OsCYP2, a chaperone involved in degradation of auxin-responsive proteins, plays crucial roles in rice lateral root initiation


For correspondence (e-mail xiaorong@zju.edu.cn).


Auxin plays a pivotal role in many facets of plant development. It acts by inducing the interaction between auxin-responsive [auxin (AUX)/indole-3-acetic acid (IAA)] proteins and the ubiquitin protein ligase SCFTIR to promote the degradation of the AUX/IAA proteins. Other cofactors and chaperones that participate in auxin signaling remain to be identified. Here, we characterized rice (Oryza sativa) plants with mutations in a cyclophilin gene (OsCYP2). cyp2 mutants showed defects in auxin responses and exhibited a variety of auxin-related growth defects in the root. In cyp2 mutants, lateral root initiation was blocked after nuclear migration but before the first anticlinal division of the pericycle cell. Yeast two-hybrid and in vitro pull-down results revealed an association between OsCYP2 and the co-chaperone Suppressor of G2 allele of skp1 (OsSGT1). Luciferase complementation imaging assays further supported this interaction. Similar to previous findings in an Arabidopsis thaliana SGT1 mutant (atsgt1b), degradation of AUX/IAA proteins was retarded in cyp2 mutants treated with exogenous 1-naphthylacetic acid. Our results suggest that OsCYP2 participates in auxin signal transduction by interacting with OsSGT1.


Lateral root organogenesis is among the best-studied auxin-regulated events. Auxin (AUX) is required for the initial cell division during the development of a lateral root primordium (De Smet et al., 2006, 2007). Auxin transport inhibitors suppress lateral root formation, suggesting that polar auxin transport (the regulated transport in plants of the plant hormone auxin) is required for lateral root initiation (Reed et al., 1998; Casimiro et al., 2001). Many Arabidopsis mutants impaired in auxin signaling exhibit a paucity of lateral roots, including tir1 (Ruegger et al., 1998), axr1 (Lincoln et al., 1990), msg2-1/iaa19 (Tatematsu et al., 2004), shy2/iaa3 (Tian and Reed, 1999), iaa28-1 (Rogg and Bartel, 2001), and axr5/iaa1 (Yang et al., 2004). In fact, slr/iaa14 lacks lateral roots completely (Fukaki et al., 2002). By contrast, the gain-of-function mutant iaa7/axr2-1 has an excess of lateral roots (Nagpal et al., 2000). Moreover, abnormal lateral root development is observed in auxin transporter mutants, including aux1 (Marchant et al., 2002) and multiple pin mutants (Benkova et al., 2003), as well as in mutants such as alf1 (Boerjan et al., 1995), ydk1 (Takase et al., 2004), and dfl1 (Nakazawa et al., 2001) with altered auxin homeostasis.

Studies on auxin signal transduction mechanisms have revealed two auxin receptor systems. The first involves the AUXIN-BINDING PROTEIN 1 (ABP1), which binds auxin and then rapidly activates two antagonistic cytoplasmic Rho-related GTPase (ROP-GTPase) pathways that orchestrate the planar morphogenesis of pavement cells (Xu et al., 2010). The second is mediated by TRANSPORT INHIBITOR RESPONSE 1 (TIR1), which is well established as both an auxin-binding receptor and a ligand-activated E3 ligase (Mockaitis and Estelle, 2008).

Signaling by TIR1 involves two classes of antagonistic regulatory transcription factors: AUX/IAA (indole-3-acetic acid) proteins and auxin response factors (ARFs) (Guilfoyle and Hagen, 2007). There are 31 AUX/IAA proteins and 25 known ARFs in rice; this diversity of the numbers of AUX/IAA proteins and ARFs suggests different roles in different tissues and/or developmental stages (Wang et al., 2007). At low auxin concentrations, AUX/IAA transcriptional repressors are relatively stable and dimerize with ARFs. The ARF-bound AUX/IAA repressors recruit the transcriptional co-repressor TOPLESS, which inactivates ARF. At higher concentrations, auxin induces an interaction between AUX/IAA repressors and TIR1 proteins. This stimulates AUX/IAA ubiquitination by the SKP1-Cullin-F-box protein (SCFTIR1) E3 ligase and subsequent proteolysis by the 26S proteasome. In the absence of AUX/IAA repression, ARF functions as a transcription factor (Vanneste and Friml, 2009).

Cyclophilins (CYPs) were originally identified as intracellular targets for the immunosuppressant drug cyclosporine A (Handschumacher et al., 1984; Wang and Heitman, 2005). Members of the immunophilin family, CYPs have peptidyl-propyl, cistrans isomerase (PPIase) activity and catalyze the cis–trans conversion of X-Pro peptide bonds (Schreiber et al., 1991). The basic function of CYPs is to facilitate protein folding. Cyclophilins have been found in diverse organisms, including bacteria, fungi, plants, animals, and humans (Wang and Heitman, 2005). Plant CYPs comprise a wide variety of isoforms in various cellular locations (Schiene-Fischer and Yu, 2001). In tomato, the lecyp1 mutant (also known as dgt) exhibits a subset of auxin-insensitive phenotypes, including a slow gravitropic response, lack of lateral roots, reduced apical dominance, altered vascular development, and reduced fruit growth (Zobel, 1973, 1974; Coenen and Lomax, 1998; Rice and Lomax, 2000; Balbi and Lomax, 2003; Ivanchenko et al., 2006; Oh et al., 2006). However, the molecular role of CYPs in auxin signaling is unknown. The closest homolog of LeCYP1 in rice is OsCYP2 (Oh et al., 2006), the expression of which is upregulated during salt, drought, heat, and cold stresses. Moreover, OsCYP2 overexpression in transgenic rice seedlings confers improved tolerance to salt stress (Ruan et al., 2011); however, it remains unknown whether OsCYP2 is related to auxin signaling.

