Ive De Smet was a finalist for the 2011 New Phytologist Tansley Medal for excellence in plant science, which recognises an outstanding contribution to research in plant science by an individual in the early stages of their career; see the Editorial by Dolan, 193: 821–822.
Plant growth relies heavily on a root system that is hidden belowground, which develops post-embryonically through the formation of lateral roots. The de novo formation of lateral root organs requires tightly coordinated asymmetric cell division of a limited number of pericycle cells located at the xylem pole. This typically involves the formation of founder cells, followed by a number of cellular changes until the cells divide and give rise to two unequally sized daughter cells. Over the past few years, our knowledge of the regulatory mechanisms behind lateral root initiation has increased dramatically. Here, I will summarize these recent advances, focusing on the prominent role of auxin and cell cycle activity, and elaborating on the three key steps of pericycle cell priming, founder cell establishment and asymmetric cell division. Taken together, recent findings suggest a tentative model in which successive auxin response modules are crucial for lateral root initiation, and additional factors provide more layers of control.
The formation of a plant root system takes place post-embryonically and relies on de novo formation of organs. Typically, these lateral root organs are initiated close to the root tip and emerge in the differentiation zone (Fig. 1a). Over the past few years, knowledge about the regulatory mechanisms behind many aspects of lateral root formation has increased considerably (De Smet et al., 2006; Péret et al., 2009). The development of a lateral root from a limited number of cells requires tightly coordinated asymmetric cell divisions to generate cell diversity and tissue patterns. This characteristically involves the specification of founder cells, followed by a number of cellular changes until the cells divide and give rise to unequally sized daughter cells (De Smet & Beeckman, 2011). This review summarizes our understanding of the early steps that control lateral root initiation and positioning in the model plant Arabidopsis thaliana. Key aspects such as the cellular context, auxin and cell cycle activity will be introduced, before considering the three major steps of priming, founder cell establishment and asymmetric cell division.
The cellular context of lateral root initiation
Marker-based cell lineage analyses in A. thaliana have revealed that lateral roots develop from three files of pericycle cells at the xylem pole, which are part of the stele tissues and are deeply embedded within the root below the epidermis, cortex and endodermis (Dubrovsky et al., 2000; Beeckman et al., 2001; Kurup et al., 2005) (Figs 1b, 2). In A. thaliana, lateral roots can develop at either of the xylem poles, but never develop at the phloem poles (Dubrovsky et al., 2000; Beeckman et al., 2001; De Smet et al., 2006; Parizot et al., 2008). This strong association with the xylem pole is further supported by the lonesome highway (lhw) mutant, which displays only one xylem pole and unilateral lateral root formation (Parizot et al., 2008). Nevertheless, a differentiated protoxylem element within the xylem pole is not required as impaired vasculature development (ivad) and arabidopsis histidine phosphotransfer protein 6 (ahp6) mutants still display lateral root initiation (Parizot et al., 2008). However, the wooden leg (wol) mutant, which lacks phloem and in which all pericycle cells have xylem pole-associated identity, hardly produces any lateral roots (Parizot et al., 2008). At present, it is unclear how the interaction with xylem and phloem poles determines the distinct pericycle identities with differing competence. Nevertheless, pericycle cells at these poles are characterized by differences in size, by ultrastructural features and by specific proteins and gene expression (see De Smet et al., 2006; Parizot et al., 2008 and references therein).
