Auxin-Induced SaARF4 Downregulates SaACO4 to Inhibit Lateral Root Formation in Sedum alfredii Hance

Lateral root (LR) formation promotes plant resistance, whereas high-level ethylene induced by abiotic stress will inhibit LR emergence. Considering that local auxin accumulation is a precondition for LR generation, auxin-induced genes inhibiting ethylene synthesis may thus be important for LR development. Here, we found that auxin response factor 4 (SaARF4) in Sedum alfredii Hance could be induced by auxin. The overexpression of SaARF4 decreased the LR number and reduced the vessel diameters. Meanwhile, the auxin distribution mode was altered in the root tips and PIN expression was also decreased in the overexpressed lines compared with the wild-type (WT) plants. The overexpression of SaARF4 could reduce ethylene synthesis, and thus, the repression of ethylene production decreased the LR number of WT and reduced PIN expression in the roots. Furthermore, the quantitative real-time PCR, chromatin immunoprecipitation sequencing, yeast one-hybrid, and dual-luciferase assay results showed that SaARF4 could bind the promoter of 1-aminocyclopropane-1-carboxylate oxidase 4 (SaACO4), associated with ethylene biosynthesis, and could downregulate its expression. Therefore, we concluded that SaARF4 induced by auxin can inhibit ethylene biosynthesis by repressing SaACO4 expression, and this process may affect auxin transport to delay LR development.


Introduction
An optimal root system is essential for plant health and productivity [1]. Under abiotic stress in particular, the survival of plants is determined by the plasticity of root growth [2]. Promoting the formation of lateral roots (LRs) and adventitious roots (ARs) is an adaptive strategy used by plants in response to environmental changes [3,4]. These activities are referred to as the stress-induced morphogenic response (SIMR) [5]. LRs are an important component of SIMR [6]. It is reported that more LRs are induced under phosphate-deficiency or a salty environment [7,8]. Moreover, LR formation appears to be dependent on the strength of the abiotic stress. The generation of LRs will be promoted under low salt stress, whereas the LR number will be seriously decreased under high salt stress [3]. Significant differences in the structure of the root system can also be found when culturing plants using low and high nutrient solutions [9]. These results indicate that LR formation should be the result of an interaction between the in vivo and in vitro signals of plants. Furthermore, local auxin accumulation (LAA) is a prerequisite for LR formation [10]. However, the detailed mechanisms of this prerequisite remain to be further elucidated. In particular, which genes function in the process of LAA formation, and how do these genes incorporate environmental and endogenous signals to control LR development? The interaction between auxin and ethylene is the basis for root development [11]. Ethylene can adjust auxin transport and can affect the LAA process [12,13]. Therefore, low levels of ethylene can be helpful for LR initiation [11]. However, abiotic stress can induce large amounts of ethylene, while high levels of ethylene will repress LR formation [11]. In other words, genes restricting ethylene synthesis may contribute to LR development and LAA is a prerequisite for LR development. Thus, the auxin-induced genes limiting ethylene biosynthesis may play important roles in LR formation under abiotic stress. These types of genes should have an ability to connect the auxin pathway and ethylene pathway. According to previous studies, auxin response factor (ARF) genes and 1-aminocyclopropane-1-carboxylate oxidase (ACO) genes (transforming ethylene precursor to ethylene) [14] may be a bridge between auxin and ethylene. For example, MdARF5 from Malus domestica can promote ethylene biosynthesis to initiate apple fruit ripening by regulating MdACO1 [15]. Similarly, SlARF7 from Solanum lycopersicum can trigger SlACO4 expression to influence the set and early development of tomato fruits [16]. These studies mainly focused on how ARFs adjust ethylene production to control fruit development by regulating ACO expression. Although ethylene is also important for LR development, whether ARFs can dominate ethylene synthesis to modify LR development by regulating ACO expression has not been thoroughly elucidated.
The miR390/TAS3/ARFs module is well-known for its functions in controlling LR development [17][18][19]. Although this module is highly conserved among different species, slight differences can also be found. For instance, the changes in LR density in Arabidopsis did not reach a significant level when overexpressing miR390 [18] whereas LR number was highly increased in poplar [19]. However, it is undeniable that this module plays important roles in LR development. ARF2, ARF3, and ARF4 are involved in this module, and these three genes are all transcription repressors [20]. Among them, the overexpression of ARF4 inhibits LR development under salt stress, indicating that ARF4 can function in root development under abiotic stress [19]. Moreover, in our previous studies, SaARF4 was found to be an important hub gene in a co-expression network under cadmium (Cd) stress [21]. Thus, SaARF4 was selected as the candidate gene for elucidating the mechanisms behind its functions in LR development.
In this study, the Cd/zinc co-hyperaccumulator Sedum alfredii Hance (HE) [22] was selected as the plant material to confirm (1) whether SaARF4 can regulate LR development in HE; (2) whether there is an ACO gene that is a downstream gene for SaARF4; and (3) the detailed mechanisms by which SaARF4 regulates ACO to change LR development.

