Functional analysis of a dihydroflavonol 4-reductase gene in Ophiorrhiza japonica (OjDFR1) reveals its role in the regulation of anthocyanin

Plant materials

O. japonica materials used in this study were grown on the mountain in Shibing, Guizhou Province. For gene isolation and tissue-specific expression analysis, different samples including flowers of four developmental stages (1–4) and six vegetative tissues, i.e., flowers (Fl), roots (Ro), stems (St), leaves (Le), scapes (Sc) and calyxes (Ca) were collected as described in our previous paper (Sun et al., 2019). Arabidopsis mutant (tt3-1, AT5G42800, ABRC stock number: CS84) was in Landsberg-0 (Ler) ecotypic background and grown at 22 °C with a 16 h light/8 h dark photoperiod. For anthocyanin measurement and RT-PCR analysis, Arabidopsis seedlings of wild type, mutant and transgenic plants cultivated on 1/2 MS medium containing 3% sucrose were obtained. Tobacco plants (Wild-type and transgenic plants) were grown in the greenhouse, the flowers at full-bloom stage were harvested and then used for later analysis. All the samples above were frozen without delay by liquid nitrogen and kept at −80 °C.

Chemical standards

Dihydroquercetin (DHQ), dihydromyricetin (DHM) and dihydrokaempferol (DHK) were bought from Sigma-Aldrich (St. Louis, MO, USA) and prepared as 10 mg/mL solutions in methanol (chromatographic grade). Cyanidin 3-O-glucoside for drawing standard curve was also purchased from Sigma-Aldrich and diluted as one mg/mL solutions in methanol (St. Louis, MO, USA).

Gene cloning and sequence analysis

Flowers of O. japonica were selected for RNA extraction by using RNA pure Plant Kit (CWBIO, China) according to the instructions. Subsequently, one μg total RNA was used to synthesize the cDNA through the method described previously (Sun et al., 2019). Specific primers for cloning the full-length gene coding sequence were designed based on the assembled transcriptomic information. After PCR, the products of appropriate length were sub-cloned into the pMD18-T vector (Takara, Japan) and verified by sequencing. The list of primers used for gene cloning is provided in Table S1. Multisequence alignment of deduced protein sequences was analyzed using DNAMAN 6.0. And the Neighbor-Joining phylogenetic tree was constructed with MEGA version 7.0 software using 1,000 bootstrap replicates.

Quantitative real-time PCR analysis

RNA extraction and cDNA synthesis of flowers and other vegetative tissues were performed using the methods above. Primers for real-time amplification were designed by IDT1 and shown in Table S1. Then, qRT-PCR reactions were carried out with the BioRad CFX96 Real-Time PCR System (BIO-RAD, Hercules, CA, USA) and TransStart^® Green qPCR SuperMix (TRANSGEN, China). Thermal cycling conditions were 95 °C for 60 s, followed by 40 cycles of 95 °C for 5 s and 60 °C for 60 s. The OjActin gene and NtTubA1 gene were chosen as the internal reference for O. japonica and tabacum samples respectively. Each experiment sample was conducted in triplicate, and the relative transcript levels of target genes were analyzed by 2^−∆∆Ct method. Meanwhile, melting curve analysis and agarose gel electrophoresis were also employed for confirming the purity of PCR products.

Plasmid construction

In order to obtain His-tagged fusions of OjDFR1, its coding sequence containing the EcoR I and Hind III restriction enzymes sites was cloned into the pET-32a expression vector. Simultaneously, the complete ORF of OjDFR1 was also inserted into the binary vector pBI121 that previously digested with Xba I and BamH I. Then above resulting plasmids were transformed into competent cells and verified by sequencing. Primers used to generate the recombinant are present in Table S1.

Production of recombinant OjDFR1 protein and in vitro enzyme assay

The recombinant construct pET-32a-OjDFR1 and empty vector were introduced into E. coli strain BL21 (DE3) by the heat shock method. Next day, a single colony containing expression plasmid was inoculated in Luria–Bertani (LB) medium and grown at 37 °C with shaking 200 rpm until OD₆₀₀ reached 0.6. For induction, 0.2 mm of isopropyl-β-d-thiogalactopyranoside (IPTG) was added and the cells were further incubated at 15 °C for 24 h. After that, the cells were harvested centrifugation at 5,000 rpm for 10 min at 4° and resuspended in phosphate-buffered saline (PBS) for sonication, then the debris was removed by centrifugation at 12,000 rpm for 5 min at 4°. Subsequently, the His-tagged OjDFR1 proteins were purified by Ni-NTA pre-packed column (TransGen, China) and its purity was examined by using sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Protein concentration was estimated by NanoDrop 1,000 (Thermo scientific, Waltham, MA, USA) Spectrophotometer before in vitro enzymatic assays.