The plant hormone auxin is a major regulator of growth and development, and, while the involvement of various phytohormones and other regulators has also been shown, auxin is critical in lateral root development (Benkováet al., 2003; De Smet et al., 2006; Laplaze et al., 2007; Ivanchenko et al., 2008; Negi et al., 2008; Péret et al., 2009). Briefly, the general mechanism by which auxin acts is as follows (reviewed in Calderon-Villalobos et al., 2010). Under low auxin concentrations, AUXIN/INDOLE-3-ACETIC ACID (AUX/IAA) proteins (AUX/IAAs) form dimers with AUXIN RESPONSE FACTOR (ARF) transcription factors, thereby blocking the activity of at least the activating ARFs. Intracellular auxin is perceived by the TRANSPORT INHIBITOR RESPONSE1/AUXIN SIGNALLING F-BOX PROTEIN1–3 (TIR1/AFB1–3) receptors, which are an integral component of a complex that mediates the ubiquitination of AUX/IAAs and thereby destines them for 26S proteasome-dependent degradation. Once freed from the AUX/IAAs, the ARFs regulate the expression of auxin-responsive genes. For example, in the case of ARF7–ARF19, members of the LATERAL ORGAN BOUNDARIES-DOMAIN/ASYMMETRIC LEAVES-LIKE family have been identified as direct targets during lateral root initiation (Okushima et al., 2007). As will become apparent in this review, different auxin response modules – an AUX/IAA–ARF pair and the specific transcriptional targets – act sequentially or in a partly overlapping manner to control the asymmetric cell division of a few pericycle cells that will develop into a complete organ (Fig. 2). Consequently, several gain-of-function aux/iaa or loss-of-function arf mutants display altered lateral root initiation and/or development (De Smet et al., 2006; Péret et al., 2009). For example, prominent mutant phenotypes are: no lateral roots (solitary root (slr)/iaa14 and arf7arf19; Fukaki et al., 2002; Okushima et al., 2005; Wilmoth et al., 2005); irregularly positioned lateral roots (monopteros (mp)/arf5 and bodenlos (bdl)/iaa12; De Smet et al., 2010); and defects in pericycle cell development and primordium morphology (tir1 afb2 afb3; Dubrovsky et al., 2011).
Several elements suggest that tightly controlled cell cycle progression is essential for asymmetric division of a subset of pericycle cells during lateral root initiation, as otherwise pericycle cells will undergo either no division at all or uncontrolled proliferation rather than giving rise to a new organ (Himanen et al., 2002; DiDonato et al., 2004; Vanneste et al., 2005; De Smet et al., 2010). For instance, the aberrant lateral root formation 4 (alf4) mutant displays no lateral root initiation and it has been suggested that the conserved ALF4-encoded protein maintains pericycle cells in a mitotically competent state (DiDonato et al., 2004; Dubrovsky et al., 2008). Furthermore, progression of pericycle cells through the cell cycle – especially the onset of the G1–S transition – is prevented in the absence of auxin (achieved through NPA-mediated auxin transport inhibition) or when the early auxin response is blocked (in a gain-of-function slr mutant), resulting in the absence of lateral root initiation (Himanen et al., 2002; Vanneste et al., 2005). Nevertheless, elevated expression of G1–S-related cell cycle genes, such as E2FA/DPA or the D-type CYCLIN D3;1 gene (CYCD3;1), is not sufficient to pattern a new lateral root; instead, pericycle cells start to undergo proliferative cell division, giving rise to stretches of short pericycle cells (Vanneste et al., 2005; De Smet et al., 2010). Interestingly, this lack of organogenesis can be overcome by auxin (De Smet et al., 2010), suggesting a complex auxin-dependent regulation of cell cycle progression, asymmetric cell division and differentiation at various stages of lateral root initiation. In addition, proteolysis of cell cycle transcription factors through the auxin-binding F-box protein S-PHASE KINASE-ASSOCIATED PROTEIN 2A (SKP2A) is involved in lateral root initiation, as overexpression of SKP2A increases the number of lateral root primordia. This occurs independently of the TIR1-AFB pathway, as overexpression of SKP2A induces lateral root initiation in the tir1 background and the skp2a mutant displays auxin-resistant root growth that is additive to the tir1 mutant phenotype (Jurado et al., 2008, 2010).
