PHABULOSA Mediates an Auxin Signaling Loop to Regulate Vascular Patterning in Arabidopsis

Abstract

Plant vascular tissues, xylem and phloem, differentiate in distinct patterns from procambial cells as an integral transport system for water, sugars, and signaling molecules. Procambium formation is promoted by high auxin levels activating class III homeodomain leucine zipper (HD-ZIP III) transcription factors (TFs). In the root of Arabidopsis (Arabidopsis thaliana), HD-ZIP III TFs dose-dependently govern the patterning of the xylem axis, with higher levels promoting metaxylem cell identity in the central axis and lower levels promoting protoxylem at its flanks. It is unclear, however, by what mechanisms the HD-ZIP III TFs control xylem axis patterning. Here, we present data suggesting that an important mechanism is their ability to moderate the auxin response. We found that changes in HD-ZIP III TF levels affect the expression of genes encoding core auxin response molecules. We show that one of the HD-ZIP III TFs, PHABULOSA, directly binds the promoter of both MONOPTEROS (MP)/AUXIN RESPONSE FACTOR5, a key factor in vascular formation, and IAA20, encoding an auxin/indole acetic acid protein that is stable in the presence of auxin and able to interact with and repress MP activity. The double mutant of IAA20 and its closest homolog IAA30 forms ectopic protoxylem, while overexpression of IAA30 causes discontinuous protoxylem and occasional ectopic metaxylem, similar to a weak loss-of-function mp mutant. Our results provide evidence that HD-ZIP III TFs directly affect the auxin response and mediate a feed-forward loop formed by MP and IAA20 that may focus and stabilize the auxin response during vascular patterning and the differentiation of xylem cell types.

Figure 1.

Alterations in the levels of HD-ZIP III TFs lead to changes in auxin responses in the Arabidopsis root. A, Schematic representation of the Arabidopsis root meristem depicting the general pattern of stele cell type organization in longitudinal and radial views. B, Xylem pattern in pCRE1>>MIR165A roots after 0 and 24 h of β-estradiol induction. m, Metaxylem; p, protoxylem. C, Quantitative real-time PCR (qPCR) analysis of AHP6 and ACL5 expression in root tips after induction of pCRE1>>MIR165A. *, P < 0.05, Student’s t test; error bars indicate sd. D, Radial (top) and longitudinal (bottom) confocal images of pAHP6::GFP in pCRE1>>MIR165A root meristems after 0, 15, and 40 h of β-estradiol induction. E, DR5::GUS staining in pCRE1>>MIR165A roots after 0 and 24 h of β-estradiol induction. F and G, DR5rev::GFP expression in Columbia (Col-0; F) and athb8 cna phb phv (G) as seen in longitudinal and radial confocal images. Dotted lines indicate positions of the radial cross sections at 20 and 60 µm above the QC. H and I, DR5::GUS staining in 3-d-old seedlings of Col-0 (H) and phb-7d (I) following 24 h of growth on mock or 1 µm IAA-containing plates. Cell walls are stained with propidium iodide and appear magenta in D, F, and G. Fractions indicate the frequency of the observed staining or fluorescence pattern. Yellow or black arrowheads mark the QC (longitudinal images), and white arrowheads mark the xylem axis (radial images). Bars = 25 µm, except in B, where bars = 10 µm.

Figure 2.

IAA20 and IAA30 are under the control of the HD-ZIP III TFs. A and B, qPCR analyses of IAA20 and IAA30 expression in root tips of pCRE1>>MIR165A after β-estradiol induction (A) or the cna phb phv and athb8 cna phb phv mutants (B), compared with the wild type (Col er-2). C, pIAA30::n3GFP in longitudinal and radial confocal images at 20 and 80 µm above the QC. Dotted lines indicate positions of the radial cross sections. The yellow arrowhead marks the QC (longitudinal image), and white arrowheads mark the xylem axis (radial images). Cell walls are stained with propidium iodide and appear magenta. Bars = 25 µm. D, At top is a schematic of the IAA20 gene and promoter region with amplified ChIP-qPCR fragments A, B, and C in base pairs relative to the ATG start site (not to scale). Untranslated regions are indicated in black, exons in white, and introns in grey. At bottom is a ChIP-qPCR analysis showing the means of four independent experiments. Anti-GFP and anti-IgG antibodies were used to precipitate chromatin prepared from pPHB::PHBd:GFP seedlings. Enrichment is shown relative to the signal of negative control sequences within the non-HD-ZIP III targets ABI2 and OTC. The anti-IgG was used as an antibody control to quantify nonspecific antibody binding. Error bars indicate the se from four independent biological replicates. *, P < 0.05 between the enrichment for the control IgG and the GFP enrichment (general linear model ANOVA test). E, qPCR analysis of IAA20 and IAA30 expression in root tips of pCRE1>>phb-1d after β-estradiol induction. F, qPCR analysis of IAA20 and IAA30 expression in root tips of the wild type (C24) and phb-7d. In A, B, E, and F, error bars indicate sd; *, P < 0.05, Student’s t test.

Figure 3.

IAA20 and IAA30 influence xylem patterning and affect the expression of auxin-responsive genes. A, Frequency of xylem phenotypes observed in roots of each of the genotypes listed. B to E, Representative images of xylem phenotypes observed in Col-0 × Landsberg erecta (Col-0 × Ler; B), iaa20-2 iaa30-1 (C), 35S::IAA30 line 17 (D), and mp^S319 (E). Asterisks indicate lack of protoxylem. F, qPCR analysis of MP and TMO5 expression in wild type (Col-0) and 35S::IAA30 line 17 root tips. G, qPCR analysis of the five HD-ZIP III genes in wild type (Col-0) and 35S::IAA30 line 17 root tips. H, qPCR analysis of IAA2, AHP6, and TMO5 in wild-type (Col-0 × Ler) and iaa20-2 iaa30-1 root tips. I, Representative images of xylem phenotypes observed in phb-7d and iaa30-1 phb-7d mutants. Arrowheads indicate a break or switch in xylem strand identity. J, Frequency of xylem phenotypes observed in wild-type (C24 × Col-0), phb-7d (C24 × Col-0), and iaa30-1 phb-7d roots. m, Metaxylem; p, protoxylem. Error bars indicate sd; *, P < 0.05, Student’s t test. Bars = 25 µm.

Figure 4.

HD-ZIP III TFs mediate MP expression level, and PHB is localized to the MP promoter in vivo. A, qPCR analysis of MP expression in pCRE1>>MIR165A root tips at the indicated induction times. B, Expression of pMP::n3GFP in Col-0 in longitudinal and radial confocal images at 20 and 80 µm above the QC. The yellow arrowhead marks the QC (longitudinal image), and white arrowheads mark the xylem axis (radial images). Dotted lines indicate positions of the radial cross sections. Bars = 25 µm. C, qPCR analysis of MP expression in three double loss-of-function HD-ZIP III mutants relative to the respective wild type (Col er-2) collected in parallel (wt^A–wt^C). D, qPCR analysis of MP expression in pCRE1>>phb-1d root tips at the indicated induction times. In A, C, and D, error bars indicate sd; * P < 0.05, Student’s t test. E, At top is a schematic of the MP gene and upstream region indicating the ChIP-qPCR fragment A, positioned at −450 bp, and fragment B, positioned at −700 bp upstream of the ATG start site (not to scale). Untranslated regions are indicated in black, exons in white, and introns in grey. At bottom is a ChIP-qPCR analysis showing means of three or four independent experiments. Anti-GFP and anti-IgG antibodies were used to precipitate chromatin prepared from pPHB::PHBd:GFP seedlings. Enrichment is shown relative to the signal of negative control sequences within the non-HD-ZIP III targets ABI2 and OTC. The anti-IgG was used as an antibody control to quantify nonspecific antibody binding. Error bars indicate the se from four (for position A) or three (for position B) independent biological replicates; * P < 0.05 between the enrichment for the control IgG and the GFP enrichment (general linear model ANOVA test). F, A model summarizing our results in light of known interactions. Previously, it was shown that MP directly regulates its own expression (Lau et al. 2011) as well as the expression of IAA20 (Krogan et al. 2014), that IAA20 and IAA30, which are stable in the presence of auxin (Dreher et al. 2006), can interact with MP (Vernoux et al. 2011), and, therefore, that IAA20 and IAA30 likely repress MP function in their overlapping activity domains. Here, we show that HD-ZIP III TFs are needed for normal expression of IAA20/IAA30 and MP, and we hypothesize that PHB and MP cooperatively may activate IAA20 and IAA30. In conditions of changing auxin levels in the root meristem, a requirement for HD-ZIP III TFs may be to create a system that is relatively stable to perturbations, dampening high auxin signaling peaks to ensure correct, continuous xylem development.