A cladistic analysis was performed in Arabidopsis to find the homologous genes of CmNACP1; for this purpose, we used the conserved N-terminal domain of the NAC protein family as a bait. First, a BLAST alignment was carried out with the CmNACP1 NAC domain reported in UniProt Knowledgebase¹ to select similar sequences in Arabidopsis. All sequences obtain displaying homology were aligned by MUSCLE multiple sequence alignment software. The phylogenetic tree was constructed with MEGA7² using the neighbor-joining (NJ) method and bootstrap test carried out with 1,000 iterations.

Seeds of Arabidopsis thaliana ecotype columbia-0 were sown on sterilized soil (agrolite-peat-soil, 1:2:2) and grown in a controlled climate chamber (16-h light/8-h dark) at 22 ± 2°C. For transgenic plants, seeds were germinated in soil and then in the seedling stage were sprayed every 6 days with the herbicide containing ammonium glufosinate at the recommended dilution of 1:6000 (Invictus, Allister; Zapopan, Mexico) for one month. Fifteen plants per independent Arabidopsis line were grown in pots containing sterilized soil in the controlled climate chamber AR715 (Percival; Perry, Iowa). The phenotype of transformed plants was assessed by measuring branch number and parameters related to leaf senescence.

To analyze and study ANAC087, we obtained two genetic constructs with different regions of the ANAC87 gene (GenBank accession number NM_121832.4). The ANAC087 open reading frame (ORF, 1005 bp) was cloned into a binary vector (pB7FWG2,0; Plant Genetic Systems, Ghent, Belgium) under the control of the CaMV 35S promoter to analyze the effect of its overexpression on plant development. The analysis of ANAC087 promoter region was carried out using 3.0 kb upstream region of the ORF, which includes the 5′ untranslated region. For this purpose, the promoter was subcloned into the pBGWFS7,0 Gateway vector (Supplementary Figure 1). To obtain recombinant plasmids, the following procedures were carried out. First, total RNA and genomic DNA were extracted from rosette leaves of A. thaliana wild type using Direct-zol™ RNA Miniprep Kit (Zymo Research, Irvine, United States) and CTAB method (Weigel and Glazebrook, 2002), respectively. The primers designed to amplify the promoter region and the ORF are reported in Supplementary Table 1. To obtain the ORF of ANAC087, cDNA was synthesized employing SuperScript III First-Strand Synthesis System (Invitrogen, La Jolla, CA, United States) using 50 ng of total RNA, oligo (dT) and SMAGGG primers at 10 μM each, a mixture of dNTPs 10 mM (New England Biolabs, Beverly, MA, United States), and sterile distilled water to a final volume of 13 μL. The reaction mixture without enzyme was incubated at 72°C for 10 min and kept on ice for 5 min. Then, 4 μL of 5′ buffer solution, 1 μL of DTT (0.1 M), 1 μL of RNase inhibitor, and 1 μL of SuperScript III RT enzyme (200 U/μL) were added to the reaction. The mixture was gently homogenized, incubated at 50°C for 1 h, and inactivated by incubation at 75°C for 15 min. Then, 1 μL of synthesized cDNA was used as a template to amplify ANAC087 ORF with TaKaRa Ex Taq DNA Polymerase (Takara Bio Inc.) and specific primers, following the manufacturer’s instructions in a Biometra T1000 Thermocycler (Biometra AG, Germany). The same DNA polymerase was used to amplify 3.0 kb upstream sequence from the start codon of the ANAC087 gene (promoter region) from genomic DNA. The amplicons obtained were cloned into the pCR8/GW/TOPO vector (Invitrogen, La Jolla, CA, United States) and sequenced to confirm the correct orientation. Then, the products were recombined into the corresponding binary vectors by Gateway LR reactions to generate the following genetic constructs: (1) PromANAC087:GFP-GUS construct or PROM-ANAC087 and (2) 35S:ANAC087 ORF-GFP construct or OE-ANAC087. Recombinant plasmids were analyzed by PCR and then introduced into Agrobacterium tumefaciens strain AGL1 by electroporation. Positive clones were cryopreserved and used for plant transformation.

