Phosphorus (P), an essential element for plants, is one of the most limiting nutrients for plant growth. A few transcription factor (TF) genes involved in P-starvation signalling have been characterized for Arabidopsis thaliana and rice. Crop production of common bean (Phaseolus vulgaris L.), the most important legume for human consumption, is often limited by low P in the soil. Despite its agronomic importance, nothing is known about transcriptional regulation in P-deficient bean plants. We functionally characterized the P-deficiency-induced MYB TF TC3604 (Dana Farber Cancer Institute, Common Bean Gene Index v.2.0), ortholog to AtPHR1 (PvPHR1). For its study, we applied RNAi technology in bean composite plants. PvPHR1 is a positive regulator of genes implicated in P transport, remobilization and homeostasis. Although there are no reports on the regulatory roles of microRNAs (miRNA) in bean, we demonstrated that PvmiR399 is an essential component of the PvPHR1 signalling pathway. The analysis of DICER-like1 (PvDCL1) silenced bean composite plants suppressed for accumulation of PvmiR399 and other miRNAs suggested that miR399 is a negative regulator of the ubiquitin E2 conjugase: PvPHO2 expression. Our results set the basis for understanding the signalling for P-starvation responses in common bean and may contribute to crop improvement.
Phosphorus (P) is an essential element for plant growth and development. One of the most important constraints for plant growth is low P availability because of its association with cations (Fe and Al) or organic compounds that create insoluble inaccessible complexes. Plants have evolved several morphological, physiological and biochemical adaptations to cope with P-deficiency (−P) that include mycorrhizal symbiosis, decreased growth rate, modification of root architecture for increased surface area, remobilization of internal inorganic phosphate, modification of carbon metabolism bypassing P requiring steps, synthesis and secretion of acid phosphatases (AP), exudation of organic acids, and enhanced expression of phosphate transporters (for reviews, see Raghothama 1999; Smith 2001; López-Bucio, Cruz-Ramírez & Herrera-Estrella 2003; Vance, Udhe-Stone & Allan 2003; Shulaev et al. 2008).
Common bean (Phaseolus vulgaris L.) is the world's most important grain legume for direct human consumption. Environmental factors such as low soil nitrogen and P levels, and acid soil conditions are important constraints for bean production in most areas of Latin America and Africa where the crop is grown (Brougthon et al. 2003; Graham et al. 2003). Despite the agronomic importance of bean, there is little information on its regulation of P-starvation responses. In an attempt to understand these responses, we have performed a macroarray analysis of −P bean roots from a Mesoamerican bean genotype and identified 124 differentially expressed genes (Hernández et al. 2007). Tian et al. (2007) identified 240 differentially expressed genes from P-deficient roots of an Andean bean genotype by screening a suppression subtractive hybridization library. Furthermore, using real-time RT-PCR profiling, we identified three MYB TF induced in P-deficient bean roots (Hernández et al. 2007). We hypothesized that these induced TF may be involved in the transcriptional regulation of P-deficiency-induced genes.
The aim of this work was to investigate if the common bean MYB TF [TC3604, tentative consensus assigned by the Computational Biology and Functional Genomics Laboratory, Dana Farber Cancer Institute (DFCI), Common Bean Gene Index, v.2.0], orthologous to AtPHR1, hereby denominated as PvPHR1, is a regulator of the P-deficiency response. We used a novel bean transformation method (Estrada-Navarrete et al. 2007) in conjunction with the RNAi gene-silencing technology, utilized for the first time for this species. In order to investigate if miR399 participates in the PvPHR1 signalling pathway, we analysed the expression of P-deficiency response genes in PvDCL1-RNAi silenced bean composite plants.
MATERIALS AND METHODS
Plant material and growth conditions
The common bean (P. vulgaris) Mesoamerican ‘Negro Jamapa 81’ cultivar was used in this study. Seeds were surface sterilized and germinated in sterile conditions for 2 d and then planted in pots with vermiculite. Plants were grown in a greenhouse with a controlled environment (26–28 °C, 16 h photoperiod) and were watered with nutrient solution (Summerfield, Huxley & Minchin 1977) with 1 mm (P sufficiency, +P) or 5 µm PO4 (P-deficiency, −P). Transcript expression analyses were performed in roots and leaves of plants grown for 21 d. Bean composite plants with transgenic roots were generated as reported (Estrada-Navarrete et al. 2007, see further discussion). Hairy roots (3–6 cm long) emerging from the Agrobacterium rhizogenes infection site were observed during the second week post-infection. After confirming the presence of reporter gene (see further discussion), their normal (untransformed) root system was cut, and composite plants were replanted in pots with vermiculite. These were grown for 21 d under controlled environmental conditions watered with +P or −P nutrient solution, as described previously. After this period, transgenic roots and non-transformed aerial tissues were collected in liquid nitrogen and were stored (−80 °C) until used for phenotypic analyses. Free P contents were determined from different organs of 21-day-old composite plant as reported (Hernández et al. 2007). Data from Table 1 represent the average ± standard error of nine replicates from three independent experiments using composite bean plants with high silencing level (80–95%).
