SUPPRESSOR OF LLP1 1-mediated C–terminal processing is critical for CLE19 peptide activity



Cell-to-cell communication is essential for the coordinated development of multicellular organisms. Members of the CLAVATA3/EMBRYO-SURROUNDING REGION-RELATED (CLE) family, a group of small secretory peptides, are involved in these processes in plants. Although post-translational modifications are considered to be indispensable for their activity, the detailed mechanisms governing these modifications are not well understood. Here, we report that SUPPRESSOR OF LLP1 1 (SOL1), a putative Zn2+ carboxypeptidase previously isolated as a suppressor of the CLE19 over-expression phenotype, functions in C–terminal processing of the CLE19 proprotein to produce the functional CLE19 peptide. Newly isolated sol1 mutants are resistant to CLE19 over-expression, consistent with the previous report (Casamitjana-Martinez, E., Hofhuis, H.F., Xu, J., Liu, C.M., Heidstra, R. and Scheres, B. (2003) Curr. Biol. 13, 1435–1441). As expected, our experiment using synthetic CLE19 peptide revealed that the sol1 mutation does not compromise CLE signal transduction pathways per se. SOL1 possesses enzymatic activity to remove the C–terminal arginine residue of CLE19 proprotein in vitro, and SOL1-dependent cleavage of the C–terminal arginine residue is necessary for CLE19 activity in vivo. Additionally, the endosomal localization of SOL1 suggests that this processing occurs in endosomes in the secretory pathway. Thus, our data indicate the importance of C–terminal processing of CLE proproteins to ensure CLE activities.


Multicellular organisms utilize intercellular communication to coordinate cellular differentiation and proliferation precisely, thus achieving organized development. In the case of higher plants, phytohormones, such as auxin and cytokinin, are known to function as intercellular signaling molecules (Moubayidin et al., 2009). In addition to these conventional phytohormones, small secretory peptides have also emerged as important factors that mediate cell-to-cell signaling. Amongst such peptides, the CLAVATA3/EMBRYO-SURROUNDING REGION-RELATED (CLE) peptides have been extensively studied because of their various important functions in plant development and plant–microbe interactions (Hirakawa et al., 2008, 2010; Miwa et al., 2008; Müller et al., 2008; Stahl et al., 2009; Bleckmann et al., 2010; Kinoshita et al., 2010; Kiyohara and Sawa, 2012; Miyawaki et al., 2013; Yamada and Sawa, 2013).

In Arabidopsis thaliana, 32 CLE genes exist in the genome and are expressed in various tissues (Oelkers et al., 2008; Jun et al., 2010). These genes encode proproteins comprising approximately 100 amino acid residues, which carry an N–terminal signal peptide, the conserved 14 amino acid C–terminal CLE domain, and a less conserved variable domain between them (Cock and McCormick, 2001; Rojo et al., 2002). In addition, 16 CLE proproteins possess functionally uncharacterized extension sequences that are C–terminally attached to the CLE domain (Olsen and Skriver, 2003). However, the functional roles of these C–terminal extensions are poorly understood. The biological relevance of the CLE domain was indicated by a number of experiments using chemically synthesized peptides (Fiers et al., 2005, 2006; Ito et al., 2006; Kondo et al., 2006; Kinoshita et al., 2007). Exogenous application of 12 amino acid synthetic peptides corresponding to the CLE domain of other CLE proteins reduced the size of the shoot apical meristem or the root length, or inhibited tracheary element differentiation (Ito et al., 2006; Kondo et al., 2006; Kinoshita et al., 2007). Furthermore, several CLE peptides have been shown to actually function as 12/13 amino acid peptides corresponding to the CLE domain (Ito et al., 2006; Kondo et al., 2006; Ohyama et al., 2009). These results strongly indicate that CLE peptides act as 12/13 amino acid peptides in vivo, and that processing steps are crucial for producing fully active CLE peptides from the CLE proproteins. However, the mechanisms underlying proteolytic processing of CLE peptides remain poorly understood.

Biochemical studies using extracts from cauliflower (Brassica oleracea) detected serine protease activity that cleaves the CLV3 proprotein at the 70th arginine residue, which is located at the N–terminus of the CLE domain (Ni and Clark, 2006; Ni et al., 2011). A few amino acid residues, especially the arginine located at the N–terminus of the CLE domain, are thought to be crucial for cleavage (Ni et al., 2011). The cauliflower extract also exhibited carboxypeptidase activity, which may have a role in C–terminal processing (Ni et al., 2011). Similarly, xylem fluid from soybean (Glycine max) and suspension culture fluid from barrel medic (Medicago truncatula) showed serine protease activity and carboxypeptidase activity that was able to produce a functional peptide from the 31 amino acid CLE36 proprotein (Djordjevic et al., 2011). Despite these findings, the molecular entities responsible for these protease activities are unknown. In addition, the functional relevance of these protease activities in CLE signaling is not well understood.

