Arabidopsis SMALL AUXIN UP RNA63 promotes hypocotyl and stamen filament elongation




Auxin regulates plant growth and development in part by activating gene expression. Arabidopsis thaliana SMALL AUXIN UP RNAs (SAURs) are a family of early auxin-responsive genes with unknown functionality. Here, we show that transgenic plant lines expressing artificial microRNA constructs (aMIR-SAUR-A or -B) that target a SAUR subfamily (SAUR61–SAUR68 and SAUR75) had slightly reduced hypocotyl and stamen filament elongation. In contrast, transgenic plants expressing SAUR63:GFP or SAUR63:GUS fusions had long hypocotyls, petals and stamen filaments, suggesting that these protein fusions caused a gain of function. SAUR63:GFP and SAUR63:GUS seedlings also accumulated a higher level of basipetally transported auxin in the hypocotyl than did wild-type seedlings, and had wavy hypocotyls and twisted inflorescence stems. Mutations in auxin efflux carriers could partially suppress some SAUR63:GUS phenotypes. In contrast, SAUR63:HA plants had wild-type elongation and auxin transport. SAUR63:GFP protein had a longer half-life than SAUR63:HA. Fluorescence imaging and microsomal fractionation studies revealed that SAUR63:GFP was localized mainly in the plasma membrane, whereas SAUR63:HA was present in both soluble and membrane fractions. Low light conditions increased SAUR63:HA protein turnover rate. These results indicate that membrane-associated Arabidopsis SAUR63 promotes auxin-stimulated organ elongation.


Stems, leaves and other plant organs typically form by cell division in organ primordia, which then enlarge by extensive cell expansion to reach their final size. Several hormones regulate expansion growth, including auxin, brassinosteroids, ethylene and gibberellins (Depuydt and Hardtke, 2011). Environmental signals, such as temperature and light, can affect hormone abundance and responsiveness, and the extent of growth depends on integration of hormonal and environmental signals (Wolters and Jurgens, 2009; Depuydt and Hardtke, 2011).

Auxin regulates plant development in part by changing gene expression (reviewed in Mockaitis and Estelle, 2008; Quint and Gray, 2006; Woodward and Bartel, 2005). Auxin response transcription factors (ARFs) bind to short promoter elements, and activate genes in the presence of auxin. In the absence of auxin, ARF proteins instead dimerize with Aux/indole-3-acetic acid (IAA) proteins, which inhibit transcription. Auxin causes Aux/IAA protein turnover by promoting interaction with TIR1/AFB F-box subunits of SCF ubiquitin ligases. Phenotypes of mutants affecting this pathway indicate that auxin promotes hypocotyl, inflorescence stem and flower organ elongation by changing gene expression.

Primary auxin response genes include members of three multigene families: IAA, GH3 and SAUR (Hagen and Guilfoyle, 2002). IAA genes encode Aux/IAA proteins, and auxin-inducible GH3 genes encode IAA-amido synthetases that convert the active auxin IAA to an inactive conjugated form (Staswick et al., 2005). Thus, these two auxin-inducible gene classes mediate negative feedback on auxin response.

The SAUR genes were first isolated as primary auxin-inducible genes, but the biochemical and developmental functions of SAUR proteins are largely unknown (Hagen and Guilfoyle, 2002). Arabidopsis has over 70 SAUR genes, Oryza sativa (rice) has 56 and the moss Physcomitrella patens has 18 (Hagen and Guilfoyle, 2002; Jain et al., 2006; Rensing et al., 2008). A Zea mays (maize) SAUR2:GUS fusion protein had a half-life of just 7 min (Knauss et al., 2003), suggesting that it may mediate transient responses to auxin. Some maize and Arabidopsis SAUR proteins can bind to calmodulin, suggesting that they might mediate responses to changes in intracellular calcium level (Yang and Poovaiah, 2000; Reddy et al., 2002; Knauss et al., 2003). However, the N-terminal residues required for this binding are not highly conserved among SAUR proteins (Yang and Poovaiah, 2000). Several protein fusions to Arabidopsis, rice or maize SAUR proteins localized to the nucleus and/or the cytoplasm of transgenic plants or transfected cells (Knauss et al., 2003; Park et al., 2007; Kant et al., 2009).

Several Arabidopsis and Glycine max (soybean) auxin-inducible SAUR genes are expressed in growing hypocotyls or other elongating tissues (McClure and Guilfoyle, 1989; Gee et al., 1991; Li et al., 1991; Gil and Green, 1997). In contrast, the SAUR32/AAM1 gene was not auxin induced, and was expressed in the epidermis of the concave side of the apical hook in dark-grown seedlings, where growth is slow (Park et al., 2007; Paponov et al., 2008). A T-DNA insertion in SAUR32/AAM1 caused no obvious phenotype, suggesting redundant function with other SAUR genes, but when overexpressed, SAUR32 caused apical hook opening and a short hypocotyl (Park et al., 2007). Similarly, OsSAUR39 overexpression in transgenic rice plants decreased shoot growth and auxin transport, and exogenous auxin could rescue shoot growth phenotypes of these plants (Kant et al., 2009).

The auxin level in growing cells is determined by auxin biosynthesis and breakdown, and by auxin transport into and out of cells by influx (AUX1 and LAX1–LAX3) and efflux (PIN1–PIN7, ABCB1/PGP1 and ABCB19/MDR1/PGP19) transporters (Zazimalova et al., 2010). Regulation of auxin synthesis and/or transport mediates some environmental growth responses such as shade avoidance and asymmetric growth during phototropism and gravitropism (Stepanova et al., 2008; Tao et al., 2008; Zazimalova et al., 2010).

We have studied developmental functions of Arabidopsis SAUR63, a member of a subclass of nine closely related SAUR genes. Using plants expressing silencing constructs or SAUR63:reporter fusion proteins, we have characterized phenotypes caused by decreasing or increasing SAUR63 activity, and determined the tissue-specific and subcellular localization of SAUR63 protein expression.


