The Columbia (Col-0) accession of Arabidopsis thaliana was used as wild type. The mutants acl5-1, sac51-d (Imai et al., 2006), sac51-1, sacl3-d (sac57-d), sacl3-1 (Cai et al., 2016), sacl1-d (sac504-d) (Mutsuda et al., 2025), have been described previously. lhw (SALK_079402C) and lh13 (SALK_126132) were obtained from the Arabidopsis Biological Resource Center (www.arabidopsis.org). Multiple mutant combinations were generated by crosses and their genotypes were confirmed by PCR with gene-specific primers ( Supplementary Table S1 ).

Plants were grown under 16-h light/8-h dark conditions at 22°C on rockwool bricks supplemented with vermiculite in the growth chamber. For seedling growth observation and RNA extraction, seeds were surface-sterilized with bleach solution containing 0.01% (w/v) Triton X-100 for 3 min, rinsed three times in sterile water, germinated and grown on 0.8% agar plates containing MS salts (Wako, Tokyo, Japan) and 1% sucrose at pH5.7. For examining the sensitivity to thermospermine, seeds were sown on MS agar plates containing 30 μM thermospermine (Santa Cruz, CA, USA). For examining the response of promoter-GUS fusions to hormones, each transgenic line was grown for 7 days on MS plates, transferred to MS solutions containing 1 μM 2,4-D, 1 μM kinetin, or 10 μM bikinin, and incubated for 24 h.

A transgenic line carrying the 990-bp SAC51 promoter fused to the GUS gene was described previously (Ishitsuka et al., 2019). For constructing promoter-GUS fusions of SACL1, SACL2, and SACL3, each gene promoter was amplified by PCR with gene-specific primers ( Supplementary Table S1 ), digested with restriction enzymes, and inserted upstream of the GUS reporter gene in pBI101 (Jefferson et al., 1987). For constructing CaMV 35S promoter-driven SACL2 5’-GUS fusions, the SACL2 5’ region was amplified by PCR ( Supplementary Table S1 ) from wild-type and sacl2-d1 genomic DNA and cloned into pBI121 (Jefferson et al., 1987). The resulting constructs were introduced into wild-type Col-0 plants by the floral dip method (Clough and Bent, 1998).

A CRISPR/Cas9 construct for the conserved uORF of SACL2 was made by using pKIR vector as described (Tsutsui and Higashiyama, 2017) with a pair of oligonucleotides SACL2edit-F and SACL2edit-R ( Supplementary Table S1 ). The construct was introduced into wild-type Col-0. Cas9-induced mutations were identified by PCR amplification of the SACL2 target region using the primer pair, uL2-F and uL2-R ( Supplementary Table S1 ), followed by sequencing. The edited lines in the acl5 background were generated by crosses.

Fluorometric quantitative GUS assays and histochemical GUS staining were performed according to a standard protocol (Jefferson et al., 1987). Samples embedded in Technovit (Heraeus Kulzer, Wehrheim, Germany) were sectioned using a rotary microtome RM2245 (Leica Microsystems, Wetzlar, Germany).

Y2H experiments were performed using the Matchmaker Gold Yeast Two-Hybrid System (Clontech, CA, USA). Full-length coding sequences of each gene were amplified by PCR with gene-specific primers ( Supplementary Table S1 ) and cloned into pGADT7 and pGBKT7 vectors, respectively. Yeast transformants were tested for interactions in synthetic dextrose (SD) media lacking Leu, Trp, His, and adenine in the presence of 3 mM 3-amino-1,2,4-triazole (3AT).

Total RNA was extracted from whole seedlings by the SDS-phenol method followed by LiCl precipitation (Imai et al., 2006). Reverse-transcription was carried out using the PrimeScript RT reagent Kit (Takara, Kyoto, Japan) with the oligo(dT) primers. qPCR reactions were performed using KAPA SYBR FAST qPCR Kit (KAPA Biosystems, MA, USA) and the Thermal Cycler Dice TP760 (Takara) with gene-specific primers ( Supplementary Table S2 ). ACTIN8 (At1g49240) was used as an internal control.