Further evidence of the CLE processing machinery emerged from a suppressor mutant screening for the CLE19 over-expression phenotype. The suppressor of LLP1 1 (sol1) mutant is insensitive to root-specific CLE19 over-expression, which normally causes root meristem consumption and a short root phenotype in the wild-type background (Casamitjana-Martínez et al., 2003). SOL1 encodes a putative Zn2+ carboxypeptidase with high sequence similarity to animal carboxypeptidase D (CPD) and carboxypeptidase E (CPE) (Casamitjana-Martínez et al., 2003). In animals, these carboxypeptidases are known to play a role in processing of peptide hormones such as insulin (Docherty and Hutton, 1983; Varlamov et al., 1997; Dong et al., 1999; Davidson, 2004). In addition, biochemical analyses demonstrated that both CPD and CPE are capable of cleaving C–terminal arginine and lysine from polypeptides in vitro (Fricker and Snyder, 1983; Greene et al., 1992; Sidyelyeva and Fricker, 2002). The strong suppressor effect of the sol1 mutation on the CLE19 over-expressing phenotype suggests the importance of SOL1 in proteolytic modification of the CLE19 proprotein, presumably by targeting arginine residues, to produce the active CLE19 peptide (Casamitjana-Martínez et al., 2003). Thus, functional dissection of SOL1 is required to decipher its role in CLE19 signaling, and this should further advance our understanding of the complex regulation of CLE peptide signaling.

In this study, we performed a series of detailed analyses of SOL1 to unravel the processing machinery of CLE peptides. We demonstrate that two SOL1 T–DNA insertion lines are not defective in CLV3 and CLE19 signal transduction pathways per se. We further show that SOL1-dependent cleavage of the C–terminal arginine is critical for CLE19 activity in planta, whereas SOL1 is not involved in generating active CLV3. Additionally, the results of the SOL1 localization study imply that C–terminal processing of CLE19 by SOL1 is likely to occur in endosomes. Our biochemical analyses revealed that SOL1 cleaves not only the C–terminal arginine but also lysine, implying a role in processing of CLE peptides other than CLE19. These findings provide insight into our understanding of the CLE peptide maturation process, which may act as an additional regulatory step for CLE signaling.


The SOL1 expression pattern indicates its involvement in many developmental processes

To determine the function of SOL1 and its potential target CLEs, we first examined the expression pattern of SOL1. The 2370 bp SOL1 upstream sequence fused to the β–glucuronidase (GUS) reporter gene was transformed into Col–0 plants (Figure 1a). The GUS activity of the resulting transgenic T2 plants was analyzed. Microscopic observation detected GUS activity in columella cells, the lateral root cap, stipules and young true leaves, with stronger expression in the basal regions (Figure 1b–d). Although weak GUS activity was also detected in cotyledons and the basal region of developed leaves, no activity was detected in the lateral root primordia and the shoot apical meristem (Figure 1e–h).

Figure 1.

Tissue-specific expression pattern of SOL1.

(a) Structure of the SOL1pro:GUS construct. The 2370 bp upstream sequence of SOL1 was fused to the β–glucuronidase (GUS) reporter gene. The black box indicates the 5' UTR of SOL1.

(b–h) Transgenic T2 plants expressing the SOL1pro:GUS reporter gene were stained for GUS activity. (b) Primary root tip. (c) High-magnification image of the primary root tip. (d) Young leaves. (e) Cotyledon. (f) Developed true leaf. (g) Lateral root primordium. (h) Section of the shoot apical meristem (10 μm). Scale bars = 50 μm (b,c,g,h) and 200 μm (d–f).

Thus, SOL1 is expressed in various tissues, and may therefore contribute to multiple functions in Arabidopsis. This ubiquitous localization may allow SOL1 to act as a generic maturation enzyme for proproteins of CLE family members, which are also expressed in various tissues (Jun et al., 2010).

The sol1 mutation does not alter responsiveness to the CLV3 and CLE19 peptides

To further explore the function of SOL1, we isolated two SOL1 T–DNA insertion lines (in the Col–0 background) from the SALK lines SALK_013659c and SALK_015449c, designated sol1101 and sol1102, respectively. These lines contain T–DNA insertions in the 8th and the 14th exons, respectively (Figure S1a,b). Transcripts for sol1101 and sol1102 were examined by quantitative RT–PCR using three primer pairs: L1/R1 (located outside the T–DNA insertion site), L2/R2 (flanking the T–DNA insertion site of the sol1101 mutant) and L3/R3 (flanking the T–DNA insertion site of the sol1102 mutant) (Figure S1b). The L1/R1 primer pair amplified as much transcript for sol1101 and sol1102 as for wild-type. No transcript was detected in the sol1 mutants using the primer pairs containing the T–DNA insertion sites for the respective mutant genes (L2/R2 for sol1101 and L3/R3 for sol1102), although the other primer pairs detected as much transcripts as wild-type even in the sol1 mutants (Figure S1c). This result indicates that the sol1101 and sol1102 genes are not correctly transcribed, and may produce longer transcripts containing the T–DNA sequence or two transcripts split by T–DNA. Therefore, both alleles were considered to produce incomplete SOL1 proteins. For further characterization, we mainly focused on sol1101 as the mutated protein produced in the sol1101 mutant is thought to lack both the conserved catalytic residues and the transmembrane domain while the sol1102 protein lacks only the transmembrane domain, which may express its residual activity (Figure S1a).

We examined the sol1 mutant in a root growth inhibition assay using chemically synthesized 12 amino acid forms of CLE19 (RVIHypTGHypNPLHN, where Hyp is hydroxyproline) and CLV3 (MCLV3, RTVHypSGHypDPLHH). CLE19 is the most likely candidate for a substrate of SOL1 as suggested by a previous report, and CLV3 is the best-studied CLE member (Casamitjana-Martinez et al., 2003). Both MCLV3 and CLE19 treatments reduced root growth in all plants tested, and no significant differences were observed among sol1101, sol1102 and wild-type (Figures 2, S2 and S3). This result indicates that the sol1 mutants are not defective in the endogenous CLV3 and CLE19 signaling pathways, consistent with the hypothesis that SOL1 is involved in the maturation of CLE peptides.