Various cell cycle genes have been assigned a role during lateral root initiation. For example, CYCDs, as regulators of the G1-to-S transition, are involved in lateral root initiation: cycd4;1 (Nieuwland et al., 2009) and cycd2;1 mutants (Sanz et al., 2011) display lateral root density defects. First, CYCD4;1 is expressed in pericycle cells adjacent to the xylem poles within the meristem. While CYCD4;1 is apparently not directly involved in lateral root initiation, it seems to affect the pericycle cell length in the basal meristem, which in turn affects the lateral root initiation number (Nieuwland et al., 2009). Secondly, CYCD2;1 appears to be rate-limiting for auxin-induced lateral root initiation, probably through contributing to the sensitivity of the pericycle cells in terms of their response to auxin (Sanz et al., 2011). Indeed, increased levels of CYCD2;1 confer an increased lateral root-related auxin response, opposite to that of the cycd2;1 mutant (Sanz et al., 2011). Furthermore, proteins of the INTERACTOR OF CYCLIN-DEPENDENT KINASE (CDK)/KINASE-INHIBITORY PROTEIN (KIP)-RELATED PROTEIN (ICK/KRP) family inhibit the G1-to-S transition, and have been shown to prevent auxin-mediated lateral root initiation (Himanen et al., 2002; Ren et al., 2008). The krp2 mutants display increased lateral root density, whilst ectopic overexpression of KRP2 results in a large reduction in lateral root density (Himanen et al., 2002; Ren et al., 2008; Sanz et al., 2011). ICK2/KRP2 interacts with the CDKA;1–CYCD2;1 complex, allowing accumulation in the nucleus of the inactive complex. In the presence of auxin, reduced ICK2/KRP2 expression and increased ICK2/KRP2 protein turnover result in a transient increase in CDKA;1–CYCD2;1 activity and subsequent cell division (Sanz et al., 2011). This mechanism explains how ICK2/KRP2 – expressed in the continuous absence of auxin in the phloem pole pericycle cells – prevents lateral root initiation at the phloem pole (Himanen et al., 2002). Finally, higher order mutants for A2-type cyclins (CYCA2s) display reduced lateral root density and deviations in lateral root primordium patterning (Vanneste et al., 2011). This coincides with a lack of expression of mitotic regulators during auxin-induced lateral root initiation, while early auxin signalling and G1–S regulation proceed normally (Vanneste et al., 2011). This indicates involvement of CYCA2s in lateral root initiation early in the G2–M transition (Vanneste et al., 2011).
Priming the xylem pole pericycle cells
Priming, or the specification of a subset of xylem pole-associated pericycle cells to provide the competence to form a lateral root, takes place in the basal meristem (De Smet et al., 2007) (Fig. 2). This coincides with the very first visible event associated with lateral root initiation, namely, an auxin response maximum – visualized by the synthetic auxin reporter DR5::GUS– in the protoxylem cells adjacent to those pericycle cells that will be able to form a lateral root (De Smet et al., 2007) (Figs 1b, 2). It is not clear if pericycle cells at both xylem poles are primed, but, based on the fact that DR5::GUS is expressed in both protoxylem strands in the basal meristem, it is likely that this is the case (De Smet et al., 2007) (Fig. 1b). Similarly to the initiation of the primary root meristem (Schlereth et al., 2010), it is possible that a (yet to be identified) signal moves from the xylem into the neighbouring pericycle cells to prime these cells for lateral root initiation (De Smet & Beeckman, 2011).
Intriguingly, this DR5-visualized response in the basal meristem manifests itself as periodic pulses, as visualized using a DR5::GUS time series (De Smet et al., 2007; De Rybel et al., 2010) and in vivo, real-time DR5::LUCIFERASE expression (Moreno-Risueno et al., 2010). Furthermore, this DR5-visualized response is TIR1–AFB-, IAA28- and ARF7-dependent, as interfering with the activity of these proteins disturbs DR5 expression and/or oscillation in the basal meristem (De Smet et al., 2007; De Rybel et al., 2010; Moreno-Risueno et al., 2010) (Figs 2, 3a). Further analyses will reveal the basis of this oscillation, but it appears that fluctuation in auxin is probably not sufficient to drive these pulses of DR5 expression in the basal meristem, as other auxin signalling reporters such as IAA19 do not oscillate in the basal meristem (Moreno-Risueno et al., 2010). In this context, it has been reported that DR5 also responds to other hormones (Nakamura et al., 2003) or the peaks in DR5 expression might merely reflect competence influenced by internal cues.
Recently, a complete oscillatory network located in the basal meristem was put forward as the endogenous developmental mechanism triggering branching in the A. thaliana root. About 3400 genes oscillate in phase or antiphase in the oscillatory zone and both phases are required to position lateral roots along the growing primary root axis, which is supported by lateral root initiation defects when genes encoding transcription factors that oscillate both in phase and antiphase are mutated (Moreno-Risueno et al., 2010). It has been proposed that this time-keeping mechanism is based on self-sustained oscillations in gene expression or autonomous oscillators driving periodic pulses of gene expression (Moreno-Risueno et al., 2010), in contrast to periodic maxima of auxin in the basal meristem (De Smet et al., 2007).
