found loaded onto chromatin. Ploidy levels in developing leaves of ABAP1^OE plants were similar as control, suggesting that ABAP1 might have a function in mitotic replication but not in endocycles (Supplementary Figure 7). Altogether, the data on the characterization of plants with modified levels of ABAP1 indicated that ABAP1 exert a negative role in cell proliferation in leaves, possibly by inhibiting mitotic DNA replication.

To unravel protein complexes in which ABAP1 takes part, a yeast two‐hybrid screen under high stringency selection conditions was performed with an Arabidopsis cDNA library as a bait. Several transcription factors, belonging to the NAC, AP2 and TCP families (unpublished results) were identified. One of the ABAP1‐associated transcription factors was AtTCP24 (At1g30210) that belongs to the class‐II TCP transcription factor family (Cubas et al, 1999), the members of which negatively regulate plant cell proliferation and leaf morphogenesis (Nath et al, 2003; Palatnik et al, 2003).

To address a possible cooperation of ABAP1 and AtTCP24 in the control of gene transcription, the association between both proteins was characterized. The AtTCP24–ABAP1 interaction was confirmed in a GST‐pulldown assay with AtTCP24–GST and radiolabelled ABAP1^35S. ABAP1 bound to AtTCP24–GST but not to GST alone (Figure 4A). In vivo interaction was observed in immunoprecipitation experiments of protein extracts of Arabidopsis cell suspension LMM‐1 protoplasts that produced ABAP1 and AtTCP24–FLAG, with the anti‐ABAP1 antibody (Figure 4B). The ability of the ABAP1–AtTCP24 heterodimer to recognize the consensus sequence for class‐II TCP (TGGGCC/T) (Trémousaygue et al, 2003) was confirmed in electrophoretic mobility shift assays (EMSAs) (Supplementary Figure 8). The data demonstrated that the association of ABAP1 with AtTCP24 did not affect the DNA‐binding properties of the latter and that ABAP1 could participate in mechanisms of transcriptional regulation. Other class‐II TCP members (AtTCP3, AtTCP5, AtTCP13 and AtTCP17) did not interact with ABAP1 in yeast two‐hybrid assays (Supplementary Figure 9A), indicating some specificity in the ABAP1 association with class‐II TCP family.

To investigate a possible role of ABAP1 in the control of gene expression, putative target genes regulated by ABAP1–AtTCP24 complex were examined by searching for class‐II TCP‐interacting motifs in the promoters of Arabidopsis genes. Surprisingly, four boxes for class‐II TCP were found in the 500 bp upstream of the initial methionine of the AtCDT1a promoter region, making it a good candidate for AtTCP24‐mediated expression. One class‐II TCP box was also found in the AtCDT1b promoter. In EMSA performed with the TCP consensus motif of AtCDT1a (ATGGGCCT, −215 to −223) and its homologue AtCDT1b (ATGGGCCC, position −129 to −137), both AtTCP24 and AtTCP24–ABAP1 were associated with the wild type (Figure 4C, lanes 1, 5, 10 and 14) but not with the mutated probes (Figure 4C, lanes 4, 9, 13 and 18). ABAP1 alone was associated with neither of the probes (Figure 4C, lanes 6 and 15). The presence of AtTCP24–GST and ABAP1 in the protein–DNA complex was confirmed by a supershift assay with anti‐ABAP1 and anti‐GST antibodies (Figure 4C, lanes 3, 7, 8, 12, 16 and 17). The data indicate that both AtCDT1a and AtCDT1b expression might be regulated by ABAP1–AtTCP24 complex. To test this hypothesis, we first confirmed that ABAP1 associates with AtCDT1 promoters in vivo in chromatin immunoprecipitation (ChIP) experiments with anti‐ABAP1 antibody (Figure 4F). The results revealed that ABAP1 was associated with regions of the AtCDT1a and AtCDT1b promoters harbouring the TCP consensus motif. This interaction was site‐specific as ABAP1 was not detected in a 350‐bp region inside the AtCDT1a‐coding region. In addition, AtCDT1a and AtCDT1b mRNA levels were analysed in 6‐day‐old ABAP1^OE and ABAP1^ET versus wild‐type seedlings by real‐time PCR (Figure 4D and E). Both genes were downregulated in plants with higher levels of ABAP1 mRNA and protein (Figure 4D) and they were upregulated in plants where ABAP1 mRNA and protein levels were diminished (Figure 4E). These data suggest a direct effect of ABAP1 on AtCDT1a and AtCDT1b expression.

