On the evolution of developmental mechanisms: Divergent polarities in leaf growth as a case study

ABSTRACT

Most model plants used to study leaf growth share a common developmental mechanism, namely basipetal growth polarity, wherein the distal end differentiates first with progressively more proliferative cells toward the base. Therefore, this base-to-tip growth pattern has served as a paradigm to explain leaf growth and also formed the basis for several computational models. However, our recent report in The Plant Cell on the investigation of leaf growth in 75 eudicot species covering a wide range of eudicot families showed that leaves grow with divergent polarities in the proximo-distal axis or without any obvious polarity. This divergence in growth polarity is linked to the expression divergence of a conserved microRNA-transcription factor module. This study raises several questions on the evolutionary origin of leaf growth pattern, such as ‘when and why in evolution did the divergent growth polarities arise?’ and ‘what were the molecular changes that led to this divergence?’. Here, we discuss a few of these questions in some detail.

Leaf growth is characterized by an initial phase of cell proliferation followed by cell differentiation where growth is driven by the expansion of the differentiating cells.^1,2 Studies on model plants have shown that the phase of differentiation has a specific pattern in the proximo-distal axis wherein differentiation begins near the distal tip and proceeds toward the proximal base.^1,3 This pattern of growth is known as basipetal growth because the cells near the base continue to proliferate and cause leaf expansion for the longest duration. Since basipetal leaf growth has been observed in all the model plants studied so far,^1,3-5 it was considered universal and has therefore served as a paradigm to explain and model leaf growth in several species.^6-9 However, our recent survey of 75 eudicots¹⁰ revealed that basipetal growth is not universal and that at least 3 more distinct growth patterns exist in nature: (i) acropetal leaf growth where differentiation begins near the base and progresses toward the tip (opposite of basipetal growth); (ii) even or diffused growth where the cells begin to differentiate synchronously throughout the leaf; and (iii) bidirectional growth where differentiation begins from both extremities and progresses toward the middle of the leaf. Since all these growth patterns are essentially different forms of polar or differential growth, we used the law of simple allometry to classify the growth patterns as positive allometry (basipetal growth), negative allometry (acropetal growth), isometry (diffused/even growth) and complex allometry (bidirectional growth). We also found that the expression of miR396, a plant-specific, conserved microRNA that promotes cell differentiation, is closely linked to the direction of cell differentiation. These new findings on the leaf growth raise several interesting questions on the evolution of leaf growth patterns and promises to serve as a new model to study the evolution of developmental mechanisms.

The origin of leaf growth patterns: When and why did the divergence in leaf growth pattern evolve?

The case of evolution of floral symmetry has been studied extensively and has provided deep insights into the mechanisms of floral evolution as well as the adaptive advantage of certain floral forms over others. The phylogenetic distribution of zygomorphic flowers shows that this trait has evolved from the ancestral radial symmetry repeatedly in independent lineages by the co-option of the same sets of genes.¹¹ Here, we report that different leaf growth patterns have evolved in different lineages of plants probably by the modification of a conserved regulatory network of genes. The questions that therefore arise are: (1) what was the ancestral pattern of leaf growth and when during the evolutionary course did other patterns arise; (2) do certain growth patterns have an adaptive advantage over the others under given ecological conditions and are therefore positively selected in certain lineages? The limited phylogenetic analysis of leaf growth polarity reported in our original article suggests that isometry/negative allometry is an ancestral growth pattern while positive allometry is a derived character during evolution. The analysis also shows that positive allometry originated multiple times independently and was also probably lost secondarily in certain lineages (e.g., in Apiales), implying a role for adaptive selection. Much more detailed phylogenetic analyses would be needed to distinguish between the ancestral vs derived states and when exactly the different growth polarities arose during evolutionary time.

Of the 75 species included in our analysis, nearly half of them belong to positive allometry; 43% grew with isometry, 7% grew with negative allometry, and one species grew with complex allometry (Fig. 1). This natural diversity in growth pattern implies that leaf growth polarity has no obvious developmental constraint and perhaps is not under strong selection pressure. However, it is possible that certain growth allometry would be more favorably adapted to certain ecological conditions. To test this, we sorted the leaf growth polarities into 2 groups according to their growth habits– (i) annuals that usually grow close to the ground level and for limited duration and (ii) perennials that usually grow away from the ground and through multiple seasons. Interestingly, all 25 annual species included in the study showed positive allometry, whereas the species with perennial growth habit showed diverse allometry (Fig. 1). A bias toward positive allometry in annual species may imply some adaptive advantage of base-to-tip growth gradient in these plants because of their unique ecological niche. For example, positive allometry of annual leaves would have an obvious growth advantage over other forms if young leaves are grazed upon by animals or damaged by insect preferably from the tip. This can be tested by studying leaf growth allometry of annual species which are not grazed by mammals (for example, those with obnoxious odor) or which grow in isolated islands devoid of grazing animals. When the distal half of a leaf of the winter annual Arabidopsis thaliana was excised at an early growth stage under laboratory condition, the remaining basal half grew to nearly normal size, lending some credence to the hypothesis discussed above.^7,12

Figure 1.