To examine whether disrupting the conserved uORF of SACL2 suppresses the dwarf phenotype of acl5, we generated the mutants using the CRISPR-Cas9 method and isolated three alleles: two with 1-bp insertions and one with a 4-bp deletion ( Figures 1A, B ), all of which result in a premature stop codon within the uORF ( Figure 1C ). In the progeny of the cross with acl5, these mutant alleles were shown to partially suppress the dwarf phenotype of acl5 in a dominant manner, and were named d1, d2, and d3, respectively ( Figures 1D, E ). In the following experiments, we used the sacl2-d1 allele. To confirm that this mutation enhances its downstream translation, we placed the wild-type and d1 mutant versions of the SACL2 5’ leader region between the CaMV 35S promoter and the GUS reporter gene. Several transgenic lines carrying these constructs were obtained. All lines carrying the d1 construct showed significantly higher GUS activity than those carrying the wild-type construct. Results from a representative line homozygous for the GUS gene are shown in Figure 1F . The effect of sacl2-d1 on gene expression was examined by quantitative RT-PCR. In 10-day-old seedlings, the transcript level of genes up-regulated in acl5, including LHW and ATHB8, which regulate ACL5 (Baima et al., 2014; Katayama et al., 2015), as well as a point-mutated form of acl5, were reversed in acl5 sacl2-d1 ( Figure 1G ). The increased accumulation of the mutant acl5 transcript in acl5 is due not only to the expansion of the expression domain in the xylem, but also to the release from the negative feedback regulation mediated by the SAC51 family. In sacl2-d1, excess SACL2 may suppress LHW function, thereby leading to a partial reduction in ACL5 expression. This result is similar to other uORF mutants in the SAC51 family and is consistent with their morphological phenotypes. In contrast, the level of the one-base-inserted sacl2-d1 transcript was several times higher than the wild type, suggesting that the sacl2-d1 transcript is stabilized by ribosomes translating the main coding sequence, even in the absence of thermospermine. At least SAC51 and SACL3 have been listed as a target of nonsense-mediated mRNA decay due to the presence of conserved uORFs (Rayson et al., 2012).

Expression patterns of all SAC51 family members were examined by using transgenic lines carrying each promoter-GUS fusion construct. In the root section, although the intensity varied, the entire central cylindrical vasculature, except for the protoxylem, was stained by all GUS constructs, with the highest staining observed in the procambium ( Figure 2A ). In cotyledons, staining was restricted to the veins in each construct ( Figure 2B ). In the inflorescence stem section, the GUS staining was consistently detected in the cambium and around the xylem vessels ( Figure 2C ). Together with the fact that uORF mutants of all SAC51 family members can suppress the acl5 phenotype, these results suggest that the regulatory systems governing the expression of all SAC51 family genes are largely conserved. In addition to SACL3, SACL2 has been shown to be regulated by LHW-TMO5 and LHW-T5L1 dimers (Katayama et al., 2015). The involvement of other members of the LHW and TMO5 families in the regulation of SAC51 family expression still needs to be explored. We also examined the hormone response of these promoter-GUS fusions using 2,4-D, kinetin, and bikinin, an activator of brassinosteroid (BR) signaling, as these are particularly important for vascular differentiation. Although no significant increase was detected in the GUS activity after 24-h treatment of each transgenic seedling with 2,4-D and kinetin, the SACL2 promoter was shown to be responsive to bikinin ( Figure 2D ), suggesting a possibility that BR-dependent induction was incorporated into the regulation of SACL2 expression during molecular evolution.

We further examined the physical interaction between SAC51 family proteins and LHW family proteins with Y2H. The result showed that all members of the SAC51 family can interact with all members of the LHW family ( Figure 3 ). Interaction between SACL3 and LHW in Y2H has been shown previously (Ohashi-Ito and Bergmann, 2007; Vera-Sirera et al., 2015). In addition, interactions of LHW with SAC51, SACL1, and SACL2 were observed in bimolecular fluorescence complementation experiments (Katayama et al., 2015). These results support the possibility that suppression of acl5 by sacl2-d mutants is also due to the functional repression of LHW and/or other members of the LHW family.

Because the double knockout of SAC51 and SACL3 is highly insensitive to thermospermine (Cai et al., 2016), we examined whether uORF mutants of the SAC51 family are hypersensitive to thermospermine. In the presence of 30 μM thermospermine, leaf expansion of all mutant seedlings, sac51-d, sacl1-d (sac504-d), sacl2-d1 and sacl3-d (sac57-d), was severely inhibited compared to the wild type ( Figures 4A, B ). Given that SAC51 family proteins bind to LHW family proteins and interfere with their function in xylem cell proliferation, this hypersensitivity to thermospermine could be attributed to the functional inhibition of the LHW family. The tiny seedling phenotype is also reminiscent of higher-order loss-of-function mutants in LHW and TMO families (De Rybel et al., 2013; Ohashi-Ito et al., 2013; Vera-Sirera et al., 2015). We then examined the thermospermine sensitivity of loss-of-function mutants of LHW and LHW LIKE3 (LHL3). lhl3 shows no obvious phenotype but enhances the lhw phenotype in the root vascular development and causes a severe defect in the seedling growth in lhw lhl3 (Ohashi-Ito et al., 2013). We confirmed that they are hypersensitive to thermospermine ( Figures 4A, B ). Furthermore, the insensitive phenotype of sac51–1 sacl3–1 was unaffected in the triple mutant of sac51–1 sacl3–1 lhw. The result can be interpreted that, while the SAC51 family acts as a brake on vascular xylem proliferation in contrast to the LHW family, which functions as an accelerator, xylem proliferation is not suppressed in sac51–1 sacl3–1 even in the absence of LHW. This may be because SACL1 and SACL2 are not sufficient to repress the function of other LHW family members as accelerators, despite the loss of the primary one, LHW. In contrast, overproduction of one SAC51 family member by its uORF mutation may be sufficient to enhance thermospermine sensitivity. A simplified scheme shown in Figure 4C indicates that the thermospermine sensitivity, and thus the degree of xylem proliferation, is regulated by interactions between all members of the SAC51 and LHW family proteins.