The SCFTIR1 complex is a central regulator of the auxin response pathway in plants (Mockaitis and Estelle, 2008). Here, we investigated whether other cofactors and chaperones participate in auxin signaling. We characterized rice mutants cyp2-1 and cyp2-2 with mutations in the cyclophilin gene OsCYP2. We found that they were defective in a subset of auxin responses. Detailed examination of the cyp2-1 mutant revealed a block in lateral root initiation at the nuclear migration stage. Furthermore, we found that OsCYP2 interacted with co-chaperone OsSGT1 and promoted degradation of AUX/IAA.


Isolation of an auxin-resistant rice mutant

To identify novel factors involved in auxin signaling, we screened ethyl-methanesulfonate (EMS)-mutagenized M2 seedlings for mutants with reduced sensitivity to the exogenous auxin 1-naphthylacetic acid (NAA; 0.1 μm). We identified 30 M2 lines that exhibited reduced auxin inhibition of seminal root length compared with wild-type (WT) plants. Among these, two independent M2 lines with similar pleiotropic phenotypes were selected for further analysis. These lines lacked lateral roots, were insensitive to exogenous auxin, and showed low fertility. After map-based cloning (see below), we found that these two mutants were deficient in the same gene, OsCYP2, and we named them cyp2-1 and cyp2-2. The cyp2-1 mutant was generated from the japonica cultivar ShiShouBaiMao (SSBM; Figure 1a), and the cyp2-2 mutant was generated from the indica cultivar Kasalath (KAS; Figure S1) These two cyp2 mutants showed defective lateral root development, reduced shoot length, decreased fertility, and a delayed heading stage (Table S1).

Figure 1.

The cyp2-1 mutant is insensitive to exogenous auxin (1-naphthylacetic acid, NAA). (a) Seven-day-old wild-type (SSBM) and cyp2-1 mutant seedlings, grown in rice culture medium without (left) and with (right) 0.1 μm NAA. Scale bar = 2 cm. (b), (c) Root hairs of 7-day-old SSBM (left) and cyp2-1 mutant (right) seedlings grown in rice culture solution without (b) and with (c) 0.1 μm NAA. Scale bar = 1 cm. (d) Lateral root which break through the epidermis number in 7-day-old wild-type and cyp2 mutants. (e) Lateral root which break through the epidermis length in 7-day-old wild-type and cyp2 mutants. (f) The relative transcript levels of two auxin-response genes (OsIAA9 and OsIAA20) in SSBM and cyp2-1 mutant roots. The white and light gray bars indicate SSBM and cyp2-1 under normal (untreated) conditions; the dark gray and black bars indicate SSBM and cyp2-1 after 3 h treatment with 10 μm exogenous auxin (IAA) in the culture medium. (g) DR5-driven GUS reporter expression (blue) indicates local auxin accumulation in SSBM and cyp2-1 mutant roots. M indicates meristem region. Scale bar = 200 μm. All values in this legend are means ± SD of three replicates. The asterisk indicates statistical significance between the wild type and mutants by t-test (*< 0.05; **< 0.005).

OsCYP2 mutants are less sensitive to exogenous auxin than wild type

In WT plants, NAA inhibited seminal root growth: roots were approximately 20% the length of those in untreated plants, and the density of lateral roots was increased. In mutant plants, inhibition of root growth by NAA was much weaker (Figure 1a). Previous studies have shown that exogenous auxin induces lateral root formation in rice (Casimiro et al., 2001). However, even after 7 days of culture with NAA, cyp2-1 lacked lateral roots (Figure 1a). This suggests that the cyp2-1 mutant completely lacked the ability to develop lateral roots and was insensitive to exogenous auxin in terms of lateral and seminal root growth. In contrast to cyp2-1, cyp2-2 showed a weaker phenotype, with the ability to grow a few lateral roots in the presence or absence of NAA (Figure S1). Quantitative analysis showed that cyp2-2 mutants were affected in lateral root number but not length (Figure 1d,e). In rice, root hair formation is promoted by exogenous auxin (Pitts et al., 1998). We observed that 0.1 μm NAA did not significantly induce root hair formation or root hair elongation in cyp2-1 plants, in contrast to WT (Figure 1b,c).

Among the rice AUX/IAA gene family members, the OsIAA9 and OsIAA20 early response genes were identified as highly responsive to auxin (Jain et al., 2006). We measured the transcript levels of these two genes by quantitative RT-PCR before and after treatment with exogenous auxin (IAA in this case). The cyp2-1 mutants showed significantly less induction of these genes than did WT plants (Figure 1f).

