Results

Root Growth Assays Define Two Functional Classes of CLE-Derived Peptides.

To investigate the respective role of CLE genes in root development, Arabidopsis seedlings were grown on medium supplemented with one of 22 synthetic peptides encoded by the corresponding CLE domain. Measured root-growth rates defined two classes: 18 peptides arrested growth and 4 (CLE41p, CLE42p, CLE43p, and CLE44p) did not. They are designated A-type and B-type peptides, respectively [supporting information (SI) Fig. S1A and Table S1]. These quantitative results confirmed independent studies (11, 19, 20) with the caveat that the synthetic peptides CLE1 to CLE7 do not arrest root apical meristem (RAM) growth at 1 μM (19, 20), but do at 10 μM (compare Fig. S1 A and B).

To confirm genetically the results obtained with exogenous application of synthetic CLE peptides, we characterized available transgenic plants that overexpress representative A- and B-type genes, CLE6^OE and CLE41^OE, respectively, under the control of the CaMV 35S promoter. In brief, CLE6^OE T1 plants exhibited a short-root phenotype and microscopic analysis revealed that the activity of the RAM ceased gradually over time, similarly to its arrest observed after A-type peptide treatment. The shoot apical meristem (SAM) of CLE6^OE seedlings also ceased activity gradually. In contrast, CLE41^OE T1 plants had normal roots, but grew as compact dwarf plants and produced numerous small leaves, although the size and structure of their primary SAM seemed unaffected. Our results agree with separate studies of closely related CLE genes [(10, 21–23); for additional details, see SI Text].

Structure of Primary Meristems Treated with CLE Peptides.

To better understand how the CLE peptides triggered root-growth arrest, representatives of either classes were applied to Arabidopsis transcriptional marker lines reporting mitotic divisions [CYCB1,1_pro:GUS; (24)] and auxin response [DR5_pro:GUS; (25)]. Compared with control plants (without added peptide), the primary root basal meristem of CLE41p-treated seedlings had identical structure and GUS pattern in both lines (Fig. S2 A, B, G, H, L, and M). In contrast, CYCB1,1_pro:GUS seedlings treated with CLE6p, CLV3p, or CLE19p (all A-type peptides) had very low GUS levels in the shorter meristem, indicating that the gradual cessation of root growth correlated with the inhibition of cell division within the RAM (Fig. S2 C–E). CLE6p-treated roots also had a disorganized cellular structure with unusually large cells and differentiated vascular strands close to the meristem center (SI Text; Fig. S2N), in agreement with the previously observed terminal differentiation of the RAM induced by CLV3p, CLE19p, and CLE40p (11). None of the tested CLE-peptide treatments altered the position of the GUS maximum in DR5_pro:GUS root tips, although GUS activity was lowered after A-type peptide treatments, suggesting that RAM arrest may be associated with decreased auxin response (Fig. S2 I–K).

To determine whether the initial distinction between A- and B-type CLE peptide activity is also reflected in other organs, we studied how CLE peptides affected the SAM. SAM size or structure of plants grown in liquid culture and treated with or without CLE41p did not differ significantly, but all seedlings treated with CLE6p had a reduced SAM size, resulting in a smaller and flatter region separating emerging leaf primordia (Fig. S3).

To summarize, the RAM and SAM of plants treated with the A-type CLE6p were consumed, whereas those treated with the B-type peptide CLE41p were indistinguishable from the wild type.

A-Type and B-Type CLE Peptides Act Synergistically on Vasculature Development.