Table 1. Free phosphate concentrations in different organs of PvPHR1-RNAi bean composite plants with high silencing level (80–95%) and control composite plants grown under P deficiency
Control (empty vector) (µmol g−1 FW)
PvPHR1-RNAi silenced plants (µmol g−1 FW)
Data represent the average ± SE of nine replicates from three independent experiments.
2.48 ± 0.131
0.96 ± 0.125
3.74 ± 0.516
0.34 ± 0.03
1.01 ± 0.029
0.78 ± 0.105
Cloning of full-length PvPHR1 cDNA and of mature PvmiR399
PvPHR1 full-length cDNA sequence was cloned based on ESTs partial sequences obtained from nodule and roots cDNA libraries (Ramírez et al. 2005), assigned as TC3604 (DFCI/Common Bean Gene Index, v. 2.0). Two primers were designed for PvPHR1 PCR gene amplification by 5′ rapid amplification of cDNA ends (RACE) (GSP5: CTGCCAC CACAATCTATGTGCTGACCATGA) and 3′ RACE (GSP3: CCGGATTCCTCATCTGATGAAGGGAAAA AGG). The SMART-RACE cDNA Amplification Kit (Clontech Laboratories, Inc., Mountain View, CA, USA) was used. PCR-amplified cDNA fragments were cloned into pCR 2.1-TOPO vector (Invitrogen, Carlsbad, CA, USA) and were sequenced. For overlaps of new sequences SeqManII program in the DNASTAR software package was used (DNASTAR, Inc., Madison, WI, USA). The amino acid sequence alignment between AtPHR1 and PvPHR1 was performed using MultAlign software developed by Corpet (1988). The identification of putative sumoylation sites was carried out using the online tools SUMOplot provided by Abgent primary antibody Company (Flanders Court, San Diego, CA, USA) and for predictions of coiled-coil regions, COILS at http://www.ch.embnet.org/software/COILS/form.html was used. The mature miR399 was cloned from a small RNA library of bean seedlings prepared as reported (Elbashir, Lendeckel & Tuschl 2001). This library was used as template for PCR amplification of the AtmiR399 using consensus sequence-specific primers (5′ADmiR399: ACG GAA TTC CTC ACT TGC CAA AGG AGA; 3′ADmiR399: CTG GAA TTC GCG GTT CAG GGC AA) and adaptor sequence primers. The PCR product was cloned in pCR 2.1-TOPO vector (Invitrogen) and was sequenced.
Plasmid construction and plant transformation
In order to obtain the RNAi constructs, we first constructed the pTDT-DC-RNAi vector. This was derived from pBA-DC-RNAi (Jang et al. 2007) by replacing the BASTA gene (cut with StuI and BglII) with the tdTomato gene (cut with EcoR1, blunted, and BamHI) from the pRSET-BtdTomato vector (Shaner et al. 2004).
The design of primers for gene amplification was based on the reported EST sequences from bean cDNA libraries (Ramírez et al. 2005). To generate RNAi constructs, fragments unique to PvPHR1 (290 bp) or PvDCL1 (228 bp) coding sequences were amplified using gene-specific forward primers (PvPHR1: 5′ACCTGAAAAAGATAAT TGAAGAA; PvDCL1: GGATGATGAAAACGGAAAA AGAA) and caccT7 reverse primer (CACCTAATAC GACTCACTARAGGG). The amplified fragments were cloned in the pENTR/SD/D-TOPO vector and were sequenced (Invitrogen). The resulting pENTR-PHR1 or pENTR-DCL1 plasmids were recombined into the pTDT-DC-RNAi binary vector. The correct orientation was confirmed by PCR using the WRKY-5-Rev (GCAGA GGAGGAGAAGCTTCTAG) or WRKY-3-Fwd primer (CTTCTCCAACCACAGGAATTCATC) and caccT7 primer (Supporting Information Fig. S1). The resulting pTDT-PHR1-RNAi and pTDT-DCL1-RNAi plasmids were introduced by electroporation into A. rhizogenes K599 and were then used for plant transformation. Bean composite plants were generated as reported (Estrada-Navarrete et al. 2007). Putative transgenic hairy roots were confirmed by checking for the presence of red flourescence, resulting from the expression of the tdTomato reporter gene by confocal microscopy. The original root system and all the transgenic roots, except one, were removed to avoid root chimeras, and the composite plants were grown in the greenhouse for 21 d in +P or −P conditions, as described.