The DR5:GUS plasmid construct contains a beta-glucuronidase (GUS) reporter gene driven by a synthetic auxin-inducible enhancer (DR5). Thus, the level of reporter gene expression reflects the local accumulation of auxin and correlates with the auxin response in root tips and lateral root primordial in Arabidopsis (Ulmasov et al., 1997). We found that GUS staining levels were similar in the steles of WT and cyp2-1 mutant plants, while in cyp2-1 mutant root tips, GUS expression did not extend into the Quiescent Center (QC), initials, and lateral root cap as it did in WT (Figure 1g). After NAA was added, the GUS staining increased at the stele in WT, while in the cyp2-1 mutants, the GUS staining was remarkably lower, especially in the meristem region (Figure 1g, Figure S2d). This indicates that cyp2-1 is impaired in responses to auxin.

Lateral root initiation is blocked in cyp2-1

The most striking developmental phenotype of cyp2-1 was its lack of lateral roots. To determine which stage of lateral root development was affected by the cyp2-1 mutation, we analyzed the expression of the cell cycle marker, CycB1;1p:GUS, in cyp2-1 primary roots. The CycB1;1 gene is only expressed in the G2/M transition (Doerner et al., 1996); thus, it serves as a marker for dividing cells in both the root pericycle and the root apical meristem (Ferreira et al., 1994; Dubrovsky et al., 2000; Casimiro et al., 2001). In 7-day-old WT roots, we detected GUS in the root apical meristem, the pericycle, and the lateral root primordia (Figure 2a). In cyp2-1 mutant roots, the lateral root primordia were missing, but GUS was detected in the root apical meristem and in single line pericycle cells, even in the root hair zone (Figure 2a). Closer observation by longitudinal section around the root hair zone showed pairs of stained pericycle cells in each staining site of the cyp2-1 mutant (Figure 2b). Notably, the density of GUS staining sites in the pericycle was not significantly altered in the mutant: a similar frequency of founder cells was identified in WT and cyp2-1 mutant plants. Upon treatment with 0.1 μm NAA, GUS staining sites in cyp2-1 roots were not significantly increased as they were in WT (Figure 2c). These results indicate that cell cycle activation under auxin is restricted in the mutant line. Although CycB1-driven GUS staining indicated that cyp2-1 mutants maintained proliferative capacity in the pairs of pericycle cells in the root hair zone, none of these cells formed short lateral root initial cells, which represent the first asymmetric anticlinal division. Interestingly, with longitudinal section microscopy, we observed a high frequency of nuclear migration in the GUS-stained sites of the cyp2-1 mutant (Figure 2b). Of the 10 slices analyzed, eight showed displaced nuclei but none exhibited anticlinal divisions (Figure 2d). Therefore, we conclude that the pairs of lateral root founder cells in cyp2-1 mutants were maintained in the G2/M phase, and their nuclei had migrated toward the common cell wall. However, normal development was blocked just before the first asymmetric anticlinal division.

Figure 2.

Lateral root primordium initiation is blocked in cyp2 mutant plants. (a) Cell cycle (CycB1)-driven GUS reporter activity (blue) indicates cell division in wild-type (SSBM) and cyp2-1 mutant 7-day-old roots. Root meristems were stained for 5 min (bottom panel); the elongation zones (middle panel) and the region of lateral root primordium emergence were stained for 1 h (top panel). Scale bar = 200 μm. (b) Longitudinal sections around root hair zones of cyp2-1 mutant plant. Nuclear migration in pericycle cells is indicated by displacement of the nuclei from the central location (normal site) between cell walls. CycB1-driven GUS reporter staining (blue) indicates dividing cells; basic fuchsin (red) indicates cell walls. *, nuclei; arrowheads, cell walls. Scale bar = 100 μm. (c) The density of sites with CycB1-driven GUS reporter activity indicates the number of lateral root initial events per centimeter in SSBM and cyp2-1 mutant roots. Values are means ± SD of three replicates. CK, control, without 1-naphthylacetic acid (NAA) treatment. (d) Displaced nuclei events observed in 10 slices of cyp2-1 mutant roots.

