Notice: Wiley Online Library will be unavailable on Saturday 27th February from 09:00-14:00 GMT / 04:00-09:00 EST / 17:00-22:00 SGT for essential maintenance. Apologies for the inconvenience.
Timing is everything: what determines when proliferation gives way to expansion?
Anisotropic growth and the importance of polarity
How does organ identity and developmental patterning modulate growth behaviour?
Coordination of growth at different scales
The growth of plant organs is under genetic control. Work in model species has identified a considerable number of genes that regulate different aspects of organ growth. This has led to an increasingly detailed knowledge about how the basic cellular processes underlying organ growth are controlled, and which factors determine when proliferation gives way to expansion, with this transition emerging as a critical decision point during primordium growth. Progress has been made in elucidating the genetic basis of allometric growth and the role of tissue polarity in shaping organs. We are also beginning to understand how the mechanisms that determine organ identity influence local growth behaviour to generate organs with characteristic sizes and shapes. Lastly, growth needs to be coordinated at several levels, for example between different cell layers and different regions within one organ, and the genetic basis for such coordination is being elucidated. However, despite these impressive advances, a number of basic questions are still not fully answered, for example, whether and how a growing primordium keeps track of its size. Answering these questions will likely depend on including additional approaches that are gaining in power and popularity, such as combined live imaging and modelling.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Every day we observe a spectacular array of diverse plant forms. Although organ size and shape can be modified by environmental factors, the genotype of an individual sets the limits within which such modification of growth and development can occur. Thus, different genotypes will condition different final sizes and shapes when grown together across a range of environments. This review will describe our current knowledge about the genetic control of plant organ growth at several levels, highlighting open questions along the way. Although research into the evolution of size and shape and the increasing use of modelling have made important contributions to our knowledge about growth control, they cannot be covered here because of space limitations and the reader is referred to relevant recent reviews (Alonso-Blanco et al., 2009; Grieneisen & Scheres, 2009; Chickarmane et al., 2010; Sicard & Lenhard, 2011).
The basic body plan of the plant is established during embryogenesis, resulting in a simple structure with a radial and an apical–basal axis (Steeves & Sussex, 1989). The apical pole is defined by the shoot apical meristem (SAM), a small group of stem cells that continuously divide and replenish themselves. The key functions of the SAM are to maintain itself as a source of cells and to generate daughter cells that are displaced towards the meristem periphery and its base, where they enter specific differentiation pathways to form lateral organs (leaves and flowers) and the stem, respectively. How the meristem is organized and maintained has been extensively reviewed (see, for example, Braybrook & Kuhlemeier, 2010). After initiation, leaf and floral organs grow by two basic cellular processes: cell proliferation and cell expansion (Steeves & Sussex, 1989). During the cell-proliferation phase, the cells increase in size and accumulate cytoplasmic mass, and subsequently divide to yield two daughters of comparable size to the original cell. Cell proliferation occurs initially throughout the entire primordium and the relatively small increase in organ size (in absolute terms) is directly related to the increase in cell number. Gradually cell proliferation becomes restricted, with proliferation arresting first at the distal tip and then in progressively more proximal parts, until all of the cells have stopped dividing (Donnelly et al., 1999). Once a cell has stopped proliferating, it begins to expand, with the often dramatic increase in overall organ size occurring largely through this post-mitotic expansion (Fig. 1a).
Changes in final organ size could be caused by a change in the number of organ founder cells or by an altered rate or duration of growth. In Arabidopsis and other dicotyledonous plants, no clear cases of changes in organ size resulting from a different number of founder cells have been documented, and we will therefore focus on post-initiation growth. Insight into growth control has largely been achieved through the characterization of mutants with specific defects in growth by proliferation and/or expansion (summarized in Fig. 1b–f). A large number of factors with diverse predicted molecular functions influence organ growth, with most of them – based on current knowledge – apparently acting in independent pathways.
II. The cell biology and biophysics of growth
The problem of growth control touches on a range of scales, from the cellular, via the tissue and organ level to the whole plant. At the cellular level individual cells need to increase their cytoplasmic content and divide during the proliferation phase, as well as to expand their cell wall throughout all of organ growth.
1. Cytoplasmic growth
Making more cytoplasm, or the synthesis of new proteins and other macromolecules, underlies cellular growth (Hall, 2004). Genes that encode components of the plant translational machinery, such as ribosomal proteins, are strongly expressed in proliferating cells (McIntosh & Bonham-Smith, 2006). However, little is known about how the expression of ribosomal-protein genes is regulated and coordinated with the synthesis of ribosomal RNAs (rRNAs) to ensure their proper stoichiometries (McIntosh & Bonham-Smith, 2006). In Arabidopsis, most mutants in cytoplasmic ribosomal genes have a pointed leaf shape (Byrne, 2009; Szakonyi & Byrne, 2011). When one of several genes encoding proteins of the large subunit of cytoplasmic ribosomes is mutated in an asymmetric leaves1 mutant background, ectopic lamina outgrowth results, revealing a role for ribosomes in regulating the expression of leaf-patterning genes (Pinon et al., 2008). How ribosomal function mechanistically affects patterning-dependent organ growth remains to be determined.
For cells to maintain a constant size while proliferating, cytoplasmic growth and cell division have to be precisely coordinated. A possible link between the two processes is suggested by analysis of the Arabidopsis TCP20 protein, a transcription factor belonging to the class I of TEOSINTE BRANCHED1, CYCLOIDEA, PCF1 (TCP) proteins (Li et al., 2005a). TCP20 binds in vivo to the GCCCR element in the promoters of genes encoding ribosomal proteins and the mitotic cyclin CYCB1;1 gene, and this element is required to activate their expression. Thus, TCP20 could coordinate ribosome biogenesis and cytoplasmic growth with progression through the cell cycle. Manipulating the activity of TCP20 in Arabidopsis seedlings suggests that depending on the organ and the developmental stage, TCP20 influences either cell proliferation or cell expansion, with many cell-wall biosynthetic genes among the putative direct targets of TCP20 in roots and hypocotyls (Herve et al., 2009). Thus, TCP20 appears to influence all three basic cellular growth processes: cytoplasmic growth, division and expansion.
