Auxin perception in Agave is dependent on the species’ Auxin Response Factors

Auxins are one of the most important and studied phytohormones in nature. Auxin signaling and perception take place in the cytosol, where the auxin is sensed. Then, in the nucleus, the auxin response factors (ARF) promote the expression of early-response genes. It is well known that not all plants respond to the same amount and type of auxins and that the response can be very different even among plants of the same species, as we present here. Here we investigate the behavior of ARF in response to various auxins in Agave angustifolia Haw., A. fourcroydes Lem. and A. tequilana Weber var. Azul. By screening the available database of A. tequilana genes, we have identified 32 ARF genes with high sequence identity in the conserved domains, grouped into three main clades. A phylogenetic tree was inferred from alignments of the 32 Agave ARF protein sequences and the evolutionary relationship with other species was analyzed. AteqARF 4, 15, 21, and 29 were selected as a representative diverse sample coming from each of the different subclades that comprise the two main clades of the inferred phylogenetic reconstruction. These ARFs showed differential species-specific expression patterns in the presence of indole-3-acetic acid (IAA) and 2,4-dichlorophenoxyacetic acid (2,4-D). Interestingly, A. angustifolia showed different phenotypes in the presence and absence of auxins. In the absence of auxin, A. angustifolia produces roots, while shoots are developed in the presence of IAA. However, in the presence of 2,4-D, the plant meristem converts into callus. According to our results, it is likely that AteqARF15 participates in this outcome.

Auxins are a very simple group of plant growth regulators with a very complex set of activities 1 . Auxins appear to be a set of molecules with limitless powers; they give rise to a vast range of specific developmental outputs 2 . The canonical auxin signaling allows auxin to respond in different ways depending on concentration and context 3,4 . The auxin binds to the TIR1/AFB family of auxin receptors, where it works as a cement between the Aux/IAA transcriptional repressor family and the TIR1/AFB receptor 5 . There are different possible Aux/IAA TIR1/AFB combinations that give rise to the wide range of effects produced by auxins 2,3 . Auxins are involved in numerous physiological and developmental processes, such as organogenesis, cellular elongation, apical dominance, gravitropism, root formation, flower morphogenesis, embryo development, and others [6][7][8] . The auxins indole-3-acetic acid (IAA), indole-3-butyric acid (IBA), phenyl acetic acid (PAA) and 4-chloro-indole-3-acetic acid (4-Cl-IAA) [9][10][11] are all auxins previously characterized in plant tissues. However, there are also auxin analogues, such as naphthalene-1-acetic acid (NAA) and 2,4-dichlorophenoxyacetic acid (2,4-D) 1 . At different concentrations, auxins can be used to induce dedifferentiation and redifferentiation of tissues, and also to promote rooting in plants [12][13][14][15][16] . The auxinic herbicides, such as dicamba (3,6-dichloro-2-pyridinecarboxylis acid), dichloroprop [(±)-2-(2,4-dichlorophenoxy) propionic acid], 2,4-D [ (2,4-dichlorophenoxy) acetic acid], and several more are described as synthetic auxins or growth regulators with herbicidal action or herbicides with growth regulator activity 11,17,18 . It has been reported that the effect of 2,4-D is 300 times more potent than IAA on the growth of plants 11 . The major characteristic shared by auxins and auxinic herbicides is the fact that the positive charge of the indole nitrogen is separated by 5.5 Å from a negatively charged substituent (the carboxyl group) on a flat lipophilic group/ring system 11 .
