The gradual colonization of the land through plant evolution has been possible thanks to the differentiation of complex vascular tissues that provide mechanical support and allow long-distance transport of water and nutrients. In land plants, vascular tissues are comprised of two conducting tissues the phloem and the xylem, and intervening cambium that can generate the different conducting cell types (Scarpella and Helariutta, 2010). Whereas the xylem, whose vessels are dead cells with highly lignified walls, ensures the transport of water, mineral salts, and some hormones, the phloem controls the transport and distribution of photosynthetic products from leaves to meristems and other sink organs such as fruits, tubers, and roots. The sieve elements (SE) are the phloem conducting cells and are associated with one or a few CCs. In general the meristematic mother cell of a CC-SE complex divides unequally in the longitudinal axis. One daughter cell develops into one or several metabolically hyperactive CCs with a dense cytoplasm and numerous mitochondria. The second one goes through a controlled disintegration process and differentiates into the SE, which is a highly specialized and unique developmental program (Juergensen et al., 2003). In the last 15 years, major insights have been obtained in the identification of the regulatory actors controlling the establishment of the vascular system as well as the fate of the vascular cell types (reviewed in Zhou et al., 2011). Many transcription factors including bHLH, HD-ZIP, ARF-GAP, AT-Hook, and Dof proteins families were shown to be involved, suggesting the existence of complex transcriptional regulatory networks in which transcription factors are likely to physically interact as shown recently by several bHLH and Dof TFs (Zhang et al., 1995; Kang and Singh, 2001; Skirycz et al., 2008). Some of these families of transcription factors are conserved among eukaryotes but the Dof TFs family, which is the focus of this review, is of particular interest for being specific to plants.

The first DNA-binding with one f inger (Dof) transcription factor has been identified in maize (ZmDof1) and was shown to be involved in light response and transcriptional regulation of genes involved in carbon metabolism (Yanagisawa and Sheen, 1998; Yanagisawa, 2001). Dof TFs are plant-specific and are characterized by a particular zinc finger domain, comprising a conserved region of 50 amino acids with a C₂–C₂ finger structure, associated to a basic region, that binds specifically to DNA sequences with a 5′-(A/T)AAAG-3′ core (Figure 1). A particular feature of these TFs is that the nucleus localization signal (NLS), which directs Dof proteins to the nucleus in Arabidopsis thaliana (Krebs et al., 2010), is an atypical bipartite NLS with a 17 amino acid long linker between its flanking basic regions (Figure 1). This bipartite NLS is highly conserved in plant Dof transcription factors (Krebs et al., 2010) and its feature allowed identifying Dof genes in a various set of plant species (for details, see below). Along plant evolution, Dof transcription factors might have originated from a common ancestor, likely represented by the single Chlamydomonas reinhardtii gene, and then expanded in the different taxonomic groups of vascular plants through recurrent duplication events (Moreno-Risueno et al., 2007). In Gymnosperms and lower plants (Selaginella moellendorffii and Physcomitrella patens) so far analyzed eight to nine Dof TF genes are found (Moreno-Risueno et al., 2007). In Angiosperms an average of 30 genes are found with 27 in Brachypodium distachyon (Hernando-Amado et al., 2012), 30 in rice (Oryza sativa) (Gaur et al., 2011), 36 Dof genes in Arabidopsis (reviewed in Yanagisawa and Schmidt, 1999; Moreno-Risueno et al., 2007), 37 in tomato (Solanum lycopersicon) (Cai et al., 2013), or 41 in poplar (Populus trichocarpa) (Yang and Tuskan, 2006), indicating additional recent duplications in higher plants (Moreno-Risueno et al., 2007). This diversification appears to be associated to the establishment and differentiation of the vascular system during the same period of time, which raises the question whether members of this plant-specific transcription factor family are involved in the formation of the vascular system or its functioning. Indeed besides their involvement in the regulation of different processes such as carbon assimilation (Yanagisawa, 2001; Tanaka et al., 2009), light signaling (Park et al., 2003; Ward et al., 2005; Gabriele et al., 2010), seed development or germination (Papi et al., 2000; Gualberti et al., 2002; Dong et al., 2007; Rueda-Romero et al., 2012), flowering (Sawa et al., 2007; Fornara et al., 2009), stomata functioning (Gardner et al., 2009; Negi et al., 2013), or response to phytohormones (Kang and Singh, 2001; Kang et al., 2003; Nakano et al., 2006) most of them are strikingly expressed in the vascular tissues suggesting a function in long-distance signaling (Gualberti et al., 2002; Papi et al., 2002; Ward et al., 2005; Fornara et al., 2009; Rueda-Romero et al., 2012). Recently several Dof TFs were shown to have a central role in the development of the vascular system (Konishi and Yanagisawa, 2007; Guo et al., 2009; Gardiner et al., 2010; Kim et al., 2010). Based on these observations it is possible to speculate that the plant-specific Dof TFs are likely to regulate directly or indirectly a wealth of processes associated to the development of the vascular system or to its role in long-distance signaling. In the present review we aim at summarizing the current knowledge about the Dof TF family highlighting their role in the control of vascular development and functioning.

