The signalling peptide EPFL9 is a positive regulator of stomatal development


Author for correspondence:
Julie E. Gray
Tel: +44 1142224407


  • The putative secretory peptides epidermal patterning factor 1 (EPF1) and EPF2 act as negative regulators of stomatal clustering and density early in Arabidopsis leaf development.
  • Here, we investigated whether the related peptide gene epidermal patterning factor-like 9 (EPFL9), which is coexpressed with EPF1 and stomatal density and distribution 1 (SDD1), also plays a role in controlling stomatal development.
  • Plants manipulated to constitutively overexpress EPFL9 showed increased stomatal density and clustering, and those manipulated to have reduced EPFL9 expression showed reduced stomatal density with no clustering, confirming that EPFL9 is a regulator of stomatal development. Genetic analysis was consistent with EPFL9 acting independently of EPF1 to control stomatal clustering, independently of EPF2 to regulate stomatal density, and independently of SDD1 to control both stomatal clustering and density.
  • These findings demonstrate that at least three secretory peptides independently regulate stomatal development. Surprisingly, EPFL9 acts to increase, rather than decrease, stomatal density and clustering. However, in common with EPF1 and EPF2, EPFL9 is unlikely to be a substrate for proteolysis by SDD1.


Peptides have been regarded as an important class of intercellular signalling molecules in animals since the characterization of insulin in the early 1920s (Banting & Best, 1922). More recently, extracellular peptide signals have emerged as intermediates in a variety of plant signalling pathways, including those mediating wound responses, pollen incompatibility and the maintenance of stem cell populations at the root and shoot meristems (Matsubayashi & Sakagami, 2006). Analysis of the Arabidopsis thaliana genome sequence suggests that it encodes at least 1000 secretory peptides (Lease & Walker, 2006), indicating that peptides may feature as regulators in an even wider range of plant developmental and environmental response pathways than currently characterized. Indeed, two Arabidopsis putative secretory peptides, epidermal patterning factor 1 and epidermal patterning factor 2 (EPF1 and EPF2), have recently been shown to regulate the density and spacing of stomata that develop on the leaf surface (Hara et al., 2007, 2009; Hunt & Gray, 2009). These peptides act independently, with EPF2 acting slightly earlier in leaf development than EPF1, to restrict the asymmetric divisions of stomatal precursor cells, known as meristemoids, and to prevent the inappropriate formation of stomata in positions adjacent to an existing stoma. Thus, the two EPF peptides inhibit stomatal formation and clustering, and it has been shown that they both require the putative plasma membrane receptor components too many mouths (TMM) and members of the ERECTA family (ER, ERL1 and ERL2) to fulfil their roles (Yang & Sack, 1995; Nadeau & Sack, 2002; Shpak et al., 2005; Bergmann & Sack, 2007; Hara et al., 2007; Hunt & Gray, 2009). A putative signal processing peptidase stomatal density and distribution 1 (SDD1) also regulates stomatal development and clustering (Berger & Altmann, 2000; Von Groll et al., 2002), but, surprisingly, EPF1 and EPF2 appear to act independently of SDD1, and the substrate(s) for this peptidase remains unknown (Hara et al., 2007, 2009; Hunt & Gray, 2009).

EPF1 and EPF2 are members of a family of 11 related putative secretory peptides which all contain six conserved cysteine residues towards their C-terminus. The additional nine family members are known as EPF-like peptides (EPFL1–EPFL9). Recently, it has been reported that EPFL4 and EPFL5, like EPF1 and EPF2, can inhibit stomatal development when ectopically overexpressed, but, as they are not normally expressed in cells of the stomatal lineage, they are unlikely to be involved in controlling stomatal development (Hara et al., 2009). Here, we present a study of the previously uncharacterized peptide EPFL9 and demonstrate that, in contrast with the other members of this peptide family, EPFL9 is a positive rather than negative regulator of stomatal density and clustering.