Overexpression of SaARF4 Inhibited the Development of the Vessels, LRs, and ARs
In the miR390/TAS3/ARFs module, microRNA390 (miR390) can effectively decrease ARF4 (the target of miR390) expression, based on previous studies [17][18][19]. Therefore, MIR390 ( Figure S1B) and SaARF4 were cloned from HE DNA and cDNA sequences, respectively. They were then separately introduced into HE to generate the overexpression transgenic lines miR390-OE and SaARF4-OE. We verified the cleavage site of miR390 in SaARF4 ( Figure S1A), and the overexpression of miR390 reduced the SaARF4 expression ( Figure S1C). However, we did not observe any significant differences in the total length or number of roots in miR390-OE compared with the wild-type (WT) plants ( Figure S1D-G). In sharp contrast, overexpressing SaARF4 seriously delayed the development of ARs and LRs ( Figure 1A). In detail, the generation of ARs in the WT was earlier than that in SaARF4-OE ( Figure S2). Moreover, the number of ARs in SaARF4-OE was lower than that of WTs after the plants were cultured for about 21 d ( Figure 1B, p < 0.05). The number of LRs in SaARF4-OE was only 35.4% of that in the WT ( Figure 1C, p < 0.01). Similarly, the lengths of the ARs and LRs in SaARF4-OE were smaller than those in the WT ( Figure 1D,E, p < 0.01). Cross sections of the SaARF4-OE stems were made and stained with HClphloroglucinol to identify whether any changes could be observed in the structures of the SaARF4-OE stems. The results indicated that aberrant vessels were formed in SaARF4-OE stems ( Figure 1F). The diameter, number, and total area of the SaARF4-OE vessels were all decreased compared with those of the WT ( Figure 1G-I). For example, the diameters of the SaARF4-OE vessels were only 60.0% of those in the WT ( Figure 1G).
LRs in SaARF4-OE was only 35.4% of that in the WT ( Figure 1C, p < 0.01). Similarly, the lengths of the ARs and LRs in SaARF4-OE were smaller than those in the WT ( Figure 1D, E, p < 0.01). Cross sections of the SaARF4-OE stems were made and stained with HClphloroglucinol to identify whether any changes could be observed in the structures of the SaARF4-OE stems. The results indicated that aberrant vessels were formed in SaARF4-OE stems ( Figure 1F). The diameter, number, and total area of the SaARF4-OE vessels were all decreased compared with those of the WT ( Figure 1G-I). For example, the diameters of the SaARF4-OE vessels were only 60.0% of those in the WT ( Figure 1G). Figure 1. Changes in vessels, and lateral and adventitious roots of the SaARF4-OE lines: (A-E) the influence of SaARF4 on the number and length of the lateral and adventitious roots (three biological replicates of the wild-type (WT) and SaARF4-OE each); (F) cross sections of the stems stained by HCl-phloroglucinol to identify the differences between the WT and SaARF4-OE lines (three biological replicates each); and (G-I) the changes in total area, number, and diameter of the vessels between the WT and SaARF4-OE lines. * p < 0.05; ** p < 0.01.

SaARF4 Is Expressed in Vascular Tissue and Is Upregulated by Auxin
LRs are generally generated at the LAA sites. The main characteristic of these sites compared with other locations in the rots is a higher auxin concentration. Therefore, analyzing how auxin affected the SaARF4 expression is important for clarifying its roles at these sites. Meanwhile, as the expression sites of a gene are closely related to its functions, determining the expression sites of SaARF4 was beneficial for the following functional analyses. Therefore, we fused the SaARF4 promoter to a β-glucuronidase (GUS) reporter gene, and the recombinant genes were introduced into HE to produce ProARF4::GUS transgenic lines. GUS expression was observed in the roots, stems, and leaves of transgenic plants after histochemical staining ( Figure 2A). Meanwhile, GUS signals were also observed in the vascular tissue of the stems ( Figure 2B). Specifically, strong preferential expression was present in the surrounding areas of the vessels in the primary xylem (Figures 2C and S3). However, GUS was weakly expressed in the phloem ( Figure 2B, the red arrow). A similar pattern was also found in the leaves ( Figure 2D,E). This expression site SaARF4-OE each); (F) cross sections of the stems stained by HCl-phloroglucinol to identify the differences between the WT and SaARF4-OE lines (three biological replicates each); and (G-I) the changes in total area, number, and diameter of the vessels between the WT and SaARF4-OE lines. * p < 0.05; ** p < 0.01.

