Membrane Sterol Composition in Arabidopsis thaliana Affects Root Elongation via Auxin Biosynthesis

Plant membrane sterol composition has been reported to affect growth and gravitropism via polar auxin transport and auxin signaling. However, as to whether sterols influence auxin biosynthesis has received little attention. Here, by using the sterol biosynthesis mutant cyclopropylsterol isomerase1-1 (cpi1-1) and sterol application, we reveal that cycloeucalenol, a CPI1 substrate, and sitosterol, an end-product of sterol biosynthesis, antagonistically affect auxin biosynthesis. The short root phenotype of cpi1-1 was associated with a markedly enhanced auxin response in the root tip. Both were neither suppressed by mutations in polar auxin transport (PAT) proteins nor by treatment with a PAT inhibitor and responded to an auxin signaling inhibitor. However, expression of several auxin biosynthesis genes TRYPTOPHAN AMINOTRANSFERASE OF ARABIDOPSIS1 (TAA1) was upregulated in cpi1-1. Functionally, TAA1 mutation reduced the auxin response in cpi1-1 and partially rescued its short root phenotype. In support of this genetic evidence, application of cycloeucalenol upregulated expression of the auxin responsive reporter DR5:GUS (β-glucuronidase) and of several auxin biosynthesis genes, while sitosterol repressed their expression. Hence, our combined genetic, pharmacological, and sterol application studies reveal a hitherto unexplored sterol-dependent modulation of auxin biosynthesis during Arabidopsis root elongation.

In this study, we used the cpi1-1 mutant to dissect the function of sterols in modulating auxin biosynthesis. The CPI1 gene encodes the cyclopropylsterol isomerase, which is responsible for opening the 9β,19-cyclopropane ring of cycloeucalenol and converting it into obtusifoliol [29,30] (Supplemental Figure S1B). The cpi1-1 mutant exhibits a short root, agravitropism, and stomatal developmental defects [8,31]. Expression of the auxin responsive reporter DR5:GUS [32,33] in the cpi1-1 root tip was markedly enhanced and expanded to the lateral root cap and epidermal cells [8]. We found that expression of auxin biosynthesis genes including TAA1, YUC8, and YUC9 was upregulated in cpi1-1. TAA1 mutation reduced DR5:GUS expression in cpi1-1 root tip and partially rescued the short root phenotype of cpi1-1. Since sterol composition in cpi1-1 was markedly altered, with accumulation of cycloeucalenol at the expense of almost complete reduction of the major sterols such as sitosterol, 24-methylcholesterol (campesterol), and stigmasterol [8], we analyzed the effect of these sterols on the expression of DR5:GUS and auxin biosynthetic genes and on the root length of wild type (WT) and cpi1-1. We found that cycloeucalenol upregulated the expression of DR5:GUS and auxin biosynthetic genes, whereas sitosterol repressed the expression of these genes and partially rescued the short root phenotype of cpi1-1.