Materials and Methods

Plant material

The overexpression gene construct p35S:EPF9 was created by amplifying the predicted coding region of EPFL9 from cDNA with 5′-CACCAATTAGAGCAAGAAGAAGAAG-3′/5′-TAACAACTAATATCTATGACAAACAC-3′ primers and KOD polymerase (Merck Biosciences, Nottingham, UK), and recombining into pENTR/D/TOPO, then into pCTAPi (Rohila et al., 2004) with LR clonase II (Fisher Scientific, Leicester, UK). Primers were designed so that the TAPi-tag was out of frame and not translated. The p35S:EPFL9RNAi construct was generated by amplifying a PCR fragment with 5′-CACCATTCAAGGAGGAGGCATATGAT-3′/5′-TAACAACTAATATCTATGACAAACAC-3′ primers as above and recombining into pENTR/D/TOPO, then into pHellsgate12 (Wesley et al., 2001) with LR clonase II Plus. Constructs were transformed into Agrobacterium tumefaciens C58 by electroporation, and Arabidopsis was transformed by floral dipping (Clough & Bent, 1998). Transformants were selected with kanamycin, hygromycin or Basta (Liberty; Agrevo, Cambridge, UK) as appropriate. epf1-1 (SALK_137459), epf2-1 (GABI_673E01) and sdd1 (GABI_627_D04) T-DNA disruption mutants were as described previously (Hara et al., 2007; Hunt & Gray, 2009). All transgenic plants were in the Arabidopsis thaliana (L.) Heynh Col-0 background and were grown in a controlled environment chamber under fluorescent lighting in a 9 h (22°C) : 15 h (18°C) light : dark cycle. For EPFL9 overexpression, experiments were carried out using three or four independently transformed lines in each of the epf1-1 and epf2-1 backgrounds, respectively. sdd1EPF9OE was created by crossing an independent line of EPFL9OE with sdd1 and confirmed by PCR on a segregating F2 population.

Mutant plant analysis

For each experiment, all plants were grown under the same conditions at the same time. Stomatal counting and epidermal cell counting were carried out on impressions of the abaxial surface of fully expanded leaves which were selected as comparable between genotypes. Dental resin (Coltene Whaledent, Altstätten, Switzerland) was applied to the leaf and nail varnish peels taken from set resin. Cell counts were taken from three 0.0625-mm2 areas of three leaves from at least three separate T2 plants of each genotype. The stomatal index was calculated as the stomatal density × 100%/(stomatal density + epidermal cell density). Unpaired t-tests were performed on the data and < 0.05 was regarded as significantly different. DIC images were captured using an Olympus BX51 microscope connected to a DP51 digital camera.


Expression pattern of EPFL9

Gene expression analysis using the Arabidopsis Co-expression Tool (Manfield et al., 2006) identified EPFL9 as having a highly similar gene expression pattern to EPF1 (Table S1, see Supporting Information). Publicly available transcriptomics data (Schmid et al., 2005) indicated relatively high expression of these genes, and others involved in stomatal development, in the shoot apical meristem and developing leaves (Fig. 1). This expression pattern suggested that EPFL9, like EPF1 and EPF2, might be involved in the regulation of stomatal development.

Figure 1.

 Coexpression pattern of EPFL9. Relative expression patterns of EPF1, EPF2, EPFL9 and SDD1 are similar across a range of Arabidopsis samples. Data were extracted from AtGenexpress (

Effects of manipulating the level of EPFL9 expression

Independently transformed lines of Arabidopsis were produced, ectopically overexpressing EPFL9 under the control of the CaMV35S promoter (Fig. S1a, see Supporting Information). All four EPFL9OE lines produced (EPFL9OE-1 to EPFL9OE-4) exhibited significant increases in the stomatal index on fully expanded leaves in comparison with Col-0 controls in the T2 generation (Fig. 2a). The EPFL9OE lines also showed significant increases in stomatal density. We were unable to identify any Arabidopsis lines with T-DNA disruptions within the coding region of EPFL9, but expression of an EPFL9RNAi construct in transgenic Arabidopsis (Fig. S1b,c) produced plants with significantly reduced numbers of stomata. A significant reduction in stomatal density was seen in four of the six independent EPFL9RNAi lines recovered. The three EPFL9RNAi lines analysed in the T2 generation had 1.8- to 1.6-fold reductions in stomatal density and 25–32% reductions in the stomatal index (Fig. 3a). The epidermis of fully expanded EPFL9RNAi leaves contained what appeared to be arrested stomatal precursor cells (small oval-shaped guard mother cells), or abnormally large stomata-free areas of epidermal pavement cells (Fig. 3c,d), suggesting that reduced levels of EPFL9 caused a subset of stomatal lineage cells to either arrest their development or to differentiate into epidermal pavement cells.

