We screened the T-DNA insertion mutant pools of zh11 in japonica rice, and identified a lateral rootless mutant lrl3 (lateral rootless 3) compared with the wild-type plant zh11, and further functional studies were conducted using homozygous mutants and BC₁F₂ generations.

In order to study the function of lrl3 mutant in details, we performed phenotypic analysis on the wild-type plant zh11 and the mutant lrl3 cultured in normal rice (Oryza sativa) medium for 7 days, and found that compared with the wild-type plant zh11, the mutant lrl3 had short shoot and long primary roots, but a few adventitious roots and few lateral roots (Fig. 1A). Further stereoscopic observation revealed that no significant differences were showed between lrl3 and zh11 in root hairs, root tips and root caps, but lateral roots were significantly absent in lrl3 (Fig. 1B). After statistical analysis, it was found that significant or extremely significant differences between lrl3 and zh11 were detected in shoot height, primary root length, number of lateral roots and number of adventitious roots (p < 0.05 or p < 0.01) (Fig. 1D). These differences were further increased when lrl3 and zh11 were cultured for 14 days (Fig. 1C).

These results suggest that the mutation of lrl3 controls shoot elongation, primary shoot, and adventitious root elongation, as well as the number of adventitious roots and lateral roots, and that the differences are more significant along with the growth process.

To investigate the root gravitropism, we horizontally placed the primary root of lrl3 T2 lines and zh11 two days after sowing, and found that lrl3 was slow in response to gravity and barely bent compared with zh11 (see Supplementary Fig. 1). This suggests that the mutation site of lrl3 regulates the response of rice root to gravity.

To understand the physiological mechanism of root development, we treated lrl3 and zh11 with auxin polar transport inhibitor N-1-naphthylphthalamic (NPA) and Synthetic hormones 1-naphthlcetic acid (NAA). The results presented that lrl3 and zh11 showed similar responses to NPA, while lrl3 showed significant resistance to NAA compared with zh11 under the treatment of 0.1 μM NAA (Fig. 2). These results suggest that the mutation locus of lrl3 is involved in auxin signaling during lateral root formation in rice, rather than auxin polar transport.

Genetic analysis of BC₁F₂ populations of lrl3 and zh11 showed that the ratio of wild-type individuals to mutant individuals was in accordance with the expected value of 3:1, suggesting that lrl3 lateral rootless mutation is controlled by a single recessive gene (Table 1).

We employed 30 mutant individuals in F₂ genetic population to detect polymorphisms by SIS2 (M2) and RM12368 markers in close linkage with OsCYP2 gene regulating lateral root formation. The lrl3 band was detected in all the 30 mutant individuals at SIS2 (M2) and in all the 29 mutant individuals at RM12368. The results indicated that the LRL3 gene, the candidate gene OsCYP2 for lrl3 mutation, was preliminarily located near the molecular markers SIS2 (M2) and RM12368 on the chromosome 2 (Fig. 3A). As reported, the gene OsCYP2 encoding rice cyclophilin, participates in the protein folding, which is thought to be associated with auxin signaling [35], so the OsCYP2 gene is regarded as the candidate gene for lrl3 mutation. The coding region of the LRL3/OsCYP2 gene is 519 bp between 1116066 bp and 1116971 bp on chromosome 2. Agarose gel electrophoresis and sequencing analysis on the mutant lrl3 indicated that 59 basepairs of nucleotides were missing at LRL3/OsCYP2 (Figs. 3B–3C). Further bioinformatics analysis showed that the 59-bp deletion at LRL3/OsCYP2 led to the deletion of glutamate from 15^th to 35^th and frame-coding mutations, so that the LRL3/OsCYP2 gene in lrl3 encodes only 94 amino acids (ΔOsCYP2 in Fig. 3C) because of frame-shift mutation. To confirm that lrl3 is an OsCYP2 allelic mutant, we crossed lrl3 (deletion of 59 bp) with lrt2 (deletion of 50 bp), and obtained positive double-mutant F₁ plants. Phenotypic analysis of the wild-type plant zh11, lrl3, double mutant F₁ between lrl3 and lrt2, lrt2 and the wild-type plant Nipponbare showed that the double mutant F₁, lrl3 and lrt2 all exhibited lateral rootless phenotype. This result of genetic complementarity suggests that lrl3 lateral rootless phenotype is caused by OsCYP2 deletion mutation (Figs. 3D–3E). In addition, the homology analysis showed that the four amino acids of OsCYP2 were only present in the 12 plants analyzed, but not in human sapiens and yeast, suggesting that these four conserved amino acids are unique to plants during the evolution of the Cyclophilin family (see Supplementary Fig. 2).