RNA extraction and analysis
Total RNA was isolated from 1 g frozen roots or leaves of bean plants and transgenic roots of bean composite plants, grown under similar +P or −P conditions in independent experiments, as reported (Ramírez et al. 2005). Isolated RNA preparations were used for semiquantitative RT-PCR (sRT-PCR) and low-molecular-weight RNA-blot analyses. Quantification of transcript levels from the bean genes was performed by two-step RT-PCR using polythymine deoxynucleotide primer following the manufacturer's instructions (Clontech). Bean gene annotation, primer oligonucleotide sequences and sRT-PCR conditions used are shown in Supporting Information Table S1. 32P-radiolabelled probes for miRNAs were synthesized using T4 polynucleotide kinase (New England Biolabs, Beverly, MA, USA) and the primers shown in Supporting Information Table S1. miRNAs were detected in roots of bean plants and in transgenic roots of bean composite plants grown in +P or −P by low-molecular-weight RNA-gel hybridization. Gels, hybridization and washing conditions were performed as reported (Reyes & Chua 2007). The intensity of bands from sRT-PCR amplification or RNA-blot hybridization was quantified by densitometry using the ImageQuant 5.2 software (Molecular Dynamics, Sunnyvale, CA, USA), and the −P/+P expression ratios were obtained.
Identification of putative PHR1 signalling pathway bean genes
We have identified 10 P-starvation-induced common bean genes encoding for putative TF from different gene families (Graham et al. 2006; Hernández et al. 2007). In this work, we further characterized the TC3604 gene (DFCI/Common Bean Gene Index, v.2.0, http://compbio.dfci.harvard.edu/tgi/) classified in the MYB superfamily and induced twofold in −P roots (Hernández et al. 2007). TC3604 is proposed as the bean ortholog of the AtPHR1 (AJ310799) gene, TF implicated in signalling of the response to P starvation in Arabidopsis and rice (Rubio et al. 2001; Zhou et al. 2008), hence designated PvPHR1. The full-length PvPHR1 cDNA clone (1.38 kb) (accession number EU500763) contained a 984 bp open-reading frame (ORF) with 63% identity to AtPHR1. It encodes for a deduced 328 amino acid protein (Mr = 35.63 kDa). The characteristic MYB domain for DNA binding, coiled-coil domain for protein–protein interaction, and three putative sumoylation sites could be identified from the deduced amino acid sequence (Fig. 1a). PvPHR1 was induced in leaves (twofold) and roots (1.6-fold) of P-deficient bean plants (Fig. 1b).
In Arabidopsis, miR399 is part of the AtPHR1 signalling pathway and plays an important role in P homeostasis (Bari et al. 2006; Chiou et al. 2006). In an attempt to identify miRNAs possibly involved in bean stress responses, we prepared a miRNA library from bean seedlings and identified sequences that were homologous to Arabidopsis miRNAs (Arenas-Huertero et al., unpublished data). From this library, we cloned a miRNA identical in sequence to AtmiR399b and c, hereby designated as PvmiR399 (Supporting Information Fig. S2a). Moreover, an in silico analysis of the National Center for Biotechnology Information (NCBI) dbEST led us to identify clone EG594372 as the reverse complement sequence containing the mature PvmiR399. RNA folding of this sequence using the mfold program developed by Zuker (2003) is consistent with it being the PvmiR399 precursor (data not shown). PvmiR399 accumulation was highly induced (80-fold) in roots of P-deficient bean plants when compared with control samples under P sufficiency (+P) (Fig. 2a).