SAUR genes are expressed in growing shoot tissues

Members of a clade of nine SAUR genes (SAUR6168 and SAUR75) were induced by auxin in seedlings or hypocotyls (Paponov et al., 2008; E. Chapman and M. Estelle, personal communication), and were underexpressed in arf6 arf8 double mutant flowers (Nagpal et al., 2005; Reeves et al., 2012). These expression characteristics suggest that these SAUR genes may be downstream targets of the auxin response factors ARF6 and ARF8, which regulate hypocotyl, inflorescence stem and flower organ elongation (Nagpal et al., 2005). SAUR61–SAUR68 genes are situated in a cluster on chromosome 1, together with three unrelated genes, whereas SAUR75 is on chromosome 5. SAUR68 encodes a shorter open reading frame than the others, and might be a pseudogene. These nine SAUR genes each share high similarity in their 5′ upstream regions, including several putative auxin response elements (AuxREs) that could potentially mediate ARF binding (Figure S1). The genes have over 80% pairwise cDNA sequence identity and the predicted proteins have over 69% amino acid sequence identity (Figure S2), suggestive of substantial functional redundancy. Indeed, several available T-DNA insertion lines with mutations in these genes appeared very similar to wild-type plants.

Several of these genes were previously found to be regulated in seedlings by the circadian rhythm (Covington and Harmer, 2007). Using a probe predicted to cross-hybridize with SAUR62, SAUR63, SAUR67 and SAUR75, we found that SAUR genes were also diurnally expressed in flowers, with the highest level expressed in the morning (Figure S3). To determine tissue-specific expression patterns of these genes, we generated ProSAURx:SAURx:GUS lines for SAUR61-68 using approximately 500 bp of 5′ upstream sequence and the coding sequence of each SAUR gene fused to GUS (abbreviated as SAURx:GUS). Preliminary analyses of multiple T1 plants for each construct indicated that SAUR62:GUS and SAUR63:GUS lines had the strongest X-Gluc staining levels, and had similar phenotypes (Figures 1, 2 and S3), and we chose SAUR63 for detailed study.

Figure 1.

 X-Gluc staining of SAUR63:GUS seedlings and flowers.
(a) Seedling grown for 5 days in short days. Scale bar: 2 mm.
(b) Seedling grown for 3 days in complete darkness. Scale bar: 2 mm.
(c) Root tip. Scale bar: 200 μm.
(d) Inflorescence apex and mature flowers of 5-week-old plants. Scale bar: 5 mm.
(e–g) Flowers at stages 12 (e), 13 (f) and 14 (g). Scale bars: 1 mm.
(h) Close-up of stamen filament. Scale bar: 100 μm.
(i–k) Cauline leaves and lateral branches at the nodes of the primary inflorescence stem (i) or on branches (j, k). Scale bars: 2 mm.

Figure 2.

 Hypocotyl and inflorescence stem growth of SAUR63 transgenic plants.
(a) DNA constructs for silencing SAUR61SAUR68 and SAUR75 (see also Figure S4a), and protein fusions to the C terminus of SAUR63.
(b) Seedlings of the indicated genotypes grown for 5 days in short days. Scale bars: 1 mm.
(c–e) Seedlings of the indicated genotypes grown in short days for 5 days on MS plates containing 10 μm IAA or 10 μm NPA.
(c) Average hypocotyl lengths. Data are shown as means ± SDs (n = 20). Letters indicate results of anova for different genotypes and treatments.
(d) Wild-type and SAUR63:GUS seedlings grown with exogenous IAA or NPA. H indicates the linear height of the hypocotyl (white bidirectional arrow). L indicates the contour length of the hypocotyl (black bidirectional arrow).
(e) Hypocotyl waviness index. The waviness index was calculated for each seedling as the percentage of shortening of the hypocotyl caused by wavy growth [= 100% × (L − H)/L]. Data are shown as means ± SDs (n = 20). Letters indicate results of anova for different genotypes and treatments.
(f) Inflorescence stems of 5-week-old plants of indicated genotypes.

SAUR63:GUS was expressed in hypocotyls, cotyledons, petioles and young rosette leaves of seedlings grown in day–night cycles (Figures 1a,b and S5a), and in hypocotyls of dark-grown seedlings (Figure 1b). No SAUR63:GUS activity was detected in roots (Figure 1c). In mature flowering plants, SAUR63:GUS was expressed in the apical portion of the inflorescence stem and in flowers (Figure 1d). In the oldest unopened flower buds and in recently opened flowers (stage 12–14 flowers, according to Smyth et al., 1990), expression was found at the junction of the pistil and the pedicel, and in stamen filaments and petals (Figure 1e–g). In hypocotyls, cotyledons, petals and stamen filaments, X-Gluc staining was strongest in vasculature, but could also be seen in other cell layers (Figure 1a,b,e–h). In contrast to expression in young leaves and the inflorescence apex, no SAUR63:GUS activity was detected in emerging lateral inflorescence stems, or in the cauline leaves subtending them (Figure 1i–k). Thus, diurnal and spatial expression patterns of SAUR63 correlated with elongation growth in hypocotyls, cotyledons, primary inflorescence stems, petals and stamen filaments.

SAUR proteins stimulate organ elongation

Because of the apparent redundancy among SAUR genes in this clade, to determine their functions we generated multiple transgenic plant lines overexpressing either of two artificial microRNA constructs designed to target all nine genes (aMIR-SAUR-A or aMIR-SAUR-B, Figures 2a and S4). A subset of lines expressing either construct had slightly shorter hypocotyls and stamen filaments than did wild-type plants (Figures 2b,c, 3e–g, S4e and S5b,d). The short stamen filaments of affected amiR-SAUR plants caused inefficient self-pollination, resulting in small siliques with only a few seeds (Figure 2f). Analyses of representative lines by RNA hybridization blots or RT-PCR assays revealed at best only modest decreases of SAUR gene expression, suggesting that the artificial microRNA constructs were partially effective (Appendix S1; Figures S4c,d and S5b,d,e). An aMIR-SAUR-A line in the SAUR63:GUS background also only partially decreased X-Gluc staining (Figure S5a). Although these constructs decreased but did not eliminate the function of these SAUR genes, these results suggest that SAUR proteins promote elongation growth in both hypocotyls and stamen filaments.

Figure 3.