Transformation of Arabidopsis wild-type Co-1 ecotype was performed using the floral dip method (Clough and Bent, 1998) with some modifications: A. tumefaciens strain AGL1 harboring the recombinant plasmids was grown on plates containing three antibiotics (spectinomycin 50 mg/L, kanamycin 25 mg/L, and carbenicillin 50 mg/L). Then, immature flower buds were immersed in the bacterial suspensions; after the recovery of the plants, seeds were collected. Finally, transgenic plants were selected by herbicide resistance in soil with ammonium glufosinate, as previously described. T2- and T3-independent plant lines were used for further analysis.

Samples of different tissues (rosette leaves, flowers, apexes, roots, stem, and petioles) were dissected from transgenic plants harboring PROM ANAC087:GFP-GUS construct. Histochemical detection of GUS activity was carried out according to Weigel and Glazebrook (2002). All tissues were covered with GUS solution and incubated at 37°C for 2 days and finally maintained in glycerol 50%. The tissues were visualized on a Nikon SMZ 745T stereoscope, and images were captured with a Nikon Digital Sight DS-U3 camera using NIS Elements D software.

Arabidopsis wild-type and transgenic lines harboring PROM-ANAC087 and OE-ANAC087 constructs were grown in soil under controlled conditions. Fifteen plants of each construct were analyzed, and images were captured at 20, 28, 36, and 44 days post-germination (dpg). Samples of rosette leaves were employed to quantify chlorophyll and anthocyanin content as described below.

Rosette leaves with 50% or more of yellowed, chlorotic leaf area were counted as senesced. The senescence rate was calculated by the ratio of rosette senescent leaves to the total number of leaves (Kim et al., 2018).

Rosette leaves at positions five and six were used for chlorophyll and anthocyanin quantification. Chlorophyll was extracted from rosette leaves at 44 dpg with phosphate buffer and 80% acetone, and absorbances were measured at 645 and 663 nm using spectrophotometer Smart Spec plus (Biorad, Hercules, CA, United States) following the protocol described by Arnon (1949). Anthocyanins were extracted with 45% methanol and 5% acetic acid, and absorbance measurement was carried out by triplicate at 530 and 657 nm (Nakata and Ohme-Takagi, 2014). Anthocyanins, total chlorophyll, chlorophyll a, and chlorophyll b concentrations were normalized to the fresh weight of each leaf.

Relative expression of the ANAC087 transcript was determined in wild-type plants as well as in the transgenic lines for overexpression and promoter analysis. The quantification of genes associated with leaf senescence and relative ANAC087 expression levels in all transgenic lines was performed using a set of primers within the 3′end of the ANAC087 ORF, because this region is highly variable among the NAC gene family members (Ernst et al., 2004). Total RNA was extracted from different Arabidopsis tissues (root, rosette leaf, caulinar leaf, stem, flower, and siliques) using Direct-Zol™ RNA Miniprep kit (Zymo Research, United States) at 44 dpg. RNA extraction of each tissue was obtained from a pool of four plants. qRT-PCR with KAPA SYBR^® FAST One-Step kit was performed with an Applied Biosystems StepOnePlus™, following the manufacturer’s recommendations and 30 ng of total RNA in a 10 μL reaction. Specific primers for ANAC087, UBQ10, RBCS1A, ORE1, SAG13, and ANAC046 were used (Supplementary Table 1). The Ct value for each product was determined by triplicate for each transgenic line. UBQ10 was used as endogenous genes to normalize gene expression, and the comparative ΔΔCt method was used to determine relative transcript abundance (Livak and Schmittgen, 2001).

Arabidopsis rosette leaves were collected to extract total proteins or nuclear proteins. For total protein extraction, leaves were mixed with extraction buffer (100 mM Tris–HCl pH = 6.8, 2% SDS, 200 mM DTT, 20% glycerol, and 0.5 mM PMSF). Extraction of nuclear proteins was performed employing the Plant Nuclei Isolation/Extraction Kit (CelLytic™ PN, Sigma-Merck, Darmstadt, Germany) following the manufacturer’s recommendations. Equal amounts of protein extracts were separated on SDS-polyacrylamide gels. The separated proteins were transferred onto Amersham Hybond-P PVDF Membrane (GE Healthcare, Chicago, United States) for 1 h at 100 volts. Membrane was blocked with TBST buffer, 0.1% Tween 20, and 5% skim milk for overnight at 4°C. Blots were incubated with antibody against ANAC087 (Genscript; Piscataway, NJ, United States) (1:2500) for 4 h at 4°C, washed, and incubated with a secondary antibody against rabbit IgG (whole molecule; Sigma-Merck) conjugated to goat peroxidase (1:10,000) for 2 h at room temperature. The washed blots were transferred to ECL Prime Western Blotting Detection Reagent (Pierce, Rockford, IL, United States) and then exposed to X-ray film.