In order to investigate the regulatory role of PvmiR399, we searched for bean orthologs of Dicer-like (DCL) genes implicated in miRNAs biogenesis. Four DCL genes involved in the generation of small RNAs have been identified in plants. Specifically, the DCL1 protein is the major player involved in miRNA biogenesis; however, in its absence, it can be partially substituted by DCL2 (Deleris et al. 2006; Jones-Rhoades, Bartel & Bartel 2006). Searching the Common Bean Gene Index (DFCI), we identified an EST, CV544176, which has 80% identity to AtDCL1, hereby designated as PvDCL1 (Supporting Information Fig. S2b). Importantly, this gene did not respond to −P treatment (data not shown).
In addition, we identified orthologs of two genes also implicated in P homeostasis: At4 (TC7206, Pv4) and PHO2 (TC8137, PvPHO2) (Aung et al. 2006; Shin et al. 2006) (Supporting Information Fig. S2c,d). At4 and its ortholog from the legume Medicago truncatula (Mt4) belong to the IPS1 (induced by phosphate starvation) gene family and lack a long ORF, containing instead a series of short overlapping ORFs. These genes play an important role in the translocation of P from roots to shoots (Shin et al. 2006). We found a clear induction of Pv4 in leaves (18-fold) and roots (29-fold) of P-deficient bean plants when compared with +P plants (Fig. 2b). The ubiquitin E2 conjugase AtPHO2 is a negative regulator of the AtPHR1 signalling pathway (Fujii et al. 2005; Chiou et al. 2006). In agreement, PvPHO2 was repressed in leaves (twofold) and roots (1.8-fold) of P-deficient bean plants when compared with +P plants (Fig. 2c).
PvPHR1 participates in P-deficiency signalling in beans
In order to demonstrate the possible role of PvPHR1 in the regulation of gene expression in P-starved bean roots, we undertook the RNAi gene-silencing approach in composite plants. The protocol for the generation of composite bean plants was recently reported as an alternative for stable transformation, especially for species recalcitrant to transformation such as common bean (Estrada-Navarrete et al. 2007).
The pTDT-DC-RNAi, with the Gateway cassette driven by the CaMV 35S promoter and the tdTomato (red fluorescent protein) reporter gene (Shaner et al. 2004) for identification of the transgenic roots by confocal microscopy, was constructed. This vector was used for the PvPHR1-RNAi gene construct consisting of inverted repeats of a PvPHR1 gene fragment (Supporting Information Fig. S1a). The expression of PvPHR1-RNAi was verified in putative transgenic roots from composite bean plants (Supporting Information Fig. S1b).
In order to confirm PvPHR1gene silencing, sRT-PCR analysis was performed in composite bean plants expressing the PvPHR1-RNAi construct, as compared with control composite plants transformed with an empty vector. Each transgenic root analysed showed a specific level of PvPHR1 transcript, indicating different degrees of gene silencing, resulting from a different transformation event in each root. The comparison of PvPHR1 transcript levels of RNAi versus the control plants grown in −P revealed a reduction ranging from 15 to 95% among the analysed composite plants (Fig. 3a). The level of PvPHR1 gene silencing showed an inverse correlation with the transcript level of two P-responsive target genes PvPHT1 (phosphate transporter) and PvAP5 (Fig. 3a).
To identify the PvPHR1 target genes, we evaluated the transcript level of selected P-deficiency-induced genes (Hernández et al. 2007) (Supporting Information Table S1) in the PvPHR1-RNAi plant with the highest degree of gene silencing (95%, Fig. 3a) as compared with control (empty vector) composite plants under +P and −P (Fig. 4a). While the −P control plant showed an evident induction of P-responsive genes, the silenced plant showed a reduction of target gene transcript levels (Fig. 4a). Transcript levels of three genes involved in P remobilization, namely PvAP, PvAPC5 and PvRNS (RNase), showed a reduced response to P deficiency in the PvPHR1 silenced plant as compared with the control plant. In addition, the −P induction of PvPHT1 and aquaporin (PvAQ), both involved in transport, decreased significantly. The expression of the PvPHO1 gene, a P translocator involved in loading root P into the xylem vessel, was not affected in PvPHR1-RNAi plants. A highly increased expression of PvPHO2, involved in P homeostasis, was observed in the PvPHR1 silenced plant under P deficiency. In addition, the expression of PvmiR399 in −P was significantly reduced despite the unaffected expression of PvDCL1. The Pv4 transcript level was not detected in +P conditions, while in −P, the silenced plant showed a complete inhibition as opposed to the control plant. We determined the transcript levels of 1-aminocyclopropane-1-carboxylate oxidase (PvACCO) for ethylene biosynthesis and an indole acetic acid-induced protein gene (PvIAAIP). Because both phytohormone-related genes are induced in P-deficient bean roots (Hernández et al. 2007), they could be implicated in the root architecture modification response under P starvation (Bonser, Lynch & Snapp 1996; Lynch & Brown 2001; López-Bucio et al. 2003). As shown in Fig. 4a, the −P response of these genes was not affected in the PvPHR1-RNAi composite bean plants.