The genes that encode the enzymes and proteins needed for synthesis, transport, and signaling of auxins have been identified mainly in maize, rice and Arabidopsis thaliana [19][20][21][22] . Some of the early genes transcribed due to auxin response are Auxin/IAA (Aux/IAA), Gretchen Hagen 3 (GH3) and SMALL AUXIN UP RNA (SAUR), which regulate the physiology of the plant by modulating the interaction of the auxin-responsive elements (AuxRes) in Results identification of ARf in Agave spp. revealed by gene structure and protein analysis. Identification of the ARFs present in Agave was done from an Agave transcriptome database 51 by homology to known ARFs. Using as probes the 23 well-characterized members of the A. thaliana ARF family, we found 32 putative Agave ARF transcripts (Table 1). These sequences were aligned to find the conserved domains ( Fig. 1). Three characteristic ARF domains were found: the DNA binding domain (DBD; B3), the auxin response domain, and the AUX-IAA domain. According to Zhang et al. 53 , sequences that do not have the DBD domain are considered to be pseudogenes and cannot be functional. Therefore, we scanned the proteins encoded by the transcripts to annotate the presence of conserved domains and discarded those that did not contain a DBD domain. Of the 32 putative ARF sequences, only 23 have the complete DBD domain, three have an incomplete DBD sequence and six do not have a DBD domain. ARFs with incomplete or missing domains could be pseudogenes or be due to technical problems of the transcriptome assembly. phylogenetic analysis reveals ARf homology with other plants. A phylogenetic analysis divided 30 of 32 transcripts into three major clades (Fig. 2). AteqARF18 and AteqARF13 were not used in the analysis due to their truncated structure (Fig. 1). The phylogenetic analysis of the ARF of Agave was performed by comparing these to the ARF from seven others already reported from different plant species, including 15 ARF of Cucumis sativus 54 , 19 in Vitis vinifera 55 , 20 in Solanum lycopersicum 56 , 22 in A. thaliana 57 , 20 in Oryza sativa 58 , and 31 in Zea mays 59 (Fig. 3). Interestingly, 13 of the Agave ARF genes were grouped with the ARF of V. vinifera (Vv). Specifically, AteqARF12 and AteqARF21 were mainly grouped with VvARF15; AteqARF4 and AteqARF25 were grouped in the same clade with VvARF13; AteqARF24 and AteqARF28 were grouped with VvARF6; AteqARF22 and AteqARF30 were grouped with VvARF1; AteqARF6 was grouped with VvARF9; AteqARF29 was grouped with VvARF3; and AteqARF5, AteqARF16, and AteqARF17 were grouped with VvARF18 (Fig. 3).
Auxins display phenotypic differences in Agave spp. In order to know how different auxins affect the growth and phenotype of different Agave species (A. angustifolia, A. fourcroydes and A. tequilana), all plants were exposed to the same conditions during days 3 and 21 (see Materials and Methods). These concentrations and time points were selected based on previous experiments done in our group 60,61 . All Agave plantlets at the beginning of the experiment (control, day 0), were selected to have the same height, on average 2.5 cm (Fig. 4A). After being in contact with 0.5 μM of IAA and 2,4-D for days 3 and 21, the weight, height and number of leaves were recorded (Fig. 4B). Due to its robust nature, plantlets of A. fourcroydes have more leaves than A. angustifolia or A. tequilana. After 3 days in contact with IAA and 2,4-D, Agave species started to show important differences in comparison with the control (without auxins). For instance, while A. fourcroydes started to develop roots in the absence of auxins, in the presence of either IAA or 2,4-D there was no root development, although these plants were thinner than the control (Fig. 4A). On the other hand, A. angustifolia showed substantial growth variation after 3 days without auxins. It was found that while some A. angustifolia plantlets did not grow, others grew taller, reaching more than 5 cm in height (Fig. 4B). In the presence of 2,4-D, A. angustifolia and A. fourcroydes plantlets grew better and had a greener phenotype than in the presence of IAA. The opposite result happened in A. tequilana, where the plants grew better in the presence of IAA than in 2,4-D. The most interesting results were observed at 21 days in the presence and absence of the auxins. For instance, in the absence of auxins, A. angustifolia developed roots, in the presence of IAA the plantlets developed shoots, and in the presence of 2,4-D, the plantlets developed callus (Fig. 4A). It is very likely that the formation of callus contributed to this condition, having more weight (0.586 g) than in the presence of IAA (0.479 g) or the control (0.351 g) (Fig. 4B). These phenotype characteristics did not appear in A. fourcroydes or A. tequilana, where roots only developed in the absence of auxins. Although the roots in the control of A. fourcroydes emerged at day 3, by day 21 A. tequilana presented the longest roots. It is worth noting that A. tequilana in the presence of 2,4-D developed plants with wide leaves.
Gene expression of ARF revealed species-specific sensibility to auxins. To determine the possible effects over time of IAA and 2,4-D on gene expression, we analyzed the expression pattern of two ARF with domains related to repression, AteqARF4 and 21, and two ARF related to activation of expression, AteqARF15 and 29, after 3 and 21 days in the absence and presence of auxins (Fig. 5). These ARFs were selected as a representative diverse sample coming from each of the different subclades that comprise the two main clades inferred from the phylogenetic reconstruction (Fig. 2).