Phenotypes resulting from single CLE gene overexpression, heterologous complementation of clv3 mutations by other CLE genes, or single CLE peptide treatment suggested that subfamily members have redundant functions (Fig. S1) (23, 26). In addition, we postulated that CLE peptides might have antagonistic or synergistic modes of action. To test this hypothesis, we assayed root growth after binary treatment with A/B CLE peptides. CYCB1,1_pro:GUS seedlings transferred to solid media containing both CLE6p (at a constant concentration of 10 μM) and CLE41p (at concentrations ranging from 1–100 μM) had the same root-growth arrest as seedlings treated with CLE6p alone (Fig. 1A). In the presence of CLE6p, whether or not combined with CLE41p, GUS activity gradually decreased in the primary root meristem, until it disappeared in arrested plants (Fig. S2 C and F). Therefore, CLE41p did not influence CLE6p-induced RAM arrest. In contrast, the binary peptide treatment resulted in a striking CYCB1,1_pro:GUS activity within the stele of the mature portion of the primary root: GUS level increased with CLE41p concentration and was the highest close to the base of extended lateral roots or the collet (Fig. 1B). Furthermore, cell proliferation induced by the CLE6p/CLE41p combination resulted in radial enlargement of the mature portion of the primary root (Fig. 1 B and D).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Combined effect of CLE6p (A-type) and CLE41p (B-type). All CYCB1,1_pro:GUS plantlets were transferred to media supplemented with the indicated peptide(s), 3 dag (n = 10 for each treatment). (A) Root growth rate after transfer. (B) GUS expression patterns in mature primary root tissue (7 dat). (C) First true leaves (10 dat; Top), cotyledons (7 dat; Middle), and hypocotyls (7 dat; Bottom). (D) Hypocotyl sections showing vascular GUS expression (7 dat in liquid medium) and counterstained with ruthenium red. In C and D, initial concentration of each added peptide was 10 μM. [Scale bars: 100 μm in B, C (Bottom), and D; 1 mm in C (Top and Middle).]

Stele radial enlargement induced by binary peptide treatment was not restricted to the root. Indeed, ectopic cell proliferation was also observed in the vascular cells of shoot organs (cotyledons, leaves, and hypocotyls) when CYCB1,1_pro:GUS seedlings were grown in liquid culture supplemented with all binary A/B combinations tested: CLE6p/CLE41p, CLV3p/CLE41p, CLE19p/CLE41p, CLE6p/CLE42p, CLV3p/CLE42p, and CLE19p/CLE42p (Fig. 1 C and D; data not shown). CLE peptides appeared to function locally, because transcriptional marker changes and vascular cell proliferation were restricted to the root (the sole organ in direct contact with the added peptides) in plantlets grown on solid medium only, but occurred in both root and shoot tissues in plantlets treated in liquid medium. The quantification of stele radial enlargement observed in wild-type hypocotyls grown in liquid medium with different combinations of CLE6p and CLE41p confirmed their synergistic activity on vascular development (Fig. 2 A and B).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Radial enlargement of the hypocotyl stele after CLE6p and CLE41p treatments. (A) Hypocotyl stele width of the wild type (Col-0) transferred 3 dag and grown for 15 days in liquid medium supplemented with the indicated peptide(s) (n = 15). (B) Representative hypocotyls. Samples are aligned according to labels of abscissa in A. Arrows indicate the measured width. (Scale bar, 100 μm.) (C) Hypocotyl stele width of wild-type (Ler), clv1–1, and clv2–1 seedlings transferred 3 dag and grown for 10 days in liquid medium with or without combined CLE6p/CLE41p (n = 10). The asterisk indicates samples significantly different from control without peptide. (Student t test P < 0.05.)

Components acting downstream in the pathway(s) relaying the CLE6p/CLE41p signal are not known. However, we determined that CLV1 and CLV2, coding for receptor-like proteins located in the plasma membrane and required for the perception of the CLV3 peptide in the SAM (27), are not necessary for CLE-induced hypocotyl enlargement, because clv1–1 and clv2–1 mutants had a response similar to that of the wild type after CLE6p/CLE41p binary treatment (Fig. 2C).

CLE Peptides Induce Vascular Proliferation and Delay Differentiation.

Because the increase in GUS activity induced by A/B peptide combinations in CYCB1,1_pro:GUS plantlets was homogeneous along the vasculature of the hypocotyl but not in the root (Fig. 1 B and C), we chose to use methodically the hypocotyl stele enlargement phenotype to examine the events leading to ectopic cell divisions.