To identify the protein homologs for CmNACP1 in Arabidopsis, we performed a BLAST alignment with the CmNACP1 NAC domain reported in UniProt Knowledgebase (see text footnote 1), since the complete virtually translated sequences are of varying lengths, and sequences outside the NAC domain are highly divergent in members of the family, even within the same species. A total of 100 sequences were selected in Arabidopsis. To study the evolutionary relationship between the CmNACP1 and NAC proteins (NACs) from Arabidopsis, a phylogenetic tree was constructed from alignments of these NACs and CmNACP1 sequences (Figure 1). The results indicate that CmNACP1 homologs form a separate clade, which includes ANAC059 (At3g29035), ANAC092 (At5g39610), ANAC079 (At5g07680), ANAC100 (At5g61430), ANAC046 (At3g04060), and ANAC087 (At5g18270). The members of this clade are involved in leaf senescence, programmed cell death, and developmental regulation (Laufs et al., 2004; Raman et al., 2008; Balazadeh et al., 2010; Oda-Yamamizo et al., 2016; Lee et al., 2017; Huysmans et al., 2018), which suggests that the members of this clade are orthologs. The present analysis focused on ANAC087, since the sequence similarity with CmNACP1 considering the NAC domain and the whole protein are slightly higher than the rest of the members of this clade.

To gain insight into the function of ANAC087, its expression pattern was studied. Quantitative real-time RT-PCR using RNA from different tissues of wild-type Arabidopsis was carried out. The results show that the highest ANAC087 transcript levels are found in root and rosette leaf followed by flower, stem, caulinar leaf, and silique, albeit at similar levels (transcript levels were normalized to rosette leaf = 1) (Figure 2A). These results are similar to those observed for CmNACP1 (Ruiz-Medrano et al., 1999) in terms of the tissue showing the highest accumulation levels of both transcripts, lending support to the notion that these genes are orthologs. The expression pattern of the ANAC087 gene was then studied with more detail through histochemical staining of tissues from four independent transgenic lines (L31, 150, L168, and L170) carrying the ANAC087 promoter (3.0 kb upstream of the start codon, including the 5′UTR) fused to the uidA-GFP reporter gene, termed PROM-ANAC087. Strong GUS activity was observed in leaf vascular tissue in first-, second-, and third-order veins (Figure 2B), and also in trichomes of caulinar leaf petioles (Figure 2C). GUS signal was detected in stem sections (Figure 2D). Also, a strong signal was visible in shoot apices, which decreases sharply in underlying tissue (Figure 2E). This pattern is also reminiscent of other NAC genes that are expressed in the SAM and have a role in its maintenance, as in the case of Petunia NAM and Arabidopsis CUC2. Of note, GUS activity is evident in developing flowers, but absent in more mature open flowers, except in the papilla (Figure 2F). Histochemical GUS staining was also observed in mature seeds, which suggests a role in its development (Figure 2G). GUS activity was very strong in the main root and becomes weaker in lateral roots, albeit more pronounced activity was observed in sites of lateral root formation (Figure 2H). The staining pattern at the base of the leaf indicates expression of this promoter mostly in the vasculature, and also in the base of trichomes (Figure 2I). Finally, transverse section of rosette leaf petiole evinces expression in the vasculature, including phloem, protoxylem, cambium, and cortex (Figure 2J). In all, the expression pattern of ANAC087 is reminiscent of NAM and CUC2 in the shoot apex and of CmNACP1 in the vascular tissue. This reinforces the notion that these genes may be, to some extent, functional homologs.

To determine the localization of the ANAC087 protein, a western blot was carried out with total and nuclear protein extracts from Arabidopsis rosette leaves and an antibody against ANAC087. The immunoblotting assay showed that this protein is present in total protein extracts and also in nuclear protein extracts (Supplementary Figure 2).

To infer the function of this gene, the phenotype of transgenic plants in which ANAC087 was expressed constitutively was analyzed. ANAC087 overexpression lines (OE-ANAC087: OE L5, OE L11, OE L24, and OE L170) were obtained and their development was evaluated at different time points. Images of plants at 20, 28, 36, and 44 days post-germination (dpg) were taken. We observed that at 36 dpg anthocyanin accumulation was stronger in OE-ANAC087 lines relative to WT plants. Furthermore, at 44 dpg, we observed early leaf senescence and anthocyanin accumulation in OE-ANAC087 plants (Figure 3). These observations confirm that ANAC087 is a positive regulator of leaf senescence.