SaARF4 Is Expressed in Vascular Tissue and Is Upregulated by Auxin
LRs are generally generated at the LAA sites. The main characteristic of these sites compared with other locations in the rots is a higher auxin concentration. Therefore, analyzing how auxin affected the SaARF4 expression is important for clarifying its roles at these sites. Meanwhile, as the expression sites of a gene are closely related to its functions, determining the expression sites of SaARF4 was beneficial for the following functional analyses. Therefore, we fused the SaARF4 promoter to a β-glucuronidase (GUS) reporter gene, and the recombinant genes were introduced into HE to produce ProARF4::GUS transgenic lines. GUS expression was observed in the roots, stems, and leaves of transgenic plants after histochemical staining ( Figure 2A). Meanwhile, GUS signals were also observed in the vascular tissue of the stems ( Figure 2B). Specifically, strong preferential expression was present in the surrounding areas of the vessels in the primary xylem ( Figure 2C and Figure S3). However, GUS was weakly expressed in the phloem ( Figure 2B, the red arrow). A similar pattern was also found in the leaves ( Figure 2D,E). This expression site of SaARF4 was consistent with its function of changing vessel diameter ( Figure 1F). Moreover, as changes in vascular tissues may alter polar auxin transport [23,24], the effect of SaARF4 on changing the vessel diameter indicated that its functions may affect auxin transport. transport using NPA, the activity of the SaARF4 promoter was only restricted to the root meristem zone that can produce auxin by itself ( Figure 2H, the red arrow). By contrast, the application of IAA enhanced the GUS signals ( Figure 2I, the red arrow). Consistent with this result, the use of IAA significantly increased the GUS expression in ProARF4::GUS. ( Figure 2F, p < 0.05). These results indicated that SaARF4 can be induced by auxin. In other words, SaARF4 expression at the LAA sites with a higher auxin concentration should be higher than those in the other parts of the roots.

SaARF4 Decreased Ethylene Content and Adjusted PIN Expression
ACO genes are the key enzymes that convert ACC to ethylene. Thus, before identifying the relationship between SaARF4 and ACOs, we wondered whether SaARF4 could Naphthylphthalamic acid (polar auxin transport inhibitor, NPA) [25] and indoleacetic acid (IAA) were used to confirm the relationship between SaARF4 and auxin. We found that, in the control plants (without any treatment), GUS was expressed in the tips and other places of the roots ( Figure 2G, the red arrow). However, when inhibiting polar auxin transport using NPA, the activity of the SaARF4 promoter was only restricted to the root meristem zone that can produce auxin by itself ( Figure 2H, the red arrow). By contrast, the application of IAA enhanced the GUS signals ( Figure 2I, the red arrow). Consistent with this result, the use of IAA significantly increased the GUS expression in ProARF4::GUS. ( Figure 2F, p < 0.05). These results indicated that SaARF4 can be induced by auxin. In other words, SaARF4 expression at the LAA sites with a higher auxin concentration should be higher than those in the other parts of the roots.