An Increased Auxin Response in cpi1-1 is Further Enhanced by Defective Polar Auxin Transport
We initially observed the auxin response reporter DR5:GUS whose expression correlates with auxin levels [32,33]. Compared to the WT, cpi1-1 displayed markedly enhanced and ectopic DR5:GUS expression ( Figure 1A,B) [8]. In WT cotyledons, DR5:GUS expression was detected at the tip, whereas in cpi1-1, DR5:GUS expression was expanded to the whole cotyledon where the cotyledon tips and the vasculature revealed strong GUS signals ( Figure 1A). In the WT root tip, DR5:GUS expression was detected in the stem cell niche and columella root cap cells, whereas in the cpi1-1 root tip, DR5:GUS expression was markedly enhanced and expanded to the lateral root cap and epidermal cells ( Figure 1B). In addition, roots of 2-week-old cpi1-1 seedlings displayed more lateral root primordia (which were marked by strong DR5:GUS signal) and lateral roots than that of the WT (Supplemental Figure S2). Polar localization of PIN2 has previously been reported to be defective in cpi1-1 root epidermal and cortex cells [8]. To address whether elevated auxin activity in cpi1-1 was caused by defective PAT, we crossed the loss-of-function pin2-T [8], aux1-T [34], and aux1-T pin2-T [35,36] alleles to cpi1-1, and compared the DR5:GUS expression patterns among the single, double, and triple mutants. The cpi1-1 pin2-T double mutant displayed higher DR5:GUS activity than either cpi1-1 or pin2-T, whereas there was no obvious difference of DR5:GUS activity between cpi1-1 aux1-T and cpi1-1 ( Figure 1A-C).
We further checked the effect of auxin transport inhibitor 1-N-naphthylphthalamic acid (NPA) on DR5:GUS activity in WT and cpi1-1 roots. Compared to the mock, obviously enhanced GUS staining was observed in both WT and cpi1-1 root tips after NPA treatment ( Figure 2A). Quantitative GUS activity assays showed that after NPA treatment, DR5:GUS activity was significantly increased in both WT and cpi1-1 seedlings ( Figure 2B), but cpi1-1 mutant displayed lower fold induction compared with WT ( Figure 2C), suggesting that it is less sensitive to NPA. This result is consistent with the previous report that cpi1-1 displayed defective polar localization of the PIN2 auxin efflux protein [8]. Whereas higher concentrations of NPA (500 nM) inhibited root elongation of both WT and cpi1-1 seedlings, there was no significant difference in relative root length between WT and cpi1-1 ( Figure 2D,E). Together, these results suggest that other factors in addition to defective PAT contribute to the elevated auxin activity in cpi1-1.