Mutations in the OsCYP2 gene

Genetic analysis of F2 progeny derived from a cross between cyp2-1 and the indica cultivar KAS revealed that the cyp2-1 genome carried a recessive mutation at a single nuclear locus. The cyp2-1 locus was mapped to chromosome 2 between the markers Lrt2P2 and RM12368. Three polymorphic markers (M1, M2, and M3) were designed for fine mapping between RM6938 and RM12368. In 1020 F2 mutant progeny, we observed three, zero, and six recombination events between the OsCYP2 locus and M1, M2, and M3, respectively (Figure 3a). The cyp2-1 gene was mapped to a 57-kb region between M1 and M3 that was covered by two contiguous bacterial artificial chromosomes (Figure 3a). DNA sequence comparison revealed that single nucleotide mutations existed in both mutant alleles of the OsCYP2 gene. cyp2-1 contained a nucleotide mutation (G457T) that resulted in a premature stop codon, and cyp2-2 had a nucleotide mutation (G215C) that changed an amino acid from Gly (72) to Ala (72) (Figure 3a). The OsCYP2 gene has no intron and its full-length cDNA recorded in the Knowledge-based Oryza Molecular Biological Encyclopedia (KOME) database (http://cdna01.dna.affrc.go.jp/) is 887 bp long, with a 94 bp 5'-untranslated region (UTR), 519 bp coding sequence, and 274 bp 3'-UTR (Figure 3a). OsCYP2 encodes a protein of 178 amino acids, and it is classified as a Group I-a, single-domain, plant CYP (Oh et al., 2006). The RT-PCR analysis showed that full-length transcripts of OsCYP2 were expressed in the cyp2-1 and cyp2-2 mutants (Figures 3b and S2b). Western blotting with antibody specific for the carboxy-terminal end of the OsCYP2 protein revealed that the cyp2-1 mutant did not express the full-length OsCYP2 protein (Figure 3c), whereas cyp2-2 did (Figure S2c). The premature stop in cyp2-1 and the amino acid substitution in cyp2-2 were consistent with the stronger phenotype (absence of lateral roots) in cyp2-1 and the weaker phenotype (development of a few lateral roots) in cyp2-2, respectively.

Figure 3.

Genomic analysis of OsCYP2. (a) The OsCYP2 locus mapped to chromosome 2 between markers M1 and M3, and was covered by two bacterial artificial chromosomes (OSJNBb0088N06 and OJ1020_C02). Two allelic mutants (cyp2-1 and cyp2-2) are indicated. Rec = recombination frequency; CDS = coding sequence. (b) The RT-PCR analysis shows OsCYP2 mRNA expression in roots of wild type (SSBM) and cyp2-1 mutant plants. (c) Western blot analysis shows distinct OsCYP2 protein expression in roots of SSBM and cyp2-1 mutant plants. Total proteins were extracted from cultured cell suspensions. Ten micrograms were loaded and separated by SDS PAGE. Western blots were probed with antibodies against OsCYP2 (18 kDa). (d) Quantitative RT-PCR analysis of OsCYP2 in roots of SSBM, cyp2-1 mutant, and two cyp2-1 mutant plant lines that overexpressed the OsCYP2 transgene (O1 and O2). Values are means ± SD of three replicates. (e) Seedling phenotypes of SSBM, cyp2-1, O1, and O2 plants. Scale bar = 2 cm. (f) Panicles from SSBM, cyp2-1, O1, and O2 plants. Scale bar = 2 cm.

A complementation test was carried out to verify that the phenotypes of the cyp2 mutants were caused by a mutation in the OsCYP2 gene. The 35S promoter-driven 519 bp open reading frame (ORF) of OsCYP2 was introduced into cyp2-1 mutant plants by Agrobacterium tumefaciens infiltration. Overexpression of OsCYP2 in transgenic lines (designated as O1 and O2) was confirmed by quantitative RT-PCR analysis (Figure 3d). The defects in lateral root growth and seed fertility were completely rescued in transgenic seedlings (Figure 3e,f), whereas the overexpressing lines did not show increased sensitivity to exogenous auxin treatment (Figure S2e).

OsCYP2 protein was localized to the cytosol and nucleus

To elucidate the subcellular localization of OsCYP2 proteins, we constructed plasmids that carried the OsCYP2 sequence fused to the green fluorescent protein (GFP) sequence, driven by the native promoter of OsCYP2. This vector (OsCYP2p:OsCYP2-GFP) was transiently expressed in rice protoplasts, and green fluorescent signals were observed by confocal microscopy. The results indicated that OsCYP2 was localized to the cytosol and nucleus (Figure 4a).

Figure 4.

Subcellular localization of OsCYP2 and expression pattern of OsCYP2. (a) Subcellular localization of OsCYP2 promoter (POsCYP2)-driven OsCYP2-GFP fusion protein and 35S promoter-driven red fluorescent protein (RFP) in rice protoplasts observed by confocal microscopy and transmission microscopy (bright field, BF). Scale bar = 20 μm. (b) Quantitative RT-PCR analysis of OsCYP2 tissue expression patterns. Total RNA was extracted from the root, stem base (SB), stem, and leaf of 15-day-old wild type (WT) seedlings, and from the flower at the heading stage. Values are means ± SD of three replicates. (c), (d) OsCYP2p:GUS reporter activity indicates OsCYP2 expression patterns in WT plants; (c) GUS staining (blue) in primary root tip (left panels) and formation of lateral root primordium (right three panels); (d) GUS staining in crown root primordia (top left), leaf vein (top right), lemma vein (bottom left), and stamen (bottom right). Scale bar = 500 μm. (e) Confocal microscopy images of the OsCYP2-GFP fusion protein (green) shows OsCYP2 expression in primary root tip (left panels), lateral root primordium (middle four panels), and lateral root (right panels). Comparisons of GFP fluorescence (top row), propidium iodide fluorescence (middle row), and merged images (bottom row). Scale bar = 400 μm.