The ErbB-3 epidermal growth factor receptor-binding protein EBP1 is required for both ribosome biogenesis and proliferation in animals (Okada et al., 2007). This protein is conserved in plants and decreased levels of EBP1 expression reduce organ growth by limiting cell number as well as cell size, whereas increased levels of EBP1 activity lead to larger organs (Horvath et al., 2006), suggesting that EBP1 may similarly link ribosome biogenesis and proliferation in plants.
In yeast and animals cellular growth is modulated by the nutrient or energy status, which is sensed and transduced via the TARGET OF RAPAMYCIN (TOR) pathway and signalling lipids (Ma & Blenis, 2009). The TOR kinase regulates numerous growth-related biological processes, including the transcription and translation of ribosomal components (Ma & Blenis, 2009). Plants also contain a functional TOR kinase pathway. Loss of AtTOR function in Arabidopsis results in embryo lethality, while more limited downregulation leads to reduced leaf size as a result of smaller cells (Menand et al., 2002; Deprost et al., 2007). Conversely, overexpression of AtTOR increases leaf-cell and overall leaf size (Deprost et al., 2007). The growth arrest observed after strong AtTOR inactivation postembryonically is linked to a reduction in translationally active polysomes (Deprost et al., 2007). Recent work has also demonstrated that AtTOR promotes the transcription of rRNA by binding to the rRNA gene promoter, indicating that AtTOR enhances the translational capacity of cells by acting at different levels (Ren et al., 2011).
AtTOR kinase acts in a protein complex including AtRaptor that activates the ribosomal protein S6 kinase (S6K) (Deprost et al., 2005; Mahfouz et al., 2006). While in animals, S6K promotes cellular and organismal growth (Shima et al., 1998), the two Arabidopsis S6K genes are required to repress cell proliferation under nutrient-limiting conditions and to safeguard genome stability and prevent polyploidization (Henriques et al., 2010). The negative effect on proliferation may be mediated via the RETINOBLASTOMA-RELATED1 (RBR1)–E2F pathway, as S6K promotes nuclear localization of RBR1 that inhibits the cell cycle promoting transcription factors of the E2F family and promotes entry of cells into differentiation (Borghi et al., 2010; Henriques et al., 2010). Leaf cell size is reduced upon downregulation of S6K function; thus in addition to constraining cell division, S6K appears to promote elongation-driven growth.
The growth of all organisms is under mechanical constraints, and this is particularly relevant for plants where the cells are ‘boxed in’ and ‘glued’ to each other by their cell walls. Thus, cell wall extension needs to be coordinated with cytoplasmic growth and cell division. Intriguingly, as described earlier, components of the TOR-S6K pathway, whose counterparts in animals are major regulators of cell proliferation, have strong effects on cell expansion in plants. At least for AtTOR these effects do not seem to be only indirect consequences of reduced cytoplasmic growth. Rather, a more direct link between TOR signalling and cell-wall synthesis is suggested by the observations that inhibition of TOR signalling changes the composition of the cell walls (Leiber et al., 2010). In addition, one can speculate that as the TOR pathway modulates mitochondrial activity and the production of reactive oxygen species (ROS) in animals (Schieke et al., 2006; Cunningham et al., 2007), it may also influence cell expansion in plants via this route, as ROS can increase or reduce cell wall extensibility in a context-dependent manner (Gapper & Dolan, 2006; Leiber et al., 2010). In addition to the established role of ROS in the loosening of cell walls in tip-growing cells (Foreman et al., 2003), ROS-dependent cell expansion during leaf growth (Rodriguez et al., 2002) and effects on leaf shape and morphogenesis (Sagi et al., 2004) have also been reported. With their potential to function as intercellular signalling molecules and to integrate both expansive growth and the stiffening of cell walls as growth ceases, a role for ROS and the plant NADPH-oxidases that generate them (Foreman et al., 2003) in cell-wall regulated organ-size control is an interesting possibility.
2. Cell wall synthesis, monitoring and expansion
Cell division and cell expansion in plants require loosening of the cell wall matrix and deposition of new wall components (Cosgrove, 2005). The type of cell wall and the shape of the cell will have a direct effect on growth; for example, the heavily thickened walls of xylem cells and the interdigitating epidermal leaf cells will have different growth dynamics. The different cell wall matrices surrounding cells with different functions suggest that biosynthesis and differentiation of cell walls are precisely regulated in both a spatiotemporal and developmental manner (Freshour et al., 2003). It has been estimated that > 2000 genes are required for the synthesis and metabolism of cell wall components (McCann & Rose, 2010), so understanding how the production and accumulation of wall components are regulated is not a trivial problem.
As cells grow, their walls and membranes are deformed and stretched. To avoid compromising cell integrity, feedback mechanisms are thought to exist that sense the status of the cell wall and control wall deposition and modification in response to need (Seifert & Blaukopf, 2010). Little is known, however, about how mechanical stimuli such as wall or membrane strain are sensed in plant cells and whether or how this information is used to alter growth, cell-wall synthesis and modification. Members of a family of MscS-Like (MSL) transmembrane channels have recently been identified as potential mechanoreceptors in plants (Haswell et al., 2008), yet no function in growth control could be defined from multiple mutants. Other candidates for sensing cell-wall strain are the receptor-like kinases (RLKs) of the wall-associated kinase (WAK) family (Anderson et al., 2001). The five members of this family contain extracellular domains that bind pectin molecules of the cell wall (Kohorn et al., 2009); they span the plasma membrane and have a cytoplasmic kinase domain. A reduction in WAK protein levels limits cell expansion (Wagner & Kohorn, 2001), yet it is not known whether this effect results from reduced cell-wall synthesis. From studies on root cells it has been proposed that WAKs sense cell-wall strain by virtue of their attachment to pectin, and transduce this information to intracellular enzymes that modulate solute concentrations, which could in turn alter the turgor pressure as a major driver of expansion growth (Kohorn et al., 2006).