The Expression of Auxin Biosynthesis Genes is Upregulated in cpi1-1
Since auxin biosynthesis and transport contribute to maintain appropriate auxin levels during plant development [37], we analyzed the expression patterns of auxin biosynthesis genes in cpi1-1. The TAA1 (also known as TRANSPORT INHIBITOR RESPONSE2, TIR2 or WEAK ETHYLENE INSENSITIVE8, WEI8) gene encodes tryptophan aminotransferase, which catalyzes the tryptophan (Trp) to indole-3-pyruvic acid (IPA) conversion [38][39][40]. In order to analyze TAA1 expression patterns in cpi1-1, we crossed ProTIR2:GUS and ProTIR2:TIR2-GUS (a fusion of TAA1/TIR2 cDNA sequence with the GUS reporter gene under control of the TAA1/TIR2 promoter, which rescued the phenotype of the tir2 mutant) [40] to cpi1-1. In 5-day-old WT seedlings, ProTIR2:GUS signal was detected in cotyledons, hypocotyl, and the root vascular tissue (Supplemental Figure S5A,B; Figure 3A). Compared to WT, obviously increased ProTIR2:GUS expression was observed in cpi1-1 (Supplemental Figure Figure 3D). In addition, the expression pattern of ProTIR2:TIR2-GUS was altered in the root tip of cpi1-1; it was not detected in the root proximal meristem, but was strongly expressed in the epidermal cells of the root elongation zone ( Figure 3D). Since the TAA1/TIR2 cDNA and genomic constructs displayed different expression patterns in the root tip [38,40,41], we also analyzed the expression pattern of ProTAA1:GFP-TAA1 (a fusion of TAA1 genomic sequence with the GFP reporter gene) [38] in cpi1-1. ProTAA1:GFP-TAA1 was strongly expressed in the quiescent center (QC) cells and in columella initials, and its expression was not different between WT and cpi1-1 (Supplemental Figure S6). Together, these results indicate that CPI1 mutation leads to enhanced expression of TAA1 in root vascular tissue and elongation zone, but do not affect its expression in the QC cells. These results are consistent with the previous report that cpi1-1 roots have normal sized meristem but a shorter elongation zone [8].
YUC flavin monooxygenase proteins catalyze the conversion of indole-3-pyruvic acid to indole-3-acetic acid (IAA), which may be considered the rate-limiting step of the main auxin biosynthesis pathway [42][43][44]. To investigate whether the expression levels of YUC genes were altered in cpi1-1, we crossed ProYUC2:GUS, ProYUC3:GUS, ProYUC4:GUS, ProYUC8:GUS, and ProYUC9:GUS to cpi1-1 [45,46]. In 5-day-old WT seedlings, ProYUC2:GUS was mainly expressed in cotyledons, the vascular tissue in the hypocotyl, in the shoot apical meristem, and in leaf primordia (Supplemental Figure S7A,B). Compared to WT, an enhanced expression of ProYUC2:GUS was observed in cpi1-1 (Supplemental Figure S7C,D). ProYUC3:GUS was mainly expressed in the columella root cap cells, and its expression displayed no obvious difference between WT and cpi1-1 (Supplemental Figure S7E,F). ProYUC4:GUS expression was detected in the tips of the cotyledons, in leaf primordia, and in emerging young leaves, and its expression displayed no obvious difference between WT and cpi1-1 (Supplemental Figure S7G-J). In 5-day-old WT seedlings, ProYUC8:GUS expression was detected mainly in the QC cells and in the columella root cap cells ( Figure 3E; Supplemental Figure S8A-C). Of interest, the ProYUC8:GUS signals in the columella initials as well as in the L1 and L2 layers of the columella root cap cells was lower than that in the QC cells and in the L3 and L4 layers of the columella root cap cells (Supplemental Figure S8C). Compared to WT, increased staining of ProYUC8:GUS was observed in the root vascular tissues of cpi1-1 ( Figure 3F; Supplemental Figure S8E), but not in its root QC cells and columella root cap cells ( Figure 3F; Supplemental Figure S8F). In 5-day-old WT seedlings, expression of ProYUC9:GUS was observed mainly in the columella root cap cells, in the root tip elongation zone, as well as in the root vasculature ( Figure   wei8-1, cpi1-1, and cpi1-1 wei8-1 seedlings. Shown are representative images of three independent experiments (employing 18 to 32 seedlings per experiment); (J) Quantitative DR5:GUS activity assay in 5-day-old seedlings. The presented data are means ± SD of n = 3 independent experiments. * p < 0.05, ** p < 0.01 (one-way ANOVA with Tukey's test); (K,L) Phenotypes (K) and root length (L) of 7-day-old seedlings. The presented data are means ± SD of n = 3 experiments (employing 13 to 28 seedlings per experiment). ** p < 0.01 (Student's t-test, one-tailed, two-sample equal variance); (M,N) Expression patterns of DR5 rev :GFP in root tips (M) and quantification of GFP fluorescence (N). The presented data are means ± SD of n = 3 independent experiments (employing 6 to 11 roots per experiment). * p < 0.05 (one-way ANOVA with Tukey's test); (O) Root length of 7-day-old seedlings. The presented data are means ± SD of n = 3 independent experiments (employing 9 to 28 seedlings per experiment). No significant difference between cpi1-1 single and cpi1-1 yuc8 and cpi1-1 yuc9 double and cpi1-1 yuc8 yuc9 triple mutants by one-way ANOVA (p < 0.05). Bars = 100 µm in (A-I,M) and 2 mm in (K).
We also analyzed the expression of auxin biosynthesis upstream genes including ANTHRANILATE SYNTHASE α1 (ASA1, also known as WEAK ETHYLENE INSENSITIVE2, WEI2) and ANTHRANILATE SYNTHASE β1 (ASB1, also known as WEAK ETHYLENE INSENSITIVE7, WEI7), which encode the αand β-subunits of a key enzyme of tryptophan biosynthesis [47]. Compared to WT, enhanced expression of ProASB1:GUS was observed in cpi1-1, whereas expression of ProASA1:GUS was not obviously different (Supplemental Figure S9A).
Together, these results reveal that expression of a number of auxin biosynthesis pathway genes is upregulated in the cpi1-1 mutant, and among them TAA1 and YUC8 genes show cell file specific upregulation.

