Cytokinin Signaling and Its Inhibitor AHP6 Regulate Cell Fate During Vascular Development

See allHide authors and affiliations

Science  06 Jan 2006:
Vol. 311, Issue 5757, pp. 94-98
DOI: 10.1126/science.1118875


The cell lineages that form the transporting tissues (xylem and phloem) and the intervening pluripotent procambial tissue originate from stem cells near the root tip. We demonstrate that in Arabidopsis, cytokinin phytohormones negatively regulate protoxylem specification. AHP6, an inhibitory pseudophosphotransfer protein, counteracts cytokinin signaling, allowing protoxylem formation. Conversely, cytokinin signaling negatively regulates the spatial domain of AHP6 expression. Thus, by controlling the identity of cell lineages, the reciprocal interaction of cytokinin signaling and its spatially specific modulator regulates proliferation and differentiation of cell lineages during vascular development, demonstrating a previously unrecognized regulatory circuit underlying meristem organization.

The root vascular cylinder has a central axis of xylem cell files consisting of protoxylem at marginal positions and metaxylem at central positions. This axis is flanked by phloem and intervening procambial cell files. A proportion of these intervening procambial cell files becomes mitotically active during secondary development and forms the lateral meristem, cambium, through periclinal divisions (1, 2) (Fig. 1A). Cytokinins have been implicated in controlling vascular morphogenesis (25). The wooden leg (wol) allele of CRE1 and the triple-knockout mutant for all three genes encoding CRE-family receptors (CRE1/WOL/AHK4, AHK2, and AHK3) display a markedly reduced number of cell files within the vascular bundle, because the periclinal procambial cell divisions required to proliferate the vascular cell files do not occur. This is associated with specification of all the vascular cell files in the root as protoxylem (2, 6, 7) (fig. S3B). This phenotype can be copied through depleting cytokinins by expressing the CYTOKININ OXIDASE 2 gene (8) under the control of the procambium-specific CRE1 promoter (fig. S3B), indicating that cytokinin signaling through the CRE-family receptors is required for proliferation and/or maintenance of the procambium.

Fig. 1.

Cytokinin signaling inhibits protoxylem differentiation. (A) Schematic of Arabidopsis root vasculature from a mature (transverse section, top) and meristematic procambial regions (three-dimensional representation, bottom). The pattern of periclinal cell divisions of the intervening procambial tissue characteristic to secondary development is shown with dotted red lines. (B) Histological analysis of root vasculature stained with toluidine blue (top) and fuchsin red (bottom, where provided) indicates that protoxylem is more abundant in genetic backgrounds with reduced cytokinin signaling, and ahp6 is able to suppress wol and cre1 ahk3. Fuchsin staining highlights lignified regions; the helical staining pattern is characteristic of protoxylem, whereas the pitted or reticulate pattern is characteristic of metaxylem. Inducible CKX1 indicates that CKX1-YFP was induced postgermination in a tissue-specific manner with the use of CRE1prom::XVE/LexAoperator::CKX1-YFP in wild-type roots; Wt with CK, wild-type seedling germinated with 100 nM benzyladenine (a cytokinin); 2ndary, during secondary development. Yellow arrows, protoxylem elements; white arrows, absence of protoxylem; yellow arrowheads, metaxylem elements; red arrowheads, sieve elements of protophloem; asterisks, pericycle cells; black arrowheads, periclinal cell division in intervening procambial cells; and red arrow, periclinal cell division of an undifferentiated cell at the protoxylem position. Scale bar, 20 μm.

To investigate whether reduced cell number is a prerequisite for exclusive protoxylem differentiation, we depleted cytokinins post-embryonically by expressing cytokinin oxidase 1–yellow fluorescent protein (YFP) under the CRE1 promoter, in an estrogen-inducible fashion (8, 9). After induction, all cell files within the root vascular cylinder differentiated as protoxylem (Fig. 1B and fig. S1). Furthermore, when wild-type seedlings were grown on media containing cytokinin, protoxylem differentiation and green fluorescent protein (GFP) expression of the enhancer trap line J0121, which highlights the protoxylem-associated pericycle cells (10), were severely impaired or even eliminated (Figs. 1B and 2E and figs. S2 and S4B). Together, these results indicate that cytokinin signaling is required to promote and maintain cell identities other than protoxylem, and in the absence of cytokinin signaling, protoxylem is the “default” identity.

Fig. 2.