Expression pattern of OsCYP2

The tissue expression pattern of OsCYP2 was investigated by quantitative RT-PCR. The OsCYP2 gene was expressed in all tissues tested (Figure 4b). To confirm this result, we cloned the 2.8-kb promoter region of OsCYP2 from WT plants and fused it to the GUS reporter. The resulting OsCYP2p:GUS construct was used to transform rice, and two independent, stable transformants were investigated by staining for GUS activity. We observed GUS staining at the seminal root meristem and at the lateral root meristem, in the later stages of lateral root initiation (Figure 4c). We also observed GUS staining in the crown root primordia, leaf, lemma vein, and anther (Figure 4d). A second construct, OsCYP2p:OsCYP2-GFP, confirmed the expression pattern of OsCYP2 in roots (Figure 4e).

OsCYP2 directly interacts with OsSGT1

To determine the roles of OsCYP2 in auxin signaling and control of pericycle cell fate, we conducted a yeast two-hybrid screen. A bait construct DNA-binding domain (BD)-OsCYP2 was used to screen an activation domain (AD) fusion cDNA library prepared from rice roots. Approximately 4 × 106 yeast transformants were screened, and several reproducibly positive clones were identified in both HIS3 and LacZ assays. Among these candidates, OsSGT1 was the only protein related to the SCFTIR1-mediated auxin response. The clone contained a 1157-bp cDNA sequence encoding the 361 N-terminal amino acids of OsSGT1. OsSGT1 was fused to Gal4 AD in-frame, and co-transformation experiments confirmed that OsCYP2 could interact with OsSGT1 in yeast (Figure 5a). Next, we performed an in vitro pull-down assay between recombinant His-OsSGT1 and glutathione-S-transferase (GST)-OsCYP2. This also showed that OsSGT1 could directly interact with OsCYP2 (Figure 5b). We further tested the interaction between OsCYP2 and OsSGT1 in tobacco leaves with a firefly luciferase complementation imaging (LCI) assay. The results supported an interaction between OsCYP2 and OsSGT1 in living plant cells (Figure 5c,d).

Figure 5.

OsCYP2 interacts with OsSGT1. (a) Yeast clones were grown on plates that contained X-Gal. The blue color in the colony co-transformed with BD-OsCYP2 and AD-OsSGT1 indicates interaction between the proteins. (b) In vitro pull-down assays. Recombinant glutathione-S-transferase (GST) or GST-OsCYP2 was used as bait and incubated with 1 μg of prey protein (His-OsSGT1). After incubation, GST and GST-OsCYP2 were retrieved with glutathione beads, and the pulled-down proteins were detected on Western blots with antibodies to His. (c) Luciferase (LUC) image of tobacco leaves co-infiltrated with agrobacterial strains that induced expression of the following: (i) NLuc + CLuc-OsSGT1, (ii) OsCYP2-NLuc + CLuc, (iii) OsCYP2-NLuc + CLuc-OsSGT1, (iv) positive control: AtSGT1a-NLuc + CLuc-AtRAR1. Luminescence intensifies with protein interaction. The pseudocolor bar on the right indicates the luminescence intensity in each image. Three independent repetitions were carried out. (d) Quantification of LUC activity in leaves that expressed the constructs shown in panel (c). Values are means ± SD of three replicates. The asterisk indicates statistical significance between the wild type and mutants by t-test (*< 0.05, **< 0.005).

cyp2-1 shows reduced AUX/IAA protein degradation with auxin treatment

It was previously reported that the Arabidopsis atsgt1b mutant (mutated in a homolog of OsSGT1) shows reduced AUX/IAA protein degradation and fewer lateral roots (Gray et al., 2003) compared with WT plants. Thus, we wondered whether altered interaction of OsCYP2 with OsSGT1 would similarly affect auxin signaling. To investigate this, we tested whether AUX/IAA protein degradation was hampered in cyp2 mutants in a transient rice protoplast expression system. We transformed WT and cyp2-1 protoplasts with a construct containing the full-length OsIAA11 ORF fused to GFP under the control of the 35S promoter (35Sp:OsIAA11-GFP). As expected, the fluorescence was specifically located in the nucleus (Figure 6a). The degradation kinetics were monitored with in vivo fluorescence imaging of nuclei of single cells. Cycloheximide (CHX) was added to inhibit the synthesis (Munro et al., 1968) of new OsIAA11-GFP fusion protein. The time-course of fluorescence intensity indicated that the cyp2-1 mutant exhibited reduced OsIAA11-GFP degradation with NAA treatment (Figure 6b). This was confirmed in experiments with three independent protoplasts (Figure S3). After 30-min of NAA treatment, OsIAA11-GFP fluorescence decreased by about 60% in WT, and by only about 30% in cyp2-1 mutants (Figure 6c). This suggests that OsCYP2 plays a role in the degradation of AUX/IAA protein.

Figure 6.