Modulation of Auxin Biosynthesis Partially Rescues the Short Root Phenotype of cpi1-1
We then investigated whether mutation of auxin biosynthesis genes could reduce the DR5:GUS activity in cpi1-1 and could rescue the short root phenotype of cpi1-1. We found that compared to the cpi1-1 single mutant, DR5:GUS staining was obviously reduced in the cpi1-1 wei8-1 double mutant ( Figure 3I). GUS activity quantification results showed that the cpi1-1 wei8-1 double mutant exhibited a similar GUS activity level as the wei8-1 single mutant ( Figure 3J). Consistently, mutation of WEI8/TAA1 partially rescued the short root phenotype of cpi1-1 ( Figure 3K,L). We also compared the fluorescent signals of DR5 rev :GFP in cpi1-1 single, cpi1-1 yuc8 and cpi1-1 yuc9 double, and cpi1-1 yuc8 yuc9 triple mutants. We found that DR5 rev :GFP fluorescence was obviously reduced in cpi1-1 yuc8 and cpi1-1 yuc9 double and in cpi1-1 yuc8 yuc9 triple mutants ( Figure 3M,N). However, no significant difference of root length was observed in these mutants ( Figure 3O), suggesting that additional YUC genes might play a role. YUC2 and YUC3 mutation also did not affect the hypocotyl length and root length of cpi1-1 (Supplemental Figure S10).
L-Kynurenine (Kyn) is an auxin biosynthesis inhibitor that inhibits TAA1 activity [48]. We thus checked the effect of Kyn on cpi1-1 root growth by growing WT and cpi1-1 seedlings in Murashige and Skoog (MS) media supplemented with a range of concentrations of Kyn (0, 0.5, 1, 2, 3, and 5 µM) for 7 days. WT root growth was inhibited by these Kyn treatments especially at higher Kyn concentrations ( Figure 4A,B). For example, after treatment with 2, 3, and 5 µM Kyn, the WT root length was approximately 58%, 50%, and 40% of the mock, respectively. In contrast, cpi1-1 root growth was partially rescued by Kyn treatment, particularly at 0.5, 1, and 3 µM Kyn, the cpi1-1 root length was increased by more than 21% ( Figure 4A,B). Previous research showed that Kyn treatment inhibited root growth by blocking root meristem function [41]. Therefore, we measured the root meristem size of WT and cpi1-1 seedlings after Kyn treatment. The results showed that size of the meristem zone of WT root was significantly reduced under 5 uM Kyn treatment, but no statistical difference was observed in the root meristem zone length of cpi1-1 mutant ( Figure 4C,D). These results suggest that cpi1-1 is less sensitive to Kyn treatment.
Together, these results demonstrate that genetic and pharmacological inhibition of auxin biosynthesis can partially rescue the short root phenotype of cpi1-1.

cpi1-1 Responds Normally to the Auxin Signaling Inhibitor
p-Chlorophenoxyisobutyric acid (PCIB) is an auxin signaling inhibitor [49]. To investigate whether CPI1 mutation has an effect on auxin signaling, we grew WT and cpi1-1 seedlings in MS media supplemented with different concentrations of PCIB (0, 0.5, 1, 2, 3, and 5 µM) for 7 days. We found that root growth of both WT and cpi1-1 were inhibited by PCIB treatment, and their relative root length showed no significant difference ( Figure 4E,F), indicating that cpi1-1 roots respond normally to the auxin signaling inhibitor. These data suggest that cpi1-1 mutation may not affect the auxin signaling.