Figure 2.

Effects of the MCLV3 and CLE19 peptides on root growth of the sol1 mutants. (a) Primary root length of seedlings treated with various concentrations of MCLV3. (b) Primary root length of seedlings treated with various concentrations of CLE19. Values are means ± SEM (n = 9–11).

The sol1 mutations suppress the CLE19 over-expression phenotype, but not the CLV3 over-expression effect

Our root inhibition assay using the newly isolated sol1 mutants and the CLE peptides revealed that SOL1 is not required for perception of CLE19 or CLV3, although the original sol1 mutation was shown to suppress the phenotype caused by root-specific over-expression of Brassica napus CLE19 (Casamitjana-Martínez et al., 2003). Therefore, we investigated the impact of the sol1 mutation on the CLE19- and CLV3 over-expression phenotypes. For this purpose, we generated stable transgenic wild-type and sol1101 plants expressing either full-length CLV3 or CLE19 under the control of an estrogen-inducible promoter. Our quantitative RT–PCR analyses revealed that β–estradiol treatment induces the expression of respective transgenes to a similar extent in Col–0 and sol1101 background plants (Figure S4o,p and S5m, n). The conditional expression of CLE19 by β–estradiol reduced the root length of plants in the wild-type background (Figures 3c and S4b,c,i,j), but not in the sol1101 background (Figures 3d and S4e–g,l–n), consistent with the findings of a previous study (Casamitjana-Martínez et al., 2003). However, the inhibition of root growth by estrogen-induced CLV3 expression was not affected by the sol1101 mutation (Figures 3a,b and S5b,c,e,f,h,i,k,l). Thus, our data are consistent with the results of a previous study suggesting that SOL1 is involved in CLE19 maturation, but not in CLE19 perception (Casamitjana-Martínez et al., 2003). Furthermore, this result strongly suggests that SOL1 is preferentially involved in the maturation process of CLE19, but not CLV3.

Figure 3.

Effects of estrogen-induced CLV3 and CLE19 expression on root growth of the sol1101 mutant. (a) Primary root length of Col–0 plants over-expressing CLV3. (b) Primary root length of sol1101 plants over-expressing CLV3. (c) Primary root length of Col–0 plants over-expressing CLE19. (d) Primary root length of sol1101 plants over-expressing CLE19. Values are means ± SEM (n = 10–11).

In support of this idea, the sol1 mutants did not show an enlargement of the shoot apical meristem, a phenotype that is a hallmark of CLV3-related mutants (Clark et al., 1993, 1995; Kayes and Clark, 1998; Müller et al., 2008; Kinoshita et al., 2010). To uncouple SOL1 activity from the CLV3 pathway genetically, we performed phenotypic analysis of crosses between the sol1 mutant and the CLV3-related mutants clv1101, clv2101, rpk22 and clv38 (Diévart et al., 2003; Kinoshita et al., 2010). We counted the number of carpels in the ten basal flowers of the inflorescence stem of Col–0, sol1101, sol1102, clv1101, clv1101 sol1101, clv2101, clv2101 sol1101, rpk22, rpk22 sol1101, clv38 and clv38 sol1101. We found that the presence of the sol1 mutation did not significantly increase the carpel number of any double mutants significantly compared to the respective single mutants (Table S1). Together with the data for estrogen-induced CLV3 expression in sol1101, we conclude that SOL1 is not required for CLV3 maturation.

SOL1 possesses in vitro carboxypeptidase activity against the Cterminal arginine and lysine

The ability of CPD and CPE, animal homologs of SOL1, to cleave the C–terminal arginine or lysine residues from polypeptides prompted us to speculate that SOL1 may be involved in C–terminal maturation of CLE proproteins (Fricker and Snyder, 1983; Greene et al., 1992; Sidyelyeva and Fricker, 2002). To characterize the in vitro carboxypeptidase activity of SOL1, SOL1 C–terminally fused to a triple hemagglutinin (HA)/single StrepII tag (SOL1–3HS) was transiently expressed in Nicotiana benthamiana leaves and then affinity-purified (Figure 4a). Purified protein was detected as two major bands, which may reflect the presence or absence of the signal peptide (Figure 4a). To investigate carboxypeptidase activity of the purified SOL1–3HS, Dansyl-Phe-Ala-Arg, a well-established artificial fluorescent substrate of CPD and CPE, was used (Fricker and Snyder, 1983; Greene et al., 1992; Sidyelyeva and Fricker, 2002). The reaction with the purified SOL1–3HS significantly increased the fluorescence of Dansyl-Phe-Ala, the cleaved product, compared to the mock reaction and buffer alone (Figure 4b). Lower pH conditions significantly reduced the cleavage activity, suggesting that SOL1 functions in neutral pH conditions in vivo (Figure 4c). These results show that SOL1 possesses carboxypeptidase activity to remove the arginine residue from Dansyl-Phe-Ala-Arg, indicating that an additional arginine at the C–terminus of the CLE domain of CLE19 may be targeted and removed by SOL1. Therefore, we directly examined SOL1 activity using a synthetic CLE19 peptide derivative containing the C–terminal arginine after the CLE domain (CLE19 + R). After incubating CLE19 + R with the purified fractions, MALDI-TOF MS analysis detected the 12 amino acid CLE19 peptide processed from CLE19 + R only in the reaction containing purified SOL1–3HS (Figure 4d,e). This result strengthens our hypothesis. We also performed the same experiment using synthetic CLE21 + K and CLE22 + R peptides. SOL1 cleaved the C–terminal arginine or lysine of CLE21 + K and CLE22 + R (Figure 5a–d). Collectively, these results imply that SOL1 cleaves both lysine and arginine at the C–terminus, which suggests the involvement of SOL1 in processing of other CLE proproteins that harbor R or K at their C–terminus.