Figure 2.

 Overexpression of EPFL9 promotes stomatal development. (a) Stomatal indices of mature leaf abaxial surfaces of four independently transformed lines of EPFL9OE in the T2 generation. Data are represented as mean ± SEM. Values significantly different from Col-0 control: *< 0.05. (b, c) Representative images from epidermal impressions of abaxial surfaces of mature Arabidopsis Col-0 and EPFL9OE leaves. Bars, 40 μm. [Correction added after online publication (8 April 2010): Figure 2 legend transposed with Figure 3 legend]

Figure 3.

 Expression of EPFL9RNAi restricts stomatal development. (a) Stomatal densities (filled bars) and indices of mature leaf abaxial surfaces (open bars) of three independently transformed lines of EPFL9RNAi in the T2 generation. Data are represented as mean ± SEM. Values significantly different from Col-0 control: *< 0.05. (b–d) Representative images from epidermal impressions of abaxial surfaces of mature Arabidopsis Col-0, EPFL9RNAi-1 and EPFL9RNAi-2. Bars, 50 μm. *Arrested meristemoid.

Genetic interactions with EPFL9

To explore whether EPFL9 regulates the same steps in stomatal development as EPF1, EPF2 or SDD1, the effect of ectopically overexpressing EPFL9 in epf1, epf2 and sdd1 mutant backgrounds was investigated. These three mutants are each individually characterized by increased stomatal density and/or clustering in their leaf epidermis (Berger & Altmann, 2000; Hara et al., 2007, 2009; Hunt & Gray, 2009). In the wild-type (Col-0 background), EPFL9 overexpression greatly increased both the stomatal density and index (Figs 2a,4a), and also caused a moderate degree of stomatal clustering, with c. 5% of EPFL9OE stomata being found adjacent to each other in pairs, which were absent in Col-0 (Fig. 4). In each of the three mutant backgrounds, the effects of EPFL9 overexpression were additive with respect to stomatal density in the case of sdd1 and epf2 or clustering in the case of epf1. In each case, EPFL9 overexpression caused even larger numbers of stomata and/or clustering over the single mutant phenotypes (Fig. 4). These additive phenotypes suggest that EPFL9 does not require EPF1, EPF2 or SDD1 to carry out its function in promoting stomatal development. SDD1 is a regulator of both stomatal density and clustering, and stomatal development was particularly affected in the sdd1EPFL9OE plants, which showed 75% higher stomatal density than EPFL9OE, 48% higher than sdd1 and 363% higher than Col-0, resulting in an exceptionally high stomatal density of over 1000 stomata mm−2 (Fig. 4b,g,h). The overexpression of EPFL9 in the sdd1 background also caused a large increase in the proportion of stomata found in clusters, from c. 13% in sdd1 to 22% in sdd1EPFL9OE (Fig. 4d). EPF1 is principally involved in restricting stomatal clustering, and similar EPFL9 overexpression experiments in the epf1 background caused an increase in the proportion of stomata found in clusters from c. 6% in epf1 to 25–36% in epf1EPFL9OE lines (Fig. 4c), but no significant changes in stomatal density were observed compared with EPFL9OE in a wild-type background (Fig. 4a,e,f). EPF2 is predominantly involved in regulating the stomatal number, and EPFL9 overexpression in the epf2 background caused a 1.5-fold increase in stomatal number compared with expression in a wild-type background, but no significant increase in the proportion of clustered stomata (Fig. 4b,d,i,j).