These results indicate that ABAP1 might exert a function together with transcription factors to regulate cell division in plants. Various class‐II TCP genes have been shown to negatively regulate cell proliferation in leaves (Nath et al, 2003; Palatnik et al, 2003); nevertheless, their mechanism of action is still not understood. Because cell division rates were lower in ABAP1^OE plants and higher in ABAP1^ET plants, one hypothesis is that, during leaf development, ABAP1 might collaborate with AtTCP24 in mediating repression of genes that promote cell proliferation, such as AtCDT1. If so, downregulation of AtCDT1 could be one of the pathways used by class‐II TCP genes to negatively regulate cell division. To test this hypothesis, plants overexpressing AtTCP24 (AtTCP24^OE) were generated (LM Cabral, unpublished data). Remarkably, these plants exhibited a phenotype similar to that of ABAP1^OE plants showing reduced rosette areas and smaller leaves with normal cell sizes but fewer cells (Supplementary Figure 9B–D), pointing to an inhibition of cell proliferation in AtTCP24^OE leaves. Gene expression analyses of 9‐day‐old AtTCP24^OE seedlings with real‐time RT–PCR showed that AtCDT1a and AtCDT1b levels were downregulated in AtTCP24^OE plants (Supplementary Figure 9E). These data are consistent with ABAP1 and AtTCP24 functioning together during leaf development, to negatively regulate AtCDT1a and AtCDT1b expression levels.

To test whether ABAP1 could have a dual function in mechanisms controlling DNA replication and transcription, or whether it executes separate functions in these complexes, we addressed if the ABAP1–AtTCP24 complex could associate with pre‐RC members in plants. The formation of a triple complex between AtORC1a, ABAP1 and AtTCP24 was shown in vitro in a GST‐pulldown experiment with AtORC1a–GST, ABAP1^35S and AtTCP24^35S (Figure 5A). A triple complex between ABAP1, AtTCP24 and AtCDT1b was also observed in a GST‐pulldown experiment with AtTCP24–GST, ABAP1^35S and AtCDT1b^35S (Figure 5B). As AtTCP24 did not bind directly to AtORC1a or AtCDT1b (Figure 5A, lane 5 and Figure 5B, lane 1), the presence of two radioactive bands indicated that ABAP1 might work as a bridge connecting pre‐RC proteins and AtTCP24 in a complex. The ability of ABAP1 to bind both AtTCP24 and AtORC1a concomitantly was further tested in a GST‐pulldown competition assay, where increasing amounts of AtORC1a–GST were mixed with ABAP1–GST and AtTCP24^35S (Figure 5C). There were no differences in binding intensity between ABAP1 and AtTCP24 in the absence or presence of either 2.5, 10 and 50 μg of AtORC1a–GST, suggesting that there was no competition between AtORC1a and AtTCP24 for ABAP1‐binding sites. The association of ABAP1 with AtTCP24 and pre‐RC proteins in the same complex in vivo was revealed in a GST‐pulldown experiment with protein extracts from protoplasts transformed with AtORC1a–GST, AtTCP24‐CBP, AtCDT1a–FLAG and AtORC3–HA (Figure 5D, lane 4). The specificity of the interactions was demonstrated in an assay where ABAP1, AtORC1a–GST, AtTCP24–CDB, AtCDT1a–FLAG and AtORC3–HA were not rescued when an excess of free GST was added prior to column purification (Figure 5D, lane 2). In addition, an EMSA supershift assay revealed that the triple complex formed by AtTCP24, ABAP1 and AtORC1a was still able to recognize and bind to the TCP consensus motif of AtCDT1a and AtCDT1b promoters (Figure 5E). Neither AtORC1a alone nor AtORC1a–ABAP1 was associated with AtCDT1a and AtCDT1b probes (data not shown). The triple complex formed by AtORC1a, AtTCP24 and ABAP1 interacted with the wild type but not with the mutated probes of AtCDT1a and AtCDT1b promoters, as the migration of the probes in the gel was delayed, indicating the presence of the triple complex interacting with the DNA probes (Figure 5E, lanes 1 and 7). The presence of ABAP1 in the complex was confirmed by an extra shift in the gel when anti‐ABAP1 antibody was added (Figure 5E, lanes 5 and 11).