In Arabidopsis thaliana, out of the 36 Dof TFs, 20 have been found expressed in the vascular tissues (Table 1 and references there in). In silico studies also indicate the presence of cis-acting elements potentially driving specific expression in the vascular tissues in the promoter of these genes. In plants, mainly through 5′-deletion experiments, several positive cis-acting promoter elements have been identified to drive vascular-specific expression of reporter genes. The study of the rice tungro bacilliform virus (RTVB) present only in the phloem of infected plants allowed the identification of a virus promoter functional in plants (Bhattacharyya Pakrasi et al., 1993; Yin and Beachy, 1995). This promoter contains four domains allowing a strong phloem-specific expression (Yin et al., 1997a,b): a GATA motif which had already been shown to be implicated in light dependent phloem-specific gene expression in plants (Lam and Chua, 1989); an ASL box characterized by the presence of a specific GCA direct repeat sequence present only in phloem-specific promoters and missing in xylem-specific promoters; a Box II containing CCA/TGG repeat and a CCCC sequence that because of their presence in the promoter of plant genes expressed in vascular tissue might be important cis-elements for expression in vascular tissues. These three cis-elements, Box II, the ASL box, and the GATA motif were shown to act in combination (synergistically and/or additively) to confer phloem-specific expression (Yin et al., 1997a). Similarly the study of the coconut foliar decay virus (CFDV) led to the identification of a 13 bp motif sequence (Hehn and Rohde, 1998) which is highly conserved in several phloem-specific promoters as reported by Yoshida et al. (2002). Finally, the 26 bp RSE regulatory element was found to drive a vascular-specific expression of the GRP1 (Glycine-rich protein1) (Keller and Baumgartner, 1991). All these motifs act synergistically or additively and Yin et al. (1997a) have suggested that at least two of these motifs are required for phloem expression. More recently, Ruiz-Medrano et al. (2011) identified degenerate sequences containing CT/GA- and GT/GA-rich repeats within many of the promoters of 150 Arabidopsis genes homologous to members of the pumpkin phloem transcriptome. Interestingly the CT/GA motifs resemble cognate sites for the Dof family of transcription factors, which bind to the core sequence AAAG/CTTT (Higo et al., 1999; Yanagisawa and Schmidt, 1999).

Searching for these cis-elements in the 1000 bp upstream sequences of the Arabidopsis Dof TF family members we showed that at least two of these vascular-specific motifs were found in the promoter region of 31 Dof genes display (Table 2). Among the remaining five, the promoter of AtDof2.1 and AtDof1.8 genes contain three and five GATA motifs respectively and the promoter of AtDof4.4 contains two ASL boxes (Table 2). These observations suggest that the vast majority, if not all, of Arabidopsis Dof transcription factors are likely expressed in the vascular system (Table 2).

The mining of RNA profiling datasets at the resolution of the vascular tissue (phloem, cambium, xylem) and vascular cell type (CC, SE) (Hertzberg et al., 2001; Birnbaum et al., 2003; Vilaine et al., 2003; Schrader et al., 2004; Zhao et al., 2005; Lee et al., 2006; Brady et al., 2007; Ruiz-Medrano et al., 2011; Gandotra et al., 2013) further indicate an expression of most Dof genes in the vascular system. Indeed, the expression of Dof TF genes could be found, either in the xylem, or the phloem or both (Table 2). AtDof1.1, AtDof2.2, AtDof3.2, AtDof3.6, AtDof3.7, AtDof5.3, and AtDof5.5 were identified in the global phloem cells transcript databases established in Arabidopsis (Birnbaum et al., 2003; Zhao et al., 2005). In addition poplar orthologs of AtDof1.6, AtDof3.4, AtDof4.6 were found in the transcriptomic analysis of the wood forming region (Hertzberg et al., 2001; Schrader et al., 2004) and the celery orthologs of AtDof1.8 and AtDof4.6 were expressed in the phloem of celery (Vilaine et al., 2003), suggesting that AtDof1.6 and AtDof1.8 which promoter contains one and five GATA motifs respectively could also be expressed in the vascular region. Besides these genes and the ones already experimentally described as expressed as well in the vascular system (Table 1), the recent transcriptome analysis of microdissected provascular/procambial cells or complete vascular bundle by Gandotra et al. (2013) allowed identifying new AtDof genes (AtDof1.3, AtDof1.4, AtDof1.7, AtDof1.8, AtDof2.2, AtDof3.1, AtDof4.5, and AtDof4.7) potentially expressed in vascular tissue at different step of its development.