Figure 4.

Carboxypeptidase activity of SOL1 against the C–terminal arginine.(a) Immunoblot assay of affinity-purified SOL1–3HS. Lane 1, total extract; lane 2, flow-through; lane 3, purified protein. (b) Enzymatic activity of the purified fractions on Dansyl-Phe-Ala-Arg. The relative fluorescence of Dansyl-Phe-Ara, a reaction product, was detected. Reactions were performed for the indicated time periods at 30°C, pH 7.0. Values are means ± SEM (n = 3). (c) pH-dependent enzymatic activity of SOL1–3HS on Dansyl-Phe-Ala-Arg. The relative fluorescence of Dansyl-Phe-Ara was measured. Reactions were performed for 3 h at 30°C at the indicated pH levels. Values are means ± SEM (n = 3). (d,e) CLE19 + R peptides were incubated with purified SOL1–3HS (d) or mock N. benthamiana extract (e) for 24 h at 30°C, pH 7.0. The reaction products were subjected to MALDI-TOF MS analysis.

Figure 5.

Carboxypeptidase activity of SOL1 on other synthetic CLE peptides.(a) Reaction of SOL1 with CLE21 + K. (b) Mock reaction of CLE21 + K. (c) Reaction of SOL1 with CLE22 + R. (d) Mock reaction of CLE22 + R. The reactions were incubated for 24 h at 30°C, pH 7.0. The reaction products were subjected to MALDI-TOF MS analysis.

The suppressor effect of sol1–101 depends on the Cterminal arginine of the CLE19 proprotein

Having established that the recombinant SOL1–3HS produced in N. benthamiana possesses carboxypeptidase activity that removes the C–terminal arginine residue from the synthetic 13 amino acid CLE19 + R peptide, we examined whether SOL1 acts against the CLE19 proprotein in plant tissues. The sol1 mutants were insensitive to estrogen-induced CLE19 expression, presumably due to their inability to process the CLE19 proprotein in the absence of SOL1 (Figure 3d). If this is indeed the case, then the sol1 mutants should be sensitive to estrogen-induced expression of CLE19 lacking C–terminal arginine (designated CLE19ΔR; Figure 6a). To test this hypothesis, we transformed wild-type and sol1101 plants with CLE19ΔR under the control of the estrogen-inducible promoter. β–estradiol treatment clearly inhibited root growth of the transgenic Col–0 plants expressing CLE19ΔR, as it did for Col–0 plants over-expressing full-length CLE19 (Figures 6b and S6). The estrogen-induced expression of CLE19ΔR also reduced the root length of the transgenic sol1101 plants (Figures 6c and S7). This result is in contrast to the insensitivity of the sol1101 mutant to estrogen-induced expression of the full-length CLE19 (Figure 3). This result strongly suggests that SOL1 processes the C–terminal arginine of the CLE19 proprotein. It also suggests that this SOL1-mediated C–terminal processing is crucial for production of active CLE19 peptide in planta.

Figure 6.

Effects of estrogen-induced CLE19ΔR expression on Col–0 and sol1101. (a) The full-length amino acid sequence of CLE19. The CLE domain is underlined. R (in red) indicates the C–terminal arginine residue that is absent in CLE19ΔR. (b) Primary root length of Col–0 plants over-expressing CLE19ΔR. Values are means ± SEM (n = 8–11). (c) Primary root length of sol1101 plants over-expressing CLE19ΔR. Values are means ± SEM (n = 9–11).

SOL1 is localized specifically to endosomes

CLE proproteins undergo proteolytic maturation processes that produce functional CLE peptides, which are subsequently secreted into apoplastic spaces, where they function (Rojo et al., 2002). In this study, we showed that the C–terminal maturation of CLE19 is mediated by SOL1, a putative membrane-bound carboxypeptidase (Casamitjana-Martínez et al., 2003). However, it remains to be determined in which compartment of the secretory pathway CLE proproteins are processed into mature functional peptides. To address this question, we analyzed the subcellular localization of SOL1, a CLE-processing enzyme. For this purpose, we transiently expressed SOL1YFP under the control of an estrogen-inducible promoter in the leaves of N. benthamiana, together with the known organelle markers 35S:SP-GFP-HDEL, an endoplasmic reticulum (ER) marker (Mitsuhashi et al., 2000), 35S:ST-mRFP, a trans-Golgi marker (Boevink et al., 1998), 35S:mRFP-SYP61, a trans-Golgi network marker (Sanderfoot et al., 2001; Uemura et al., 2004) or 35S:TagRFP-ARA6, an endosomal marker (Ueda et al., 2001). Confocal microscopic observation revealed that SOL1–YFP localized to dot-like organelles inside the estrogen-treated cells, suggesting that SOL1 is localizes to the endomembrane system other than the plasma membrane (Figure 7). The co-expression study showed that SOL1–YFP did not co-localize with SP–GFP–HDEL or ST–mRFP, but occasionally localized with mRFP–SYP61 and more frequently with TagRFP–ARA6, indicating that SOL1–YFP is preferentially localized to the endosomes (Figure 7a–d). We therefore used TagRFP–ARA7 as another endosomal marker (Sohn et al., 2003; Kotzer et al., 2004), and found that it showed almost fully overlapping localization with SOL1–YFP (Figure 7e). These data indicate that SOL1 is mainly 7localized to endosomes, and, occasionally, to the trans-Golgi network, suggesting that C–terminal processing of the CLE19 proprotein occurs in endosomes.