SaARF4 Decreased Ethylene Content and Adjusted PIN Expression
ACO genes are the key enzymes that convert ACC to ethylene. Thus, before identifying the relationship between SaARF4 and ACOs, we wondered whether SaARF4 could affect ethylene production. Moreover, according to the above analysis, we also needed to further evaluate whether SaARF4 could regulate auxin transport as well as the relationships among SaARF4, ethylene, and auxin transport. Thus, we measured the ethylene production of WT and SaARF4-OE under Cd stress. The results showed that ethylene was increased at 0-2 h in both WT and SaARF4-OE ( Figure 3A). However, the ethylene content of SaARF4-OE did not increase significantly compared with that of WT after 2 h. To identify if overexpressing SaARF4 could change the mode of auxin distribution, ProDR5::GUS was introduced into WT and SaARF4-OE using transgenic technology. We found that the distribution of GUS signals in the DR5pro:GUS-SaARF4-OE root tips was significantly different from that of DR5pro:GUS-WT ( Figure 3B). Auxin in the root tips is partially transported by PIN proteins [26], especially PIN1, PIN3, and PIN7. In order to further identify whether SaARF4 could change auxin transport, we measured the PIN expression in SaARF4-OE and WT. A total of 16 PINs were detected in the HE genome. Compared with those in the WT, eight PINs in SaARF4-OE were repressed ( Figure 4C, p < 0.05). Among them, the expressions of SaPIN1, SaPIN2.3, SaPIN3.2, SaPIN4, SaPIN5, and SaPIN7.2 were highly decreased (p < 0.01). affect ethylene production. Moreover, according to the above analysis, we also needed to further evaluate whether SaARF4 could regulate auxin transport as well as the relationships among SaARF4, ethylene, and auxin transport. Thus, we measured the ethylene production of WT and SaARF4-OE under Cd stress. The results showed that ethylene was increased at 0-2 h in both WT and SaARF4-OE ( Figure 3A). However, the ethylene content of SaARF4-OE did not increase significantly compared with that of WT after 2 h. To identify if overexpressing SaARF4 could change the mode of auxin distribution, ProDR5::GUS was introduced into WT and SaARF4-OE using transgenic technology. We found that the distribution of GUS signals in the DR5pro:GUS-SaARF4-OE root tips was significantly different from that of DR5pro:GUS-WT ( Figure 3B). Auxin in the root tips is partially transported by PIN proteins [26], especially PIN1, PIN3, and PIN7. In order to further identify whether SaARF4 could change auxin transport, we measured the PIN expression in SaARF4-OE and WT. A total of 16 PINs were detected in the HE genome. Compared with those in the WT, eight PINs in SaARF4-OE were repressed ( Figure 4C, p < 0.05). Among them, the expressions of SaPIN1, SaPIN2.3, SaPIN3.2, SaPIN4, SaPIN5, and Sa-PIN7.2 were highly decreased (p < 0.01).

Discussion
To date, ARF2, ARF3, ARF4, ARF5, ARF6, ARF7, ARF8, and ARF19 have been reported to be associated with LR development [18,29]. Their downstream genes identified in recent years mainly function in the early developmental stages of LRs [30][31][32], while the SHY2/IAA3-ARFs module is related to auxin signals [33]. These studies indicated that ARFs may profoundly affect almost every aspect of LR development. LAA is a basis for respecifying root pericycle cells into LR founder cells [34], and auxin transport is the precondition for LAA [35]. Understanding whether ARFs are involved in auxin transport and how their functions are achieved is an important aspect for understanding the LR developmental mechanism. In this study, we found that SaARF4 may alter auxin transport to delay LR development. First, the vessel diameter in SaARF4-OE was significantly smaller than that in WT ( Figure 1G), and this phenotype was similar to those induced by a polar auxin transport inhibitor [36]. Second, differences in auxin distribution were found in the root tips by introducing ProDR5:GUS into WT and SaARF4-OE ( Figure 3B). Third, PINs related to auxin transport in the roots [26] were seriously downregulated ( Figure 3C). The decline in ethylene content caused by SaARF4 regulating SaACO4 may be an important reason for the variations in auxin transport. Two lines of evidence supported our judgement: the expression patterns of PINs in the roots of WT were similar to those of SaARF4-OE at the application of PZA ( Figure 3C,D), and moreover, PZA could effectively inhibit the generation of LRs of the WT (Figure 3). In previous studies, MdARF5 in apples effectively upregulated MdACO1 to enhance ethylene production [15], and SlARF7 in tomato positively regulated SlACO4 to influence fruit development [16]. These studies mainly focused on the roles of ethylene in fruit initiation [16] or ripening [15]. Together, our study and previous studies have demonstrated the roles of ARFs and ACOs in connecting the pathways of auxin and ethylene. . Vectors in the dual luciferase assay system were performed (B). The results of the Y1H indicated physical binding between SaARF4 and the SaACO4 promoter (C). HSa, pHIS2 + pGADT7-rec-SaARF4; HA, pHIS-AuxRE + pGADT7-rec; HASa, pHIS2-AuxRE + pGADT7-rec-SaARF4. The intensity of the fluorescence indicates the promoter activities of SaACO4 (D) in three biological replicates. Quantitative RT-PCR with Ren as the reference gene was conducted to further verify the results of the dual luciferase assay system (E,F). RML, relative mRNA level. * p < 0.05; ** p < 0.01.
SaARF4 can reduce the production of ethylene ( Figure 3A), and ethylene was previously reported to be able to regulate PIN expression [12,27]. Therefore, we next determined whether these PINs were directly influenced by SaARF4 or indirectly regulated by SaARF4 via ethylene. Pyrazinamide (PZA, ethylene biosynthesis inhibitor) can effectively inhibit the activities of ACO enzymes to block ethylene production [28]. Therefore, WT was treated with PZA to determine whether the PIN expression was affected by ethylene. The results indicated that the expression patterns of PINs in WT with PZA treatment were highly similar to those of SaARF4-OE ( Figure 3C,D), especially the PINs that were reported to be related to auxin transport in the roots [26] (the red boxes). Meanwhile, the promoter elements of these PINs were analyzed using the PlantCARE website (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/), and the results indicated that the AuxRE element (TGTCTC, the special binding site for ARFs) was not found in these PIN promoters (Table S1). Therefore, we speculated that SaARF4 regulated the PINs by influencing ethylene production. In other words, the effect of SaARF4 on the root development may be mediated by restricting ethylene production. In order to further confirm this conclusion, we need to identify whether inhibiting ethylene (using PZA) could influence root development in the WT under Cd stress. The results indicated that the root development of the WT under Cd stress was highly inhibited by PZA treatment ( Figure 3E). In detail, the LRs were 19.0% of that in the control plants (without PZA treatment) and the ARs also decreased by 49.0% under PZA treatment ( Figure 3F). In summary, in combination with the functions of ethylene in PIN expressions, we drew a basic conclusion that ethylene induced by Cd stress plays important roles in root development by regulating PIN expression while SaARF4 acted on root development by regulating ethylene production.