Pre‐RC assembly, an important event that regulates G1‐to‐S transition in eukaryotes (Blow and Dutta, 2005), has led many groups to study which elements define a replication origin and how they are activated during development (Costa and Blow, 2007). Nevertheless, few studies have been carried out to understand the molecular mechanisms involving other proteins that control pre‐RC assembly and functioning, and how it is coupled with developmental signalling pathways. In addition to the regulatory mechanisms involving protein degradation and CDK activity (Arias and Walter, 2007), few proteins were shown to interact with pre‐RC and directly regulate DNA replication (Nishitani and Lygerou, 2004; Iizuka et al, 2006). Here, we identified and characterized a novel pre‐RC interactor—ABAP1—and we provided data on a dual role of ABAP1 in DNA replication and transcription controls. Our data demonstrate that ABAP1 associates with pre‐RC members and the transcription factor AtTCP24 exert a role in a negative feedback loop regulating mitotic DNA replication during leaf development either by reducing AtCDT1 expression mediated by AtTCP24 or possibly by direct association with pre‐RC.

Several experimental evidences support a role of ABAP1 in DNA replication and cell cycle progression. ABAP1 directly interacts with pre‐RC components in vitro, and the in vivo protein interaction data suggest that ABAP1 might associate with pre‐RC subcomplexes or eventually with the full complex in plant cells. ABAP1 is exclusively located in the nucleus and is highly regulated during the cell cycle, accumulating in G1 and early S. Our results illustrated reduced cell division rates in ABAP1^OE plants and an opposite phenotype in ABAP1^ET plants, suggesting that ABAP1 has a negative function in cell proliferation in leaves. A more direct role on DNA replication was demonstrated by decreased levels of thymidine incorporation and reduced pre‐RC loading onto chromatin in ABAP1^OE plants and a reverse observation in ABAP1^ET plants. Altogether, the data indicate that ABAP1 has a function as an inhibitor of DNA replication.

Strong evidence for a role of ABAP1 in gene expression control comes from the identification of different putative transcription factors that associate with ABAP1, of which some are members of families known to be involved in the regulation of different developmental programmes, such as AtTCP24. Although ABAP1 houses two domains that are involved in transcriptional regulation, ARM repeats and BTB, it does not exhibit any clear DNA‐binding signature. Therefore, ABAP1 might exert a function by taking part in transcriptional protein complexes, together with transcription factors and other regulators. The site‐specific binding of the ABAP1–AtTCP24 complex with DNA regulatory elements further corroborates this hypothesis.

The relevance of the interaction between ABAP1 and AtTCP24 was investigated by assessing possible gene targets of the complex in vivo. The presence of TCP‐binding motifs in the promoter regions of Arabidopsis CDT1 homologues, together with the site‐specific binding of the ABAP1–AtTCP24 complex with these DNA regulatory elements in EMSA, indicates that AtCDT1a and AtCDT1b might be targets of ABAP1–AtTCP24 regulation. Confirming this hypothesis, ChIP experiments have shown that ABAP1 binds to promoter regions of the AtCDT1 homologues and the levels of AtCDT1a and AtCDT1b mRNA are lower in ABAP1^OE and higher in ABAP1^ET plants than those in wild‐type plants, indicating that ABAP1 regulates AtTCP24 targets in vivo. A similar effect on AtCDT1 gene expression occurs in plants overexpressing AtTCP24.