Cycloeucalenol Application Increases the Expression of Auxin Biosynthesis Genes
Previously, Men et al. (2008) demonstrated that sterol composition was markedly altered in the cpi1-1 plants. Cycloeucalenol, a minor sterol component in WT (1.0 µg/g fresh weight (fw)), was detected as a major sterol compound in cpi1-1 (50.3 µg/g fw) [8]. By contrast, sitosterol, the predominant sterol in WT (122.3 µg/g fw), was reduced to trace amounts in cpi1-1 (2.5 µg/g fw). 24-methylcholesterol (campesterol) and stigmasterol, the other two major sterols in WT (22.4 and 10.3 µg/g fw, respectively), were undetectable in cpi1-1 [8]. We wondered whether these alterations of sterol composition caused the upregulation of auxin biosynthesis in cpi1-1. To address this, we performed cycloeucalenol treatment on ProTIR2:GUS, ProYUC8:GUS, ProYUC9:GUS, and DR5:GUS expressing seedlings. Upon cycloeucalenol treatment (1 µM), the GUS staining of ProTIR2:GUS and DR5:GUS was obviously stronger than that of the mock (0.1% acetone (v/v), the solvent for cycloeucalenol) ( Figure 5A-D,I,J). The GUS signals were also slightly enhanced in the ProYUC8:GUS and ProYUC9:GUS lines treated with cycloeucalenol ( Figure 5E-H). To gain further evidence, we performed quantitative GUS activity assay. The results showed that after cycloeucalenol treatment, the GUS activity in ProTIR2:GUS, ProYUC8:GUS, and DR5:GUS seedlings was significantly increased, but no statistical difference was observed in ProYUC9:GUS with or without cycloeucalenol treatment ( Figure 5K). In RT-qPCR analysis, YUC8 and TAA1 genes also showed significant upregulation after cycloeucalenol treatment, whereas YUC9 gene showed a slight, but not significant upregulation upon cycloeucalenol treatment (Supplemental Figure S11A). RT-qPCR was also performed to determine whether cycloeucalenol affects the expression of PAT genes. The results showed that the expression of PIN genes and AUX1/LAX genes was not significantly different between mock and cycloeucalenol treatments (Supplemental Figure S11B). These results indicate that while cycloeucalenol can contribute to upregulation of the expression of auxin biosynthesis genes, it does not affect PIN and AUX1/LAX gene expression.

Sitosterol Application Decreases DR5:GUS and ProYUC8:GUS Expression and Partially Rescues the cpi1-1 Short Root Phenotype
To investigate whether sitosterol affects auxin biosynthesis gene expression, RT-qPCR was performed on 5-day-old WT seedlings grown on MS medium supplemented with or without sitosterol. The results showed that the expression of YUC8 was reduced by sitosterol treatment, whereas the expression of YUC9 and TAA1 was not significantly different between mock and sitosterol treatments (Supplemental Figure S12A). To gain further evidence, we compared the expression of ProTIR2:GUS, ProYUC8:GUS, and ProYUC9:GUS between mock and sitosterol treatments. Consistent with the RT-qPCR results, after sitosterol treatment the GUS staining and GUS activity of ProYUC8:GUS was obviously reduced ( Figure 6A,C), whereas ProTIR2:GUS and ProYUC9:GUS displayed no obvious difference between mock and sitosterol treatments (Supplemental Figure S12B). Consistently, sitosterol repressed expression of DR5:GUS in both WT and cpi1-1 roots ( Figure 6B,D). We then investigated whether sitosterol was involved in regulating auxin transport by examining the transcript levels of auxin transporters after sitosterol treatment. The results revealed that, except for PIN7, the expression of none of the other PIN and AUX/LAX genes examined was altered after sitosterol treatment (Supplemental Figure S12C). We then investigated whether the root length of cpi1-1 was influenced by sitosterol. Application of sitosterol slightly repressed the root growth of WT, but partially rescued the short root phenotype of cpi1-1 ( Figure 6E,F).  To determine whether stigmasterol and cholesterol affect the root growth of cpi1-1, we measured the root length of 7-day-old WT and cpi1-1 seedlings grown on MS medium supplemented with or without stigmasterol or cholesterol. Stigmasterol repressed the root elongation of both WT and cpi1-1, while cholesterol showed no detectable influence on root growth (Supplemental Figure S13).
Taken together, our results suggest that cycloeucalenol and sitosterol have opposite effects on auxin biosynthesis and activity (Figure 7). In the plant sterol biosynthesis pathway, CPI1 catalyzes the conversion of cycloeucalenol into obtusifoliol, which is further metabolized to sitosterol. Cycloeucalenol upregulates the expression of auxin biosynthesis genes; by contrast, sitosterol represses the expression of these genes. Normal content of cycloeucalenol and sitosterol is important for optimal auxin biosynthesis, and correct plasma membrane sterol composition is required for normal polar auxin transport. Auxin biosynthesis and polar auxin transport cooperatively establish an optimal auxin gradient in the root tip to regulate root growth.