The P-deficiency phenotype of bean composite plants was also assessed in relation to the concentration of free P in different plant organs. PvPHR1-RNAi composite bean plants showed a decrease of free P content in their leaves (2.5-fold), stems (11-fold) and roots (1.3-fold) as compared with control composite plants (Table 1).
PvmiR399 and PvPHO2 are components of the PvPHR1 signalling pathway in beans
With the rational that miRNA levels would be suppressed in plants with reduced levels of DCL1, we analysed PvDCL1 silenced composite plants as a tool to investigate the participation of PvmiR399 in the PvPHR1 signalling pathway through the regulation of PvPHO2. Composite plants bearing the PvDCL1-RNAi gene construct driven by the CaMV 35S promoter (Supporting Information Fig. S1c,d) showed a 20–70% reduction of PvDCL1 transcript levels as compared with the control plants (Fig. 3b). PvDCL1-RNAi composite plants showed an inverse correlation between PvDCL1 silencing levels and the abundance of three bean miRNAs, homologous to Arabidopsis miR399, miR159 and miR160, thus indicating that reduced amounts of PvDCL1 transcript affected miRNA biogenesis and accumulation (Fig. 3b). An inverse correlation between PvDCL1 transcript and the P-responsive gene PvPHT1 was also observed in PvDCL1-RNAi plants grown in −P, suggesting the participation of miR399 in signalling for P-deficient response (Fig. 3b).
The PvDCL1-RNAi composite plant with the highest degree of gene silencing (70%, Fig. 3b) was further analysed to evaluate the transcript levels of selected P-responsive genes (Hernández et al. 2007; Supporting Information Table S1) as compared with control (empty vector) composite plants under +P and −P (Fig. 4b). The PvDCL1-RNAi plant showed a significant reduction in the −P response of PvAP, PvAPC5 and PvPHT1 genes. In contrast, a high increase in PvPHO2 transcript levels was observed in the PvDCL1 silenced plant grown in −P. This behaviour was independent of PvPHR1, because the PvPHR1−P induction was not affected by PvDCL1 silencing. Also, the −P gene expression response of PvRNS and Pv4, two PvPHR1 target genes, was not affected in the PvDCL1 silenced plant (Fig. 4b).
Although these results suggest the participation of miR399 in the PvPHR1 signalling pathway, the effect of silencing PvDCL1 may not be solely mediated through this miRNA. PvDCL1-RNAi plants showed reduced accumulation of several miRNAs (Fig. 3b) probably involved in diverse biological processes. The effect of other miRNAs on gene expression of −P-responsive genes in bean roots cannot be excluded.
In contrast to Arabidopsis, common bean and other legumes are not easily amenable to stable genetic transformation, and hence, protocols for high-throughput generation of transgenic legume plants are not available. On the other hand, alternatively, fast, reproducible and efficient protocols for the generation of composite plants, with untransformed shoots using A. rhizogenes-mediated transformation, have been established for several legume species including common bean (Estrada-Navarrete et al. 2007). This approach had not been used in common bean. In this work, we used the composite plants approach in conjunction with RNAi-silencing technology and showed its feasibility to carry on functional genomics in common bean.
A model summarizing the crucial molecular events in the P-deficiency signalling controlled by PvPHR1 in bean roots is presented (Fig. 5). Once P deficiency is sensed – either locally or systemically – by unknown molecule(s), this TF positively regulates the expression of target genes involved in P transport (PvPHT1 and PvAQ), remobilization (PvAP, PvAPC5 and PvRNS) and homeostasis (Pv4 and PvmiR399). Transcript analysis of PvDCL1 silenced plant suggests that PvmiR399 is a negative regulator of PvPHO2 that can, in turn, directly or indirectly, regulate the expression of P-responsive genes like PvAPC5, PvPHT1 and PvAP. However, these effects may also be mediated by other miRNAs that are reduced in PvDCL1 silenced plants. The action of PvmiR399 is apparently regulated by Pv4 probably through the mimicry mechanism that regulates P homeostasis (Franco-Zorrilla et al. 2007). The bean At4 ortholog that we identified (Pv4, TC7206) has a miR399 target site with the same characteristics of At4, showing base pairing and a mismatched loop with the PvmiR399 sequence required for the mimicry mechanism.