Since the early leaf senescence phenotype was detected in most cases at 44 dpg stage, we collected rosette leaves at this stage to test different parameters that are related to the reduction in the photosynthetic capacity. First, we calculated the senescence rate of rosette leaves in all ANAC087 lines. High senescence rate was obtained in OE-ANAC087 and PROM-ANAC087 lines, and there was a significant difference compared to wild-type plants (Figure 4A). Then, the anthocyanin content was measured as the absorbance at 530 nm and normalized to leaf fresh weight, and the OE-ANAC087 and PROM-ANAC087 lines showed high content levels of anthocyanins compared to wild-type plants (Figure 4B). Further, the total chlorophyll, chlorophyll a, and chlorophyll b concentrations were calculated in all transgenic lines. In OE-ANAC087, the concentrations of chlorophyll (total, a and b) decreased significantly compared to wild type. In PROM-ANAC087 lines, no significant difference was observed compared to wild-type plants (Figure 4C). On the other hand, some studies report that photosynthesis is gradually inactivated during senescence and accompanied by degradation of chlorophyll and also anthocyanin accumulation (Bresson et al., 2018). Thus, the results in this work show that ANAC087 expression is related to the process of leaf senescence, since OE-ANAC087 lines showed higher anthocyanin accumulation and senescence rate relative to control plants. Also, chlorophyll concentrations (total, a, b) of OE-ANAC087 lines were reduced. To analyze whether ANAC087 is involved in senescence leaf process, we performed qRT-PCR for genes associated with senescence in Arabidopsis. The expression levels of RBCS1A from the downregulated gene set (SDGs) group during senescence were lower in OE-ANAC087 followed by PROM-ANAC087 lines than in wild-type plants. For genes that belong to senescence-associated genes (SAGs) group, we obtained the highest relative expression of ORE1, SAG13, and ANAC046 in OE-ANAC087 lines. Also, we observed that ANAC087 expression levels were high and comparable to that one obtained in genes associated with senescence. The PROM-ANAC087 lines presented higher levels of ANAC046 and ANAC087 than in wild-type plants (Figure 4D).

Another phenotype observed in different PROM-ANAC087 (Figures 5A–D) lines and OE-ANAC087 plants (Figures 5E–G) was the presence of aerial, or ectopic, rosettes. We observed the development of more than one ectopic rosette in each OE- and PROM-ANAC087 plant. For PROM-ANAC087 lines, this phenotype was more marked in PROM- than OE-ANAC087 lines in T1 and T2 generations. Interestingly, leaves from ectopic rosettes showed early senescence compared to the main rosette leaves in PROM-ANAC087 (Figure 5A). However, in a normal plant stage, senescence progression ends with the complete disintegration of leaf tissues, senescence occurring first in rosettes, which is not observed in PROM-ANAC087 plants. In OE-ANAC087 lines, we identified several ectopic rosettes that presented less foliage than PROM-ANAC087 plants (Figures 5E–G). Furthermore, a significant increase in the branching of secondary and quaternary shoots was also observed in these transformed plants (Figure 5H). Thus, the ANAC087 gene may also have an important role in regulating the Arabidopsis body plan. However, this was observed in T1 and T2 plants; T3 plants harboring either construct showed this phenotype at much lower frequencies, which suggests epigenetic silencing.

Intriguingly, in PROM-ANAC087 lines, a phenotype similar to OE-ANAC087 plants was observed after 44 dpg, in terms of anthocyanin accumulation, although early leaf senescence was not as pronounced as in overexpressing lines (Figure 6A). This suggested that the endogenous ANAC087 gene was also overexpressed in the PROM-ANAC087 lines. This in turn prompted us to assay the mRNA levels of ANAC087 in the OE and PROM lines. Thus, qRT-PCR for this RNA was carried out in different tissues from four transgenic Arabidopsis lines for each transgene (OE and PROM) and compared to wild-type plants. In OE-ANAC087 lines, the highest mRNA levels were found in caulinar and rosette leaves, followed by siliques, roots, stems, and flowers. In the case of PROM-ANAC087 lines, increased accumulation of the endogenous ANAC087 transcript occurred in all tissues (Figure 6B).