Sections of CYCB1,1_pro:GUS hypocotyls grown without added peptides or with CLE6p alone were indistinguishable 7 or 10 days after transfer (dat) in liquid medium (Fig. 1D and Fig. 3 A and B). Hypocotyls treated with CLE6p/CLE41p had unusually high GUS activity in most cells of the stele, except those at the phloem poles and in mature xylem (Fig. 1D). In these tissues, GUS staining coincided with a remarkably higher number of small cells and a disrupted arrangement of mature xylem bundles (Figs. 1D and 3C). CLE41p alone also caused the dispersion of mature xylem elements among the smaller and, presumably, actively dividing cells, but to a lesser extent than the CLE6p/CLE41p combination (Fig. 3D).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Cellular details of vascular tissues in hypocotyls treated with CLE peptides. Hypocotyls were sectioned directly above the collet and stained with ruthenium red. Seedlings were transferred 3 dag and grown for 15 days in liquid medium supplemented with (A) no peptide, (B) CLE6p, (C) CLE6p and CLE41p, and (D) CLE41p (n = 10). (A, B, and D) Black arrows indicate phloem poles. Initial concentration of each added peptide was 10 μM. (Scale bars, 50 μM.)

To better specify which cells proliferated when exposed to CLE peptides, we treated APL_pro:GUS and ATHB8_pro:GUS plants. These lines report the transcriptional activity of Altered Phloem Development (APL) marking protophloem, companion cells and metaphloem sieve elements (28), and HOMEOBOX GENE 8 marking procambium and protoxylem cells (29, 30; Y. Helariutta, personal communication), respectively. In APL_pro:GUS hypocotyls exposed to CLE41p, phloem strands lacked polar organization with respect to the stele compared to CLE6p or control samples. GUS activity was detected in hypocotyls treated with the combined CLE6p/CLE41p, but only in isolated irregular strands (Fig. 4A and Inset). On the contrary, GUS expression in ATHB8_pro:GUS hypocotyls significantly increased in the presence of CLE41p alone, and further increased and expanded to a larger domain in the presence of the combined CLE6p/CLE41p, compared with CLE6p or control samples (Fig. 4B).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Defects in vascular development after CLE peptide treatment. GUS expression in (A) APL_pro:GUS and (B) ATHB8_pro:GUS hypocotyls. Seedlings were transferred 3 dag and grown for 15 days in liquid medium supplemented with the peptide(s) indicated, each at an initial concentration of 10 μM (n = 10). (Scale bars, 50 μm.) (Inset) Longitudinal section through a hypocotyl stele showing xylem (X) and phloem strands (PS). (Scale bar, 10 μm.)

To investigate on which pathways CLE41p (in combination with CLE6p) impinges to promote proliferation in the stele of the hypocotyl, we analyzed the occurrence of key transcriptional changes, relative to vascular development and cell division, after 3 h to 10 days of peptide exposure. Compared with untreated controls, the first changes in stele width were observed 2–4 days after the start of peptide treatment, regardless of the reporter lines (Fig. S4).

The first transcriptional changes induced by CLE6p/CLE41p occurred much earlier (from 3 h onward) in the ATHB8_pro:GUS and ET1335:GUS lines marking the onset of procambium formation (31). CLE41p treatment alone induced transcriptional changes within 1 day in ATHB8_pro:GUS plants, whereas the ATHB8 and ET1335 markers were only visible at 4 and 7 days, respectively, in plants grown in the absence of added peptide (Fig. S4 A and B). A relative increase in cell proliferation reported by GUS activity within the stele of CYCB1_pro:GUS hypocotyls was detectable 7 days after CLE6p/CLE41p treatment, compared with 10 days in the untreated control (Fig. S4C). On the contrary, GUS activity reporting differentiation of phloem cells in APL_pro:GUS plants was delayed from days 7–10 by CLE6p/CLE41p (Fig. S4D).

In summary, CLE41p first increased ATHB8 and ET1335, procambium and protoxylem cell-identity markers, and this induction was enhanced and advanced by the presence of CLE6p. The characterization of hypocotyls treated for ≥7 days suggested that CLE6p/CLE41p maintained an abnormally high number of cells into a proliferative mode (Fig. 3C and Fig. 4B), delaying downstream vascular differentiation. This observation is in agreement with the promotion of Zinnia mesophyll cell division and suppression of their transdifferentiation into tracheary elements by CLE41 (19).