Figure 7.

Subcellular localization of SOL1–YFP and organelle markers transiently expressed in N. benthamiana leaves. SOL1–YFP was transiently co-expressed with (a) SP–GFP–HEDL, (b) ST–mRFP, (c) mRFP–SYP61, (d) TagRFP–ARA6 and (e) TagRFP–ARA7. Scale bars = 5 μm.


SOL1 has been implicated as a peptidase involved in processing of the CLE19 proprotein (Casamitjana-Martínez et al., 2003). Our biochemical analysis showed that SOL1 is capable of removing the C–terminal arginine or lysine from CLE proproteins in vitro. We suggest that SOL1-mediated removal of the arginine is essential for CLE19 activity in planta. These results suggest a critical role for SOL1 in CLE19-mediated signaling in planta.

SOL1 generates active CLE19 through post-translational processing

Peptide hormones regulate various aspects of animal and plant development. In many cases, peptide hormones are first translated as inactive precursor polypeptides, and become active through post-translational processing (Fricker, 1988). It is considered that such processing, in addition to transcriptional regulation, enables organisms to release peptide hormone activities at accurate times during development (Muller et al., 1999; Westphal et al., 1999). To understand the processing of CLE peptides, which are the best known signaling peptides acting in plant development, we analyzed the processing of CLE19. We showed that SOL1-dependent processing of the CLE19 proprotein is essential for CLE19 activity. Our biochemical experiments revealed that SOL1 exhibits in vitro processing activity with respect to the C–terminal arginine of the CLE19 polypeptide. In addition, conditional over-expression of CLE19 and CLE19ΔR also indicated the importance of SOL1 activity in removal of the C–terminal arginine of the CLE19 proprotein in vivo. Addition of the C–terminal arginine to TDIF, another well-studied CLE peptide, reduces its activity to one-seventh (Ito et al., 2006). Thus, these results suggest that the C–terminal arginine processing performed by SOL1 or SOL1-like enzymes is a critical step for generating active CLE peptide.

Biochemical activity may be used for identifying other SOL1-targeted CLE peptides

Several CLE proproteins harbor a C–terminal arginine, in addition to CLE19 (Figure S8). It has been reported that CPD and CPE, the animal SOL1 homologs, cleave not only C–terminal arginine residues but also C–terminal lysine residues of their substrates (Fricker and Snyder, 1983; Greene et al., 1992; Sidyelyeva and Fricker, 2002). Hence, the same activity may be predicted for SOL1, and SOL1 showed in vitro enzymatic activity against the C–terminal arginine and lysine of Dansyl-Phe-Ala-Arg, CLE19 + R, CLE21 + K and CLE22 + R peptides. Amongst Arabidopsis CLE proproteins, CLE14, CLE20, CLE21, CLE22 and CLE42 possess a C–terminal arginine or lysine residue directly after their CLE domains, and are designated RK type, suggesting that these CLE proproteins are good candidates for SOL1 in planta substrates (Olsen and Skriver, 2003). Simultaneous over-expression of SOL1 and CLE19 in sol1 mutants successfully demonstrated SOL1 activity against CLE19 proprotein. Therefore, similar methods may be used to identify additional CLE substrates of SOL1. In addition to this RK type of CLE peptide, there is a group of CLE peptides, including CLE25, CLE26, CLE40, CLE45 and CLE46, that contains a CLE domain-Arg/Lys-X motif (RK embedded type, where X represents any polypeptide). It is reasonable to assume that primary removal of the C–terminal polypeptides (X) from these CLEs by peptidases reveals the arginine or lysine residue flanking the CLE domain, which may than be processed by SOL1. Thus, these CLEs are also possible substrates for SOL1 (Olsen and Skriver, 2003). For example, CLE40, which is expressed in the columella cells of root tips, as SOL1 is, possesses a lysine residue directly after the CLE domain, and the lysine is followed by an additional six amino acid stretch (Olsen and Skriver, 2003; Stahl et al., 2009). Additionally, CLE21, CLE25 and CLE26 are expressed in the shoot apices , young leaves and tips of young leaves, respectively (Jun et al., 2010). These expression patterns also resemble those of SOL1, further supporting our hypothesis. Thus, the presence of the buried arginine or lysine residue in the RK embedded type motif implies that a two-step C–terminal processing mechanism is involved in maturation of these CLEs. Such a mechanism would strongly ensure precise regulation of these CLE activities. Further analyses of SOL1, focusing on CLE peptides containing these RK embedded type motifs, are required to provide further insights into as yet unknown functions of CLE peptides.