Founder cell specification of xylem pole pericycle cells
Following the priming event, founder cell specification occurs within a developmental window that is located in a well-defined zone along the primary root axis where auxin content and response are minimal (Dubrovsky et al., 2011) (Fig. 2). Founder cells are those pericycle cells at the xylem pole that give rise to a lateral root primordium. They are characterized by static points of DR5 expression, namely fixed expression along the growing primary root axis (Fig. 3b). This local activation of auxin response precedes the initiation of lateral root primordia, hence predicting the position of lateral roots along the primary root axis (Dubrovsky et al., 2008; Moreno-Risueno et al., 2010) (Fig. 3b). Specifically, this auxin response occurs in either one or two longitudinally adjacent xylem pole pericycle cell(s) per file, which is probably associated with an increase in auxin concentration (Dubrovsky et al., 2001, 2008). Indeed, as shown through locally stimulated auxin production, a local, cell-specific increase of auxin concentration in a few pericycle cells is sufficient for founder cell establishment (Dubrovsky et al., 2008). This suggests that auxin acts as a morphogenetic trigger; in other words, a factor that induces – through a local increase in its concentration – acquisition of a new developmental fate in plant cells that were originally similar to their neighbours, and thus specifies the site at which a new organ will be formed (Dubrovsky et al., 2008; Benkováet al., 2009).
Using a bending assay, it was shown that this local accumulation of auxin possibly comes about through localized induction of the AUX1 influx carrier in a subset of pericycle cells (Laskowski et al., 2008). In addition to auxin-related signals, a transient increase in cytosolic Ca2+, associated with root bending, seems to be an additional signal leading to specification of founder cells, as blocking this change in Ca2+ also prevented lateral root formation in the curved region of the root (Richter et al., 2009). Based on the above results, it is probably a complex combination of gravitropic root bending, cytoplasmic Ca2+ changes and mechanical changes in the pericycle cell that triggers AUX1-dependent auxin influx and that determines founder cell identity and the bias towards pericycle cells at one xylem pole.
Another key player in founder cell establishment is GATA23, a member of the GATA-type transcription factors that are known to be critical players in regulatory networks determining the specification of cell fates during mammalian development. When GATA23 levels are down- or up-regulated, this results in a strong decrease or increase in early primordia stages, respectively (De Rybel et al., 2010). GATA23 is one of several transcription factors that display a fluctuating gene expression pattern over developmental time along the primary root axis (Brady et al., 2007; De Rybel et al., 2010) (Figs 2, 3b). These genes are early markers for founder cell identity and their expression is maintained for some time in these pericycle cells, clearly indicating the positions of future lateral root primordia. The fluctuating gene expression of GATA23 is dependent on a TIR1–AFB-dependent auxin response, and, specifically, AUX/IAA28 represents an important factor controlling founder cell identity upstream of GATA23, as overexpression of GATA23 in iaa28 rescues the lateral root mutant phenotype (De Rybel et al., 2010) (Fig. 2).
In addition to specific gene expression, a morphological feature of founder cells when two longitudinally adjacent pericycle cells per file are activated is that their respective nuclei migrate towards a common cell wall (De Smet et al., 2007, 2008; De Rybel et al., 2010). This AUX/IAA–ARF-dependent simultaneous migration of nuclei is essential, as migration in a single cell does not appear to lead to an anticlinal asymmetric cell division and a lateral root organ (De Rybel et al., 2010). It is currently unclear how the opposing polarity within adjacent founder cells is established, and if the auxin gradient is the driving force for this or part of the output.