Unexpectedly, A. fourcroydes did not present important expression changes in the absence or presence of IAA or 2,4-D. Interestingly, in the case of A. angustifolia and A. tequilana, the presence of IAA or 2,4-D provoked differential expression for different ARFs.  www.nature.com/scientificreports www.nature.com/scientificreports/ 2,4-D in A. angustifolia at day 21 (Fig. 5). Therefore, ARF15 is affected by 2,4-D but only in A. angustifolia. This opens a new avenue to study auxin regulation in a species-specific manner that will help to develop new strategies to understand plant development.

Discussion
The family of the ARF genes is large and its members have functional redundancy, which creates additional challenges when trying to understand ARF function. Finet et al. 62 reported a complete phylogenetic analysis covering 21 species, including A. thaliana, Carica papaya, Glycine max, Medicago truncatula, O. sativa, V. vinifera, Z. mays and others in order to find the similitude between and among the ARF of different species. In that analysis, it was found that all ARFs were grouped in three clades. We also found that the 23 ARFs in Agave can be grouped in these three major clades (Fig. 2). With a phylogenetic analysis, we found that AteqARF15 was in the same clade as ARF27 of Z. mays, ARF7 of Solanum lycopersicum, ARF7 and 19 of A. thaliana, ARF10 of V. vinifera and ARF3 of C. sativus (Fig. 3). This similitude seems to be the result of duplication events among the genes and a domain rearrangement due to alternative splicing 63,64 , which has modified the number of ARF genes in different species 53,62 . An expression analysis in silico of ZmARF27 showed that this gene in maize is mainly expressed in the seed, tassel and ear 59 , tissues affected by auxin concentration [65][66][67] . For instance, Chen et al. 65 found that there is a positive correlation between IAA concentration and ear height. On the other hand, Erkoyuncu et al. 67 reported that high concentrations of 2,4-D increase callus formation in maize embryos. We found that 2,4-D was associated with callus formation in A. angustifolia after 21 days, the day at which AteqARF15 was downregulated (Fig. 5). A similar conclusion was made in a study related to callus formation in Arabidopsis and rice, in which it was found that AtIAA14-ARF7/19 in Arabidopsis is not required for callus initiation but in rice a homolog gene is strictly required 68 .
The number of ARF genes in Agave (Table 1) is consistent with the number of ARF found in other monocotyledonous species 62 . Studies in Arabidopsis have revealed that ARFs show gene redundancy 69 . Therefore, we selected four representative ARF, two repressor-related and two activator-related, to investigate whether different species exposed to auxin could interfere with ARF gene expression (Fig. 5). Here, we show that plants from the same genus but different species not only can be susceptible to a different degree in the presence or absence of auxins but also depending on the auxin (IAA or 2,4-D). In all cases, the effect on the expression of the ARFs would be different (Fig. 5).
In vitro propagation of Agave has solved many of the problems that conventional Agave farming has faced, such as a long maturation period and lack of high-quality materials to work with and genetic improvement programs [70][71][72] . In vitro culture represents an excellent alternative for producing a massive number of plants with superior market qualities in a short period of time. However, plant growth regulators have been a challenge to work with in order to determine the best doses for plant growth and propagation, and auxin concentrations have been particularly challenging. The use of concentrations of 2,4-D between 0.05 and 0.5 μM have been related to phenotypic variation in commercial cultivars such as strawberry 73 , soybean 74 and cotton 75 . In the case of Agave, the fourcroydes it promoted robustness of the leaves, a finding that indicates different auxin regulation functions in different species. In A. thaliana, it has been reported that ARF19 increases its expression as a function of the concentration -the greater the concentration, the higher the expression -and the time in contact with external IAA 42 .
Given that only A. angustifolia showed callus formation in response to the same auxin stimulus experienced by A. tequilana and A. fourcroydes, it is clear that a different mechanism of callus initiation must exist between species. It is important to answer how the mechanism for auxin response differs in one species versus another species and how ARFs can be regulated in order to achieve high-quality plants resistant to developing challenges.