Distribution of allometric growth patterns in 75 plant species sorted according to their growth habits, i.e., perennial or annual. The percentages within the circles indicate the fraction of total species belonging to a specific allomertry and the numbers inside the parentheses indicate the number of species.

The growth polarity of individual leaflets is independent of the direction of their initiation on compound leaves

Plant biologists have already noted that the direction of leaflet initiation on a compound leaf can be basipetal (younger leaflets are formed near the proximal end while the terminal leaflets are more mature), acropetal (younger leaflets are formed toward the distal end) or divergent (younger leaflets are formed at both ends).¹³ This of course is an example of the polarity of leaflet initiation on a compound leaf, which is essentially a patterning event on a leaf primordium, and does not reveal the growth polarity of individual leaflet lamina on it.

In our report in Plant Cell,¹⁰ we addressed how the individual leaflets on a compound leaf mature. Our analysis included several species with pinnately compound leaves (Table 1). Leaflets in 3 of them matured from tip-to-base and leaflets in the remaining 6 species show apolar (isometric) maturity. Initiation of leaflets on these compound leaves also showed divergent polarity; 3 of them showed basipetal while 5 showed acropetal leaflet initiation. For example, the leaflets of Solanum lycopersicum and Tecoma stans are formed in a basipetal fashion with the terminal leaflets being the most mature with successively younger leaflets toward the proximal end. The leaflets themselves, however, grow with a basipetal gradient such that the maturity starts near the tip of the leaflet and progresses toward the base (Fig. 2). In other examples, such as Cassia spectabilis, the leaflets are formed in an acropetal fashion with the terminal leaflets being the youngest, while individual leaflets mature evenly without a growth gradient along their proximo-distal axes. An example of acropetal leaflet initiation had been reported earlier in Eschscholzia californica,¹⁴ but the growth polarity of its individual leaflets was not addressed. Some of the compound leaves included in our study, e.g., Azadirachta indica, showed acropetal formation of the leaflets while the individual leaflets showed basipetal growth gradient (Fig. 2). This analysis implies that the polarity of the maturation of individual leaflets is independent of the polarity of their initiation on a compound leaf, and therefore these two polarities are likely regulated by different molecular mechanisms. We have some evidence that the expression difference of the conserved miR396-GROWTH REGULATING FACTOR module, which has been shown to pattern the lamina growth along the proximo-distal axis of simple leaves,^15,16 is linked to the divergent growth polarity of leaves/leaflets. The polarity of leaflet initiation on compound leaves, on the other hand, is possibly regulated by genes involved in meristem programs such as KNOTTED1-like homeobox (KNOX) genes^17,18 and their repressors ASYMMETRIC LEAVES1/ROUGHSHEATH2/PHANTASTICA (ARP).¹⁹ The redeployment of regulatory modules (such as the miR396-GRF) that affect lamina growth to leaflets and the fact that the individual leaflets grow as simple leaves with various growth polarities, perhaps indicate that leaflets are serial homologs created by piecemeal movement of regulatory modules.

Table 1.

A list of plant species with pinnately compound leaves that are included in our study and their leaf maturation polarities.

	Species	Polarity of leaflet formation	Growth polarity of leaflet
1	Tecoma stans	Basipetal	Basipetal
2	Solanum lycopersicum	Basipetal	Basipetal
3	Samanea saman	Basipetal	Isometry
4	Azadirachta indica	Acropetal	Basipetal
5	Cananga odorata	Acropetal	Isometry
6	Cassia spectabilis	Acropetal	Isometry
7	Cedrela odorata	Acropetal	Isometry
8	Spathodea campanulata	Acropetal	Isometry

Figure 2.

Schematic representation of growth polarities in simple and compound leaves. (A) Schematics showing the progression of cell proliferation arrest (yellow arrows) in simple leaves growing with different polarities. There is no directionality in cell division arrest in leaves growing evenly (the leaf schematic on the right). A similar schematic also appeared in our original Plant Cell article. (B) Schematic representation of the direction of leaflet formation and the growth gradient within individual leaflet on compound leaves. The black arrows indicate the direction of formation of leaflets; the yellow arrows indicate the direction of the progression of cell division arrest in individual leaflets; the red dots represent dividing cells. Note that there is no directional cell division arrest in the case of Cassia spectabilis since the cells throughout the individual leaflets exit cell proliferation simultaneously.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed