Isolation and characterization of drd2-D displaying short and twisted root phenotypes
We isolated a dominant mutant developing short and twisted roots named defective root development 2-D (drd2-D) by introducing the 5x UAS tag into a Q2610 enhancer trap line (Fig. 1B). The drd2-D displayed the compromised root epidermal patterning as shown by the non-hair cell markers (Supplementary Fig. S1). By thermal asymmetric interlaced polymerase chain reaction (TAIL-PCR) analysis, we found the insertion site of the UAS tag at the 132 base pairs (bps) upstream of the start codon of a C2H2-type zinc-finger protein gene, ZINC-FINGER OF ARABIDOPSIS THALIANA (ZAT1, At1g02030) (Fig. 1C) that was only partially characterized previously (22). The transcripts of ZAT1 increased in drd2-D compared with the WT control, as shown by the reverse transcription (RT)-PCR analysis (Fig. 1D), and the transgenic expression of UASp: ZAT1 under the control of Q2610 (Q2610>>ZAT1) induced severe growth inhibition in the root of transgenic seedlings (Fig. 1E). When UASp:ZAT1 was expressed during the embryogenesis by a J1721 enhancer mimicking the expression of MONOPTEROS (MP) (23-25), the transgenic seedlings exhibited rudimentary roots reminiscent of the mp phenotype (Fig. 1F). The organization of RAM of the Q2610>>ZAT1 seedling was compromised, as shown by the diffusively extended expression pattern of WOX5p:GUS, a QC marker (Supplementary Fig. S2A, B). In addition, the cell division was suppressed in the RAM of Q0990>>ZAT1, displaying defective growth and a reduced meristematic zone (Supplementary Fig. S2C, D), as shown by the reduced expression of CYCB1;1p:GUS (Supplementary Fig. S2E, F). These results indicate that ZAT1 can inhibit the growth of Arabidopsis, leading to the complete loss of specific organs where it is highly expressed.
In the Arabidopsis genome, there are two ZAT1-related genes, ZAT4 and ZAT9. ZAT1 possesses 43.4% and 43.8% amino-acid (a.a.) sequence identity with ZAT4 and ZAT9, respectively, and ZAT4 and ZAT9 exhibit 60.4% identity with each other (Supplementary Fig. S3). They share a conserved EAR motif at the C-terminal ends together with three conserved C2H2 zinc-finger motifs. Both ZAT4 and ZAT9 likely shared the function of ZAT1 as Q2610>>ZAT4 and Q2610>>ZAT9 exhibited severe growth inhibition in the primary root (Fig. 1G, H) comparable to Q2610>>ZAT1 (Fig. 1E). These results suggest that the three related ZATs might function redundantly to suppress the growth of Arabidopsis in the specific tissues.
To find out whether ZAT1 can affect the growth of aerial organs as well, we compared the organ size and the ZAT1 transcription level in a weak transgenic line, J0571>>ZAT1, that can complete the whole life cycle by developing reproductive organs. The growth inhibition in the roots (Fig. 1I, J) and siliques (Fig. 1K, L) was well correlated with the transcription level of ZAT1 (Fig. 1M).
The outmost cell-specific expression patterns of ZAT1, ZAT4, and ZAT9
To examine the overall expression patterns of the ZATs in various organs, we did RT-PCR analysis with gene-specific primers. The transcripts of ZATs accumulated in the various tissues at the various developmental stages (Supplementary Fig. S4). To visualize the tissue-specific expression patterns of ZATs, a β-glucuronidase (GUS) reporter was driven by each ZAT promoter in the transgenic plants. The three ZATs commonly expressed in the outermost layer of the root cap (Fig. 2A, H, O). To identify the earliest developmental stage when ZATs begin to express, we examined the GUS activity in the transgenic embryos. ZATs did not express at the torpedo (Fig 2B, I, P) and bent cotyledon (Fig 2C, J, Q) stages when the three discrete root-cap layers are present.
The ZAT1p:GUS and ZAT4p:GUS began to express around 24 h after germination (Fig. 2D, K), whereas ZAT9p:GUS did around 36 h after germination (Fig. 2S). In addition to the root cap, the ZAT1-specific GUS expression was observed in the stele and endodermis in the root (Supplementary Fig. S5A-B) and hydathode (Fig. 2F). The ZAT4-specific expression was found in the guard cells (Fig. 2M) and pollen grains (Fig 2N), where the ZAT9p:GUS was active as well (Fig. 2U). These results indicate that the promoters of ZATs are inactive during the embryogenesis and then are activated post-embryonically after the root tips are exposed to growth media.
The expression pattern of ZAT1p:GUS was not apparently altered in drd2-D, where ZAT1 is highly expressed (Supplementary Fig. S5C) and in the Q2610>>SMB line, in which the root-cap layers increased (Supplementary Fig. S5D). In contrast, the number of root-cap layers expressing ZAT1p:GUS increased in tyrosylprotein sulfotransferase (tpst)-1 developing an array of partially detached root-cap layers attached at the columella cells (26) (Supplementary Fig. S5E, F). These results suggest that ZAT1 expression is not under the regulation of ZAT1 and is activated when the root-cap cells are exposed to growth media. It was reported that the expression of ZAT1 is responsive to nitrate (20, 27); however, this interaction might be indirect, because nitrate affects the development of lateral roots and thereby the ‘ZAT1-expressing’ root caps as well (28).
The outermost layer-specific expression of ZAT1 is not altered in the background-accumulating additional root-cap layers
To more precisely examine the ZAT1-expressing cell layers in the root cap, the GFP5-ER reporter and ZAT1-GFP were driven by the ZAT1 promoter in the transgenic plants. Both ZAT1p: GFP5-ER and ZAT1p:ZAT1-GFP expression was specifically localized in the outermost cell layer of the root caps of wild-type seedlings (Fig. 3A, B, H-K). This layer-specific expression was not altered even in the background of smb-3 and smb-3 brn1-1 brn2-1 triple mutant, where additional root-cap layers accumulated (Fig. 3C-G, 3L, M). Interestingly, the underlying cells began to express the ZAT1 as they were exposed to growth media, even though the ex-outermost cells was not completely detached (Fig. 3F). In addition to the root cap, ZAT1-GFP was localized in the nuclei of the endodermal cells (Fig. 3J, K). These results suggest that ZAT1 promoter might not be gradually activated as the root-cap cells move away from the initial position, but are rather abruptly activated by the exposure to the growth media. Furthermore, ZAT1 expression is likely regulated independently of SMB, BRN1 and BRN2.
Both ZAT4p:GFP5-ER and ZAT9p:GFP5-ER were specifically expressed in the outermost layers of the root cap as well (Fig. 3N-O, 3R-S). Unlike ZAT1-GFP, ZAT4-GFP and ZAT9-GFP were broadly distributed to the inner root-cap cells even close to the CSCs (Fig. 3P, Q, 3T, U) together with the outermost cells. These results suggest that ZAT4 and ZAT9 proteins or ZAT4 and ZAT9 transcripts might be mobile, unlike ZAT1 or ZAT1. The discrepancy between the expression sites and the location of ZAT4-GFP and ZAT9-GFP fusion proteins implies that these proteins could transfer the growth inhibition signals from the outermost cells that can directly sense the environmental cues.
The expression of PASPA3 was induced by the ectopic expression of ZAT1
To examine whether ZAT1 regulates the expression of the root-cap-specific genes such as SMB, CEL5, and PASPA3, each GUS reporter driven by the respective promoter was introduced into the Q2610>>ZAT1 background. The expression of SMBp:GUS and CEL5p:GUS was maintained at the root tip but reduced, probably because of the growth inhibition by ZAT1 (Fig. 4A, B, 4G, H). In contrast, the expression of PASPA3p: GUS in the Q2610>>ZAT1 seedlings was increased in the root-hypocotyl transition zone, where Q2610 is actively expressed (Fig. 4C-F, 4I-L). These results suggest that SMB and CEL5 are not under the regulation of ZAT1, whereas ZAT1 might be involved in the regulation of the PASPA3 expression triggering dPCD.
The growth inhibition by ZAT1 is dependent on the EAR motif
EAR motif is known to mediate the interaction between transcription repressors and TPL/TPR co-repressors to recruit histone deacetylase complexes (15). ZAT1 possesses two conserved EAR motifs, motif 1 (LSLML) present in the 130-134th a.a. and motif 2 (DLNLPA) in the 250-255th a.a. (16). Because motif 2 is well conserved among ZAT1, ZAT4, and ZAT9 (Supplementary Fig. S3), we introduced the single or double mutation into motif 2 by site-directed mutagenesis to investigate whether the EAR motif 2 is required for the growth inhibition (Fig. 4M).
Overexpression of the UASp:ZAT1-L251A or UASp:ZAT1-L253A single mutant driven by Q2610 induced the growth inhibition of transgenic seedlings comparable to that of Q2610>>ZAT1 (Fig. 4N-P). However, the overexpression of UASp:ZAT1-mEAR2 (L251; 253A) failed to inhibit the growth when we examined 30 independent transgenic lines (Fig. 4Q). These results indicate that the conserved EAR motif 2 is required for the growth inhibition by ZAT1.
It was reported that transgenic plants overexpressing ZAT7, a C2H2 zinc-finger protein gene with an EAR motif, displayed salinity resistance together with growth-inhibition phenotypes. The EAR motif of ZAT7 was essential for the salinity resistance but not for the growth inhibition by ZAT7 (29). In contrast, the growth inhibition by ZAT1 in this study is dependent on the EAR motif of ZAT1, suggesting that the function of an EAR-motif differs for each zinc-finger protein.
Considering the growth inhibition activities, the roles of ZAT1 in the outmost cells might be to prevent them from further growth and to induce cell maturation after the exposure to the media. Indeed, the ZAT1 overexpression induced the ectopic expression of PASPA3, a marker for dPCD (30), in the hypocotyl-root transition zone, although it is not clear whether this interaction is direct or not. Because the endogenous expression patterns of ZAT1 and PASPA3 partially overlap in the columella and bottom of LRC, ZAT1 might regulate the PASPA expression in the corresponding tissues (13).
The shared functions of ZATs might not be essential for normal growth, because a zat1 zat4 zat9 triple mutant did not display any discernible phenotypes (Supplementary Fig. S6). However, in a specific environmental condition that we have not yet investigated, ZATs might be required for the proper growth of Arabidopsis.