Phenotypic analysis of ahp6 mutants. (A) From left to right: wild type, wol, wol ahp6-1, wol ahp6-2, and wol ahp6-1 complemented with AHP6prom::AHP6. (B) The adventitious root formation assay indicates that ahp6 mutants confer increased cytokinin responsiveness in wol background. Cytokinins normally inhibit the formation of adventitious roots near the cut end of a hypocotyl. The percent of explants producing adventitious roots when grown on media with increasing concentrations of t-zeatin (a cytokinin) is presented. Significant differences existed between wol and wol ahp6-1 (blue asterisk) or between wol and wol ahp6-2 (red asterisk), according to the Fisher's exact probability test (P < 0.05). (C) In situ hybridization with an antisense probe reveals ARR15 expression at the protoxylem position in ahp6-1. A sense probe showed no detectable signal (12). (D) Loss of protoxylem differentiation is accompanied with loss of ZCP4prom::GUS expression in ahp6-1. (E) Expression of J0121, a GFP marker line with expression specifically in protoxylem-associated pericycle cells (Fig. 1A and fig. S2) is down-regulated after a 48-hour treatment of benzyladenine. Panels shown are representative of analyses performed with 15 to 20 individuals. Black arrow, protoxylem position; asterisks, pericycle cells. Scale bars, (A) 5 mm; (C) to (E), 20 μm.

To identify molecules regulating cytokinin-mediated vascular morphogenesis, we performed a suppressor screen for the determinate root growth associated with wol (6). Two allelic, recessive, extragenic suppressor mutations were identified, ahp6-1 and ahp6-2 (Fig. 2A). Compared with wol, wol ahp6 roots display an increased number of vascular cell files with intervening procambial and phloem cell files present (Fig. 1B and fig. S3, A and C). Neither ahp6-1 nor ahp6-2 plants display an obvious seedling phenotype. However, histological studies revealed that protoxylem differentiation occurred sporadically along the root (Fig. 1B and figs. S3D and S4). Quantitative analysis showed that this phenotype is less prominent in ahp6-2, indicating that ahp6-1 is a stronger allele (fig. S4B). Expression of a molecular marker highlighting immature xylem cells (ZCP4prom::GUS) (11) coincides with visible xylem differentiation in wild-type and ahp6-1 roots (Fig. 2D). Furthermore, expression of the GFP transcript in the enhancer trap line J0121 (10) is slightly down-regulated in ahp6-1 roots (Fig. 2E) (12). To further investigate cell fate at the protoxylem position in ahp6-1 roots, we analyzed the vascular pattern during secondary development when the intervening procambial cell files undergo periclinal cell divisions in wild-type and ahp6-1 roots (Fig. 1A). Notably, these cells underwent periclinal divisions simultaneously with the neighboring procambial cells files (Fig. 1B and fig. S5). In wild-type roots, this position is invariably differentiated as protoxylem (fig. S4B). In conclusion, the ahp6 phenotypes in wol and in wild-type roots reveal a role for the AHP6 locus in promoting the specification of protoxylem.

To characterize the AHP6 locus at the molecular level, we determined through positional cloning that both ahp6-1 and ahp6-2 contain mutations in the gene ARABIDOPSIS HISTIDINE PHOSPHOTRANSFER PROTEIN 6 (AHP6), At1g80100 in the Arabidopsis Genome Initiative database (Fig. 3A). A genomic DNA fragment containing the AHP6 locus complemented ahp6-1 (Fig. 2A and fig. S3C). Current evidence supports a model of signal transduction in which cytokinins first activate phosphotransfer from adenosine 5′-triphosphate (ATP) to the three CRE-family receptors, and then the phosphoryl group is transferred to the conserved histidine residue of the AHP proteins (AHP1 to AHP5) and finally to response regulators governing physiological responses (13, 14). AHP6 lacks the conserved histidine residue (fig. S6) (Asn83 in AHP6b), which is required for phosphotransfer and is present in the other AHPs (13, 15). We identified cDNA for two independent transcripts (AHP6a-DQ093642 and AHP6b-DQ093643), differing in the length of the first exon (Fig. 3, A and B). The mutation in ahp6-1 resulted in a premature stop codon in the first exon (CAG to TAG, at Gln35), and the mutation in ahp6-2 (G to A) was located in the first intron, 5 base pairs from the 5′ border of the AHP6b splice variant. In ahp6-2 seedlings, only the AHP6a transcript is present (Fig. 3B). Given that this represents a weak allele, it is likely that both transcripts are functional. We also identified a transferred DNA (T-DNA) insertion allele (at Ser111 of AHP6b), ahp6-3, which phenocopies ahp6-1 (Fig. 3A and figs. S3D and S4B). Because the open reading frame is terminated early in ahp6-1, and because only reduced level of AHP6 message was detected upstream and no message was detected downstream of the T-DNA insertion in ahp6-3, these are likely to represent null alleles (Fig. 3B).