Asymmetric cell division – starting to build a new organ
Once founder cell specification has occurred, primordium initiation follows relatively soon thereafter in the differentiation zone (Dubrovsky et al., 2011) (Fig. 2). The founder cells undergo a series of anticlinal asymmetric cell divisions, and this is followed by a 90° rotation of the cell division plane and periclinal asymmetric cell divisions, giving rise to a second layer (Malamy & Benfey, 1997; De Smet et al., 2008) (Fig. 1c,d). Following the first asymmetric cell divisions, an auxin response maximum can be observed in the central core of small cells (Benkováet al., 2003). Indeed, at this stage, the auxin response module centred around the MP/ARF5–BDL/IAA12 pair plays an essential role. When MP/ARF5–BDL/IAA12-dependent signalling is disrupted, lateral roots develop as closely spaced or even as fused lateral root primordia. Analysis of transcriptional activation of ARF19 and MP/ARF5 during lateral root initiation, together with the lateral root rescue of the slr mutant by overexpression of MP, demonstrated that ARF5 is downstream of SLR and clearly highlighted the sequential nature of different auxin response module activities required to initiate a lateral root (De Smet et al., 2010) (Fig. 2).
The distinct identities of the large and small daughter cells produced following the first anticlinal asymmetric cell divisions are reflected in differential gene expression. For example, only when divided pericycle cells will give rise to a lateral root, the receptor-like kinase ARABIDOPSIS CRINKLY4 (ACR4) is expressed in the small daughter cells. In other cases, for instance when pericycle cells are stimulated to proliferate by overexpression of CYCD3;1, ACR4 expression is absent (De Smet et al., 2008, 2010). Based on acr4 mutant phenotypes, such as additional cell divisions in pericycle cells adjacent to the lateral root initiation site and the formation of lateral root primordia at uncommon positions (such as opposite each other), ACR4 appears to play a role in lateral inhibition during lateral root initiation (De Smet et al., 2008) (Fig. 2). Lateral inhibition occurs along the longitudinal pericycle axis where ACR4 appears to control the division potential of cells neighbouring those in which it is expressed, delimiting the lateral root initiation site, which is a prerequisite for the generation of a discrete primordium. Simultaneously, there is also lateral inhibition along a radial axis, where ACR4 is involved more broadly in preventing lateral root initiation at the opposite xylem pole. Interestingly, MP/ARF5 and BDL/IAA12 appear to act in a similar way to ACR4, at least along the longitudinal axis, but it is at present unclear how these pathways are connected. Further work will be required to elucidate the mechanisms behind ACR4-dependent signalling, to identify the (peptide) ligands involved, and to determine how this is integrated with auxin and other hormone signalling pathways.
Conclusions and outstanding questions
In conclusion, our knowledge of the molecular mechanisms controlling anticlinal asymmetric cell division during lateral root initiation is increasing, but there are still a number of pressing biological questions that need to be addressed. It is now clear that lateral root initiation comes about through the sequential activities of independent and/or overlapping auxin response modules. To gain insight into the complexity and interactions of these modules, it will be essential to identify their common and/or distinct targets during lateral root initiation, and using techniques such as ChIP-on-chip (combining chromatin immunoprecipitation with microarray technology) this is now feasible.
While various developmental systems have proved to be useful in understanding aspects of asymmetric cell division (reviewed in De Smet & Beeckman, 2011), lateral root initiation occurs deep within the primary root, hampering analysis of what controls the polarity of the pericycle cells that will undergo this division, the coordination of the simultaneous nuclear migration, and the determination of the position of the division plane. However, with the improved imaging technology and the availability of new markers, elucidating these aspects of lateral root initiation and identifying the key players involved in these processes is possible.
At present, the dominant role of auxin and associated factors in lateral root development is striking. However, elucidation of the lateral root regulatory network is revealing interactions with many more signalling pathways, such as short-distance communication via peptide hormone-receptor kinase signalling (De Smet et al., 2008), and other hormones, such as cytokinin and ethylene (Laplaze et al., 2007; Ivanchenko et al., 2008; Negi et al., 2008). Collectively, these signals and their interactions hold the key to understanding lateral root initiation. Unlocking this complexity is the challenge that we face in the coming years.
I thank M. Bennett, J. Dubrovsky and S. Smith for critical reading of the manuscript, insightful discussions and useful suggestions. I.D.S. is supported by a BBSRC David Phillips Fellowship (BB_BB/H022457/1), a Marie Curie European Reintegration Grant (PERG06-GA-2009-256354) and the Research Foundation Flanders (FWO09/PDO/064 A 4/5 SDS). I apologize to those colleagues whose work could not be incorporated because of space restrictions.