Auxin Activity and Auxin Biosynthesis are Increased in the cpi1-1 Mutant
The cpi1-1 mutant displayed enhanced and ectopic DR5:GUS expression in the root tip [8], suggesting that either auxin signaling is enhanced in the cpi1-1 mutant or its root tip accumulates more auxin. We found that the cpi1-1 mutant responds normally to the auxin signaling inhibitor, so it is unlikely that auxin signaling is enhanced in cpi1-1. The DR5:GUS expression patterns in cpi1-1 root tip resembled that of pin2-T mutant and polar PIN2 localization was disrupted in cpi1-1 root tip epidermis and cortex cells [8], suggesting that defective PAT caused auxin accumulation in the cpi1-1 root tip. However, our genetic experiments showed that DR5:GUS signals in cpi1-1 pin2-T double mutant root tips was stronger than either cpi1-1 or pin2-T single mutant, indicating an additive effect. This finding suggests that defective PAT alone cannot explain the increased levels of auxin in the roots of cpi1-1 mutants. Indeed, we found that expression of auxin biosynthesis genes including TAA1, YUC8, and YUC9 was upregulated in the cpi1-1 mutant. These results suggest that altered sterol composition in the cpi1-1 mutant not only affects PAT but also affects auxin biosynthesis.
A recent report showed that in rice, low membrane sterol levels compromised the interaction between the ethylene receptor OsERS2 and the inhibitory kinase OsCTR2 in the ER membrane, thus activates the ethylene response [50]. It has been known that ethylene upregulates auxin biosynthesis [51][52][53][54]. Thus, reduced membrane sterol content might upregulate auxin biosynthesis through activation of ethylene signaling. The cpi1-1 mutant displays an almost complete replacement of the usual ∆ 5 -sterols including sitosterol, campesterol, and stigmasterol by 9β,19-cyclopropyl sterols, but its total sterol content is not reduced compared to that of the WT (285 µg/g fw in cpi1-1 versus 190 µg/g fw in WT) [8]. The next step should analyze whether ethylene signaling is affected in the cpi1-1 mutant to determine if altered sterol composition also has an effect on ethylene signaling.

Cycloeucalenol Might Play a Role in Plant Development
In vertebrates and fungi, lanosterol synthase (LAS) cyclizes 2,3-oxidosqualene to the tetracyclic lanosterol, which is further metabolized to tetracyclic cholesterol and ergosterol, respectively, whereas in plants, 2,3-oxidosqualene is mainly cyclized to the pentacyclic cycloartenol (which contains a 9β,19-cyclopropane ring) by cycloartenol synthase (CAS) [3,55,56]. After C-24 methylation and C-4 demethylation, cycloartenol is converted to cycloeucalenol [3]. Then, CPI1 catalyzes the opening of the 9β,19-cyclopropane ring of cycloeucalenol, and converts it to tetracyclic obtusifoliol [29], which is further metabolized to tetracyclic end-products such as sitosterol and stigmasterol (Supplemental Figure S1). Why plants use this two-enzyme route instead of cyclizing directly to tetracyclic sterols and whether those 9β,19-cyclopropyl sterols have physiological roles are long-standing questions that remain to be answered. Several 9β,19-cyclopropyl sterols, including cycloartenol, 24-methylene cycloartanol, and cycloeucalenol, are indeed present as minor sterol compounds in plants [2,[6][7][8][9][10]57]. During male gametophyte development, the tetrads have similar sterol species as vegetative organs, but mature pollen cells accumulate esterified 9β,19-cyclopropyl sterols [58,59]. CAS1 loss of function leads to male-specific transmission defect [55]. These results suggest that 9β,19-cyclopropyl sterols might have a role in male gametophyte development and transmission. Interestingly, cycloeucalenol was found to be the major newly synthesized sterol during pollen cells germination and its accumulation increased with pollen tube elongation [58,59]. Furthermore, inhibition of cycloeucalenol synthesis inhibited pollen tube elongation [58]. Mutants upstream of CPI1 such as hmg1 also showed defects in pollen cells germination and pollen tube elongation [60]. These findings strongly suggest that cycloeucalenol is required for pollen cells germination and pollen tube growth. How cycloeucalenol contributes to pollen cells germination and pollen tube elongation is so far unknown. Like pollen tubes, root hairs also grow in a polarized fashion. Whether cycloeucalenol is also involved in root hair growth is unknown. Recently, it has been reported that induced accumulation of a 9β,19-cyclopropyl sterol, 4-carboxy-4methyl-24-methylenecycloartanol (CMMC), by loss-of-function of ERGOSTEROL BIOSYNTHETIC PROTEIN28 (ERG28) led to phenotypes resemble of those caused by defective PAT, including inhibited root growth, pin-like inflorescence, and fused leaves [10]. Furthermore, exogenous application of CMMC inhibited PAT [10]. Since CMMC was not detectable in WT at normal growth conditions, further investigations are required to determine at what circumstances it is released to regulate PAT. In this study, we found that increased content of cycloeucalenol in cpi1-1 was associated with enhanced auxin biosynthesis and response. Furthermore, in vitro application of cycloeucalenol upregulated the expression of auxin responsive reporter DR5:GUS and auxin biosynthesis genes such as TAA1/TIR2 and YUC8. These results imply that cycloeucalenol might play a role in plant development and growth by regulating auxin biosynthesis. Further investigations are required to understand how cycloeucalenol upregulates auxin biosynthesis.