 Floral organ elongation of SAUR63 transgenic plants.
(a–f) Mature open flowers of SAUR63 transgenic lines at floral stage 14. Orange arrowheads indicate stamens with elongated filaments in SAUR63:GFP and SAUR63:GUS flowers. Black arrowheads indicate stamens with short filaments in aMIR-SAUR-A and aMIR-SAUR-B flowers. Scale bars: 1 mm.
(g) Average organ lengths of SAUR63 transgenic flowers at floral stage 14. Data are shown as means ± SDs (n = 7–10); *P < 1 × 10−3 in a Student’s t-test, compared with the wild type.
(h, i) Developmental comparison of stamen (h) and petal (i) elongation relative to gynoecium length of wild-type (orange diamonds), SAUR63:GUS (blue squares) and SAUR63:GFP (green triangles) flowers. Stage 12 flowers are the last closed buds before flower opening, and stage 13 flowers are the first open flowers (Smyth et al., 1990).

In contrast to the phenotypes of aMIR-SAUR plants, SAUR63:GUS plants, as well as SAUR63:GFP plants expressing a SAUR63:GFP fusion protein behind the PSAUR63 promoter, had phenotypes consistent with increased SAUR63 activity. These plants had longer hypocotyls, stamen filaments and petals than wild-type plants (Figures 2b,c and 3c,d,g–i, Figure S5d). In SAUR63:GUS and SAUR63:GFP flowers, stamen filaments and petals began to elongate rapidly at stage 12, just before flower opening, when wild-type stamens and petals also elongate rapidly. However, the transgenic flower organs elongated at a faster rate at stages 12–13, and still elongated somewhat at stage 14, when wild-type flower organs had ceased to elongate (Figure 3h,i). In each of these elongating organs, SAUR63:GUS plants had more severe phenotypes than SAUR63:GFP plants.

In addition, SAUR63:GUS and SAUR63:GFP hypocotyls and inflorescence stems grew less straight than wild-type hypocotyls and stems (Figure 2d–f). SAUR63:GUS and SAUR63:GFP seedlings had a higher hypocotyl waviness index, calculated from the difference between the contour length and the height of the hypocotyl, than wild-type seedlings (although these differences were only statistically significant in the presence of exogenous IAA; Figure 2d,e). Analogously, growing inflorescence stems of SAUR63:GUS and SAUR63:GFP plants grew in a corkscrew-like pattern rather than straight (Figure 2f). Circadian expression of the PSAUR63 promoter driving these transgenes may account for this periodic inflorescence growth pattern (Covington and Harmer, 2007).

The opposite hypocotyl and stamen filament elongation phenotypes of amiR-SAUR lines and SAUR63:GFP or SAUR63:GUS lines suggest that the SAUR63:GUS and SAUR63:GFP fusion proteins have a hypermorphic gain of function (i.e. increase a normal SAUR63 activity), reinforcing the conclusion that SAUR63 promotes elongation of multiple organs.

In contrast to the SAUR63:GUS and SAUR63:GFP lines, plants expressing a PSAUR63:SAUR63:HA construct had normal hypocotyl length, hypocotyl waviness, inflorescence stem growth and flower organ lengths (Figures 2b,c,e,f and 3b,g). These results suggest that the SAUR63:HA fusion protein lacks the hypermorphic activity shown by the SAUR63:GUS and SAUR63:GFP proteins, perhaps because the HA tag is smaller than GUS or GFP.

SAUR63:GUS and SAUR63:GFP affect auxin response and transport

To explore whether SAUR63 affected hypocotyl auxin response, we grew seedlings of aMIR-SAUR lines and SAUR63 protein fusion transgenic lines on MS-agar plates containing the auxins IAA or 2,4-dichlorophenoxyacetic acid (2,4-D), or the polar auxin transport inhibitor N-1-naphthylphthalamic acid (NPA; Figures 2c–e, S6 and S7). Exogenous IAA did not affect hypocotyl lengths or waviness in wild-type, SAUR63:HA or amiR-SAUR seedlings. However, IAA did increase both elongation and waviness of SAUR63:GFP and SAUR63:GUS seedling hypocotyls (Figures 2c–e, S6 and S7). Thus, increased SAUR63 activity caused increased responsiveness to exogenous auxin. Growth at higher temperature (28°C), which increases endogenous auxin level (Gray et al., 1998), increased hypocotyl elongation in all genotypes, and increased hypocotyl waviness in wild-type, aMIR-SAUR-A and SAUR63:GFP seedlings (Figure S7). The shorter hypocotyls of aMIR-SAUR-A relative to wild-type seedlings was more obvious at 28°C than at 22°C.

The synthetic auxin 2,4-D, which is a poor substrate for auxin efflux carriers (Delbarre et al., 1996), had a similar effect on hypocotyl elongation as IAA (Figure S7). However, whereas IAA increased hypocotyl waviness in SAUR63:GUS and SAUR63:GFP seedlings, 2,4-D did not. These results suggest that auxin efflux is required for the increased waviness caused by IAA in SAUR63:GUS and SAUR63:GFP seedlings, but not for stimulation of hypocotyl elongation by exogenous auxin. Consistent with a requirement for auxin efflux for wavy hypocotyl growth, the auxin transport inhibitor NPA decreased hypocotyl elongation and waviness of all genotypes (Figure 2d,e). NPA also caused seedlings to grow agravitropically (Figure S6).

These physiological results suggest that SAUR proteins might affect auxin transport or accumulation. Indeed, after application of [3H]IAA to the hypocotyl apex, SAUR63:GFP and SAUR63:GUS seedlings accumulated more [3H]IAA in a basal hypocotyl segment than did wild-type seedlings (Figure 4a). SAUR63:HA and aMIR-SAUR lines accumulated similar levels of [3H]IAA as wild-type seedlings (Figure 4a). NPA suppressed the increased basipetal [3H]IAA accumulation of SAUR63:GUS and SAUR63:GFP hypocotyls (Figure 4a).

Figure 4.