In conclusion it is likely that almost all the members of the Arabidopsis Dof TFs family are at least expressed in the vascular system, being so far one of the TF family displaying such a strong tissue-specificity.

Coordinated growth and development of multi-cellular organisms require mechanisms that allow for extensive cell-to-cell communication which can be mediated by signaling molecules directly transported from one cell to the other. Sessile nature of plant growth also necessitated the evolution of mechanisms that rapidly transmit signaling molecules in response to environmental changes or pathogen attacks. These mobile signals in plants can be proteins, RNAs, small RNAs, and other small molecules. They are transmitted through the vascular system in a long-distance (between organs), or through plasmodesmata in a short distance (between cell types) (Lucas et al., 2009; Dinant and Suárez-López, 2012).

By comparing the translational and transcriptional patterns of several TFs, including Dofs, expressed in a cell-type specific manner in Arabidopsis root, Lee et al. (2006) found that the translational pattern was broader or different than the transcriptional pattern for two Dof TFs, AtDof3.7/DAG1 and AtDof4.1, suggesting that these TFs could move out from the cell/tissue they were expressed. In the case of AtDof4.1 the transcriptional expression domain was contained in the translational one indicating that the expansion in the expression pattern must occur after the synthesis of the mRNA (from pericycle to endodermis, Figure 2A). Thus it is likely that it occurred via cell-to-cell protein movement, as shown for the CAPRICE protein (Kurata et al., 2005). For AtDof3.7/DAG1, because the transcriptional AtDof3.7_prom:GFP pattern (in the stele, Figure 2A) was shown expressed in a different tissue compared the one described in the published root mRNA expression map (Birnbaum et al., 2003) the protein might not be expressed in its native tissue. Nevertheless, the expansion of the pattern is unlikely to be because of mRNA movement since the mRNA and the translational expression pattern were also different (transcriptional and translation fusions expressed in the stele and endodermis cell layer respectively; Figure 2A) (Lee et al., 2006).

In addition, cell-type specific expression maps from Arabidopsis root and cell-type specific translatome data were produced by several groups (Birnbaum et al., 2003; Nawy et al., 2005; Lee et al., 2006; Brady et al., 2007; Mustroph et al., 2009) and represent an extremely valuable source of information for predicting putative mobile transcription factors. By analyzing these data sets Lee and Zhou (2012) further proposed that seven Dof TFs (AtDof1.1, AtDof2.1, AtDof2.2, AtDof2.4, AtDof3.2/DOF6, AtDof3.6/OBP3, AtDof3.7/DAG1) are potential mobile transcription factor RNAs.

A direct experimental proof of the intercellular movement of a non-cell-autonomous Dof TFs has recently been brought by Chen et al. (2013) who showed that AtDof4.1/IDT1 (INTERCELLULAR TRAFFICKING DOF 1, IDT1) moves between the root stele and the endodermis in a plasmodesmata-selective pathway. This movement seems to be possible in either ways but more efficient from the endodermis to the stelar tissues (Figure 2A) (Chen et al., 2013). Moreover, the authors demonstrated that IDT1 movement requires an intercellular trafficking motif (ITM) comprising the Zinc Finger motif and the nuclear localization signal (NLS (Figure 1) (Chen et al., 2013). By looking for the existence of an ITM sequence in the different members of the Dof TF family, Chen et al. (2013) also showed that in contrast to AtDof5.4 and AtDof2.2, AtDof3.7/DAG1, and AtDof4.5 were also able to move from cell to cell.

In conclusion, some of the Dof TFs are able to move from cell to cell, however, although several Dof genes have been found expressed in the phloem, there is so far no evidence for long-distance transport since neither RNAs nor Dof proteins have been found in transcriptome and proteome datasets of phloem sap (Giavalisco et al., 2006; Deeken et al., 2008; Kehr and Buhtz, 2008).

The vascular cambium is the lateral meristem source of secondary xylem and secondary phloem. The fascicular cambium, initiated from the procambium, is present between the xylem and the phloem of a vascular bundle. Additionally an interfascicular cambium can also appear during secondary growth, giving rise to interfascicular fibers in the floral stem in species like Arabidopsis. By studying the Arabidopsis gain-of-function mutant high cambial activity2 (hca2), which shows a precocious formation of the interfascicular cambium and its subsequent cell division, Guo et al. (2009) showed that the phenotype was due to the overexpression of AtDof5.6 gene. Repression of AtDof5.6/HCA2 activity led to the disruption of interfascicular cambium formation and development in inflorescence stems. HCA2 promotes interfascicular cambium formation at a very early stage of inflorescence stem development. In addition Guo et al. (2009) showed that AtDof5.6/HCA2 is preferentially expressed in the vasculature of all the organs, but more particularly in the cambium, the phloem, and the interfascicular parenchyma cells of inflorescence stems. In the hca2 mutant high cambial activity and high phloem proliferation activity were detected not only in inflorescence stems, but also in petioles and leaf major veins suggesting that, in addition to interfascicular cambium formation/development, AtDof5.6/HCA2 may also be involved in other developmental aspects during vascular tissue development, for instance, (pro)cambium initiation and activity (Figure 2C).