During expansion, cell walls need to be loosened in a controlled manner to allow their extension without compromising their integrity (Cosgrove, 2005). Cellulose is the major load-bearing component of cell walls; long cellulose microfibrils that are connected by flexible matrix polysaccharides provide cell-wall strength, while gel-forming pectins act as ‘plasticizers’ between the microfibrils (Szymanski & Cosgrove, 2009). Cell-wall extensibility can be modified by adjusting the deposition of matrix polysaccharides that associate with cellulose, as well as altered secretion of wall-modifying enzymes that influence the cellulose–xyloglucan network (Szymanski & Cosgrove, 2009). Proteins of the expansin family promote cell-wall loosening by breaking the noncovalent bonds between cellulose microfibrils and matrix polysaccharides and allow for cell and organ growth by expansion (Cho & Cosgrove, 2000; Zenoni et al., 2004). They are most effective in doing so at the transition from cell proliferation to cell expansion (Sloan et al., 2009), yet how the expression and activities of expansins and other wall-loosening enzymes are regulated remains largely unknown.
Much of the information about the role of cell walls in growth relates to postmitotic expansion, and investigating whether signalling from the cell wall also influences cell division will be an interesting area for future research.
III. Timing is everything: what determines when proliferation gives way to expansion?
As outlined in the Introduction, at the level of individual organs overall growth is initially underpinned by cell proliferation, which gradually gives way to growth by postmitotic cell expansion. A number of factors act to modify the timing of this transition, and this appears to be a major determinant of final organ size, setting the cell number and thus future expansion potential of the organ (Fig. 1). The process of proliferation arrest has been described as an arrest front that travels from the tip to the base of the leaf (Donnelly et al., 1999; Nath et al., 2003); however, recent measurements in Arabidopsis suggest that the distal region merely grows out of a field with proliferative competence that has a relatively constant size and is anchored at the leaf base (Kazama et al., 2010).
1. Genes controlling the timing of proliferation arrest
The TCP transcription factors are key regulators of cell proliferation in growing organs and the balance between the growth-promoting class I factors and the negatively acting class II TCPs has been proposed to regulate the arrest of proliferative growth (Li et al., 2005a).
The role of class II TCP genes was identified by the cincinnata (cin) mutant in Antirrhinum and jaw-D mutant in Arabidopsis. These plants show defects in leaf shape, size and curvature, with larger, rounder and more crinkled laminas than wild type (Nath et al., 2003; Palatnik et al., 2003). In mutant leaves, arrest of proliferation in distal leaf regions is delayed, and particularly the cells at the leaf margin proliferate for too long, leading to the wavy leaf phenotype due to buckling of the excess tissue out of the plane of the lamina. CIN encodes a class II TCP protein that is expressed proximal to the arrest front, while in jaw-D five TCP genes are downregulated owing to overexpression of the microRNA miR319 that targets the affected TCP mRNAs for degradation. In addition reducing the expression of three more class II TCPs by a second miRNA leads to even more severe leaf overgrowth in Arabidopsis (Efroni et al., 2008). Conversely, gain-of-function mutations that render TCP4 mRNA resistant to miR319 degradation result in severe developmental defects, including seedling lethality (Palatnik et al., 2003), and disruption of the miR319a gene in Arabidopsis reduces petal growth because of overaccumulation of TCP4 mRNA (Nag et al., 2009). Mutations in the mRNA of the class II TCP member Lanceolate of tomato that prevent its cleavage by miR319 reduce leaf size and complexity, owing to premature differentiation of leaf margin cells (Ori et al., 2007). Based on comparative studies on Solanaceae species with different leaf morphologies, it has been proposed that variation in the timing and spatial dynamics of class II TCP expression during leaf maturation underlies phenotypic variation of leaf sizes and shapes (Shleizer-Burko et al., 2011).
The effect of class II TCPs on cell proliferation in leaves appears to be mediated at least in part by inducing the expression of another microRNA, miR396 (Rodriguez et al., 2010). miR396 targets seven of the nine members of the GROWTH-REGULATING FACTOR (GRF) gene family, encoding plant-specific transcription factors involved in organ size control (van der Knaap et al., 2000). The GRF genes regulate leaf size largely by modulating cell proliferation, as shown by the reduced cell number and organ size in grf5 and grf1 grf2 grf3 triple mutants; by contrast, overexpression of GRF1, GRF2 or GRF5 increases leaf and cotyledon area (Kim et al., 2003; Kim & Kende, 2004; Horiguchi et al., 2005). miR396 accumulates during leaf development and restricts GRF expression and proliferation activity to increasingly more proximal regions of the leaf, thus contributing to the observed pattern of proliferation arrest (Rodriguez et al., 2010).
The GRF proteins bind to members of a putative transcriptional coactivator family called GRF-INTERACTING FACTORs (GIFs) (Kim & Kende, 2004; Horiguchi et al., 2005). One of the three Arabidopsis GIF genes is ANGUSTIFOLIA3 (AN3), whose loss-of-function phenotype is characterized by narrow leaves consisting of fewer cells (Horiguchi et al., 2005). Like the GRFs, GIF genes act redundantly in organ size control. Loss of GIF function in multiple mutants reduces both the rate and period of proliferation, leading to smaller leaves and flowers (Kim & Kende, 2004; Horiguchi et al., 2005; Lee et al., 2009).