TCP family of transcription factors has a central function in leaf development, transcriptionally regulating genes involved in the cell cycle and thus defining marginal configuration and the overall leaf geometry (Li et al, 2005; Barkoulas et al, 2007). Class‐II TCP genes from Antirrhinum majus and Arabidopsis have been shown to be negative regulators of cell proliferation (Nath et al, 2003; Palatnik et al, 2003). This observation, together with our findings, corroborates the hypothesis that ABAP1 and AtTCP24 might work together to repress cell division in specific developmental contexts. Because class‐I TCP factors have been reported earlier to stimulate growth and positively regulate expression of the cell cycle regulator CYCB1;1 (Kosugi and Ohashi, 2002; Li et al, 2005), organ growth might be controlled by the balance of opposite activities of class‐I and ‐II TCPs (Li et al, 2005). The mechanism by which class‐II TCP factors negatively regulate cell proliferation might involve the repression of pre‐RC gene expression such as CDT1. Both class‐I and ‐II TCP factors probably require interaction with other proteins to activate or repress transcription (Kosugi and Ohashi, 2002; Li et al, 2005). Interaction with ABAP1 might represent one mechanism to assist repression of gene expression mediated by class‐II TCP factors. During leaf development, ABAP1–AtTCP24 complex might be part of a mechanism to maintain homoeostasis of AtCDT1a and AtCDT1b levels, and assure proper rates of cell division.

In Arabidopsis, as mitotic divisions cease in leaves from tip to base, subsequent growth depends primarily on cell expansion and differentiation accompanied by endocycles of DNA replication (Donnelly et al, 1999; Beemster et al, 2005). Ectopic expression of AtCDT1a was shown to increase ploidy levels in leaves (Castellano et al, 2004). Nevertheless, ABAP1^OE leaves did not show reduced ploidy levels compared with wild‐type plants, suggesting that ABAP1 has a role controlling only mitotic DNA replication in leaves. Corroborating this hypothesis, ABAP1 expression in leaves decreases from tip to base and is opposite to the increase in rates of cell expansion and endocycles.

Our data raise the intriguing observation that ABAP1 function might exemplify a new negative feedback loop between cell proliferation and control of DNA replication by both repressing transcription of pre‐RC genes and possibly by regulating pre‐RC utilization through direct association with pre‐RC components (Figure 6). The conception of this dual role in the control of pre‐RC could represent a strategy to ensure an efficient repression of cell cycle progression at G1‐to‐S transition. This model for the action mechanism of ABAP1 has similarities to another G1‐to‐S regulator, the retinoblastoma (RB). RB negatively regulates cell cycle progression through interaction with the transcription factor E2F/DP, inhibiting transcriptional activation of E2F/DP target genes, including pre‐RC genes (Bracken et al, 2004). In Drosophila, RB‐E2F associates with DmORC at replication origins, limiting DNA replication (Bosco et al, 2001). RB was also linked to the Myb transcription factor, forming a protein complex that participates in cell differentiation through transcription regulation (Lewis et al, 2004).

In multicellular eukaryotes, proteins called geminins have also been implicated in the regulation of cell proliferation and cell differentiation by interacting with components of the pre‐RC (CDT1) and with transcription factors (Six and Hox) (Luo et al, 2004; del Bene et al, 2004). The interaction of geminin with CDT1 inhibits replication licensing, by recruiting this protein and avoiding the complete pre‐RC assembly. Binding with the transcription factors can displace geminin from CDT1, allowing proliferation to occur. Curiously, so far no putative geminin homologue could be found in plants. In Arabidopsis, a CDT1 interactor GEM was identified, the mechanism of action of which shares similarities with that of geminins (Caro et al, 2007). GEM mediates histone H3 modifications and binds to TTG1—member of a transcription regulator complex that modulates the expression of the homeobox GLABRA (GL2) gene involved in root epidermal cell fate. Competition of GEM for CDT1 and TTG1 might determine GL2 expression and cell division control (Caro et al, 2007). Our data indicated that transcription factors do not compete with pre‐RC components for binding ABAP1, but could possibly participate in the same complex, implying different mechanisms for ABAP1 and geminin function.