Fig. 3.

Molecular characterization of the AHP6 locus. (A) Structure of the AHP6 gene. White boxes, untranslated regions; gray lines, the alternative splicing between AHP6a (a) and AHP6b (b) variants. Locations of the ahp6 mutations are shown by arrows (single nucleotide polymorphisms) or arrowheads (T-DNA insertion, stock number SALK_058085). (B) Analysis of both AHP6 transcripts by reverse transcription polymerase chain reaction (RT-PCR). UBQ10 is an internal control. The smaller transcript only becomes apparent in wild-type and ahp6-1 roots after a greater number of cycles (12). (C) AHP6 is a pseudo-HPt because of the lack of the conserved His. Upper panels demonstrate that AHP6 inhibits phosphorelay from the SLN1 histidine kinase domain (SLN1HisK) to the SLN1 receiver domain (SLN1Rec). Lower panels indicate that AHP6 inhibits phosphorelay from phosphorylated AHP1 to ARR1. Radioactivity is measured in relative units and each value was quantified from four independent reactions (standard deviation indicated). (D and E) AHP6 is expressed in protoxylem and adjacent pericycle cells as shown by in situ hybridization and with the AHP6prom::GFP reporter construct. The sense AHP6 probe did not reveal any signal (inset). px, protoxylem; pc, pericycle. (F) AHP6prom::GFP expression during embryogenesis in early torpedo stage of wild-type and wol embryos. Scale bars, 20 μm.

To investigate the biochemical nature of AHP6 function, we based an in vitro assay around the already well-characterized phosphotransfer activity of the yeast SLN1 histidine kinase (16). The histidine kinase domain (SLN1HisK) and the receiver domain (SLN1Rec) were prepared as separate peptides, enabling examination of interdomain phosphotransfer (16). In this assay, the phosphoryl group is transferred from 32P-labeled ATP to a conserved His residue of SLN1HisK, then to a conserved Asp residue within the SLN1Rec, and subsequently to a conserved His residue within an HPt protein. Using this in vitro phosphotransfer system, we demonstrated that AHP1 to AHP3, AHP5, and a mutant version of AHP6b (Asn83 replaced with the conserved His required for phosphotransfer activity) were able to accept a phosphoryl group from SLN1 (Fig. 3C and fig. S7). In contrast, native AHP6 was unable to accept a phosphoryl group. These results suggest that AHP6 does not function as a phosphotransfer protein. Although AHP6 does not appear to have phosphotransfer activity, it was able to inhibit phosphotransfer from the SLN1HisK to SLN1Rec. We also examined the effect of AHP6 on phosphotransfer from AHP1 to an Arabidopsis response regulator, ARR1 (Fig. 3C). When phosphorylated AHP1 and ARR1 were co-incubated, phosphotransfer occurred from phosphorylated AHP1 to ARR1, and this was inhibited by AHP6 (Fig. 3C). Taken together, the evidence suggests that AHP6 acts as an inhibitor of cytokinin signaling by interacting with the phosphorelay machinery, potentially at multiple steps. Database searches using the AHP6 sequence reveal homologs in other plant species (fig. S6), indicating that the negative regulation of the His-to-Asp phospho-relay system by pseudo-HPts may be widespread in the plant kingdom.

Using in situ RNA hybridization, and the AHP6prom::GUS and AHP6prom::GFP reporter constructs, we observed expression of AHP6 in developing protoxylem and associated pericycle cell files, revealing a spatially specific expression pattern consistent with the mutant phenotype (Figs. 3, D and E, and 4A; fig. S8A). GFP signal is ubiquitous in the cotyledons of heart-shaped embryos and is gradually restricted to two poles by the late heart stage/early torpedo stage of embryogenesis (Fig. 3F and fig. S8C). By the mature stage, GFP expression is refined to the cotyledon apices and two poles in the embryonic root which later form protoxylem. AHP6 is also expressed in the shoot apex and young leaves (fig. S8A). The existence of the AHP6 protein was confirmed by complementing ahp6-1 with AHP6prom::AHP6-GFP and observing GFP signal (fig. S8B).

Fig. 4.

Cytokinin negatively regulates AHP6 expression. (A) The expression domain of AHP6 broadens in genetic backgrounds with decreased cytokinin signaling. White asterisks, pericycle cells expressing AHP6; black asterisks, other pericycle cells. (B) AHP6 expression is down-regulated following cytokinin treatments as shown by real-time quantitative PCR. Error bars show means ± SD. (C) With intermediate cytokinin concentrations AHP6 is expressed in a sporadic manner. Panels shown represent analyses performed with 15 to 20 individuals. (D) A model showing the reciprocal interaction of cytokinin signaling and AHP6 in regulating the balance between the maintenance of procambial cell identity (PC) and the differentiation of protoxylem elements (PX). Scale bars, 20 μm.

