Dispatch R555

Plant development: Keeping your distanceColin Brownlee

The spatial distribution of stomatal cells in the plantepidermis is generated by stereotyped cell divisions.Recent evidence suggests that a subtilisin-like proteaseis involved in controlling these cell-division patterns; itmay play a role in processing a signal peptide orreceptor involved in cell-to-cell communication.

Address: Marine Biological Association, The Laboratory, Citadel Hill,Plymouth, PL1 2PB, UK.E-mail: cbr@mba.ac.uk

Current Biology 2000, 10:R555–R557

0960-9822/00/$ – see front matter © 2000 Elsevier Science Ltd. All rights reserved.

The outer skin of a higher plant consists of a single layerof cells known as the epidermis. The epidermal layer isestablished during embryogenesis and is formed continu-ally by cell division in meristems in the developing apicesof the root and shoot. The epidermis expands by celldivision and cell elongation as the plant grows. In additionto the relatively unspecialised epidermal cells, morehighly differentiated cell types can be found in this tissuein both root and shoot organs. In the root epidermis,specific cells can differentiate into root hairs that playimportant roles in nutrient acquisition. In the aerial partsof the plant, specific epidermal cells can differentiate intoeither polarized hair-like structures known as trichomes orstomatal guard cells, which form bicellular pores in theleaf surface through which water and gasses can beexchanged with the atmosphere. A recent study [1] hasshed new light on the mechanisms by which the non-random spatial pattern of stomatal cells in the plantepidermis is generated.

The correct distribution of these different epidermal celltypes is obviously important for their physiologicalfunction. The relatively simple, two-dimensional patternsof cells in a single layer also present attractive systems forstudying patterning mechanisms in plants generally. It hasbeen known for some time that the distribution of special-ized epidermal cells is non-random and that a minimumspacing exists between these cells [2], but until recentlythe mechanisms underlying spatial patterning have beenobscure. It is also clear that there are fundamental differ-ences in the ways that different epidermal cell types arise(Figure 1). Whether or not a cell in the root epidermisbecomes a root hair is governed by its position relative tounderlying cells in the cortical layer, and while epidermalcell lineages are established during embryogenesis thefate of the epidermal cells can be manipulated by alteringtheir position following ablation of neighbouring cells.

This is thought to involve short-range signalling fromcortical to epidermal cells [3]. Trichome cells in Arabidop-sis leaf epidermis arise from a change in fate of single epi-dermal cells. A developing trichome cell is not necessarilyclonally related to its surrounding cells [4], and spacingappears to result from lateral inhibition of differentiationby existing or developing trichome cells.

Specific cell lineages thus do not appear to be involved inthe production of trichomes. In contrast, Arabidopsis stom-atal guard cells are produced by ordered division patternsof individual precursor (protodermal) cells, giving rise todifferentiated guard cells surrounded by clonally relatedepidermal cells. This generates the normal pattern ofguard cell distribution, in which each guard cell pair is

Figure 1

Regulation of the spatial pattern of epidermal cell types in Arabidopsis.(a) Trichome spacing appears to involve lateral inhibition of newtrichome formation by existing or developing cells. (b) Root hairs formin precise locations relative to the intercellular space betweenunderlying cortical cells. (c) Stomatal guard cells form from ordereddivision patterns of precursor epidermal cells, so that each guard cellpair is surrounded by clonally related epidermal cells.

(a)

(b)

(c)

Current Biology

separated from another pair by at least one epidermal cell.The key to understanding how the normal non-randomdistribution of stomatal guard cells is generated thus liesin understanding the factors controlling the pattern of celldivision and differentiation that gives rise to muticellularstomatal lineage units from single protodermal cells.

Several external factors can influence stomatal density,defined as the stomatal index — the number of stomatarelative to the total number of epidermal cells. Theseinclude humidity, the atmospheric CO2 concentration andlight. Varietal differences within a species suggest thatendogenous factors also control stomatal density. Disrup-tion of stomatal patterning, resulting in the production ofguard cell pairs in direct contact without interveningepidermal cells (clustering), can be brought about by highhumidity or by interfering with production of the planthormone ethylene, suggesting that signalling may beinvolved in stomatal patterning.

Arabidopsis mutants, such as too many mouths (tmm) andfour lips (flp), have been identified that also give rise tostomatal clustering [5]. A recent paper [1] describing anew stomatal clustering mutant, stomatal density anddistribution (sdd1-1), has provided some new clues to themechanisms controlling the cell-division patterns that leadto the differentiation of stomatal cells and their correctspacing. During stomatal development, asymmetricdivisions from a single protodermal precursor cell arefollowed by a symmetrical division that forms the twoguard cells of the stomata (Figure 2). Satellite stomata canform from further division of cells adjacent to a stomatalpore giving rise to secondary or, less frequently, tertiarystomatal complexes.