Figure 4.

 The ability of EPFL9 to enhance stomatal development in Arabidopsis is not dependent on EPF1, EPF2 or SDD1. Stomatal densities of mature leaf abaxial surfaces of epf1-1 in combination with overexpression of EPFL9 (a) and sdd1 and epf2-1 in combination with overexpression of EPFL9 (b). (c, d) Stomatal cluster sizes: 2-mer, filled bars; 3-mer, open bars; 4-mer, hatched bars. Data are represented as mean ± SEM. Representative images of impressions of abaxial surfaces of epf1-1 (e), epf1EPFL9OE-1 (f), sdd1 (g), sdd1EPFL9OE (h), epf2-1 (i) and epf2EPFL9OE-1 (j). Images (e, f) from developing leaves (c. 5 mm long) and (g–j) from mature leaves. Bars, 50 μm.


We have identified EPFL9 as the first putative secreted positive regulator of stomatal development. Previous studies have shown that several related members of the EPF/EPFL secretory peptide family are able to regulate stomatal development (Hara et al., 2007, 2009; Hunt & Gray, 2009), and it is now clear that at least three related peptide signals act independently during the stomatal developmental programme. EPF1 and EPF2 negatively regulate stomatal clustering and density, whereas EPFL9 positively regulates both of these processes, suggesting that it may antagonize the action of EPF1 and/or EPF2. The existence of activators and inhibitors of stomatal development makes it tempting to speculate that stomatal patterning is controlled by an activator–inhibitor system of diffusible morphogens (Meinhardt & Gierer, 1974). However, the genetic analysis presented here and previously (Hara et al., 2007, 2009; Hunt & Gray, 2009) indicates that each of these three peptides acts independently.

The emerging picture of a family of secretory peptides that regulates stomatal fate in the developing epidermis has parallels with the action of the CLE (CLAVATA/ESR) family of peptides (Fletcher et al., 1999; Gray et al., 2008). There are over 30 members of the Arabidopsis CLE peptide family, many of which have been shown to regulate stem cell number in shoot, floral and root meristems via a leucine-rich repeat-receptor mediated pathway (Ogawa et al., 2008). CLE peptides also regulate tracheary element differentiation during development of the vasculature, with three CLE peptides having been shown to inhibit, and one to promote, tracheary element differentiation (Ito et al., 2006; Hirakawa et al., 2008). Thus, patterning mechanisms involving families of secretory peptides acting as positive or negative regulators of cellular differentiation may feature as a common theme in plant development. Many signalling peptides, including those of the CLE family, are proteolytically cleaved to release an active fragment (Kondo et al., 2006). Although the putative peptidase SDD1 is a well-characterized regulator of stomatal development (Berger & Altmann, 2000; Von Groll et al., 2002), our results suggest that it cannot be an activator of EPFL9, and it remains unknown whether any members of the EPF/EPFL peptide family are subject to proteolytic cleavage.

The ability of stomata to regulate gas exchange in and out of leaves allows plants to inhabit a range of environments, and affects global water and carbon cycles (Hetherington & Woodward, 2003). The frequency of stomata that develop in the leaf epidermis is affected by several environmental variables, including light intensity and atmospheric CO2 levels. An improved understanding of the network of interacting signals that regulate stomatal differentiation has emerged from recent studies of Arabidopsis mutants (Bergmann & Sack, 2007), but we still know relatively little about how such environmental factors modulate stomatal frequency (Gray et al., 2000; Casson & Gray, 2008; Casson et al., 2009). The work presented here suggests that stomatal frequency may be regulated by a balance of positively and negatively acting mobile peptide signals. Thus, it will be interesting to determine whether the relative levels of EPF/EPFL peptides are involved in the environmental modulation of stomatal development.


During the final revision of the manuscript, the article ‘Stomagen positively regulates stomatal density in Arabidopsis’ was published, which also describes the properties of EPFL9 (Sugano et al., 2010)