The synergistic action of A- and B-type peptides was confirmed genetically in F1 plants resulting from crosses between CLE6^OE, CLE41^OE, and wild-type lines. When compared with wild-type plants, hemizygous CLE41^OE and CLE6^OE plants had a retarded rosette growth and a typical A-type CLE gain-of-function SAM arrest phenotype, respectively. The F1 plants carrying both the CLE6^OE and CLE41^OE transgenes displayed a strongly stunted and bushy phenotype (Fig. S5A). This dramatic synergistic effect extended to vascular development: When compared to the wild-type, CLE6^OE/-, or CLE41^OE/- plants, hypocotyl sections from CLE6^OE/-;CLE41^OE/- F1 plants had a mass of cytoplasm-dense small cells and a disrupted arrangement of secondary xylem with dispersed clusters of mature xylem vessels, as also observed in vascular bundles of hypocotyls treated with CLE6p/CLE41p (Fig. S5B).

However, the SAM arrest coupled with the activation of axillary meristems observed in CLE6^OE/-;CLE41^OE/- plants suggested that CLE6-induced SAM arrest was not suppressed by CLE41 overexpression. This result is in agreement with the inability of synthetic CLE41p to suppress CLE6p-induced SAM or RAM arrest, and confirms that CLE6 and CLE41 are not antagonistic with regard to apical meristem function (for additional details, see SI Text).

Auxin Mediates A/B CLE-Induced Proliferation.

Polar auxin transport is a key process in vascular development (32). Therefore, we investigated the potential interplay between auxin and CLE actions. A shoot-derived signal was found to be required for CLE-induced cell proliferation, because stele enlargement was not enhanced in wild-type seedlings decapitated 5 days after germination (dag) and treated with CLE6p/CLE41p for 10 days (Fig. 5A). We compared peptide-treated seedlings grown in liquid culture supplemented with either the synthetic auxin 1-naphthaleneacetic acid (NAA, 1 μM) or the polar auxin transport inhibitor 1-naphthylphthalamic acid (NPA, 1 μM). CLE6p/CLE41p-driven enlargement was markedly enhanced by NAA, but reduced by NPA (Fig. 5B), consistent with the hypothesis that auxin is involved in CLE-induced vascular proliferation.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

Interaction between CLE peptide and hormone signaling. (A) Hypocotyl stele width of whole or decapitated wild-type seedlings transferred at 3 and 5 dag, respectively, and for 10 days into liquid medium supplemented with the indicated peptide(s) (n = 15). (B) Treatments combining CLE peptides, auxin, and auxin transport inhibitor. Ten seedlings were transferred 3 dag and grown for 10 days either with NAA (1 mM) or NPA (1 μM), in combination with no peptide, CLE6p, CLE41p, or CLE6p/CLE41p. Initial concentration of each added peptide was 10 μM in both experiments. The asterisk marks samples significantly different from mock-treated controls. (Student t test P < 0.05.)

Four additional reporter lines were tested to study auxin-related transcriptional changes during CLE-induced radial enlargement of the hypocotyl stele: DR5_pro:GUS, IAA2_pro:GUS (33), PIN1_pro:GUS (34), and PIN3_pro:GUS (35). In the absence of peptide, GUS staining was detected only in the stele of PIN3_pro:GUS hypocotyls at 7 and 10 days after treatment. Upon CLE6p/CLE41p treatment, GUS staining was observed earlier (day 4) in PIN3_pro:GUS hypocotyls, and was detectable at day 7 in the other 3 lines. The transcriptional activation of the DR5, IAA2, PIN1, and PIN3 genes in the enlarged steles (Fig. S4 E–H) hinted at an auxin response concomitant with increased cell proliferation.

Auxin stimuli are relayed transcriptionally by auxin response factors (ARFs). In particular, MONOPTEROS(MP)/ARF5 mediates auxin responses involved in vascular development (36). Therefore, we tested its involvement in CLE6p/CLE41p-induced cell proliferation. In wild-type leaves, numerous small cells parallel to mature xylem appeared after 10 days of binary peptide treatment, but under the same conditions these cells are absent in leaves homozygous for mpG92, a weak monopteros mutant allele, even though such leaves form rudimentary vascular-cell files (Fig. S6).

To conclude, we propose that our observations can be explained by the following sequence of events: A and B-type peptides act synergistically to prime specific vascular cell types for division; auxin level or auxin sensitivity increase in these cells, resulting in proliferation and delaying vascular differentiation into phloem and xylem. This model is in agreement with the observation that, in trees, auxin concentration peaks within the actively dividing zone of the cambial meristem and that the radial width of the auxin concentration gradient correlates with cambial growth rate (37). Further information, including Figs. S7–S11, is available in the SI Text.