In addition to the primary TCP-dependent arrest front that terminates cell proliferation in the bulk of leaf epidermal and mesophyll cells, a second arrest front targets populations of cells that continue proliferating for longer, such as the vascular procambium cells and the meristemoids in the stomatal lineage. This second front requires the activity of the duplicated PEAPOD (PPD) genes encoding putative DNA-binding proteins (White, 2006). While overexpression of PPD leads to a reduction in lamina size by promoting the early arrest of meristematic cell proliferation, deletion of both genes increases leaf lamina size and results in dome-shaped rather than flat leaves. In this case, the curvature results from excess growth of the lamina compared with the leaf margin. How the PPD genes themselves are regulated is not known and identification of both upstream and downstream components will be of great interest. The arrest of proliferation and the entry into differentiation by cells in the stomatal lineage also requires Arabidopsis RBR1, as downregulation of RBR1 leads to overproliferation of leaf epidermal cells expressing a marker of the stomatal lineage (Borghi et al., 2010). Whether and how PPD and RBR1 interact to effect the second proliferation arrest front remains an open question.
The auxin-induced ARGOS (AUXIN-REGULATED GENE INVOLVED IN ORGAN SIZE) gene contributes to regulating the timing of proliferation arrest (Hu et al., 2003). ARGOS encodes a novel, plant-specific protein, that acts upstream of AINTEGUMENTA (ANT), encoding a member of the AP2/ERF transcription factor family. ARGOS promotes growth by stimulating ANT expression; ANT activity maintains the proliferative potential of cells in leaves and floral organs, with loss or gain of function leading to reduced or increased lateral organ size, respectively (Krizek, 1999; Mizukami & Fischer, 2000; Hu et al., 2003). ANT expression is negatively regulated by AUXIN RESPONSE FACTOR2 (ARF2). ARF2 limits organ growth by proliferation, with arf2 mutants forming thicker stems and larger organs with persistent ANT expression (Schruff et al., 2006).
The period of proliferative growth is also limited by the BIG BROTHER (BB) gene whose overexpression leads to a reduction in growth that is tightly correlated with BB mRNA levels (Disch et al., 2006). Plants lacking BB activity show a strong increase in floral organ size because of a prolonged period of cell proliferation. BB encodes an E3 ubiquitin ligase. Loss of function of the DA1 gene encoding a putative ubiquitin-binding protein leads to a very similar phenotype with larger organs consisting of more cells (Li et al., 2008), suggesting that both proteins limit cell proliferation in lateral organs by targeting positive growth stimulators for protein degradation; however, the substrates for either of these proteins are unknown at present.
2. How is proliferation terminated at the right time?
Despite this progress in identifying factors that modify the timing of proliferation arrest, it remains an open question how the timing of this transition is set, and whether and how proliferation arrest is coupled to the primordium reaching a certain size. In principle, at least three scenarios are conceivable for how the arrest could be timed, that is, how proliferative growth might be measured and used to decide for or against further proliferation: The time since primordium initiation might be measured and proliferation terminated after a given period; the number of cell divisions might be counted; or primordium size itself might be monitored. Intuitively, a mechanism based on a defined ‘proliferation period’ seems less attractive than one of the other two scenarios, as it would be sensitive to variation in the rate of proliferation, and constructing a robust regulatory mechanism would then require additional assumptions about how to modify the ‘proliferation period’ in accordance with the rate of proliferation. By contrast, the number of divisions and thus of cells generated is tightly correlated with the primordium size (assuming that variation in cell size is reasonably small). Thus, measuring either cell number or the target parameter (primordium size) itself would be a more elegant and potentially more robust solution. The difference between these last two scenarios touches on a long-standing debate in plant biology about the relative importance of control at the cellular or at an organ-wide level for plant morphogenesis (Kaplan & Hagemann, 1991; Tsukaya, 2002, 2003).
Molecular mechanisms for how cell divisions might be counted have been proposed, for example involving the gradual accumulation of a cell-cycle inhibitor (Conlon & Raff, 1999). For plants this could mean for example that the activity of factors such as class II TCPs, BB or DA1 increases with successive cell divisions; however, there is currently little experimental evidence for such a counting mechanism. An elegant model for how overall primordium size could be measured and used to determine proliferation arrest has been developed for the Drosophila wing (Hufnagel et al., 2007); at a certain primordium size, the peripheral cells would no longer receive sufficient amounts of a growth factor secreted from the centre, and the mechanical stress resulting from their proliferation arrest – while the centre was still growing – would feed back to terminate proliferation throughout the primordium. For plants, a related scenario has been proposed based on the progressive dilution of a growth signal generated at the organ’s periphery (Anastasiou et al., 2007); yet subsequent characterization of the range of action of this signal argues against its use for the organ-autonomous measurement of primordium size (Eriksson et al., 2010). Thus, answering how a plant ensures that proliferation only arrests once an appropriate number of primordium cells have been generated remains an important challenge in the field. Given the added complexity that proliferation arrests in a defined spatial pattern, combining genetics with live imaging and modelling will likely be required to answer this question.
Although in many mutants a precocious or delayed arrest of proliferation translates into corresponding changes in final organ size, this is not always true, and variation in final size is often smaller than differences in cell numbers (Tsukaya & Beemster, 2006; Tsukaya, 2008). Thus, a change in cell number can be at least partially offset by an opposite change in cell expansion. This phenomenon has been termed ‘compensation’ or ‘compensated cell enlargement’ (Tsukaya & Beemster, 2006; Tsukaya, 2008). Representative examples include mutants lacking the entire CYCLIN D3 gene family in Arabidopsis, where a premature arrest of cell proliferation and reduced leaf cell number are compensated for by more extensive endoreduplication and cell expansion, so that final organ size is unchanged (Dewitte et al., 2007). Similarly, loss of FIZZY-RELATED2 (FZR2) function, which encodes a stimulator of endoreduplication, results in normal-sized leaves composed of more, but smaller cells, while moderate overexpression causes the opposite cellular changes, again without a difference in overall organ size (Larson-Rabin et al., 2009).