Rigid control of the division pattern normally ensures thatstomatal pores in these complexes are separated byintervening cells. Berger and Altmann [1] found that the

sdd1-1 mutant has a relatively high proportion of secondaryand tertiary complexes, and even quaternary complexes,indicating that the wild-type SDD1 gene product may be anegative regulator of stomatal complex determination.These mutants also exhibit relaxed control of the pattern ofcell division, whereby the division that gives rise tosecondary or tertiary stomatal pairs can occur in a cell imme-diately adjacent to an existing stomata or simultaneously intwo adjacent cells. This suggests that the SDD1 geneproduct is also involved in control of cell division pattern.

Map-based cloning of the SDD1 gene revealed that itencodes a 750 amino-acid protein that has clear homologyto a class of serine proteases known as subtilisins orsubtilases [6]. Subtilisins possess a substrate-binding siteand a catalytic domain comprising three regions — D, Hand S — together with an endoplasmic reticulum target-ting signal sequence. The sdd1-1 mutant arises from asingle recessive point mutation resulting in loss of the Sregion and predicted to abolish catalytic activity. Mutantplants could be rescued at least partially by complementa-tion with the SDD gene.

Subtilisins are found in prokaryotes and eukaryotes. Inprokaryotes, subtilisins are involved in protein degrada-tion, while in eukaryotes the presence of the largeP domain is associated with their roles as processingproteins. Plant subtilisins do not possess a P domain, butthere is good evidence that they also have a processingrole in signal transduction pathways. Thus, the active formof the extracellular plant signalling peptide known as sys-temin is produced from a precursor, prosystemin, by theaction of SBP50, a pro-systemin-binding protein that hasimmunological similarities to a subtilisin [7].

Subtilisin-like proteins may also be involved in regulationof cell-to-cell signalling in the control of plant meristemdevelopment. Thus, the CLAVATA genes have been

R556 Current Biology Vol 10 No 15

Figure 2

(a) (b)

Current Biology

(a) Stomatal guard cell pairs (green) normally form from a symmetricaldivision (green line) that follows asymmetric divisions (red line) startingfrom a protodermal cell. (b) In sdd1-1 mutants, this division pattern is

disrupted so that symmetrical divisions can occur in adjacent cellsgiving rise to clustered stomata.

shown to be negative regulators of meristem activity inArabidopsis shoot development [8]. The CLAVATA genesinteract with positive regulators of cell division, such asSHOOTMERISTEMLESS and WUSCHEL [9]. CLAVATA1encodes a putative receptor kinase, CLAVATA2 encodes aproposed accessory protein and CLAVATA3 encodes aputative signalling peptide with a dibasic subtilisin recog-nition processing site.

Subtilisins may thus play a more general role in regulationof plant development. It is tempting to speculate thatstomatal patterning involves similar peptide signals thatmay act over short ranges to regulate cell division and guardcell differentiation. If this turns out to be the case, then anobvious question is whether these signals emanate from theguard cells themselves to suppress the differentiation ofsurrounding epidermal cells. Identification of proteins thatinteract with the SDD1 gene product is eagerly awaited.

AcknowledgementsI am grateful to Alison Taylor for critical reading of the manuscript.

References1. Berger D, Altmann T: A subtilisin-like serine protease involved in

the regulation of stomatal density and distribution in Arabidopsisthaliania. Genes Dev 2000, 14:1119-1131.

2. Larkin JC, Marks MD, Nadeau J, Sack F: Epidemal cell fate andpatterning in leaves. Plant Cell 1997, 9:1109-1120.

3. Berger F, Haseloff J, Schiefelbein J, Dolan L: Positional informationin root epidermis is defined during embryogenesis and acts indomains with strict boundaries. Curr Biol 1998, 8:421-430.

4. Larkin JC, Young N, Prigge M, Markes MD: The control of trichomespacing and number in Arabidopsis. Development 1996,122:997-1005.

5. Yang M, Sack FD: The too many mouths and four lips mutationsaffect stomatal production in Arabidopsis. Plant Cell 1995,7:2227-2239.

6. Siezen RJ, Leunissen JAM: The superfamily of subtilisin-like serineproteases. Protein Sci 1997 6:501-523.

7. Schaller A, Ryan CA: Identification of a 50 kDa systemin-bindingprotein in tomato plasma membranes having Kex2p-likeproperties. Proc Natl Acad Sci USA 1994, 91:11802-11806.

8. Fletcher LC, Brand U, Running MP, Simon R, Meyerowitz EM:Signalling and cell fate decision by CLAVATA3 in Arabidopsisshoot meristem. Science 1999, 283:1911-1914.

9. Schoof H, Lenhard M, Haeker A, Mayer KFX, Jurgens G, Laux T:The stem cell population of Arabidospsis shoot meristems ismaintained by a regulatory loop between the CLAVATA andWUSCHEL genes. Cell 2000, 100:635-644.

Dispatch R557