 Auxin flux in hypocotyls, and interactions between SAUR63:GUS and mutations affecting auxin efflux transporters.
(a) Basipetal [3H]IAA flux in hypocotyls of SAUR63 transgenic seedlings, with (black bars) or without (white bars) 10 μm NPA. Data are shown as means ± SDs (n = 20–40); *P < 0.05 in a Student’s t-test, compared with the wild type. Two additional replicate experiments gave similar results.
(b, c) Average hypocotyl lengths (b) and waviness index (c) of seedlings of the indicated genotypes. Data are shown as means ± SDs (n = 20, except for pin1-9 SAUR63:GUS, for which n = 5). Letters indicate results of anova for different genotypes.
(d, e) Mature stage 14 flowers of seedlings of the indicated genotypes. The arrowhead indicates stamens with short filaments, asterisks indicate stamens with filaments longer than the pistil and a circle indicates normalized stamen elongation in the pin3-4 pin7-2 SAUR63:GUS flower. Scale bars: 1 mm.

To explore the extent to which SAUR63:GUS and SAUR63:GFP phenotypes depend on particular auxin transporters, we assessed interactions between mutations affecting ABCB or PIN auxin efflux transporters and SAUR63:GUS. pgp1 pgp19 double mutant plants deficient in ABCB1/PGP1 and ABCB19/PGP19/MDR1 had short hypocotyls and reduced fecundity because of short stamen filaments, and mdr1-3 single mutant seedlings also had shorter hypocotyls than wild-type seedlings (Figure 4b,e, Figure S9; Noh et al., 2003, 2001). mdr1-3 SAUR63:GUS and pgp1 pgp19 SAUR63:GUS seedlings had hypocotyl lengths intermediate between those of abcb mutant and SAUR63:GUS seedlings (Figure 4b). pgp1 pgp19 SAUR63:GUS hypocotyls appeared slightly less wavy than SAUR63:GUS hypocotyls, although this difference was not statistically significant (Figure 4c). pgp1 pgp19 SAUR63:GUS plants were dwarfed like pgp1 pgp19 plants, but had longer stamen filaments than pgp1 pgp19 plants and were fully fecund, without requiring manual pollination (Figures 4e and S9).

Among well-characterized PIN genes, PIN1, PIN3 and PIN7 are expressed in hypocotyls, stamen filaments and/or inflorescence stems (Noh et al., 2003; Zadnikova et al., 2010; Christie et al., 2011; Ding et al., 2011; Gallego-Bartolome et al., 2011). pin1–9 (SALK_097144) SAUR63:GUS plants had elongated twisty hypocotyls and twisty (rather than straight) pin-formed inflorescences, suggesting that the pin1-9 mutation and the SAUR63:GUS transgene interacted additively (Figures 4b,c, and S8b). Similarly, pin3-4 (SALK_038609) and pin7-2 (SALK_044687) mutations each had only small effects on the hypocotyl elongation, twisted inflorescence or elongated stamen filament phenotypes conferred by SAUR63:GUS (Figures 4b,c,d, S8b and S9). X-Gluc staining revealed that the SAUR63:GUS transgene was expressed in these pin and pgp mutant backgrounds (Figures S8a and S9a).

Auxin and light conditions affect SAUR63:HA protein abundance

Attempts to recognize native SAUR63 protein with polyclonal anti-SAUR antibodies were unsuccessful. Instead, we used a SAUR63:HA line and anti-HA antibodies to characterize SAUR63 protein stability and localization. The absence of gain-of-function phenotypes in SAUR63:HA lines suggests that the SAUR63:HA protein may mimic the wild-type protein more closely than SAUR63:GUS or SAUR63:GFP proteins do.

We evaluated SAUR63:HA protein levels in seedlings after treatment with 10 μm IAA, 30 μm cycloheximide (CHX, a translation inhibitor) or 10 μm MG132 (a proteasome inhibitor) for 1 h (Figure 5a). In the absence of exogenous IAA, SAUR63:HA protein disappeared after CHX treatment. IAA treatment increased the SAUR63:HA protein level, and CHX largely blocked this increase, indicating that induction of SAUR63 transcription by auxin has a large impact on SAUR63 protein levels. Consistent with the circadian regulation of SAUR63 transcription, in seedlings grown in long days, SAUR63:HA protein level increased after dawn, peaked at about 8 h after dawn and decreased to a baseline level at about 12 h after dawn (Figure 5b).

Figure 5.

 SAUR63:HA and SAUR63:GFP protein levels and turnover.
(a) Effects of 30 μm cycloheximide (CHX), 10 μm MG132 and 10 μm IAA on SAUR63:HA protein levels.
(b) Diurnal abundance of SAUR63:HA or SAUR63:GFP proteins. HAD indicates hours after dawn.
(c) Light dependence of protein turnover for SAUR63:HA or SAUR63:GFP. BL and DL indicate bright (>100 μmol m−2 s−2) and dim (<10 μmol m−2 s−2) light growth conditions.
(d, e) Protein half-life for SAUR63:HA and SAUR63:GFP in bright (d) and dim (e) light. Numbers indicate SAUR63 levels normalized to L-ascorbate peroxidase (APX) levels.
(f) mRNA levels for SAUR + SAUR63:HA, SAUR63:GFP and SAUR in bright light (BL) and dim light (DL). Numbers indicate relative SAUR levels normalized to ACT2 levels.

Treatment with the proteasome inhibitor MG132 increased SAUR63:HA level in both the absence and the presence of cycloheximide (Figure 5a), indicating that under the bright light conditions used, SAUR63:HA is turned over rapidly by the proteasome. In the presence of both IAA and CHX, the MG132 effect was not clearly detected, suggesting that auxin may influence SAUR63:HA protein turnover.

While performing these experiments, we observed that ambient light conditions affected SAUR63:HA level. SAUR63:HA level decreased dramatically within 30 min after shifting plants from bright (≥100 μmol m−2 s−1) to dim (≤10 μmol m−2 s−1) light (Figure 5c). Treatment with CHX accelerated SAUR63:HA disappearance, whereas treatment with 10 μm MG132 delayed it. To determine the half-life of SAUR63:HA protein under both light conditions, we pre-treated seedlings with CHX for 10 min and examined protein levels over the following 2 h. In bright light, SAUR63:HA protein level decreased to about half of its starting level in about 15 min (Figure 5d). After the shift to dim light, the half-life was less than 5 min (Figure 5e). Thus, bright light increased the half-life of SAUR63:HA protein. RNA hybridizations using a probe expected to cross-react with multiple members of the clade revealed no difference in SAUR plus SAUR63:HA transcript levels in seedlings under these two light conditions (Figure 5f).