SaARF4 Negatively Regulated Its Downstream Gene, SaACO4
It has been confirmed in previous studies that ARFs can control ACOs to influence ethylene [15,16]. Our results indicated that root development inhibition was detected in SaARF4-OE and that similar phenotypes were observed in the WT with PZA treatment (Figures 1 and 3). In order to identify which genes SaARF4 regulates in ethylene biosynthesis, we performed ChiP (chromatin immunoprecipitation)-seq of SaARF4. The results of the ChiP-seq (Table S2) indicated that the AuxRE element in SaACO4 was a possible binding site of SaARF4 (the predicted binding site is indicated in Figure 4B). The dual-luciferase assay system and yeast one-hybrid (Y1H) were used to identify the relationship between SaARF4 and SaACO4. The construction of relevant vectors is indicated in Figure 4B. The results showed that SaARF4 could effectively rescue the auxotrophic phenotype of yeast ( Figure 4C), indicating that SaARF4 can bind to the AuxRE element of the SaACO4 promoter. In the dual-luciferase assay system, the fluorescence intensity was seriously decreased after adding SaARF4 ( Figure 4D). On the contrary, when removing AuxRE from the SaACO4 promoter ( Figure 4B), the fluorescence intensity of the dual-luciferase assay system was increased ( Figure 4D,F), and these results were also verified by qRT-PCR ( Figure 4D). In addition, a 20-fold lower expression level of SaACO4 was detected in SaARF4-OE than in WT ( Figure 4A). In summary, SaARF4 can negatively regulate SaACO4 by binding to the AuxRE in the SaACO4 promoter, and this conclusion was consistent with the expression level of SaACO4 in SaARF4-OE ( Figure 4A).