Compensation is not always so ‘perfect’ and also does not operate in all genetic backgrounds, as otherwise no mutants with altered organ size could have been recovered. This has led to the suggestion that cell number in a leaf needs to fall below a threshold value for compensation to result (Fujikura et al., 2009). How the defect in cell proliferation is linked with postmitotic cell expansion is still unclear, and different hypotheses involving either active monitoring of cell numbers and signalling or not have been proposed (Tsukaya, 2008; Breuninger & Lenhard, 2010). Arguably the strongest evidence for an active signalling process that induces the excess cell enlargement comes from elegant clonal analyses of compensation seen in an3 mutants (Kawade et al., 2010). When cells with an active copy of AN3 coexist with an3 mutant cells in the same half of a chimeric leaf, they exhibit full compensation, suggesting that the proliferation defect of an3 mutant cells leads to the generation of a mobile signal that induces compensated cell enlargement; however, this signal seems to be restricted to within one-half of a leaf (Kawade et al., 2010). Several factors have been genetically identified that are required for compensation in an an3 mutant background, and analysis of single and double mutant phenotypes suggests that compensation in an3 involves both the enhanced activity of factors that are required in ‘normal’ cell expansion in wild type, but also the activity of genes that are not necessary during wild-type cell expansion (Fujikura et al., 2007). A better understanding of this fascinating, but enigmatic process is eagerly awaited.
4. Is proliferation arrest correlated with the source–sink transition in leaves?
The distal-to-proximal arrest of cell proliferation in leaves is mirrored by the physiological sink–source transition (Oparka et al., 1999). As the leaf differentiates, cells in the distal region begin to produce an excess of photoassimilates and to act as a source, exporting these assimilates to sink tissue where growth and the consumption of assimilates exceeds local photosynthesis. The boundary between source and sink regions within one leaf then moves towards the proximal leaf domain, until eventually the whole leaf acts as a source. Given this similarity, it is conceivable that the movement of these two ‘transition fronts’ is functionally connected. However, to our knowledge, it is currently not clear what the exact temporal relationship between these two transitions is within an Arabidopsis leaf.
Plants are thought to continuously monitor their sugar status to optimize the use of available sugar for growth and development (Smeekens et al., 2010). A number of key regulators in the signalling network that links sugar availability to growth have been identified. The Arabidopsis hexokinase (HXK1) glucose sensor acts to promote growth and hxk1 (glucose insensitive2, gin2) mutants have reduced shoot and root growth (Cho et al., 2006). Trehalose 6-phosphate is a novel growth-promoting signal in plants, and its levels correlate with sugar availability (Paul et al., 2010). At physiological levels T6P can inhibit the SNF1-related Protein Kinase1 (SnRK1) (Zhang et al., 2009); SnRK1 activity in turn inhibits growth and its function is important to safeguard cellular energy levels during sugar starvation. Intriguingly, activation of SnRK1 in response to sugar limitation downregulates the expression of a set of genes involved in ribosome biogenesis, providing a link between sugar signalling and cytoplasmic growth (Baena-Gonzalez et al., 2007). In addition, the C/S1 bZIP transcription factors are growth inhibitors whose activity is repressed by sucrose (Hanson et al., 2008; Smeekens et al., 2010). These findings point to complex and diverse interactions within and between the systems monitoring sugar status and growth-control pathways.
IV. Anisotropic growth and the importance of polarity
Growth of plant organs is frequently anisotropic, that is, the rate of growth in one direction exceeds that in another. Together with the geometry of the early organ primordium the extent of anisotropic growth contributes strongly to the final shape of an organ. This is exemplified by classical studies on differences in fruit shape in squashes and gourds. In Cucurbita pepo differences in mature fruit shape were already preshadowed in very young primordia, yet the ratio between growth in length and in width (i.e. the allometric constant) was the same between the two varieties (Sinnott & Durham, 1929; Sinnott, 1936). By contrast, in different strains of Lagenaria the primordia were similar in shape, and it was rather a difference in the allometric constant that led to the different mature fruit shapes between the two lines (Sinnott, 1936). However, in both cases the variation in shape was inherited in a simple Mendelian manner independently of variation in final fruit size, indicating that only a single-gene difference in a dedicated ‘shape’ gene was responsible (Sinnott, 1935, 1954).
Insight into the genetic control of organ shape and anisotropic growth has been gained from the analysis of Arabidopsis mutants. Loss of ANGUSTIFOLIA (AN) function leads to a distinctive narrow-leaf phenotype. While very young primordia are indistinguishable from wild type, a reduced ratio of growth in width vs growth in length causes a narrow leaf (Tsuge et al., 1996). At the cellular level, the changed allometric constant is caused by reduced cell expansion in the leaf-width direction, which in turn correlates with changes in the cortical microtubule cytoskeleton (Kim et al., 2002). The AN gene encodes a protein related to C-terminal binding protein (CtBP)/brefeldin A ribosylated substrate (BARS) (Folkers et al., 2002; Kim et al., 2002). The opposite defect to that seen in an mutants is shown by plants lacking ROTUNDIFOLIA3 (ROT3) function, encoding a cytochrome P450 involved in brassinosteroid (BR) biosynthesis (Kim et al., 1998, 2005). Again, the shorter, more rounded final leaf results from a change in the allometric constant (Tsuge et al., 1996), and at the cellular level reduced cell expansion in the leaf-length direction. Two novel Arabidopsis genes LNG1 and LNG2 also promote longitudinal polar cell elongation and thus leaf growth in length, acting independently of ROT3 (Lee et al., 2006).