Localization analysis links C–terminal processing of CLE19 and ARA6 endosomes

CLE peptides are secretory signaling molecules. Processing of CLE proproteins is thought to occur in extracellular spaces after secretion (Ni and Clark, 2006; Djordjevic et al., 2011; Ni et al., 2011). Recent research suggests that secreted serine proteases may be involved in N–terminal processing of CLE peptides (Djordjevic et al., 2011; Ni et al., 2011). However, our findings provides another possibility for the CLE peptide maturation pathway. As SOL1 possesses a signal peptide and transmembrane domain, we predicted that SOL1 localizes to the endomembrane system. Subcellular localization analyses indicated that SOL1 localizes mainly to ARA7-positive endosomes, and, to a lesser extent, ARA6-positive endosomes. In agreement with this, animal homologs of SOL1 have beem reported to localize to the trans-Golgi network and secretion vesicles (Docherty and Hutton, 1983; Hook and Loh, 1984; Varlamov et al., 1999). The protein structure of SOL1 predicted by SignalP ( indicates that the enzymatically active part of SOL1 is locates towards the inside of the vesicle, suggesting its role in processing of vesicle content, consistent with a previous report on CPD (Varlamov and Fricker, 1998). ARA7 is known to function in vacuolar transport from endosomes (Sohn et al., 2003; Kotzer et al., 2004). ARA6 is suggested to mediate vesicle transport from ARA7-positive endosomes to the plasma membrane (Ebine et al., 2011). Although SOL1 highly co-localized with ARA7, CLE19, a substrate of SOL1, is considered to be secreted into the extracellular space (Rojo et al., 2002). We demonstrated that SOL1 loses its activity under low pH, suggesting that SOL1-mediated C–terminal maturation of CLE19 occurs in the endosomes, but not in the acidic vacuole, before branching of the ARA6-mediated secretory pathway and ARA7-mediated vacuolar transport pathway. Thus, we propose a model for CLE19 secretion. First, the CLE19 proprotein is C–terminally processed by SOL1 in ARA7-positive endosomes. Second, the processed precursor is secreted through the ARA6-mediated secretion pathway to the apoplast, in which N–terminal processing is performed by serine proteases. After CLE19 processing in ARA6- and ARA7-positive endosomes, SOL1 may be carried via the ARA7 pathway to the vacuoles for degradation. The plant-unique ARA6-dependent secretory pathway is proposed to participate in environmental responses, such as salinity resistance (Ebine et al., 2011). In this context, CLE19 may contribute to achieving an orchestrated developmental plasticity in response to various environmental conditions. The post-translational regulation of CLE19 through SOL1 activity may add another level of regulation in response to environmental cues. Detailed studies of CLE19 activity and its SOL1-dependent regulation should provide insights into the role of the CLE19 peptide as a signaling molecule.

Post-translational control provides another layer of CLE activity regulation

The Arabidopsis genome encodes 32 CLE genes corresponding to at least 27 CLE peptides (Oelkers et al., 2008). A chemical genetics approach using 26 synthetic CLE peptides functionally categorized these peptides into four groups, indicating highly redundant activity within groups (Ito et al., 2006; Kinoshita et al., 2007; Hirakawa et al., 2011; Kondo et al., 2011). In contrast, a comprehensive analysis of their promoter activities revealed distinct and characteristic expression patterns (Jun et al., 2010). These results suggest that plant development is fine-tuned through complex and precise transcriptional regulation of various CLE genes. Here, we raised the possibility that peptidase-mediated post-translational processing may be another important layer in the control of CLE activities, at least in the case of CLE19. Our study also highlights involvement of the plant-unique ARA6-mediated secretory pathway in the CLE19 maturation process (Ebine et al., 2011). Thus, further studies of the CLE19 maturation process are required to determine the biological relevance of the plant-unique secretory pathway in CLE peptide production. Furthermore, identification of other peptidase(s) required for the N- and C–terminal processing of CLE proproteins is the next challenge in achieving a comprehensive understanding of CLE-mediated plant morphogenesis.

Experimental Procedures

Plant materials and growth conditions

Columbia–0 (Col–0), sol1101 (SALK_13659c), sol1102 (SALK_15449c), rpk22 (Mizuno et al., 2007; Kinoshita et al., 2010), clv1101 (Kinoshita et al., 2010) and clv38 ER (CS3604) were obtained from the Arabidopsis Biological Resource Center ( at Ohio State University (Diévart et al., 2003). clv2101 seeds (GK686A09) were obtained from GABI-Kat (, Kleinboelting et al., 2012). All lines used in this paper are in the Col–0 background, except for clv38 (unknown background) and clv1101 (Col–2 background).

For carpel number analysis, seeds incubated in water at 4°C for 2 days were sown on soil and grown at 22°C under continuous white light (20–50 mmol m−2 sec−1). For root length assays, GUS histochemical analysis and SOL1 expression analysis, surface-sterilized seeds were plated on growth medium containing 0.23% w/v Murashige and Skoog basal salts, 1% w/v sucrose, 0.05% w/v MES, pH 5.7, 1.5% w/v agar. For peptide treatment, MCLV3, CLE19 or 0.00001% trifluoroacetic acid were added to the medium, and, for estrogen-induced expression, 5 μm β–estradiol or 0.05% dimethylsulfoxide were added. The peptide solutions were dissolved in 0.1% trifluoroacetic acid to concentrations of 10, 1, 100 and 10 μm, respectively, and the β–estradiol solution was dissolved in dimethylsulfoxide to a concentration of 10 mm. Seedlings were grown for 9 days (root length assay), 12 days (GUS staining) or 16 days (expression analysis of SOL1) at 22°C under continuous white light (20–50 mmol m−2 sec−1) after a two-day incubation at 4°C. For estrogen-induced expression analysis, surface-sterilized seeds were sown in 9 ml of liquid growth medium containing 0.23% w/v Murashige and Skoog basal salts, 1% w/v sucrose and 0.05% w/v MES (pH 5.7), and then grown at 22°C under continuous white light (20–50 mmol m−2 sec−1) with shaking at 110 rpm.