SAUR63:GFP protein has slow turnover

To ascertain the basis for the apparent hypermorphic activity of the SAUR63:GFP transgene, we assessed SAUR63:GFP protein levels using an anti-GFP antibody. In seedlings growing in long days, SAUR63:GFP level peaked at about 10–12 h after dawn, and significant levels of SAUR63:GFP were maintained until 16 h after dawn (Figure 5b). This shift in peak protein level to later in the day suggested that SAUR63:GFP may be more stable than SAUR63:HA. In fact, the shift from bright to dim light had no effect on SAUR63:GFP levels (Figure 5c). In the presence of 30 μm CHX, SAUR63:GFP was stable for at least 2 h under both light conditions (Figure 5d,e). Thus, the half-life of SAUR63:GFP is at least eight times longer than the half-life of SAUR63:HA under the bright light conditions used for hypocotyl growth assays.

SAUR63:GFP protein accumulates stably in the plasma membrane fraction

To obtain insight into how SAUR63 might influence growth, we assessed the intracellular localization of SAUR63:HA and SAUR63:GFP proteins. In hypocotyl epidermis cells, SAUR63:HA immunofluorescence signal was present in a punctate pattern, suggestive of possible vesicle localization (Figure 6a). After microsomal fractionation of extracts from seedlings grown in bright light, SAUR63:HA was present in both soluble and membrane fractions (Figure 6d). After the shift to dim light, the SAUR63:HA protein disappeared quickly from both soluble and membrane fractions (Figure 6d).

Figure 6.

 Localization of SAUR63:HA and SAUR63:GFP proteins.
(a) Immunolocalization of SAUR63:HA in hypocotyl epidermis cells. Scale bars: 20 μm. The wild-type control image illustrates the background green fluorescence pattern.
(b) Confocal microscopy of SAUR63:GFP in epidermis cells of the hypocotyl. FM-4-64 staining was used to stain the plasma membrane. Scale bars: 20 μm.
(c) Co-localization of SAUR63:Cerulean and the plasma membrane marker Wave138 (Geldner et al., 2009) in transiently expressing cotyledon epidermal cells. Scale bars: 10 μm.
(d) Microsomal fractionation of SAUR63:HA or SAUR63:GFP seedling protein extracts at the indicated times after the shift from bright light (BL, >100 μmol m−2 s−2) to dim light (DL, <10 μmol m−2 s−2). T, S and M indicate total protein extracts, soluble fraction and membrane fraction, respectively. Anti-APX and anti-RD28 antibodies were used as soluble and membrane fraction protein controls, respectively.
(e) A model for localization and turnover of Arabidopsis SAUR63 and SAUR63 fusion proteins. Membrane-localized SAUR63 or SAUR63:GUS may affect hormone transport.

Viewed by confocal microscopy, the SAUR63:GFP protein was localized in the plasma membrane of epidermal cells of both hypocotyls and cotyledons (Figures 6b and S10). Similarly, a transiently expressed SAUR63:Cerulean fusion protein co-localized with the plasma membrane marker Wave138Y in cotyledon epidermal cells (Geldner et al., 2009; Figure 6c). In microsomal fractionation experiments, SAUR63:GFP was present almost exclusively in the membrane fraction (Figure 6d). After the shift to dim light, SAUR63:GFP protein persisted at a high level in the membrane fraction for the full time course of 1 h (Figure 6e).


SAUR63 and members of its clade are primary auxin-responsive genes, and were expressed in growing regions of hypocotyls, inflorescence stems, petals and stamen filaments. Moreover, diurnal SAUR63:HA protein levels matched diurnal phases of auxin responsiveness and hypocotyl elongation (Covington and Harmer, 2007; Nozue et al., 2011). Decreased or increased SAUR activity caused correlated changes in growth of hypocotyls and flower organs, and exogenous auxin stimulated hypocotyl growth in SAUR63:GUS and SAUR63:GFP seedlings, but not in wild-type seedlings. These results indicate that these SAUR genes promote auxin-induced growth in multiple elongating tissues.

The larger C-terminal tags of SAUR63:GUS and SAUR63:GFP proteins compared with SAUR63:HA protein likely cause their increased stability. For SAUR63:GFP, and probably also SAUR63:GUS, the C-terminal tag both increased half-life compared with SAUR63:HA protein and caused constitutive plasma membrane localization. These features of SAUR63:GFP may be mechanistically linked. Thus, either a soluble form of the protein may normally be turned over, so that constitutive membrane localization might prevent entry into the turnover pathway, or steric interference with turnover may promote membrane localization. In either case, it seems likely that SAUR63:GFP acts at the membrane (Figure 6e). That SAUR63:HA is present in both membrane and soluble fractions suggests that localization of the native protein may be more dynamic, which might allow regulated activity, localization or turnover (Figure 6e). As SAUR proteins have no obvious membrane localization signal peptide sequences or hydrophobic regions, membrane localization may occur through interaction with an integral membrane protein.

Other SAUR genes have been found to be regulated by destabilization of mRNA through a 3′ untranslated element (Gil and Green, 1996). This element is not obviously present in SAUR63, and the fusion protein transgenes used in our experiments lacked endogenous SAUR63 3′ untranslated sequences. Nevertheless, HA, GFP or GUS sequences might conceivably affect transcript stability differentially.

SAUR63:GUS and SAUR63:GFP hypocotyls accumulated more auxin after basipetal transport from the apex than did wild-type hypocotyls. Moreover, NPA decreased SAUR63:GUS and SAUR63:GFP hypocotyl elongation, and genetic experiments revealed that ABCB1/PGP1 and ABCB19/PGP19/MDR1 auxin transporters contributed to hypocotyl elongation in SAUR63:GUS seedlings. These observations are consistent with the model that SAUR63 might regulate auxin transport, either directly or indirectly. Alternatively, our results may reflect an indirect requirement for auxin transport to enable SAUR63:GUS action, if a threshold level of auxin is required for SAUR63:GUS to stimulate elongation. Further study will be required to distinguish whether SAUR63 regulates auxin transport, or activates some other pathway whose activity is seen only when auxin is present.

twisted dwarf 1 (twd1) mutants also have twisted growth of multiple organs in part because of impaired membrane sorting of ABCB auxin transporters (Geisler et al., 2003; Bouchard et al., 2006; Bailly et al., 2008; Wu et al., 2010). However, the TWD1 immunophilin protein localizes to the endoplasmic reticulum (Wu et al., 2010), suggesting that it acts differently from SAUR63:GFP.