Decreased degradation of OsIAA11 in the cyp2-1 mutant. (a) Nuclear localization of the 35S-driven, OsIAA11-GFP fusion protein in rice protoplast. From left to right, OsIAA11-GFP (green), 35S-driven mCherry (red, control marker), merged image, and corresponding transmission microscopy image. Scale bar = 20 μm. (b) Single 35Sp:OsIAA11-GFP transfected rice protoplasts, observed over time under high magnification (60×) to monitor localization and levels of IAA11-GFP fusion protein in the nucleus. Nuclei from wild type (WT; SSBM) and cyp2-1 mutant cells were maintained in auxin-free medium and imaged immediately after adding cycloheximide (CHX; time 0 min) (top two panels). Nuclei from SSBM and cyp2-1 mutant cells were treated with 20 μm 1-naphthylacetic acid (NAA) and imaged after adding CHX (bottom two panels). (c) The degradation of indole-3-acetic acid (IAA)-GFP in SSBM and cyp2-1 mutant cells. The pixel intensity of the nucleoplasm was the mean of 10 randomly selected nucleoplasms in three independent replicates. The degradation percentage was calculated as 100% – (+NAA 30'/+NAA 0')/(–NAA 30'/–NAA 0'). Values are means ± SD of three replicates.


Here, we characterized novel rice mutants with mutations in the rice cyclophilin gene OsCYP2. cyp2 mutants were defective in a subset of auxin responses, exhibiting a lack of lateral roots and insensitivity to exogenous NAA treatment. Both rice and tomato express only one member (OsCYP2, LeCYP1) of the Group I-a single domain CYPs. By contrast, Arabidopsis plants express four members of this subgroup (Oh et al., 2006). The redundancy of multiple Arabidopsis CYPs of this subgroup might explain why, to date, no Atcyp mutant has been identified as being related to auxin signal transduction. Previous studies reported that LeCYP1 was involved in auxin regulation of lateral root primordium initiation (Ivanchenko et al., 2006). Our current results suggest that OsCYP2 mediates auxin signaling by interacting with the co-chaperone OsSGT1. This finding provides insight into the molecular mechanisms of auxin regulation of plant development.

OsCYP2 is involved in auxin signal transduction

cyp2 mutants were insensitive to exogenous auxin. The DR5p:GUS staining pattern in seminal root tips showed that cyp2-1 mutants had an abnormal auxin response in root apices (Figure 1g). Moreover, exogenous auxin could not rescue the lateral root formation defect of the cyp2 mutant, despite its ability to rescue the aux1 lateral root phenotype in Arabidopsis (Marchant et al., 2002). This indicates that the cyp2 mutant phenotype is not related to auxin transport. The WT and cyp2 mutants had similar concentrations of free IAA in roots (Figure S2a), suggesting that the insensitivity of cyp2 mutants to auxin is primarily caused by defective auxin signaling, not auxin synthesis or distribution. This is consistent with observations reported for tomato mutated in Lecyp1, the OsCYP2 homolog (Oh et al., 2006).

OsCYP2 controls asymmetric anticlinal pericycle cell division

Spatiotemporal control over local auxin distribution can enable stimulation of a select group of cells to execute a specific change in a developmental program. This mechanism plays an important role in conveying positional information during organogenesis. Random stimulation of auxin biosynthesis in a single pericycle cell in Arabidopsis causes a localized increase in auxin sufficient to induce the initiation and formation of lateral roots (Dubrovsky et al., 2008). Therefore, all cells within the pericycle are able to form lateral roots, but only some cells execute the formation. It was suggested that lateral root initiation requires the coordinated action of auxin transport and signaling, cell cycle regulators, and novel root-specific proteins (Yadav et al., 2010). Our results suggest that OsCYP2 plays an important role in coordinating auxin signaling to induce lateral root initiation.

In Arabidopsis lateral roots are formed from pericycle cells adjacent to protoxylem poles, whereas in rice lateral roots are formed from pericycle cells in close proximity to phloem poles (Hochholdinger and Zimmermann, 2008). In Arabidopsis, the formation of lateral root primordia is initiated by asymmetric, anticlinal divisions in pericycle cells (Malamy and Benfey, 1997). In rice, lateral root primordia are initiated by anticlinal divisions, followed by periclinal divisions in the pericycle and endodermis (Kawata and Shibayama, 1965). In many species, asymmetric anticlinal divisions in pericycle cells result in the formation of short cells, which then leads to the initiation of a lateral root (Casero et al., 1993). In the cyp2-1 mutant, we observed no asymmetric anticlinal division of pericycle cells (Figure 2b). In WT rice, application of exogenous auxin induces lateral root formation (Casimiro et al., 2001), but in cyp2-1 mutants, 7 days of NAA treatment could not rescue lateral root formation (Figure 1a). Thus, cyp2-1 mutants completely lacked the ability to initiate asymmetric anticlinal divisions in pericycle cells. By contrast, the Lecyp1 tomato mutant could achieve anticlinal asymmetric division (Ivanchenko et al., 2006).

In Arabidopsis, auxin signaling normally primes pairs of xylem pole pericycle cells to become founder cells in the basal meristem, which is immediately behind the primary root apical meristem (Casimiro et al., 2001; De Smet et al., 2007). We found that the cyp2 mutant was insensitive to exogenous auxin, specifically in the meristem. Furthermore, our data showed that OsCYP2 was mainly expressed in the apical and basal meristem of primary roots, but not in lateral root initial cells. This implies that OsCYP2 might play an important role in the cell cycle of lateral root founder cells and may control lateral root initial cell fate in the apical and basal meristem of primary roots. We found that WT and cyp2-1 mutants had similar CycB1;1 densities along pericycle lines, suggesting that the cyp2-1 mutation did not hamper the establishment of lateral root founder cells. In cyp2-1 mutants, lateral root founder cells remained in the G2/M stage, and their nuclei could migrate to the common cell wall; however, asymmetric division was blocked. A potential explanation for these results might be that the auxin response in the basal meristem not only determines the position of the founder cells but also controls the cell fate and that this control requires the function of OsCYP2. Nevertheless, the absence of detectable OsCYP2 in lateral root initial cells does not rule out the possibility that it is present and plays a role in pericycle initial cells. In fact, the cyp2-1 mutant phenotype is similar to that observed upon disruption of the auxin response in xylem pole pericycle cells (De Smet et al., 2007). Both phenotypes involve a lack of initial asymmetric division in pairs of founder cells and a high frequency of nuclear migration.