Materials and Methods
plant material and culture conditions. Three different in vitro-propagated Agave species (Agave angustifolia Haw., A. fourcroydes Lem., and A. tequilana Weber var. Azul) were used to perform all the analyses. All Agave plantlets were cultured in multiplication media supplemented with 2.22 μM BA and 0.1 μM 2,4-D 76 under photoperiod (12/12 hr) conditions. After six weeks, plantlets 1.5-2 cm in height were selected in order to have a Figure 2. Phylogenetic analysis of the ARF of Agave tequilana. Phylogeny was calculated from a MSA of the 30 putative A. tequilana ARFs (two highly truncated ARFs were removed). The evolutionary history was inferred using the Minimum Evolution method 81 . The optimal tree with the sum of branch length = 11.98949640 is shown. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1,000 replicates) are shown next to the branches 82 . The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. The evolutionary distances were computed using the JTT matrix-based method 80 and are in the units of the number of amino acid substitutions per site. The rate variation among sites was modeled with a gamma distribution (shape parameter = 1.219). The ME tree was searched using the Close-Neighbor-Interchange (CNI) algorithm 83 at a search level of 2. The Neighbor-joining algorithm 84 was used to generate the initial tree. The analysis involved 30 amino acid sequences. All positions containing gaps and missing data were eliminated. There were a total of 74 positions in the final dataset. Evolutionary analyses were conducted in MEGA7 85 . ARFs selected for gene expression analyses are marked by solid black circles. (2020) 10:3860 | https://doi.org/10.1038/s41598-020-60865-y www.nature.com/scientificreports www.nature.com/scientificreports/ homogenous group of plants. Once plants of the same size from each species were selected, these were kept for eight weeks in Magenta culture boxes filled with 50 mL of Murashige and Skoog media with reduced nitrogen, solidified with 1.75 g/L of Gelrite and without growth regulators in order to avoid contact with exogenous growth regulators before the treatments. After a two-month period without auxins, each Agave species was cultured with either of two different types of auxins (IAA and 2,4-D) at 0.5 μM and without auxins (Control). In each condition for all Agave species the weight, height and number of leaves were evaluated at days 0, 3 and 21. These concentrations and time points were selected based on previous experiments done in our group 60,61 .  59 ). The evolutionary history was inferred using the Minimum Evolution method 81 . The optimal tree with the sum of branch length = 52.16578416 is shown. The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. The evolutionary distances were computed using the JTT matrix-based method 80 and are in the units of the number of amino acid substitutions per site. The rate variation among sites was modeled with a gamma distribution (shape parameter = 1.011). The ME tree was searched using the Close-Neighbor-Interchange (CNI) algorithm 83 at a search level of 2. The Neighbor-joining algorithm 84 was used to generate the initial tree. The analysis involved 158 amino acid sequences. All ambiguous positions were removed for each sequence pair. There were a total of 1827 positions in the final dataset. Evolutionary analyses were conducted in MEGA7 85 . (2020) 10 80 with a gamma distribution (shape parameter α = 1.011), and 1,000 bootstrap repetitions as a test of confidence level. The best amino acid substitution model was found using the best-fit test implemented in MEGA7. There were 1,827 amino acid positions on the dataset. Ambiguous data were removed for each sequence pair. For structural analysis and determination of conserved domains, the aligned sequences were tested against the databases of Pfam (http://pfam.xfam.org/), Gene3D (http://gene3d. biochem.ucl.ac.uk/Gene3D/), Interpro (http://www.ebi.ac.uk/interpro/), SMART (http://smart.embl-heidelberg. de/), PANTHER (http://www.pantherdb.org/), and Superfamily (http://supfam.org/SUPERFAMILY/hmm.html), using the InterProScan plugin for Geneious ® R7.
Highly truncated A. tequilana ARF proteins missing entire domains due to incomplete transcripts were removed from the original multiple alignment as well as all non-A. tequilana ARF proteins. A second phylogeny was inferred from the resulting dataset as above with the following changes: gamma distribution α = 1.219 and complete deletion of ambiguous or missing data (gaps). There was a total of 30 amino acid sequences with 74 positions in the final dataset. primer design. Primers for the different ARF genes of the three selected Agave species were designed based on the identified sequences in the A. tequilana transcriptome 51 . Primer sequences for RT-PCR are listed in Table S1.
Rt-pcR analysis. Total RNA was extracted from 0.3 g of leaf tissue from each plant (A. angustifolia, A. fourcroydes, and A. tequilana) and every condition (control, IAA, and 2,4-D) at day 0, 3, and 21 using the BRL Trizol reagent (Invitrogen) and re-purified with the Qiagen RNeasy Mini Kit, following the manufacturer's instructions. RT reactions were performed in a 20 µl volume containing 5 µg of total RNA and 200 units of the M-MLV Reverse Transcriptase (Invitrogen), following the manufacturer's conditions. Platinum Taq polymerase