To examine the role of AHP6 in cytokinin signaling, cytokinin responsiveness conferred by ahp6 was measured with the use of the adventitious root formation assay (17). The impaired cytokinin response in wol was partially restored in wol ahp6 (Fig. 2B, consistent with the negative role of AHP6 on cytokinin signaling. We also demonstrated that ahp6-1 could not suppress the cytokinin-insensitive cre1 ahk2 ahk3 phenotype, indicating that the suppressor phenotype in the wol background is due to residual cytokinin signaling (17) (Fig. 2B and fig. S3C). To study the status of cytokinin signaling in ahp6-1 at the position normally occupied by protoxylem files, we performed in situ RNA hybridization, with the cytokinin primary response gene ARR15 (18) as a probe (Fig. 2C). Whereas ARR15 expression in wild-type roots was restricted to the intervening procambial cells adjacent to the xylem axis (n = 4), the expression domain in four out of five ahp6-1 roots expanded to include the protoxylem position. These results indicate that AHP6 functions to facilitate protoxylem specification by down-regulating cytokinin signaling in a spatially specific manner. To confirm this, we examined whether the cytokinin-degrading enzyme CKX2 (8) could substitute AHP6 when ectopically expressed in the protoxylem. AHP6prom::CKX2 suppressed the ahp6-1 phenotype (fig. S3D). Furthermore, when ahp6-1 plants were germinated on media containing cytokinin, the loss-of-protoxylem phenotype was enhanced (figs. S2A and S4B), supporting the role of AHP6 as a negative regulator of cytokinin signaling and implicating additional unidentified factors.

cre1 ahk3 shows intermediate cytokinin responsiveness (17). Histological studies revealed that 91% of cre1 ahk3 roots (n = 35; wild type 0%, n = 50) had one or more extra protoxylem files, even though the number of procambial cell files remained normal (Fig. 1B) (cre1 ahk3: 32.6 ± 2.0 cells, n = 17; wild-type: 31.8 ± 2.2 cells, n = 20; errors are SD). Similar phenotypes are observed in weak CRE1prom::CKX2 lines (fig. S3B). To investigate whether the ectopic protoxylem files in these genotypes are due to an increase in the spatial domain of AHP6 action, we first studied the expression of AHP6 in cre1 ahk3. Consistent with this, the AHP6 expression domain is slightly broader than in wild-type roots, because it is typically found in two (as opposed to one) protoxylem cell files and in three (as opposed to two) adjacent pericycle cell files (88%; n = 17) (Fig. 4A). This expanded domain of AHP6 expression is responsible for the cre1 ahk3 phenotype, because loss of AHP6 function in ahp6-1 cre1 ahk3 was able to suppress the ectopic protoxylem (91%; n = 45) (Fig. 1B). In both wol and cre1 ahk2 ahk3, the AHP6 expression pattern expands throughout the vascular bundle (Fig. 4A) (12). The expanded expression pattern in wol is already evident by the early torpedo stage of embryogenesis when it occupies one broad domain within the embryonic root as opposed to two narrow strands in wild-type (Fig. 3F and fig. S8C). This indicates that cytokinin signaling specifies the spatial domain of AHP6 expression upstream of protoxylem differentiation, which occurs after embryogenesis. Next, we examined the effect of exogenous cytokinins on AHP6 expression. We observed down-regulation of the AHP6 transcript after a 6-hour treatment with cytokinins (Fig. 4B). Likewise, the level of fluorescence in the AHP6prom::GFP line was reduced by cytokinins, and the reduction occurred at lower levels of cytokinin in ahp6-1 than in wild-type roots. (Fig. 4C and fig. S9). In the absence of applied cytokinin, the levels of AHP6 transcript in ahp6-1 were slightly lower than in the wild type (12).

We report a regulatory circuit between cytokinin signaling and its newly identified inhibitor, AHP6, which specifies the meristematic versus differentiated nature of procambial cell files (Fig. 4D). In this sense, our results are consistent with requirement of cytokinins for transdifferentiation of xylem observed in Zinnia mesophyll cell culture (5, 19). AHP6 can be considered the founding member of a new “pseudo” subclass of HPt proteins within the wider group present in prokaryotes and eukaryotes.

Supporting Online Material

Materials and Methods

SOM Text

Figs. S1 to S9


References and Notes

View Abstract

Navigate This Article