OsCYP2 interacts with OsSGT1 and promotes degradation of AUX/IAA protein

To investigate the mechanism underlying the role that OsCYP2 plays in the auxin signaling pathway, we conducted a yeast two-hybrid screen to determine whether OsCYP2 interacts with other proteins in rice plants. Our results showed that OsCYP2 interacts with OsSGT1. In yeast, Sgt1p was simultaneously associated with the suppressor of kinetochore protein (Skp1p) and heat shock protein 90 (Hsp90p) through its TPR and CS domains, respectively. This complex was necessary for assembling an active centromere/binding factor 3/kinetochore complex during cell cycle progression (Kitagawa et al., 1999; Bansal et al., 2004; Lingelbach and Kaplan, 2004; Catlett and Kaplan, 2006). The SGT1 homolog in plants was identified as a co-chaperone in the HSP90 chaperone system. Interestingly, OsCYP2 has several characteristics in common with HSP90. Both interact with SGT1, both are stress-inducible chaperones, and both function in protein folding. A structural and biochemical analysis of the SGT1–HSP90 interaction suggested its role in the assembly of the SCF ubiquitin ligase complex (Kitagawa et al., 1999; Lingelbach and Kaplan, 2004). Moreover, in both mammals and plants, many immunophilins were identified in the HSP90-based chaperone system and assist in folding modifications of target proteins (Pratt et al., 2001). In Arabidopsis, AtSGT1b is required for the SCFTIR1-mediated degradation of AUX/IAA proteins (Gray et al., 2003). In the present study, we found that cyp2-1 mutants exhibited retarded degradation of AUX/IAA protein upon auxin treatment, similar to the retarded degradation observed in the AtSGT1b mutant eta3. This result demonstrates that the auxin insensitivity of cyp2-1 is most likely due to its role in the degradation of AUX/IAA protein, and this function might be dependent on its interaction with OsSGT1. We propose that OsCYP2, as a chaperone that modifies protein folding, associates with OsSGT1b in a co-chaperone complex that stimulates the function of the SCFTIR ubiquitin E3 ligase complex.

Experimental Procedures

Plant materials and growth conditions

Hydroponic experiments were conducted in normal rice (Oryza sativa) culture solution (Zhu et al., 2012) with the pH adjusted to 5.0 with 1 m NaOH. In all hydroponic experiments, seeds were germinated in water, then directly grown in each type of solution culture (3 L) under a photosynthetic photon flux density of approximately 200 μmol photons m–2 sec−1 with a 12-h light (30°C)/12-h dark (26°C) photoperiod. Humidity was controlled at approximately 70%. To investigate the effects of auxin on root growth and development, WT and cyp2 mutant seeds were grown in culture solutions with 0.1 μm NAA (Sigma-Aldrich, http://www.sigmaaldrich.com/). Indole-3-acetic acid was used in short time auxin treatment.

Mapping and cloning OsCYP2

The OsCYP2 gene was mapped to the short arm of chromosome 2 between simple sequence repeat markers, Lrt2P2 and RM12368, in 30 F2 mutant plants. We developed two derived, cleaved, amplified, polymorphic sequence (dCAPS) markers (M1, M2) and one single nucleotide polymorphism marker (M3) between RM6938 and RM12368 for fine mapping. The primer sequences for Lrt2P2, RM6938, and RM12368 were from Gramene (http://www.gramene.org/). The details of all mapping markers are provided in Table S2. The OsCYP2 gene was fine-mapped to a 57-kb region between the M1 and M3 markers with 1020 F2 mutant plants. All cDNAs within the 57-kb region were then amplified from cyp2-1 mutant plants and sequenced.

Plasmid constructs and plant transformation

For complementing the cyp2-1 mutation, we PCR-amplified the WT ORF (519 bp) and digested it with restriction enzymes SmaI and BglII for cloning into the Ubiquitin-pCAMBIA1300 vector. For the OsCYP2p:GUS construct, the WT genomic sequence from –2800 to –1 was amplified by PCR and introduced in front of the GUS reporter gene in the PBI101GUS vector. For the OsCYP2p:OsCYP2-GFP construct, the WT genomic sequence from –2800 to –1 and the 519 bp of the ORF were amplified by PCR and introduced, in order, in front of the GFP reporter gene of the PBI101sGFP vector. The DR5p:GUS construct was generated as reported previously (Wang et al., 2011). The OsCYCB1;1p:GUS was constructed according to a previously reported strategy for constructing AtCYCB1;1p:GUS (Ferreira et al., 1994), except that the genomic fragment comprised 2317 bp of the putative promoter sequence, and the coding sequence corresponded to the N-terminal 124 amino acids of OsCYCB1;1, and included a candidate mitotic destruction box (King et al., 1996). These fragments were fused to the GUS gene (uidA). All constructs were transformed into the WT japonica cultivar via A. tumefaciens EHA105 as previously described (Hiei et al., 1994).