Sitosterol Modulates Auxin Biosynthesis and Response
Previous studies have shown that a number of sterol mutants exhibit aberrant auxin response and polar auxin transport due to altered sterol composition [8,9,19,21,22,25,27]. However, whether a certain sterol component has a function in regulating auxin activity is largely unknown. Sitosterol, campesterol, and stigmasterol are the most abundant endproducts of the plant sterol biosynthesis pathway [2][3][4][5]. It has been found that in vitro application of sitosterol and stigmasterol enhanced the expression of genes involved in cell expansion and cell division [61]. Exogenous application of sitosterol and stigmasterol but not campesterol partially rescued the abnormal auxin-induced lateral root development in the smt2 smt3 double mutant [26]. Further research demonstrated that application of sitosterol restored the DR5:GUS signal gradient in the lateral root tips of the smt2 smt3 mutant, suggesting that directional auxin transport was amended by sitosterol [26]. These results were consistent with the altered sterol profiles in the smt2 smt3 mutant, as in this mutant the sitosterol and stigmasterol content were reduced to trace amount, whereas the campesterol content was increased [9]. Similarly, in cpi1-1, sitosterol content was reduced to trace amount, whereas stigmasterol and campesterol were undetectable [8]. We found that the aberrant sterol profile in cpi1-1 was associated with increased auxin biosynthesis in addition to defective PAT. Further analysis showed that sitosterol application reduced the expression of auxin biosynthesis genes and an auxin responsive reporter and partially rescued the short root phenotype of cpi1-1. These results suggest that sitosterol functions in regulating auxin activity by inhibiting auxin biosynthesis. How sitosterol regulates the expression of auxin biosynthesis genes remains to be determined. In animals, cholesterol stimulates the autoprocessing of the precursor protein of the Hedgehog (Hh) morphogen important for embryo development. Cholesterol covalently attaches to the N-terminus of the mature Hh protein, which is essential for the long-range activity of Hh [62,63]. Recently, Xiao et al. (2017) found that the Hh signaling pathway downstream protein Smoothened (SMO) was also covalently modified by cholesterol, which activates SMO [64]. In plants, no sitosterol-binding protein has been identified so far. Interestingly, a recent study found that the early gravistimulation induced ROSY1 protein contains a MD2-lipid binding domain and specifically binds stigmasterol in vitro [65]. Mutation of ROSY1 leads to decreased basipetal auxin transport and enhanced gravitropic response [65]. Screens for sitosterolbinding proteins should provide insights for understanding the roles of sitosterol in plant growth and development.
Overall, our findings suggest that cycloeucalenol upregulates, whereas sitosterol downregulates auxin biosynthesis. The altered sterol composition in cpi1-1 leads to enhanced auxin biosynthesis and defective polar auxin transport, resulting in an enhanced, ectopic auxin gradient in the cpi1-1 root tip accompanied by altered root growth. Hence, our findings provide an initial mechanistic basis for future studies on the molecular mechanisms by which sterols influence auxin biosynthesis in plants.