Organ shape and size appears to be regulated by both cellular and supracellular mechanisms (see section III.3). Clonal analysis of petal growth in Antirrhinum has suggested long-range polarity as a possible mechanism for the supracellular control of shape by orienting local growth directions (Rolland-Lagan et al., 2003). The question of whether genes control organ shape by affecting local rates of growth or by changing the polarity of growth across the tissue has recently been addressed using a combination of mathematical modelling with growth analysis in different genetic backgrounds of Antirrhinum (Green et al., 2010). A generative model in which shape genes only affected the local rates of tissue growth in independently determined orientations was not fully able to reproduce the mature Antirrhinum flower shape; however, this was possible when the shape genes also modified tissue polarity and thus the orientations of growth. These results highlight not only the importance of tissue polarity for the development and evolution of form, but also the power of modelling to understand complex four-dimensional processes such as growth.
V. How does organ identity and developmental patterning modulate growth behaviour?
1. Organ identity and growth
Size and shape are two of the main criteria by which we judge the identity of an organ, for example leaf vs sepal. However, little is known about how organ identity is translated into specific growth patterns to generate the characteristic organ shapes and sizes. The observation that – but for a few exceptions (Nath et al., 2003; Crawford et al., 2004; Disch et al., 2006) – most mutants in growth-regulating genes show comparable changes in leaves and floral organs indicates that the same basic growth machinery acts in both types of organs, consistent with their known homology. Thus, a plausible scenario is that organ identity factors, such as the homoeotic MADS-box genes, modulate the activity of this common growth machinery (Dornelas et al., 2011). In fact, several growth-related genes, for example, TCP and GRF genes or a NAC-type transcription factor influencing the transition from proliferation to expansion, have been identified among the potential direct targets of homoeotic transcription factors (Sablowski & Meyerowitz, 1998; Wellmer et al., 2004; Gomez-Mena et al., 2005; Kaufmann et al., 2010; reviewed in Dornelas et al., 2011); however, how these possible regulatory connections impinge on aspects of size and shape is not fully understood. A notable exception is the basic helix–loop–helix transcription factor, BIGPETAL (BPE), whose petal-specific isoform is required to limit cell expansion in petals (Szecsi et al., 2006). This isoform is generated by alternative splicing in response to the activities of A- and B-class homoeotic genes. (Note that this example does not fully fit with the hypothesis, as for the ubiquitously expressed isoform so far no role in growth control could be demonstrated.) Jasmonate biosynthesis appears to be an intermediate step, as key jasmonate biosynthetic genes are downregulated in B-function mutants, and lack of jasmonates blocks the alternative splicing leading to the petal-specific isoform of BPE (Brioudes et al., 2009).
A promising alternative approach to identifying the critical links between organ identity and growth may be the analysis of natural variation in organ size and shape, as this variation is often highly organ-specific. For example, flower size has often changed dramatically in conjunction with a change in the reproductive system, while at least in some cases leaf size has remained unaltered (Sicard & Lenhard, 2011). Also, quantitative trait loci (QTL) analyses of variation in leaf and floral-organ size in Arabidopsis and tomato indicate an independent genetic basis for changes in the size of the different organs (Frary et al., 2004; Jünger et al., 2005). It will be exciting to see what cloning the genes underlying such QTL can teach us about how organ identity determines growth patterns.
2. Regionalization, developmental patterning and growth
Lateral organ development occurs along three axes: proximo-distal, medio-lateral and adaxial-abaxial (Hudson, 2000). A number of genes involved in the patterning of these regions also show defects in organ size, suggesting they may also have a role in growth regulation. The recently identified NGATHA (NGA) family of RAV-B3 type transcription factors provides one such example (Alvarez et al., 2009). The four Arabidopsis NGA genes are redundantly required for apical–basal patterning of the gynoecium; in addition, nga mutants form larger organs because of increased cell proliferation.
The role of the BLADE-ON-PETIOLE1 (BOP1) and BOP2 genes in regulating leaf morphogenesis and patterning (Ha et al., 2003, 2004, 2007) provides another illustrative example of a link between patterning and proliferative growth. Combined loss of BOP1 and BOP2 function leads to ectopic outgrowth of blade tissue along leaf petioles, indicating a defect in proximal-distal pattern formation. BOP1 and BOP2 repress the expression of KNOTTED-LIKE HOMEOBOX (KNOX) genes and of JAGGED (JAG) and its partially redundant homologue NUBBIN (NUB). JAG and NUB are required for proliferative growth in leaves and floral organs, and JAG overexpression causes the ectopic growth of lamina tissue or entire organs (Dinneny et al., 2004, 2006). KNOX genes maintain cells in an undifferentiated state, and in species with compound organs maintain the organogenic capacity of the leaf margin (Hay & Tsiantis, 2009). Thus, by repressing key factors promoting proliferative growth, BOP activity directly links patterning and growth control.
VI. Coordination of growth at different scales
For a population of cells to generate an organ of defined shape and size or even a collection of functionally integrated organs, their growth should be coordinated. Such coordination is evident at a number of levels, and progress has been made at all of these in understanding the genetic basis.