The auxin transport assay used herein measures accumulation of basipetally transported exogenous auxin in a basal hypocotyl segment after a short assay period, and therefore reflects the capacity of the hypocotyl to transport auxin and the likely accumulation of auxin in basal tissues. The increased auxin level observed in SAUR63:GUS and SAUR63:GFP hypocotyls in this assay suggests that hypocotyls in these genotypes may also accumulate higher levels of endogenous auxin than hypocotyls of wild-type seedlings, and this may cause increased elongation. The epidermis can limit elongation, and auxin levels in epidermal cells may thus determine the extent of growth (Kutschera and Niklas, 2007; Savaldi-Goldstein et al., 2007; Christie et al., 2011). SAUR63:GUS and SAUR63:GFP were expressed in epidermal cells, and might act autonomously in those cells. However, SAUR63:GUS was expressed most strongly in central perivascular cells of hypocotyls and stamen filaments, and it is also possible that SAUR63:GUS acts in these central tissues, for example to direct auxin flux peripherally. Increased waviness of hypocotyl growth in SAUR63:GUS and SAUR63:GFP plants might arise from asymmetric auxin accumulation. That 2,4-D did not induce wavy growth in SAUR63:GUS and SAUR63:GFP plants, whereas exogenous IAA did, suggests that auxin efflux is required to induce growth asymmetries. Wavy hypocotyl growth and enhanced phototropism in abcb1 abcb19 mutants have similarly been attributed to increased lateral auxin flow from the center of the hypocotyl (or from more apical cells), because of lower basipetal auxin transport through the center of the hypocotyl (Noh et al., 2003; Christie et al., 2011).

In addition to regulation by the circadian rhythm and auxin, the ambient light level affected SAUR63:HA protein abundance and half-life. Thus, wild-type SAUR63 probably acts primarily in seedlings grown in high light. Light-grown Arabidopsis and tomato seedlings have significantly more basipetal auxin transport in hypocotyls than dark-grown seedlings, and hypocotyl elongation and auxin transport are much more sensitive to NPA in light-grown than in dark-grown seedlings (Jensen et al., 1998; Rashotte et al., 2003; Liu et al., 2011). In roots, light increases PIN2 plasma membrane localization (Laxmi et al., 2008), and it will be interesting to learn whether root-expressed SAUR proteins contribute to this effect.

Similarly to our results, other workers studying a different clade of Arabidopsis SAUR genes have found that an N-terminal GFP fusion increased SAUR19 protein stability, and that plants overexpressing GFP-SAUR19 constructs had long hypocotyls and increased auxin flux in the hypocotyl (Franklin et al., 2011; Spartz et al. 2012). From a practical standpoint, the hypermorphic activity of SAUR protein fusions enables study of their developmental and biochemical functions, despite extensive genetic redundancy. Similarly, the functions of Aux/IAA proteins were determined by biochemical studies and by the identification of gain-of-function mutations in IAA genes that stabilized the corresponding proteins (Gray et al., 2001).

Other Arabidopsis, maize and rice SAUR proteins have been reported to localize to the nucleus and/or cytoplasm (Knauss et al., 2003; Park et al., 2007; Kant et al., 2009). Although nuclear or cytoplasmic localization in some of those studies may have arisen from the use of transiently overexpressed proteins, comparison of those results with ours raises the possibility that SAUR proteins can act in multiple subcellular locations. Moreover, the findings that the rice SAUR39 gene decreased auxin transport activity (Kant et al., 2009), and that the Arabidopsis SAUR32 gene decreased hypocotyl elongation (Park et al., 2007), suggest that different SAUR proteins may have distinct or even opposite activities. SAUR32 falls in a distinct clade from SAUR63, and is not auxin inducible, suggesting that it may in fact have a distinct function (Paponov et al., 2008).

Experimental procedures

Plant growth conditions and phenotypic analyses

For hypocotyl length measurements, Arabidopsis thaliana seedlings were vertically grown on MS agar medium (without sucrose), with or without the indicated concentrations of IAA, 2,4-D or NPA in a growth chamber at 22 or 28°C under an 8-h light/16-h dark photoperiod (100–120 μmol m−2 s−2). At 5 days after germination (5 DAG), seedlings were scanned and lengths of the hypocotyl were measured using ImageJ (Abramoff et al., 2004). For a waviness index, the contour length and the linear height of the hypocotyl were separately measured for each individual seedling (n = 20). The percentage (%) of difference in the two values was calculated as a level of waviness. These data were analyzed statistically by analysis of variance (anova) with jmp pro 9 software (SAS Institute Inc., Cary, NC, USA), using default parameters and a model of Y (contour length or waviness index) = G + ε for Figure 4, or Y = G + T + G*T + ε for Figures 2 and S7 (where G = genotype, T = treatment and ε = error). Results of Tukey’s honestly significant difference (HSD) test are grouped by letters, indicating measurements that were not significantly different at 95% confidence. Mature plants were grown in a glasshouse. For floral organ elongation, consecutive flowers of stages 11–14 (Smyth et al., 1990) from between 5 and 10 apices per genotype were manually dissected, and images of the pistil, stamens, petals, and sepals were scanned and measured using ImageJ (Abramoff et al., 2004).

Generation of SAUR63 transgenic plants

To generate plants with silenced SAURs (61–68 and 75), two artificial microRNAs (aMIR-SAUR-A and aMIR-SAUR-B) were designed to target conserved sequences in the coding regions of SAURs (Figure S4). The MIR167a template was modified by site-directed mutagenesis using PCR primer sets corresponding to the two target sequences (Table S1), and cloned into pB7WG2 (Karimi et al., 2002). The binary plasmid constructs were introduced into Arabidopsis plants using Agrobacterium tumefaciens strain GV3101 by floral dip (Clough and Bent, 1998), and T1 transformed plants were selected on soil by spraying with a solution of 120 mg/l glufosinate ammonium (Southern States, Carrboro, NC, USA). The characterization of these lines is described in Appendix S1.