Semi-quantitative and quantitative real-time RT-PCR

Total RNA was extracted with TRIzol D0410 reagent, according to the manufacturer's instructions (Invitrogen, http://www.invitrogen.com/). First-strand cDNA was synthesized from 5 μg of DNaseI-treated, total RNA with SuperScript II reverse transcriptase (Invitrogen). Semi-quantitative RT-PCR and quantitative RT-PCR were performed as previously described (Zhu et al., 2012). The primers are listed in Table S2.

Sub-cellular localization of OsCYP2 and OsIAA11

The 35Sp:OsCYP2-eGFP and 35Sp:OsIAA11-eGFP expression vectors were constructed by subcloning the full-length ORFs of OsCYP2 and OsIAA11 into the PBI221eGFP vector. The resulting constructs were sequenced to verify in-frame fusion, and then used for transient transformations of rice protoplasts (Miao and Jiang, 2007). Transformed protoplasts were analyzed with a confocal microscope (Zeiss LSM 510).

Histochemical analysis and GUS assay

Histochemical GUS analysis was performed as described previously (Liu et al., 2005).

Protein extraction and western blotting

A 13-amino acid peptide (EKVGSRGGSTAKP) based on the C-terminal region of OsCYP2 was synthesized and used to immunize a rabbit (Abmart Co., http://www.ab-mart.com/) to produce an OsCYP2-specific antibody. Western blotting and immunodetection were conducted as described previously (Nishizawa et al., 2003).

Degradation of GFP

Rice protoplast cultures transfected with 35Sp:OsIAA11-GFP were treated with cycloheximide (CHX) and 20 μm NAA, then immediately transferred to 0.1% poly-l-lysine-coated slides. Single cells were observed with a confocal microscope (Zeiss LSM 510, http://www.zeiss.com/) at high magnification (60×). Images were exposed for the same amount of time in each series; therefore, sequential images reflect the fluorescence intensity relative to that observed in the initial image. Fluorescence intensity in the nucleoplasm was quantified with ImageJ software. The intensity in the nucleoplasm at 0 min was taken as 100%. The fluorescence degradation rate was the mean value of three independent repetitions. All of these processes were consulted as described previously (Tao et al., 2005).

Yeast two-hybrid cDNA library construction and screening

Yeast two-hybrid cDNA library construction and screening was conducted with a Matchmaker Two-Hybrid system kit (Clontech, http://www.clontech.com/). Poly(A)+ RNA prepared from rice root was used to synthesize cDNA. The pGADT7 two-hybrid vector (Clontech) was used to construct a GAL4 activation domain (GAD) fusion cDNA library. The entire coding region of OsCYP2 was cloned into the pGBKT7 vector to produce BD-OsCYP2. The yeast strain Y190 was used to screen and retest co-transformations.

In vitro pull-down assays

The His-OsSGT1 and GST-OsCYP2 fusion genes were cloned into the pET-28a and pGEX-4T-1 vectors, respectively, in Escherichia coli strain BL21(DE3) (Novagen, www.novogen.com/). For in vitro pull-down assays, 1 μg prey protein, His-OsSGT1, was incubated with 2 μg GST or GST-OsCYP2 bait protein for 4 h at 4°C. Then, glutathione sepharose 4B beads (20 μl) were added and incubated for 30 min at 4°C. The pulled-down proteins were extensively washed with buffer [50 mm 2-amino-2-(hydroxymethyl)-1,3-propanediol (TRIS)-HCl, pH 7.4, 400 mm NaCl, and 0.2% NP-40]. As a control, 3% of the input was removed before adding glutathione sepharose 4B beads. Input and pulled-down proteins were analyzed with SDS-PAGE, followed by Western blotting with either the anti-His or anti-GST antibody.

Firefly luciferase complementation imaging (LCI) assay

Full-length OsSGT1 was cloned into the 35Sp:NLuc vector, and full-length OsCYP2 was cloned into the 35Sp:CLuc vector (Chen et al., 2008). Transient LCI assays in Nicotiana benthamiana were performed as described previously (Chen et al., 2008). Luciferase signals were examined with an IVIS® 200 optical imaging system (Caliper Life Sciences, http://www.caliperls.com/) and quantified with Living Image® 4.0 software.


This work was supported by the National High Technology Research and Development Program of China 863 (grant 2007AA10Z118 to XM), the National Basic Research and Development Program of China (grant no. 2011CB100300) and the Fundamental Research Funds for the Central Universities.

Accession Numbers

Sequence data from this article can be found in the GenBank/EMBL data libraries under the following accession numbers: OsCYP2 (BAF07631.1), OsSGT1 (BAF05534.1), OsIAA9 (BAF10351.1), OsIAA11 (BAF12627.2), OsIAA20 (BAF18829.2).