1. Growth coordination between tissue layers and different regions within an organ
The relative importance of the different clonally distinct layers within a plant lateral organ for determining its overall growth has been a matter of some debate (Szymkowiak & Sussex, 1996; Marcotrigiano, 2001; Savaldi-Goldstein & Chory, 2008). Experimental results vary between organs, species and the developmental marker studied; yet several lines of evidence indicate that the epidermis, which forms a continuous, clonally distinct layer stretched over the organ, regulates organ size and shape by promoting or restricting growth of the underlying tissue, particularly in organs with an extended lamina (such as sepals and petals). A key study in this argument is based on manipulating BR signalling specifically in the epidermis (Savaldi-Goldstein et al., 2007). Depletion of BR from the epidermis was sufficient to cause dwarfing while, conversely, the restoration of BR sensitivity only in the epidermis of a dwarf BR-insensitive mutant was sufficient to rescue normal organ growth. This result is consistent with a number of clonal and laser ablation studies, leading to the proposal that a noncell autonomous signal from the epidermis controls the growth of underlying cells, allowing the epidermis to drive and restrict the growth of shoots and lateral organs (Marcotrigiano, 2001; Reinhardt et al., 2005). Some evidence also exists for growth coordination by signalling in the opposite direction, that is from the inner tissue to the epidermis. For example, when AN was only expressed in the epidermis of an mutants, it was unable to rescue leaf shape; conversely, AN activity only in subepidermal tissues can complement the leaf-width defect of an mutants, and this rescue is accompanied by an increase of epidermal cell number (Bai et al., 2009). Proposed molecular mechanisms underlying the coordination of plant growth across layers include the intercellular movement of developmentally important transcriptional or cell-cycle regulators (Vincent et al., 2003; Weinl et al., 2005) and the perception of diffusible growth substances or signalling molecules either at the plasma membrane or within cells.
As plant cells are ‘glued’ to each other by their middle lamellae and cell walls, local differences in growth rates will lead to the build-up of stress in a tissue and – if possible – to deformation that relieves this stress. Prominent examples of such deformations resulting from an imbalance in growth are the opposite changes to leaf shape in plants lacking either class II TCP function or the PPD genes (see section III.1). Thus, these factors are required to coordinate growth of the centre and the periphery of an organ, yet how exactly they do so is currently unclear.
Similarly, differential growth between the adaxial and abaxial side of an organ will result in upward or downward curling. A recent example illustrating this point is the dominant incurvata6 (icu6) mutant allele of the AUXIN-RESISTANT3 gene encoding an Aux/IAA family member that represses auxin response (Perez-Perez et al., 2010). Stabilization of the AXR3 protein by this mutation results in reduced cell expansion specifically in the adaxial leaf epidermis, which in turn causes upward leaf curling.
2. Growth coordination between different organs
For the majority of flowering plants reproduction depends on animal pollinators. Flowers need to appear attractive to the pollinators, and the different reproductive organs have to be positioned correctly to ensure efficient pollen transfer between individuals of the same species. Thus, growth of the individual floral organs needs to be coordinated to ensure that they all reach their appropriate relative size. By analogy to the timing of proliferation arrest in growing organs, it is conceivable that the individual organs of a flower only follow their autonomous developmental and growth programmes, which have been individually adjusted during evolution to ensure an optimal final flower shape. Alternatively, mechanisms may exist to coordinate the growth of individual organs, so as to dampen the effects of random fluctuations in growth. For example, random variation in the final size and/or shape of individual petals would render the flowers less symmetrical, which in turn would reduce their attractiveness to animal pollinators (Moller, 1995).
The presumed growth-promoting signal generated by the activity of the cytochrome P450 KLUH (KLU)/CYP78A5 has the potential to act in a regulatory mechanism that would coordinate the growth of individual organs. Increased activity of KLU causes organ overgrowth, while klu mutants form smaller aerial organs consisting of fewer cells, and this effect is further enhanced by mutating the homologous CYP78A7 gene (Anastasiou et al., 2007; Wang et al., 2008). In flowers, KLU is expressed in the receptacle, at the margins of sepals and petals, and in developing ovules, from where it stimulates proliferative growth in a noncell autonomous manner. Based on transcriptomic and pharmacological evidence, KLU does not seem to influence the amounts of classical phytohormones, suggesting that KLU and CYP78A7 contribute to generating a novel mobile growth signal. In chimaeric flowers or inflorescences containing both wild-type and klu mutant organs, both types of organ grow to the same size, indicating that they do not behave autonomously according to their genotype (Fig. 2; Eriksson et al., 2010). This suggests that the level of the putative growth signal is integrated across an inflorescence. Thus, movement of a KLU-dependent signal throughout an inflorescence could be used to coordinate the growth of individual organs and buffer the flowers and the inflorescence against asymmetries as a result of random growth perturbations (Eriksson et al., 2010).
3. Growth coordination between shoot and root
Across a wide range of seed plants, total shoot biomass is directly proportional to total root biomass (Enquist & Niklas, 2002), while at the individual level shoot growth can be dynamically modulated in response to environmental conditions perceived by the root (reviewed in Sieburth & Lee, 2010). Thus, growth is also coordinated at the whole-plant level, and root-derived long-distance signals are important for adjusting shoot growth to water and nutrient conditions in the soil. This is exemplified by the BYPASS1 (BPS1)-modulated signal in Arabidopsis (Van Norman et al., 2004; Van Norman & Sieburth, 2007). This carotenoid-derived signal molecule is transported from root to shoot through the xylem, and limits the growth of young leaves by acting on both cell division and expansion. bps1 mutants overproduce the signal in roots, leading to nonautonomously arrested shoot growth, which correlates with reduced auxin responses in the affected leaves.
4. The role of phytohormones in growth coordination
Plant hormones are prominent candidates for controlling growth in a coordinated manner (Wolters & Jurgens, 2009). Detailed information regarding hormone signal transduction and biosynthesis is available and discussed in numerous excellent reviews (Harberd et al., 2009; Vanneste & Friml, 2009; Kim & Wang, 2010; Perilli et al., 2010). Auxin has a prominent role in the positioning of new organs and establishing the polarity of growth. Also, a number of auxin-responsive genes have been identified in organ growth control, for example ARGOS as a stimulator of proliferation (see section III.1). A role for auxin in organ size regulation is indirectly supported by the leaf overgrowth in plants overexpressing the H+-pyrophosphatase AVP1, which show prolonged cell proliferation and thus higher cell numbers in leaves (Li et al., 2005b). AVP1 contributes to the regulation of apoplastic pH and to auxin transport, presumably by mediating the trafficking of the plasma membrane ATPase and associated proteins, including the PIN-FORMED1 auxin efflux carrier (Li et al., 2005b). In addition, a role for auxin in promoting expansion growth is well established and summarized in the acid growth hypothesis, according to which auxin-induced acidification of the cell wall stimulates cell wall extension by activating expansins and other cell-wall loosening enzymes (Hager, 2003).