Additionally two Dof TFs (AtDof3.4/OBP1 and AtDof2.3) are suspected to be involved in the control of the maintenance of the cambium and/or cambial activity by controlling the cell cycle (Skirycz et al., 2008) (Figure 2C). AtDof3.4/OBP1 is highly expressed in tissue with high cell proliferation activity (cell cultures, calli, developing embryo, apical, and root meristems, fascicular procambium in the stem), and while looking for direct targets of AtDof3.4/OBP1 several core cell cycle genes and transcription factors were identified (Skirycz et al., 2008). OBP1 modulates cell cycle activity by affecting the expression of CDKA regulators, S-phase specific transcription factors and components of the replication machinery, and is therefore important for cell cycle onset. OBP1 most likely operates through a direct regulation of CYCD3.3 and AtDof2.3 gene expression since it was shown to interact physically with their respective promoter of AtDof2.3 and CYCD3.3 (Skirycz et al., 2008) (Figure 2C).

Structurally, like for other zinc fingers, the Dof domain is known to be a bi-functional domain that mediates not only DNA-binding but also protein–protein interactions (Yanagisawa, 2002). When looking at cis-elements present in Dof TFs promoters, one can observe that each of the promoter displays at least one cis-element related to other transcription factor (bHLH, MADS boxes, ARF, Homeobox, MYB, bZIP, WRKY) (for more details, see the AGRIS interface, http://arabidopsis.med.ohio-state.edu/AtcisDB/). Altogether these data suggest that interactions between Dof TFs and TFs belonging to other families are theoretically possible. Experimentally, OBP1 was shown to directly interact with the promoter of a bZIP transcription factor but also with the promoter of a member of the same family AtDof2.3 (Zhang et al., 1995; Kang and Singh, 2001; Skirycz et al., 2008). Moreover Kang et al. (2003) showed that two bHLH members (bHLH038 and bHLH039) are targets of the vascular expressed OBP3 transcription factor. Interestingly, bHLH039 was also found to be highly expressed in phloem tissue during hypocotyl secondary growth in Arabidopsis (Zhao et al., 2005). In the regulation of leaf axial patterning, a direct binding of AtDof5.1 on the promoter of the HD-ZIP, REVOLUTA have also been demonstrated (Kim et al., 2010). Besides these interactions at the transcriptional level, protein–protein interactions between the Dof TFs, CDF1, CDF2, and CDF3 and the F-box FKF1 have also been observed in the frame of the photoperiodic flowering response (Imaizumi et al., 2005). These few examples illustrate the fact that Dof TFs and TFs of other families (bHLH, bZIP, HD-ZIP, ARF, F-box) have the potential to interact together at the transcriptional as well as at the protein level and therefore be part of transcriptional networks involved in different developmental processes such as vascular development. Such transcriptional network, involving a bHLH complex, has been recently shown to control the establishment of the embryonic vascular tissue (De Rybel et al., 2013). This discovery opens a new level of complexity in the molecular mechanisms underlying the vascular development. Similar complexes involving Dof TFs could also be speculated. Indeed besides TARGET OF MONOPTEROS (TMO5/bHLH135), a Dof TF (AtDof5.3/TMO6) has been also identified as a potential indirect target of MP transcription factor (Schlereth et al., 2010) and has been described as one of the transcription factors participating in Arabidopsis vein formation (Gardiner et al., 2010). TMO6 could therefore be a good candidate to explore the involvement of Dof TFs in the regulation of the early steps of procambium formation. Altogether, TMO6 as well as the other Dof TFs mentioned in this review could constitute the starting point to study the potential implication of the Dof TFs in the transcriptional networks underlying vascular development.

In the context of the vascular system development and functioning, which integrates a large variety of internal or external stimuli, regulators are needed to fine-tune the responses to changes in the environment. In this review, we present evidences that the Dof TFs family is likely to participate in such fine-tuning of the responses to the different stimuli in the vascular system. By modulating a variety of transcription factors, by protein-DNA and/or protein–protein interactions, the Dof TFs could therefore act at the cross-talk of various developmental pathways directly or indirectly linked to the vascular development and/or functioning. One of the next challenges, besides the functional characterization of the still undescribed Dofs TFs, will therefore be identifying the actors involved in an integrated regulatory network. Combination of the high-throughput data collection at the resolution of the vascular cell types and developmental stages as well as the use of the next generation sequencing technology will certainly allow identifying genes network regulating the fundamental features of the vascular development.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.