GUS Staining
GUS expression of various genotypes were detected by incubating 5-day-old seedlings of various plant lines in a staining solution (0.5 mg mL −1 5-bromo-4-chloro-3-indolylβ-D-glucuronic acid, 50 mM sodium phosphate, 0.5 mM ferricyanide, and 0.1% (v/v) Triton X-100, pH 7.0) at 37 • C for various hours (indicated below). After incubation, the samples were first fixed in ethanol:acetic acid (2:1) for 6 h, then were cleared in a clearing solution of 8:3:1 (w/v/v) chloral hydrate:distilled water:glycerol. The samples were observed using an Olympus BX63 (Olympus Corporation, Tokyo, Japan) microscope. The incubation hours for different transgenic lines in the GUS staining solution are as follows: 12 h (WT and mutants in Figures 1-3

Fluorescence Microscopy
GFP fluorescence was observed either using a fluorescent microscope (Olympus BX63, Olympus Corporation, Tokyo, Japan) or by a confocal laser scanning microscope (Leica TCS SP5, Wetzlar, Germany) using 488 nm excitation filter with 505-550 nm band pass. Images were processed with Adobe Photoshop CS5 and assembled in Adobe Illustrator CS4.

RNA Extraction and Reverse Transcription Quantitative PCR (RT-qPCR)
Total RNA was extracted from 7-day-old seedlings by TRIzol according to the manufacturer's instructions (TransGen Biotech, Beijing, China). For RT-qPCR analysis, approximately 2.5 µg of total RNA was reverse-transcribed into cDNA using the EasyScript First-Strand cDNA Synthesis SuperMix (TransGen Biotech, Beijing, China). One microliter of each cDNA sample was mixed with 7.5 µl SYBR Green Real-Time PCR Master Mix (DBI Bioscience, Shanghai, China), and then analyzed on a fluorescent quantitative PCR machine (Eppendorf, Hamburg, Germany). The TAP42 INTERACTING PROTEIN OF 41 KDA (TIP41, AT4G34270) gene was used as an internal control. The relative transcription level was calculated by the 2 −∆∆Ct (Ct, cycle threshold) value. Primers used in the RT-qPCR analysis are listed in Supplemental Table S1.

Sterol Treatment
For cycloeucalenol and sitosterol treatments, ProTIR2:GUS, ProYUC8:GUS, ProYUC9:GUS, and DR5:GUS seeds were germinated and grown for 5 days on cycloeucalenol-or sitosterolcontaining MS medium (pH 5.8) solidified by 0.8% (w/v) agar. Seedlings were then collected for GUS staining. The concentrations of sterols used were according to previous publications [26,60]. Cycloeucalenol (purchased from BioBioPha, Kunming, China) was dissolved in absolute acetone (Sangon Biotech, Shanghai, China) and added to the MS medium at 1 µM final concentration. The same MS medium supplemented with 0.1% (v/v) acetone was used as control. Sitosterol (purchased from TCI Shanghai, Shanghai, China) was dissolved in absolute chloroform (Sangon Biotech, Shanghai, China) and added to the MS medium at 1.5 or 3 µg mL −1 final concentrations. The same MS medium supplemented with 0.1% (v/v) chloroform was used as control. For stigmasterol and cholesterol treatments, WT and cpi1-1 seeds were germinated and grown for 7 days on stigmasterolor cholesterol-containing MS medium (pH 5.8) solidified by 0.8% (w/v) agar. Stigmas-terol (purchased from TCI Shanghai, Shanghai, China) and cholesterol (purchased from Sigma-Aldrich, Shanghai, China) were dissolved in absolute chloroform and added to the MS medium at 1 or 10 µM final concentrations, respectively. The same MS medium supplemented with 0.1% (v/v) chloroform was used as control.

Statistical Analyses
Three independent repetitions were conducted for all experiments. Statistical analyses were performed by using one-tailed Student's t-test or using one-way analysis of variance (ANOVA) with Tukey's test. All values are presented as means ± SD. Significant differences are noted as follows: * p < 0.05, ** p < 0.01, and *** p < 0.001.

Data Availability Statement:
The data presented in this study are available in article and supplementary material.