Brassinosteroids largely act to promote cell elongation in organs (Kim & Wang, 2010). The action of auxin and BR in growth control appears to converge on ARF2 (Vert et al., 2008). The BR-regulated kinase BIN2 can phosphorylate ARF2 in vitro, inhibiting its ability to bind DNA and repress transcription, and thus possibly potentiating auxin signalling via transcription-stimulating ARFs. In addition, the BREVIS RADIX protein appears to mediate crosstalk between the auxin and brassinosteroid pathways not only in root growth, where it was first characterized, but also in shoot growth (Mouchel et al., 2004; Beuchat et al., 2010).
Cytokinins promote cell proliferation and the growth of shoot organs (Wolters & Jurgens, 2009; Perilli et al., 2010). Recent findings indicate that cytokinins are not only required for growth, but that their level in wild-type is limiting for organ size. Increased cytokinin levels in mutants lacking cytokinin-degrading enzymes lead to significantly larger organs because of increased cell proliferation, indicating that control over cytokinin levels could be used to regulate final organ size (Bartrina et al., 2011).
The mechanisms by which GA control aspects of plant growth and development have also been extensively studied and involve the DELLA transcription factors (Harberd et al., 2009). A link between GA signalling and patterning-dependent organ growth is demonstrated by the tomato procera (pro) mutant that represents a loss of DELLA protein function. The pro mutant leaves form fewer leaflets, suggesting a reduced competence to respond to KNOX signalling (Jasinski et al., 2008). The increased cell size in pro organs raises the interesting possibility that cell growth or cell size may modulate the activity of developmental regulators. The role of GA in stress responses also has direct implications for growth (see section VI.5).
5. Coordination of growth with other physiological functions
As part of their stress response, plants actively reduce their growth through a ‘short-term adjustment’, followed by a ‘long-term adaptation’ to the new conditions (for example West et al., 2004; Fricke et al., 2006). Most studies have focused on the resulting growth changes that allow plants to save and redistribute resources that can become limiting; for example, smaller leaves lose less water because of a reduced transpiration area, while differential growth recovery leads to beneficially higher root-to-shoot ratios (Hsiao & Xu, 2000).
The underlying mechanisms regulating these changes are not fully understood, and only a few genes and metabolites are known that are involved in the regulation of leaf growth under adverse environmental conditions (Granier & Tardieu, 2009). Abscisic acid, ethylene, and DELLA proteins have been implicated in stress tolerance, with GA having a prominent role. Different abiotic and biotic stresses reduce GA levels and stabilize DELLA proteins, which in turn restrict plant growth by limiting both cell proliferation and expansion and activate tolerance mechanisms (Achard et al., 2006, 2008, 2009; Navarro et al., 2008).
At the population-genetic level, a trade-off between defence and growth was recently demonstrated by studying the allelic diversity at the ACCELERATED CELL DEATH 6 (ACD6) locus. Variation in ACD6 activity is responsible for differences in vegetative growth and resistance to microbes and herbivores among Arabidopsis accessions (Todesco et al., 2010). Plants with a hyperactive ACD6 allele are less susceptible to various pathogens; however, as a pleiotropic side-effect they form leaves more slowly, and those leaves reach a smaller final size. The intermediate frequency of this hyperactive allele among both the global and local Arabidopsis populations suggests that under certain conditions it is advantageous to stay smaller, but be more resistant to pathogen attack (Todesco et al., 2010).
The genetic control of organ size is complex, involving a large number of genes. Do these act in many independent pathways, as is the current picture, or are there functional connections still to be discovered? A recent study looked for such connections between five regulatory factors representing various different hormone and growth-control pathways. Transgenic lines overexpressing the AVP1, BRASSINOSTEROID-INSENSITIVE1, GRF5, GA20-OXIDASE1 or JAW (encoding miR319) genes all produce larger leaves. However, they do so in different ways, as judged from transcriptomic and metabolomic comparisons, suggesting that the represented pathways act independently on organ growth (Gonzalez et al., 2010). Thus, final organ size may simply be the sum of all the positively and negatively acting independent effects, with growth as the ultimate integrative phenotype.
In conclusion, work over the recent years has greatly expanded our understanding about the genetic control of organ growth in plants, with many of the factors involved having been identified by forward and reverse genetics in models such as Arabidopsis and Antirrhinum. Future directions that will be important for answering remaining questions are: the increasing use of combined imaging and modelling approaches to investigate the dynamics of growth in different genetic backgrounds (Grieneisen & Scheres, 2009; Chickarmane et al., 2010; Green et al., 2010); the development of precise high-throughput phenotyping platforms to address the interaction of genes and the environment in growth control (Jansen et al., 2009; Berger et al., 2010); and the exploitation of natural variation in organ size and shape to determine how evolution has modified size in often highly organ-specific ways without pleiotropic effects (Frary et al., 2000; Chen et al., 2007). Combined with more traditional genetic approaches, work in these areas will deepen our understanding of the genetic control of organ growth, which should serve as a solid basis for the rational manipulation of plant growth and yield in response to the manifold challenges we face.
We thank members of the Lenhard laboratory for critical reading and comments on the manuscript. We apologize to all colleagues whose work could not be discussed because of space constraints. This work was supported by grants from the Biotechnology and Biological Sciences Research Council (BB/G001421/1) and the Deutsche Forschungsgemeinschaft (DFG Le1412/3-1).