Discussion
To date, ARF2, ARF3, ARF4, ARF5, ARF6, ARF7, ARF8, and ARF19 have been reported to be associated with LR development [18,29]. Their downstream genes identified in recent years mainly function in the early developmental stages of LRs [30][31][32], while the SHY2/IAA3-ARFs module is related to auxin signals [33]. These studies indicated that ARFs may profoundly affect almost every aspect of LR development. LAA is a basis for respecifying root pericycle cells into LR founder cells [34], and auxin transport is the precondition for LAA [35]. Understanding whether ARFs are involved in auxin transport and how their functions are achieved is an important aspect for understanding the LR developmental mechanism. In this study, we found that SaARF4 may alter auxin transport to delay LR development. First, the vessel diameter in SaARF4-OE was significantly smaller than that in WT ( Figure 1G), and this phenotype was similar to those induced by a polar auxin transport inhibitor [36]. Second, differences in auxin distribution were found in the root tips by introducing ProDR5:GUS into WT and SaARF4-OE ( Figure 3B). Third, PINs related to auxin transport in the roots [26] were seriously downregulated ( Figure 3C). The decline in ethylene content caused by SaARF4 regulating SaACO4 may be an important reason for the variations in auxin transport. Two lines of evidence supported our judgement: the expression patterns of PINs in the roots of WT were similar to those of SaARF4-OE at the application of PZA ( Figure 3C,D), and moreover, PZA could effectively inhibit the generation of LRs of the WT (Figure 3). In previous studies, MdARF5 in apples effectively upregulated MdACO1 to enhance ethylene production [15], and SlARF7 in tomato positively regulated SlACO4 to influence fruit development [16]. These studies mainly focused on the roles of ethylene in fruit initiation [16] or ripening [15]. Together, our study and previous studies have demonstrated the roles of ARFs and ACOs in connecting the pathways of auxin and ethylene.
It is a common phenomenon that multiple hormones interact with each other to regulate the developmental process of plants [37,38]. An important function of the interactions is to transform in vivo/in vitro signals into developmental signals [39]. Furthermore, these regulatory processes mainly alter the transport or content of auxin to modify plant morphologies [13]. For example, ABA upregulated PIN2 under osmotic stress, resulting in decreased auxin content in the root meristem leading to root growth inhibition [40]. The interactions between auxin and ethylene act in many developmental aspects of plants, such as apical hook development [41], fruit ripening [42], and primary root elongation [43].
It is worth noting that auxin and ethylene regulate LR development based on a certain threshold level [11]. For example, a low level of ACC can promote LR initiation while higher doses of ACC will inhibit LR development [11,44]. The dose response reflects the complexity and flexibility of developmental regulation. Therefore, keeping a low ethylene content level is important for LR development, especially under abiotic stress. In this study, SaARF4 induced by auxin significantly downregulated SaACO4 expression ( Figure 4A) and thus decreased ethylene production (Figures 2 and 3). Based on these results, the ethylene content at the sites of LAA should be lower than that at other sites of the roots. As ethylene can accelerate the IAA transport rate [45,46], the decline in ethylene could be helpful for maintaining local auxin concentration. This may explain why LAA is a precondition for LR development. Ethylene increases IAA transport by enhancing the expressions of PIN3 and PIN7 to inhibit LR development in Arabidopsis [12]. In line with these studies, we also found that SaPIN1, SaPIN3.2, and SaPIN7.2 were significantly decreased in the SaARF4-OE lines compared with those in WT ( Figure 3C). Moreover, the expression patterns of these genes in SaARF4-OE were highly similar to those of WT treated with PZA ( Figure 3C,D, the red wireframes). These results supported that SaARF4 can downregulate these PINs by inhibiting ethylene production. In summary, more ethylene induced by abiotic stress will enhance PIN expression, and this process may not be conducive to LAA. SaARF4 can effectively decrease ethylene production. Therefore, the functions of SaARF4 will play important roles in LR development under abiotic stress.
In conclusion, we mainly described the roles of SaARF4 and SaACO4 in the sites of LAA of roots under abiotic stress in this study. A model involving SaARF4 and SaACO4 was built to clearly illustrate the mechanisms behind LR development. Although ethylene was induced at the beginning of Cd stress, the ethylene content was still at a low level. At this time, PIN expression upregulated by ethylene can be helpful for LAA ( Figure 5A). With auxin accumulation, a large amount of auxin enhanced SaARF4 expression ( Figure 5B), following which SaACO4 was seriously inhibited by SaARF4. Thus, ethylene production will be downregulated at the sites of LAA. The declined ethylene can repress PIN expression, and this process can effectively prevent the outflow of auxin from LAA sites ( Figure 5C). Therefore, LRs tend to be generated at the sites of LAA. For this reason, more LRs grew in S. alfredii at a low level of Cd stress [22]. However, if the content or intensity of ethylene induced by abiotic stress exceeds the range that SaARF4 could adjust to, the generation of LRs may also be inhibited. This may explain why LRs are inhibited under high levels of abiotic stress [22]. ; at the beginning of Cd stress, the production of ethylene was enhanced and promoted the process of local auxin accumulation (LAA). Then, the expression of SaARF4 was enhanced with auxin accumulation, which gradually repressed SaACO4 expression and finally inhibited ethylene production (B); the decline in ethylene production may be helpful for maintaining the auxin concentration at the sites of LAA, and this process may provide the auxin conditions for LR development (C).

Plant Materials and Growth Conditions
The S. alfredii (HE) plants were cultured in our laboratory. The growth conditions and hydroponic experiments were as described previously [47].