Construction of plasmids and transgenic plants

The primers used in this study are listed in Table S2, and all the coding sequences used in this study were amplified by PCR from cDNA derived from young Col–0 seedlings. The SOL1 promoter sequences were amplified from Col–0 genomic DNA. The Gateway Cloning System (Life Technologies, was used unless stated. The estrogen-inducible CLV3, CLE19 and CLE19ΔR constructs were generated as follows: PCR amplification products were cloned into pENTR–D/TOPO (for CLE19) or pDONR221 (for CLV3 and CLE19ΔR), and then cloned through the LR reaction into pMDC7 (Curtis and Grossniklaus, 2003) in accordance with the manufacturer's instructions for the Gateway Cloning System. A mutated primer was used to cause an arginine codon deletion into CLE19ΔR. As for the SOL1pro:GUS construct, the 2370 bp sequence upstream of the SOL1 translational initiation site and the GUS coding sequence from R4pGWB433 were amplified independently, fused by PCR, and then cloned into R4pGWB401. The estrogen-inducible SOL1–3HAStrepII (SOL1–3HS) and SOL1–YFP constructs were generated as follows: the SOL1 coding sequence without the stop codon was cloned into pXCSG–3HS and pH35GY resulting in pXCSG-SOL1-3HS and pH35GY-SOL1-YFP, respectively. Subsequently, the SOL1–3HS and SOL1–YFP sequences were amplified from these vectors, and then cloned into pMDC7. 35S:SP-GFP-HDEL, 35S:ST-mRFP and 35S:mRFP-SYP61 were provided by I. Hara-Nishimura (Department of Botany, Graduate School of Science, Kyoto University, Japan), K. Shoda (Molecular Membrane Biology Laboratory, RIKEN Advanced Science Institute, Japan) and T. Uemura (Department of Biological Sciences, Graduate School of Science, The University of Tokyo, Japan), respectively. 35S:TagRFP-ARA7 was obtained from E. Ito (Department of Biological Sciences, Graduate School of Science, The University of Tokyo, Japan) and K. Ebine (Department of Biological Sciences, Graduate School of Science, The University of Tokyo, Japan). R4pGWB401-SOL1pro:GUS, pMDC7-CLV3, pMDC7-CLE19 and pMDC7-CLE19ΔR were transformed into Agrobacterium tumefaciens strain GV3101::pMP90 and then into Col–0 and sol1101 plants using the floral-dip method (Clough and Bent, 1998). pXCSG-SOL1-3HS was transformed into Agrobacterium tumefaciens strain GV3101::pMP90RK and used for N. benthamiana transient expression.

GUS staining

Grown plants were subjected to GUS staining treatment as follows: plants were transferred to 90% acetone for 15 min on ice, washed with 100 mm NaPO4 buffer (pH 7.0), vacuum-infiltrated at room temperature for 15 min with GUS staining solution containing 100 mm NaPO4 (pH 7.0), 10 mm EDTA, 1 mm potassium ferricyanide, 1 mm potassium ferrocyanide, 0.1% Triton X–100 and 1 mg ml−1 5-Bromo-4-chloro-3-indolyl--D-glucuronide Cyclohexylammonium Salt, and incubated at 37°C for 6 h.

Microscopic analyses

GUS-stained samples were incubated at room temperature in 70% EtOH for 1 h, further incubated at room temperature in a mixture of EtOH:acetic acid (6:1 v/v) for more than 3 h, and then stored in 70% EtOH. Samples were mounted in a mixture of chloral hydrate/glycerol/water (8 g/1 ml/2 ml), and then subjected to microscopic analysis.

Sectioning of the shoot apical meristem

GUS-stained samples stored in 70% EtOH were dehydrated in a graded ethanol. series, and embedded in Technovit 7100 resin (Heraeus Kulzer,‎) according to the manufacturer's instructions. Sections (10 μm thick) were prepared using a Leica RM2165 microtome ( and then mounted in water for microscopic observation.

Quantitative RT–PCR analysis

Extraction of total RNA from whole seedlings and quantitative RT–PCR analysis were performed as described previously (Kinoshita et al., 2010). Primer/probe pairs used for quantitative RT–PCR are as follows: SOL1–L1, SOL1–R1 and probe #80 for SOL1 (L1/R1); SOL1–L2, SOL1–R2 and probe #102 for SOL1 (L2/R2); SOL1–L3, SOL1–R3 and probe #136 for SOL1 (L3/R3); CLE19–L, CLE19–R and probe #155 for CLE19; CLV3–155F, CLV3–155R and probe #155 for CLV3; TUA4–22F, TUA4–22R and probe #22 for TUA4 (tubulin α4 chain) as a control (Universal Probe Library, Roche Applied Science,

Transient gene expression in Nicotiana benthamiana

Transient gene expression in N. benthamiana was basically performed as described previously (Voinnet et al., 2003; Kinoshita et al., 2010).