In other experimental models, transcription factors have been found to be part of protein complexes that bind to replication origins that control DNA replication. Transcription factors could exert an effect redundantly to recruit the general replication machinery for site‐specific origin utilization during replication, facilitating pre‐RC localization and formation (Danis et al, 2004; MacAlpine et al, 2004). Conversely, in Drosophila, E2F–RB interaction with the pre‐RC inhibits ORC to bind origins, limiting DNA replication initiation at origins (Bosco et al, 2001). Finally, it is possible that both machineries are working together in processes other than DNA replication licensing, such as gene regulation or establishment of heterochromatin (Suter et al, 2004; Sasaki and Gilbert, 2007). In this scenario, our data did not exclude that ABAP1 can be shared between replication licensing and transcription machineries to exert a dual function in these processes. EMSA revealed that AtORC1a is also present in the AtTCP24–ABAP1 complex bound to the class‐II TCP element. Furthermore, protein interaction experiments indicated that AtTCP24–ABAP1 might associate with pre‐RC subcomplexes or eventually with the full complex in plant cells.

Therefore, ABAP1 appears to be involved in the regulation of two vital cellular processes—replication and transcription. Transcription factors are well‐known targets of signal transduction that regulates development; and pre‐RC assembly and function are essential requirements for cell cycle progression. Therefore, we propose that ABAP1 participates in a novel signalling network that controls cell cycle progression in plants, by integrating plant developmental signals with DNA replication and transcription controls (Figure 6). The identification of ABAP1 specifically in plants raises the question whether this signalling pathway has evolved to satisfy peculiar strategies of the plant development. ABAP1 probably has different functions in plant development, depending on the proteins it interacts with. Future biochemical analysis of multiprotein complexes containing ABAP1 will be important to better understand the complexity of the mechanisms that lead to the precise balance of cell division and cell differentiation in various plant morphogenetic events.

Expression analyses using real‐time PCR, transgenic plants production, GUS activity assay and subcellular localization in Arabidopsis are described in Supplementary data.

Transgenic lines were identified by selection in 50 mg/l kanamycin in germination medium (for proABAP1::GUS and GFP::ABAP1) or by enhanced green fluorescence (for ABAP1^OE and AtTCP24^OE plants). At least, 10 transgenic lines were analysed for each construct. ABAP1^ET belongs to the Cold Spring Harbor Enhancer trap bank (seed code ET13614). Lines were selected in 50 mg/l kanamycin in germination medium. Thirty plant genotypes were determined by PCR with specific primers for Enhancer Trap DS element and for ABAP1. Primers ABAP1 5 and ABAP1 3, DS5‐4 and ABAP1 3 were used for WT and DS insertion alleles, respectively (Suplementary data—constructs). For details on phenotypic analysis of ABAP1^OE and ABAP1^ET plant, see Supplementary data.

Yeast two‐hybrid screening of an Arabidopsis cDNA expression library was carried out according to De Veylder et al (1999). For details, see Supplementary data.

ABAP1–GST, AtTPC24–GST and AtORC1a–GST were produced in cells of Escherichia coli strain BL21 (Supplementary data). In vitro transcription and translation of ³⁵S‐methionine (GE Healthcare) was performed using the TNT Quick Coupled Transcription/Translation Systems (Promega) according to the supplier's instructions. GST‐pulldown analyses were carried out according to Tarun and Sachs (1996). Protoplast transformation, immunoprecipitation and protein gel blots were carried out by standard techniques, according to protocols described in the Supplementary data.

EMSAs were performed as described earlier (Kosugi and Ohashi, 1997) with modifications. For details, see Supplementary data.

Wild‐type, ABAP1^OE and ABAP1^ET plants were collected at 6 days after sowing and submitted to lysis in liquid nitrogen. Plant cell lysates and chromatin fractionation were performed as described by Mendez and Stillman (2000), with minor adaptations. For details, see Supplementary data.

Supplementary data are available at The EMBO Journal Online (http://www.embojournal.org).