For C-terminal protein fusions to SAUR63, a genomic DNA fragment containing promoter and SAUR63 coding sequences (from −400 to +450 bp) was amplified using gene-specific primers (Table S1) and cloned into Gateway binary vectors (Nakagawa et al., 2007): pGWB13 for ProSAUR63:SAUR63:HA (SAUR63:HA), pGWB4 for ProSAUR63:SAUR63:GFP (SAUR63:GFP) and pGWB3 for ProSAUR63:SAUR63:GUS (SAUR63:GUS), respectively. T1 plants were selected on MS-agar medium containing 30 μg ml−1 kanamycin.

X-Gluc staining for β-glucuronidase activity

SAUR63:GUS gene expression patterns were examined in seedlings at 5 DAG in short-day growth conditions or at 3 DAG after growth in complete darkness, and in mature flowers at various stages of floral development (Smyth et al., 1990). Tissue samples were incubated in X-Gluc reaction solution (100 mm sodium phosphate, pH 7.0, 10 mm EDTA, 0.5 mm potassium ferrocyanide, 0.5 mm potassium ferricyanide, 0.1% Triton X-100 and 1 mm 5-bromo-4-chloro-3-β-d-glucuronide) at 37°C for 24 h. X-Gluc-stained tissues were decolorized in 70% ethanol at 37°C for 24 h, and examined using a Leica M205 FA stereo microscope (Leica, Buffalo Grove, IL, USA).

Protein half-life measurements

To assess protein levels, SAUR63:HA or SAUR63:GFP seedlings were grown for 5 DAG on MS-agar plates, and 3 ml of water containing IAA (10 μm), cycloheximide (30 μm) and/or MG132 (10 μm) was applied to the surface of the plates. Seedlings were collected and frozen at the indicated time points. Experiments were performed starting 4 h after subjective dawn (defined as time 0), when SAUR63:HA protein is abundant (Figure 5b). To measure protein half-life, seedlings were pre-treated with 30 μm cycloheximide for 10 min, and either left in bright light (100 μmol m−2 s−2) or shifted to dim light (10 μmol m−2 s−2) at time 0.

Western blotting

Total protein extracts were prepared from 30 5-day-old seedlings using grinding buffer (20 mm Tris–HCl, pH 7.5, 150 mm NaCl, 1% Triton X-100, 1 mm EDTA, pH 8.0, 0.1% SDS), with the addition of 10 mm DTT and full-strength protease inhibitor cocktail (Sigma-Aldrich, St Louis, MO, USA). Proteins (30 μg) were combined with an equal volume of double-strength SDS sample buffer (125 mm Tris–HCl, pH 6.8, 4% SDS, 20% glycerol, 0.01% bromophenol blue and 10 mm DTT), heated at 95°C for 5 min, and separated by 12% SDS-PAGE. Proteins were then transferred to polyvinylidene difluoride (PVDF) membranes (Hybond™-P; GE Healthcare, Piscataway, NJ, USA) by electroblotting. Non-specific antibody interactions were pre-blocked by 3% dry-milk powder in TBST (20 mm Tris-HCl, pH 7.5, 200 mm NaCl and 0.1% Tween 20). For immunoblotting of SAUR63:HA, the membrane was incubated with a 1:10 000 dilution of rat anti-HA antibody (Roche Diagnostics, Indianapolis, IN, USA) at 4°C overnight, and then treated with a 1:10 000 dilution of goat anti-rat antibody conjugated with horseradish peroxidase (Abcam, Cambridge, MA, USA) at 22°C for 1 h. For SAUR63:GFP, the membrane was incubated with a 1:200 dilution of monoclonal mouse anti-GFP antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA) at 4°C overnight, and then treated with a 1:1000 dilution of sheep anti-mouse antibody conjugated with horseradish peroxidase (Amersham, GE Healthcare, Piscataway, NJ, USA) at room temperature for 1 h. The fluorescence signals of proteins were detected on X-ray film using Chemiluminescence Reagent Plus (Bio-Rad, Hercules, CA, USA). Rabbit anti-APX (l-ascorbate peroxidase) antibodies (Agrisera, Vännäs, Sweden) were used as protein controls in blotting.

Microsomal fractionation

Total, soluble and membrane extracts were prepared as described by Chung et al. (2011). Total protein extracts were obtained by grinding 30 5-day-old seedlings with 100 μl of extraction buffer (20 mm Tris-HCl, pH 8.0, 330 mm sucrose and 1 mm EDTA), with the fresh addition of 10 mm DTT and full-strength protease inhibitor cocktail (Sigma-Aldrich). Following the precipitatation of cell debris at 8 000 g for 10 min, the supernatant was transferred to a new microcentrifuge tube (total extract). The half volume (50 μl) of total extract was then centrifuged at 16 000 g at 4°C for 1 h. The clear supernatant was removed to a new tube (soluble fraction), and the pellet was resuspended in 50 μl of extraction buffer with the fresh addition of 10 mm DTT and full-strength protease inhibitor cocktail (membrane fraction). Protein extracts were combined with an equal volume of double-strength SDS sample buffer and subjected to western blotting using rat anti-HA antibodies or monoclonal mouse anti-GFP antibodies. Rabbit anti-APX (l-ascorbate peroxidase) antibodies and rabbit anti-RD28 antibodies (Boyes et al., 1998) were also used as blotting markers for soluble and membrane fractions, respectively.

RNA analysis

Total RNAs were extracted from 30 5-day-old seedlings using the RNeasy Mini Kit (Qiagen, Valencia, CA, USA) and treated on-column with DNase I (Qiagen). For RT-PCR, first-strand cDNAs were synthesized from 1 μg of total RNAs using the Reverse Transcription System (Promega, Madison, WI, USA). For amplification of SAUR gene transcripts, PCR was performed for 35 cycles with 50 or 60°C annealing temperature using gene-specific primer pairs (Figure S4; Table S1). For RNA blot hybridizations, a SAUR63 coding region probe was PCR-amplified from genomic DNA using flanking primers (Table S1) and labeled by random priming with [32P]dNTPs.