Plasmid Construction
The open reading frame and 2 kb promoter of SaARF4 were amplified from the cDNA and genomic DNA of HE, respectively. These products were cloned into pDONR222 and then recombined into pK2GW7.0 and pMDC164 to produce 35S::SaARF4 and ProARF4::GUS, respectively. Moreover, the miR390 precursor (MIR390) was cloned from the DNA, and the precursor was also recombined into the Pk2GW7.0 vector to produce 35S::miR390. Meanwhile, the complete coding sequence of SaARF4 was assembled into pGADT7-Rec (for theY1H) and pGreenII 62-SK (for the dual-LUC reporter system) using the sites of SmaI/BamHI and EcoRI/HindIII, respectively. The promoter sequence (approximately 500 bp, ProSaACO4) of SaACO4 was amplified from DNA. After that, overlapping PCR was used to remove the AuxRE element [15] from ProSaACO4 to produce ProSaACO4m, which was then inserted into the vector of pGreenII 0800-LUC using SalI/BamHI to create ProSaACO4::LUC. The ProDR5::GUS vector was consistent with that in previous reports [48]. The primers used are listed in Table S1.

Plant Materials and Growth Conditions
The S. alfredii (HE) plants were cultured in our laboratory. The growth conditions and hydroponic experiments were as described previously [47].

Plasmid Construction
The open reading frame and 2 kb promoter of SaARF4 were amplified from the cDNA and genomic DNA of HE, respectively. These products were cloned into pDONR222 and then recombined into pK2GW7.0 and pMDC164 to produce 35S::SaARF4 and ProARF4::GUS, respectively. Moreover, the miR390 precursor (MIR390) was cloned from the DNA, and the precursor was also recombined into the Pk2GW7.0 vector to produce 35S::miR390. Meanwhile, the complete coding sequence of SaARF4 was assembled into pGADT7-Rec (for theY1H) and pGreenII 62-SK (for the dual-LUC reporter system) using the sites of SmaI/BamHI and EcoRI/HindIII, respectively. The promoter sequence (approximately 500 bp, ProSaACO4) of SaACO4 was amplified from DNA. After that, overlapping PCR was used to remove the AuxRE element [15] from ProSaACO4 to produce ProSaACO4m, which was then inserted into the vector of pGreenII 0800-LUC using SalI/BamHI to create ProSaACO4::LUC. The ProDR5::GUS vector was consistent with that in previous reports [48]. The primers used are listed in Table S1.

Plant Transformation
The vectors containing 35S::SaARF4, 35S::miR390, ProARF4::GUS, and ProDR5::GUS were introduced into EHA105 Agrobacterium tumefaciens strain by electroporation. Subsequently, the A. tumefaciens containing the 35S::SaARF4, 35S::miR390, and ProARF4::GUS vectors was used to infect the HE calluses. After obtaining the overexpression transgenic lines of SaARF4 (SaARF4-OE), the A. tumefaciens containing the ProDR5::GUS vector was used to infect the WT and SaARF4-OE lines. The A. tumefaciens-infected method was performed as described before [49] with minor modifications. The differentiation medium was a Murashige and Skoog medium (MS) + 2 mg·L −1 6-benzylaminopurine + 0.3 mg·L −1 1-naphthaleneacetic acid, and the rooting medium was 1/2 MS + 2 mg·L −1 3-indolebutyric acid. The use of antibiotics is varied with the different vectors. The concentrations of kanamycin and hygromycin were 30 mg·L −1 and 20 mg·L −1 , respectively. WT and transgenic plants were subsequently transplanted into soil for further detection and experimentation.

qRT-PCR
Total RNAs were extracted from the roots of WT and transgenic lines (with three different lines) using an RNA extraction kit (RNAprep Pure Plant Kit, TIANGEN, Dalian, China). Next, first-strand cDNAs were generated using a cDNA synthesis kit (PrimeScript™ RT Master Mix, TAKARA, Beijing, China). These cDNAs were utilized for qRT-PCR using TB Green reagent (TB Green™ Premix Ex Taq™, TAKARA, Dalian, China). Sequence Processing and Data Extraction (SPDE) was used for the batch design of the qRT-PCR primers [50]. All of the primers, including that of the reference gene (UBC9), are listed in Table S1.

Identification of the Members of PIN Gene Family
The PIN gene family members were extracted from genomic files using the related pfam file (PF03547.18) on the SPDE software platform. These genes were retested by NCBI-Blast (https://blast.ncbi.nlm.nih.gov/Blast.cgi) and named after their homologous genes in Arabidopsis using the TAIR website (https://www.arabidopsis.org/). The genes with low expression levels (ct value > 30 in qRT-PCR experiment) in the roots were then also removed.

Histochemical Analyses and Tissue Section
The above 2nd-5th leaves and 3rd internode of the stems of ProARF4::GUS were collected for tissue section. β-glucuronidase (GUS) staining was performed as described previously [51], and three independent transgenic lines were used for GUS staining. The cross sections of WT and SaARF4-OE were made as described before [52], and the sections were stained by HCl-phloroglucinol (three biological replicates).