There was no significant difference in leaf morphology and color between zh11 and lrl3 before low temperature treatment. However, compared with the wild-type plant zh11, lrl3 was more severe curled of leaf, yellowing of leaf tip and needle-like leaf shape after low temperature treatment (Figs. 4A–4B). Furthermore, because of lrl3 low seed-setting, we investigated the response of the OsCYP2 gene to low temperature using OsCYP2-overexpressed lines SSBM-OE and the wild-type plants SSBM. To clarify the different response of the OsCYP2-overexpressed lines SSBM-OE and the wild-type plants SSBM to low temperature stress, we carried out low temperature treatment of the OsCYP2-overexpressed lines SSBM-OE and the wild-type plants SSBM for 27 days, all the wild-type plants SSBM died while some of the OsCYP2-overexpressed lines SSBM-OE were still alive (Fig. 5). These results indicated that OsCYP2 positively regulated resistance to low temperature.

In order to detect downstream genes regulated by OsCYP2, we performed transcriptome analysis on zh11 and lrl3 before and after low temperature treatment, and found that a TPS gene OsTPS38 (Os08g0139700) showed a significant difference in expression between zh11 and lrl3 after low temperature treatment, which was contrary to the expression pattern of OsCYP2 gene, confirmed by qRT-PCR (Figs. 6A–6B). This suggests that OsTPS38 is negatively regulated by OsCYP2 to mediate the mechanism of low temperature response in rice. Meanwhile, OsCYP2 regulates the low temperature response mechanism in rice through OsCDPK7-dependent pathway (Fig. 6C).

To further clarify the molecular mechanism of OsTPS38 in response to low temperature, the overexpression lines ox-2 of OsTPS38 and the wild-type plant zh11 were used to undergo low temperature treatment for 3 days at three-leaf stage, and it was found that the degree of leaf curling in ox-2 was higher than that in zh11 (Figs. 7A–7B). These results suggest that OsTPS38 may negatively regulate the low temperature response in rice.

Auxin is the first plant hormone that has been recognized and studied by humans. It almost participates in the whole process of plant growth and development, and plays an important role in the hormonal regulatory network at all stages of growth and development [44–46]. Interestingly, the auxin-signaling mutants axr1, tir1, gps2-1, gps1, gps2 and gps3, which show a decreased gravity response, responded to low temperature treatment [47–49]. In our study, lrl3 mutant was also affected by gravity response and involved in response to low temperature, suggesting that OsCYP2 may be involved in the low temperature response mediated by the auxin pathway. Low temperature stress may affect the function of the Pin-formed proteins (PINs) by inhibiting the transport or localization of intracellular proteins, and eventually cause slow root growth and gravitation anomaly [50–53]. However, we found significant differences in the expression of OsPIN10a in OsCYP2 mutant compared with the wild type (see Supplementary Fig. 3). These results may suggest that OsCYP2 also responds to low temperature through the auxin pathway. There are few studies on the mechanism of auxin pathway regulating low temperature response, but WES1, which encodes IAA amino acid synthase, is up-regulated in low temperature stress, and activates the expression of stress-related genes Pathogenesis-related protein 1 (PR-1) and C-repeat binding factors (CBFs) by inactivating IAA, this indicates that the expression of the key genes CBFs in low temperature is directly regulated by auxin [54]. It has been reported that low temperature can affect the differential expression of some auxin response genes, such as IAA20 [55,56], while we also found that the expression of IAA20 gene was obviously affected by OsCYP2 under exogenous IAA treatment (see Supplementary Fig. 4). Otherwise, another study has confirmed that OsCYP2 can interact with OsIAA11 [37]. In conclusion, we speculate that OsCYP2 may be involved in the regulation of auxin response genes so as to mediate the cold response in rice (Fig. 8). However, the molecular mechanism of OsCYP2 regulating rice response to low temperature through auxin pathway needs to be further improved.