Basipetal auxin transport assay

Basipetal IAA transport in the hypocotyl was measured as previously described (Lewis and Muday, 2009). Seedlings were vertically grown on MS-agar (without sucrose) under dim light (<10 μmol m−2 s−2) growth conditions for 5 days. Seedlings were then transferred to bright light (100 μmol m−2 s−2) growth conditions and grown for 2 days. Seedlings (n = 10) with equal hypocotyl length were transferred to fresh MS agar plates and aligned at their apices 1 h prior to starting the assay, after which shoot apices of seedlings were removed. An auxin-containing agar droplet [0.02% 2-(N-morpholine)-ethanesulphonic acid (MES), 1% sucrose, 1.25% agar and 200 nm [3H]IAA, pH 5.5] was then applied to the apical (cut) end of each hypocotyl. Control sets of hypocotyl samples were independently prepared using a control agar droplet containing auxin plus 10 μm NPA. Decapitated seedlings were left upside down to prevent non-specific auxin flow during basipetal IAA transport. After a 3-h incubation, the apical 3-mm section of each hypocotyl was removed and discarded, whereas the next 4-mm section was taken to measure radioisotope counts of basipetally transported [3H]IAA using a Beckman LS 6500 scintillation counter (Beckman, Brea, CA, USA).

Fluorescence imaging for protein localization

Five-day-old SAUR63:GFP seedlings were examined for SAUR63:GFP localization in epidermis cells of the hypocotyl. GFP fluorescence was visualized on a Zeiss 710 DUO confocal microscope (Zeiss, Thornwood, NY, USA) using a 488-nm laser line attenuated to 7.0% and a 458–514-nm band pass filter. For visualization of the plasma membrane, whole seedlings were stained by 5 μm FM4-64 (Calbiochem, EMD Millipore, Billerica, MA, USA) on ice for 5 min. Fluorescence of FM4-64 was visualized using a 514-nm laser line attenuated to 2.0% and a >690-nm band pass filter. Wild-type samples were processed in parallel and imaged using the same settings to reveal background fluorescence signals. For the transient assay for SAUR63:Cerulean localization, the SAUR63 coding sequence was cloned into pMDC7-Cerulean-HA (a gift from E. Washington (University of North Carolina at Chapel Hill, NC, USA); Akimoto-Tomiyama et al., 2012; Curtis and Grossniklaus, 2003; Rizzo et al., 2004). Agrobacterium strain GV3101 pMP90 carrying pMDC7-SAUR63:Cerulean:HA was transiently infected into Arabidopsis cotyledon epidermal cells as described by Marion et al. (2008), and viewed by confocal microscopy 24 h after induction with 10 μm estradiol (Zuo et al., 2000).

The immunofluorescence assay for SAUR63:HA protein localization was performed as described by (Sauer et al., 2006). Five-day-old SAUR63:HA seedlings were placed in fixative (4% paraformaldehyde in full-strength PBS with 0.1% Triton X-100) and incubated at room temperature in a vacuum desiccator for 1 h. The cuticle was removed by the following wash series: full-strength PBS (twice for 10 min at room temperature), 100% methanol (three times for 10 min at 37°C), ethanol/xylene (1:1 v/v, three times for 10 min, 37°C), 98% xylene (three times for 10 min at 37°C), ethanol/xylene (1:1 v/v, twice for 10 min at room temperature), 99% ethanol (twice for 10 min at room temperature), 90% ethanol (5 min at room temperature), 75% ethanol (5 min at room temperature), 50% ethanol (5 min at room temperature), 25% ethanol (5 min at room temperature) and ddH2O (twice for 5 min at room temperature). After washing, seedlings were mounted on Superfrost microscope slides (Fisher Scientific, Pittsburgh, PA, USA), covered with a cover slip and then subjected to five freeze–thaw cycles in liquid nitrogen. Samples were left to dry at room temperature overnight and stored at −20°C for 24 h. For cell wall digestion, samples were briefly rehydrated in full-strength PBS for 5 min, and then incubated with 2% (w/v) Driselase (Sigma-Aldrich) in full-strength PBS at 37°C for 45 min. For cell permeabilization, samples were washed with full-strength PBS (five times for 10 min), and then incubated in full-strength PBS plus 0.5% Tween 20 at room temperature for 1 h. Permeabilized samples were washed in full-strength PBS (seven times for 10 min) and pre-blocked in 3% BSA in full-strength PBS at room temperature for 1 h. For primary antibody reaction, rat anti-HA antibodies (Roche Diagnostics) were treated in 1:1000 dilutions at 4°C overnight and at 37°C for 4 h. After a brief wash in full-strength PBS for 5 min, samples were incubated in 1:500 dilutions of Alexa-Fluor 488 conjugated anti-rat antibody (Invitrogen, Life Technologies, Grand Island, NY, USA) at 37°C for 3 h. Samples were washed in full-strength PBS (six times for 10 min), and then mounted with 0.1%p-phenylenediamine (Acros Organics, Fisher Scientific, Pittsburgh, PA, USA) in Mowiol (Calbiochem) for fluorescence imaging.

Arabidopsis Genome Initiative (AGI) numbers

Arabidopsis Genome Initiative (AGI) numbers for the genes discussed in this article: SAUR61 (At1g29420); SAUR62 (At1g29430); SAUR63 (At1g29440); SAUR64 (At1g29450); SAUR65 (At1g29460); SAUR66 (At1g29500); SAUR67 (At1g29510); SAUR68 (At1g29490); SAUR75 (At5g27780); ACTIN2 (At3g18780).


We thank Angus Murphy for providing Arabidopsis abcb mutants, Tony Perdue for help with confocal microscopy, Eui Hwan Chung, Sara Ploense and Vandana Yadav for helpful discussions, Corbin Jones and Charles Hodgens for help with statistical analyses, Jeff Dangl for sharing research materials and equipment, E. Washington for the pMDC7-cerulean-HA vector and Esther J. Park for technical assistance. This study was supported by a grant from the National Science Foundation (IOS-0920418 to JWR and PN).