Auxin/Ethylene Inhibitors Treatment and IAA Application
The selection and application of NPA [26] and PZA [28] were based on published studies. In detail, six transgenic lines of ProARF4::GUS cultured under hydroponic conditions for two weeks were treated with 10 µM NPA and 20 µM IAA for 7 d. The WT plants were cultured hydroponically in the presence and absence of 100 µM PZA for about 14 d with three biological replicates.

Measurement of Ethylene Content
Cadmium treatment (as stated above) was conducted using equal qualities of SaARF4-OE and WT (3 g, at least three biological replicates) for 0-5 h. The ethylene content was measured using an F-900 Portable Ethylene Analyzer (Felix instruments, Camas, WA, USA) as in Lerud et al. (2019) [53].

Y1H Assay
The coding region of SaARF4 was inserted into the vector of pGADT7-Rec (Clontech) to produce pGADT7-rec-SaARF4. A 20 bp fragment (as a unit) with the AuxRE motif was cloned from the SaACO4 promoter. Four tandem copies of the unit were constructed and cloned into pHIS2 vectors to produce pHIS2-ACO4 [54]. The pGADT7-rec-SaARF4/pHIS2-ACO4 was co-transformed into AH109 (yeast strain) using the LiAc-PEG3350 method [55]. SD-Leu-Trp plates were used for transformant selection, and SD-Leu-Trp-His plates supplemented with 30 µM 3-AT (3-amino-1,2,4-triazole) were utilized for testing the interactions. Three biological replicates and three technical repeats were used for this process. The primers are listed in Table S1.

ChIP Assays
The expression and purification of the SaARF4 protein were performed as described before [56] with minor changes. After cloning the coding sequence of SaARF4, the sequence was inserted into the NotI and SbfI sites of the pMAL-c5x vector to generate pMAL-c5x-SaARF4. Then, the recombinant vector was transformed into Escherichia coli Arctic Express, and the recombinant protein was expressed under the condition of 0.5 mM β-D-thiogalactopyranoside (IPTG) at 20 • C. After that, the purified protein was used as the antigen to produce the mouse monoclonal antibody (AbMART). Western blot was performed for detecting the specificity of the antibody ( Figure S4) [57]. The basic process of the ChIP assay refers to the method stated by Landt et al. (2012) [58]. In detail, approximately 4 g of SaARF4-OE leaves were collected for ChIP-seq. The chopped leaves were placed into 1% formaldehyde and vacuum-infiltrated for 15-30 min. These materials were ground in liquid nitrogen. After that, the chromatin was broken into 100-500 bp DNA fragments by sonication. The ChIP experiment was performed using SaARF4-antibody, and protein-DNA-antibody complexes were pulled down and digested using 2 µL proteinase K to obtain the product. Then, the product was purified to obtain the immunoprecipitated DNA. After testing the concentration and purity of DNA using Q-bit, the immunoprecipitated DNA (6 ng) was used for sequencing. The sequencing results were aligned into the genomic data of HE. The final ChIP-seq result is indicated in Table S2.

Transient Transcriptional Activities Assay
The transient transcriptional activity assays were performed as described previously [59]. The corresponding vectors were constructed as stated above. A Lumazone imaging system (Mag Biosystems, Tucson, AZ, US) was used for the image acquisition of luciferase. Image processing was accomplished using the software ImageJ (https://imagej.net/Welcome). The qRT-PCR was conducted for testing the expression level of Luciferase and Rluc, and the Rluc that was driven by 35S promoter was chosen as the reference gene.

Statistical Analysis
At least three biological replicates and three technical repeats were performed for each assay. SPSS 20.0 (IBM Corp., Armonk, NY, USA) was used for statistical analysis (https://www.ibm.com/cn-zh). The method of one-way ANOVA was used for significance analysis. All data are showed as means ± SD. The least significant difference (LSD) test was applied to analyze the differences at the 0.05 or 0.01 probability levels (* p < 0.05 and ** p < 0.01). The histograms were generated by Excel (Microsoft Corp., Albuquerque, NM, USA), and images were processed by Abode Illustrator and Abode Photoshop (Abode Systems).

Conclusions
In this study, we demonstrated a possible function of SaARF4 in LR development. SaARF4 was induced by auxin at the sites of LAA. The expression of SaARF4 was enhanced, and SaACO4 was seriously inhibited. The ethylene content was thus decreased, which could be helpful in the maintenance of local auxin concentration. This regulatory process may play important roles in LR development, especially under abiotic stress.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
The data that support the findings of this study are available from the corresponding author upon reasonable request.

Conflicts of Interest:
The authors declare no conflict of interest.