Silenced PvPHR1 transgenic roots showed a 23% reduction on free P content under P deficiency as compared with control roots. This phenotype could be related to the drastic reduction of the high-affinity P transporter (PvPTH1) that should impair P uptake in the PHR1-RNAi plant during P starvation. In addition, a reduction of free P content was observed in non-transformed stems (90%) and leaves (38%) of the silenced plant as compared with plant organs of control composite plants, which can be related to the drastic reduction of Pv4 transcript, implicated in P homeostasis. Arabidopsis At4 loss-of-function mutants are impaired in P distribution between roots and shoots during P starvation (Shin et al. 2006).
Under P deficiency, bean and other plants modify their root architecture by decreasing primary root growth and increasing lateral roots and root hair formation, and by modifying their root gravitropism to develop shallower and adventitious roots (Bonser et al. 1996; Lynch & Brown 2001; López-Bucio et al. 2003). However, one of the caveats of the transgenic roots system is its limitation for morphological/physiological analyses of root phenotypes such as the one expected for −P bean plants. We observed that −P induction of the phytohormone-related genes PvACCO and PvIAAIP was not altered in PHR1-RNAi silenced plant. These genes could be involved in modification of root architecture, known to be regulated through phytohormone gradients, principally involving auxins, cytokinins and ethylene (Nacry et al. 2005). In Arabidopsis, AtWRKY75, AtZAT6 and AtBHLH32, but not AtPHR1, have been implicated in the regulation of root architecture modification (Chen et al. 2007; Devaiah et al. 2007a,b).
The regulation of common target genes by different TF, such as AtPHR1, AtWRKY75 and AtZAT6 (Rubio et al. 2001; Devaiah et al. 2007a,b), indicates that there is a complex crosstalk occurring during signalling in P-deficient response in Arabidopsis. We identified TF candidate genes from different families proposed to be involved in the regulation of P-deficient responses (Graham et al. 2006; Hernández et al. 2007). Tesfaye et al. (2007) have reported that members of the WRKY gene family are P responsive in bean, either by induction or by repression. It is possible that crosstalk among different signalling pathways during P starvation functions in bean, but this remains to be demonstrated.
Recent reviews have emphasized the essential role of TF and miRNAs in gene regulation, representing the largest families of gene regulatory molecules in multicellular organisms. Most of the genes in the eukaryotic genomes are controlled by a combination of transacting factors; miRNAs and TF are linked to one another in gene regulatory networks (Hobert 2008). In this work, we demonstrate the role of a TF, together with a miRNA, in the regulation of P-responsive genes in common bean. It has been proposed that miRNAs are involved in rapid, adaptive changes in gene expression to maintain homeostasis and respond to specific environmental signals (Hobert 2008). In this regard, the role of PvmiR399 could be essential for the regulation of P homeostasis in beans, and its transport into phloem sap, demonstrated in Brassica napus, Cucurbita maxima and Oryza sativa (Buhtz et al. 2008; Pant et al. 2008; Zhou et al. 2008), could be necessary for regulating P homeostasis in different plant organs.
The results presented here set the basis for understanding the signalling events that occur in response to P deficiency in bean, and can be used for improving bean germplasm for tolerance to abiotic stress, either by breeding or by biotechnological approaches.
This work was supported in part by PAPIIT grant IN211607 from Dirección General de Asuntos del Personal Académico/UNAM, and grant USDA-FAS MX161. O.V.L. is a PhD student from Doctorado en Ciencias Biomédicas, Universidad Nacional Autónoma de México and a recipient of a studentship from Consejo Nacional de Ciencia y Tecnología, México (200048). We are grateful to Xochitl Alvarado-Affantrager (IBt-UNAM) for assistance in confocal microscopy; Victor M. Bustos (CCG-UNAM) for plant maintenance; Roger Y. Tsien (UCSD) for providing the red fluorescent protein – tdTomato – reporter gene; and Michael Dunn (CCG-UNAM) for critically reviewing the manuscript.