Confocal microscopic analysis

Cultures of A. tumefaciens strains GV3101 MP90 (OD600 = 1.0) carrying the p19 silencing suppressor, estrogen-inducible SOL1–YFP or an organelle marker were mixed at a ratio of 5:4:1, and then infiltrated into the leaves of N. benthamiana for transient expression assays. The leaves were infiltrated with 10 μm β–estradiol 3 days after the first infiltration. Leaf disks from the infiltrated leaves were further incubated in 10 μm β–estradiol at 25°C under dark conditions for 24 h. These leaf disks were analyzed using an LSM 710 confocal microscope (Zeiss,

Affinity purification of SOL1–3HS

Cultures of A. tumefaciens strains GV3101 MP90 (OD600 = 1.0) carrying the p19 silencing suppressor and estrogen-inducible SOL1-3HS or buffer alone (Mock) was mixed at a ratio of 1:1, and then infiltrated into the leaves of N. benthamiana for transient expression assays. Estradiol treatment was performed as above using leaf disks. Total protein was extracted from the leaf disks using twice the volume of extraction buffer (50 mm Tris/HCl, pH 8.0, 150 mm NaCl, 10% glycerol, 1% Triton X–100. Five milliliters of the lysates were centrifuged at 4°C, 20 400 g for 20 min. Then supernatants were centrifuged again at 4°C, 20 400 g for 5 min. The resulting supernatants were incubated with 50 μl anti-HA affinity matrix resin (Roche Applied Science) at 4°C for 2 h with rotation. A washing step was performed using Micro Bio-Spin® chromatography columns (Bio–Rad, The resin was washed three times with 1 ml extraction buffer, and then four times with 500 μl extraction buffer. Prior to elution, the resin was equilibrated with 1 ml elution buffer (see below) without HA peptide. For elution, the resin was kept three times with 100 μl elution buffer (50 mm Tris/HCl, pH 8.0, 0.05% Triton X–100, 1 mg ml−1 HA peptide) at 37°C for 15 min, and the flow-through was pooled. The eluates were subjected to enzymatic assays (see below). SOL1–3HS was detected by immunoblot analysis using anti-HA 3F10 (Roche Applied Science, as the primary antibody and goat anti-rat IgG-HRP (Santa Cruz Biotechnology, as the secondary antibody.

Enzymatic assays

All reactions were performed in a buffer containing 50 mm Tris/HCl (pH 5.0, 6.0, 7.0 or 8.0) and 0.05% Triton X–100 at 30°C in a final volume of 62.5 μl. A 5 μl aliquot of purified extract (SOL1–3HS, mock and elution buffer) and 1.25 μl of 1 mm substrate [Dansyl-Phe-Ala-Arg for the fluorescence assay or CLE19 + R (RVIHypTGHypNPLHNR), CLE21 + K (RSIHypTGHypNPLHNK) and CLE22 + R (RRVFTGHypNPLHNR) dissolved in 0.1% trifluoroacetic acid for MALDI-TOF MS] were added to the reaction. For detection of fluorescence, the reaction was stopped with 25 μl of 0.5 m HCl after incubation for the time shown in figure. Then, 500 μl of chloroform was added to the samples. Samples were mixed and then centrifuged at 200 g for 2 min. The fluorescence of the chloroform phase was measured at 25°C using a fluorescence spectrometer (excitation 350 nm; emission 500 nm). For MALDI-TOF MS analyses, 25 μl of anti-HA affinity matrix resin (Roche Applied Science) was added to the reaction to remove SOL1–3HS and HA peptide, and the reaction was incubated at 4°C for 1 h. After centrifugation at 4°C, 20 400 g for 5 min, the resulting supernatant was collected and centrifuged again at 40°C, 20 400 g for 5 min. Then, the supernatant and α–cyano-4–hydroxy-cinnamic acid were mixed at a 1:125 ratio, and masses of molecular contents of the mixture were analyzed using an Autoflex–N MALDI-TOF/TOF mass spectrometer (Bruker,


We thank David Baulcombe (Department of Plant Sciences, University of Cambridge) and Plant Bioscience Ltd ( for the Agrobacterium strain carrying the p19 silencing suppressor, Jane Parker (Department of Plant Microbe Interactions, Max Planck Institute for Plant Breeding Research) for pXCSG–3HS, Tsuyoshi Nakagawa (Department of Molecular and Functional Genomics, Center for Integrated Research in Science, Shimane University) for the R4pGWB401 and R4pGWB433 binary vectors, Ueli Grossniklaus (Institute of Plant Biology and Zürich-Basel Plant Science Center, University of Zürich) for pMDC7, Takashi Ueda for valuable discussions and support with the confocal microscopic observations, Kazuo Ebine for support with the confocal microscopic observations and for the 35S:TagRFP-ARA7 construct, Emi Ito for the 35S:TagRFP-ARA7 construct, Ikuko Hara-Nishimura (Department of Botany, Graduate School of Science, Kyoto University) for the 35S:SP-GFP-HDEL construct, Keiko Shoda (Molecular Membrane Biology Laboratory, RIKEN Advanced Science Institute) for the 35S:ST-mRFP construct, Tomohiro Uemura for the 35S:mRFP-SYP61 construct and Rie Kurata (Plant Global Education Project, Graduate School of Biological Sciences, Nara Institute of Science and Technology) for help with the mass spectrometric analysis. This work was supported by Grants-in-Aid from the Ministry of Education, Science, Sports and Culture of Japan to S.S. (23119517, 23012034, 24114001, 24114009, 24370024, 24657035 and 24658032), S.B. and H.F., from Japan Science and Technology Agency to S.B. and from the Japan Society for the Promotion of Science to T.T. (JSPS Research Fellow). This work was also supported by the Plant Global Education Project of Nara Institute of Science and Technology.