Plants can regulate the corresponding physiological and biochemical responses to adapt to low temperature by activating signal transduction pathways, and the response to low temperature can be divided into ABA-dependent pathway, Ca²⁺ sensing pathway and ROS sensing pathway [57]. The low temperature environment causes the hardening of the plasma membrane in rice, which leads to the Ca²⁺ channel opening and the extracellular Ca²⁺ entering the cytoplasm, resulting in the difference of intracellular and extracellular Ca²⁺ concentration. However, protein kinases such as CDPKs can sense the change of endogenous Ca²⁺ level and carry out phosphorylation reaction and activate the transcription factors of each downstream family, ultimately affecting the expression of COR genes. In this experiment, OsCYP2 is involved in the pathway in response to low temperature by CDPK7-mediated Ca²⁺ signaling, which is also known as the ABA-independent ICE-CBF-COR transcriptional cascade pathway [58,59]. Under low temperature stress, the AP2 conserved domain of CBFs can bind CCGAC (CRT, C-repeat), the core elements of CORs initiation region, to activate CORs at the transcriptional level, while CBF genes are usually regulated by ICE1. The transcription factor OsDREB1F may be activated by OsICE1, which then activates the expression of downstream low-temperature responsive COR genes harboring DRE/CRT domain [60]. In addition, OsICE1 is regulated by phosphorylation, sumoylation, and ubiquitination mediated by E3 ubiquitin ligase, and simultaneously binds to cis-elements on the promoter of CBF3/DREB1 to participate in the low temperature response [61]. However, some studies have suggested that CBF3 inhibition in ice1-1 is gene silencing caused by T-DNA-triggered methylation [62]. OsICE2 over-expression is involved in low temperature response of CBF1/DREB1 [63]. CBF2 is a negative regulator of CBF1 and CBF3, but not depending on the expression of CBF1 and CBF3 [64–66]. OsMYB3R-2 and OsMYB2 involved in cold resistance up-regulates the expression of OsDREB2A, while MYBS3 negatively down-regulates the expression of OsDREB1 and OsDREB2A [67–69]. Taken together, we propose that rice response to low temperature may occur through the following pathways (Fig. 8).

The rice seeds of zh11 (the wild-type cultivar in japonica background), lrl3, Oscyp2-1, SSBM (the wild-type cultivar in japonica background), SSBM-OE, Kasalath (the wild-type cultivar in indica background) and Oscyp2-2 in cold treatment were provided by Professor Xiaorong Mo of Zhejiang University, SSBM-OE transgenic lines were obtained by introducing the OsCYP2 CDS sequence into the vector 35S-pCAMBIA1300 (see Supplementary Fig. 5) and then the transformation of the above construct into SSBM wild-type calli by EHA105, the seeds of lrt2 and Nipponbare in allelic verification were provided by Professor Jianru Zuo of Chinese Academy of Sciences. The F₁ of lrl3 crossed by lrt2 was obtained in the Growth Chamber of Northeast Agricultural University.

In the hydroponic experiment, the pH was adjusted to 5.5 with 1 N NaOH in normal rice (O. sativa) medium [70]. Rice seeds were washed with distilled water, treated to break dormancy by 0.6% HNO₃ for 16 h at room temperature, and transferred to germinate in the incubator at 37°C. The germinated seeds were planted in normal rice (O. sativa) medium (3 L) on a nylon mesh floating, in rice culture chamber (day/night: 30°C/22°C, 12/12 h; Rh80%, 450 μmol photons m⁻² s⁻¹).

In order to study the effect of auxin on root growth and development, the seeds of zh11 and lrl3 were cultured under 0.1μM NAA (Sigma Aldrich, http://www.sigmaaldrich.com/) liquid nutritious medium. IAA and NPA were used in short-term auxin treatment.

The seeds of the wild-type zh11, lrl3 mutant, OsCYP2-overexpressed lines SSBM-OE and the wild-type plants SSBM were germinated and cultured in hydroponic solution. Then the one-leaf seedlings were transplanted into 6 cm × 6 cm × 6 cm culture pots in the incubator (HPG-280, Harbin Donglian Electronic Technology Development Co., Ltd., Harbin, China). The culture conditions were as follows: the diurnal temperature was 25°C/22°C, 12/12 h, and the relative humidity was 80%. Sampling for 0 h was conducted before chilling treatment, and zh11 and lrl3 were subsequently treated at 17°C at three-leaf stage. The culture conditions were as follows: the day and night temperature was maintained at 17°C/17°C, 12/12 h, and the relative humidity was 80%.

Seedling roots of zh11 and lrl3 were sampled on 7 and 14 days after sowing, and 5 samples were taken from each treatment for statistical analysis. Shoot height, primary root length and adventitious root length of the plants were measured with a scale. The number of adventitious roots was counted visually. Transmission Scanner STD1600 Scanner (Epson, Nagano Prefecture, Japan) was used to scan the primary roots of rice. WinRhizo (Regent Instruments Inc., Quebec, Canada) image analysis system was employed to calculate the number and length of lateral roots of the root system.

The seeds of zh11 and lrl3 were germinated and cultured on a nylon mesh in 5 L rice culture solution for two days, then the seeds were transferred to a plastic plate (11.7 cm × 11.7 cm) with the nutritious solution soaked in blue phosphorus-free paper, and the main roots were placed in parallel horizontally.

F₁ individuals were obtained by crossing lrl3 mutant (japonica rice zh11) as female parent and indica rice Kasalath as male parent. F₁ individuals self-cross to obtain F₂ genetic population. In 30 F₂ mutant plants, LRL3 gene responsible for lrl3 mutation was mapped to molecular markers SIS2 and RM12368. OsCYP2 controlling lateral root formation was detected in this region and named the candidate gene LRL3/OsCYP2. Supplementary Tables 1–2 provide details of all the mapping markers.

To test the function of LRL3/OsCYP2 for lateral root formation of lrl3 mutant, we crossed lrl3 by lrt2 mutant with 50 bp deletion in OsCYP2 to obtain double mutant F₁, and to verify the allelism of lrl3 and lrt2 genetically.

ClustalX 1.81 and Genedoc3.2 software were used to analyze the protein sequences of OsCYP2 with the homologous proteins of the biological cyclophilin-ABH-like domain, and the homology of their conserved domains was compared. Then the phylogenetic tree was constructed by Neighbor-joining method with MEGA3.1 software. The evolutionary standard bootstrap was 1000, and the results were output with MEGA3.1 software.

Total RNA was extracted using Trizol D0410 reagent according to the manufacturer’s instructions (Invitrogen, http://www.invitrogen.com/). The first cDNA strand was synthesized after 5 μg of total RNA was treated with Invitrogen II reverse transcriptase. Semi-quantitative RT-PCR and quantitative RT-PCR are described above [70]. Details of primers are shown in Supplementary Table 3. RT-PCR was performed at least three times with independent biological replicates, qRT-PCR was performed at least three times with independent technical replicates.

Rice seedlings were treated at low temperature and the leaves of zh11 and lrl3 seedlings under the normal and chilling treatment were sampled and quickly frozen in liquid nitrogen and stored at −80°C for later analysis. Novogene Bioinformatics Technology Co., Ltd., China was entrusted to complete RNA extraction, quality control, database construction and Illumina HiSeQTM sequencing.

According to the known sequence of OsTPS38 gene (http://rice.uga.edu/cgi-bin/ORF_infopage.cgi?orf=LOC_Os08g04500.1) and the cloning site of plant expression vector pBWA(V)HU-45001OE-Gus, specific primers (45001OE-F: cagtCACCTGCaaaacaacatggcaacctctgttccgagtgtacta; 45001OE-R: cagtCACCTGCaaaatacattaaacagagaggatgtagatggagtg) were designed. Then the OsTPS38 gene was amplified and introduced into the vector pBWA(V)HU-45001OE-Gus (see Supplementary Fig. 6). The construction and genetic transformation of OsTPS38-overexpressed vector were entrusted to Wuhan Biorun Biological Technology Co., Ltd. (China) to obtain OsTPS38-overexpressed positive lines in zh11 background.