Three dihydroflavonols (DHQ, DHM and DHK) were selected as experimental substrates to analyze the substrate specificities of OjDFR1 following the methods depicted previously (Cheng et al., 2013). Shortly, a 500 µL reaction mixture containing 100 mm TrisHCl buffer (pH 7.0), 20 mm NADPH (50 µL), 10 mg/mL substrate (10 µL) and 35 µg OjDFR1 enzyme extract was conducted at 30 °C for 30 min. And the reaction products were identified by HPLC using a Shimadzu HPLC system with detection at 280 nm. An ACCHROM XUnion C18 column was used with the temperature at 30 °C and 20 µL samples were eluted with 1% H₃PO₄ (solvent A) and methanol (solvent B) according to the program as follows: 0 min, 15% B; 0–20 min, 15–60% B; 20–30 min, 60–15% B at flow rate of one mL/min.

Transformation of OjDFR1 into A. thaliana and tobacco

To further characterize the function of OjDFR1, the resulting vector pBI121-OjDFR1 and empty expression vector were maintained in Agrobacterium tumefaciens strain GV3101 by freeze-thaw method. Next, about 5∼6-week old tt3-1 mutant plants were used for genetic transformation followed the standard floral dipping transformation procedure (Clough & Bent, 1998). Transformants were screened on 1/2 MS medium with 50 mg L⁻¹ kanamycin to obtain T2 seeds. T2 transgenic seeds as well as wild type and mutant seeds were then cultured on anthocyanin induction media (1/2 MS medium supplemented with 3% sucrose) to set seedlings for phenotypic investigations and metabolite analysis. Meanwhile, the construct pBI121-OjDFR1 was also transformed into tobacco via the method described in previous reports (Imogen et al., 2006). The transgenic seedlings of tobacco (T1) were grown in green house, and their flower color was observed after flowering. For detecting the expression levels of OjDFR1, Arabidopsis actin-1 and NtTubA1 were selected as internal control genes.

Total anthocyanin quantifications by HPLC

The contents of anthocyanin were measured in both Arabidopsis seedlings and tobacco flowers according to previous method reported by predecessors (Li et al., 2017). Briefly, 0.1 g samples of seedlings as well as tobacco flowers were mashed into a fine powder and placed into extraction solution (water: methanol: hydrochloric acid = 75:24:1) at 4° C for 12 h. After centrifugation at 12,000 rpm for 5 min at 4° C, the supernatant was collected and filtered by a 0.22 μm nylon membrane filter for the following analysis. Chromatographic detection was performed on a Shimadzu HPLC system with the absorbance at 520 nm based on the procedure described by Li et al. (2017). Total anthocyanin content was quantified by using the external standard curve calibration of cyanidin 3-O-glucoside with three biological replicates (Fanali et al., 2001).

Statistical analysis

All experiments were repeated at least three times. The data were expressed as mean ± SD (standard deviation). We used a t-test to test for significant differences in the anthocyanin contents as well as gene expressions between wild type and transgenic tobaccos (OjDFR1-4 and OjDFR1-5). Difference with P < 0.05 was considered statistically significant.

Isolation and phylogenetic analysis of DFR-like genes from Ophiorrhiza japonica

After in situ TBLASTN search of O. japonica transcriptomic database, three putative genes which encoded NADP-binding reductases were obtained. Among them, one gene was named as OjDFR1 (cDNA sequences listed in the Supplementary Materials), which was most likely to be the bona fide DFR, because its protein sharing 69.74% identities to petunia DFR, and 67.98% identities to Nicotiana tabacum DFR. On the contrary, other two genes might encode ANR-like and FR proteins on the basis of manual BLASTX search, and thus were tentatively designated as OjANR1 and OjFR1.

Multisequence alignment with AtDFR and NtDFR showed that the three O. japonica proteins had the highly conserved NADP-binding motif and the substrate-binding domain at their N terminus (Fig. 2A). And the amino acid residue (at position 134) which is important for substrate specificity of DFR is Asn in OjDFR1 (encoding 357 amino acid residues with a calculated molecular mass of 39.87 kD) indicating that OjDFR1 may belong to the Asn-type DFR and catalyze the reduction of three dihydroflavonols (DHK, DHQ, and DHM) to leucoanthocyanidins. Moreover, phylogenetic analysis between OjFRs and other NADPH-dependent reductases (DFR, LAR, ANR, and CCR) are consistent with the above analyses. As shown in Fig. 2B, OjDFR1 protein was grouped into the eudicots subclade of DFR and exhibited most similar to the DFR from petunia. While, other two genes from O. japonica clustered outside the DFR branch, one fell into the subgroup of ANR, and another did not belong to a clear subclade, suggesting they might participate in other NADPH-dependent reduction during flavonoid biosynthesis.

Expression analysis of DFR-like genes in different organs

Before further analysis, all three genes were subjected to analyze their expressions at four flower developmental stages (1–4) and six different organs (Fig. 3A). It was observed that transcripts of three genes all exhibited significant spatial and temporal specificity. The relative levels of OjDFR1 transcripts peaked at stage 1 and decreased gradually with the flower development. For OjFR1 and OjANR1 genes, although they were also highly expressed in stage 1, but their transcript levels did not show correlation with the accumulation pattern of anthocyanin (Fig. 3B). Furthermore, expression pattern analyses in various tissues displayed that OjDFR1 showed higher transcript levels in scapes and calyxs accumulating high levels of anthocyanin than that of OjFR1 and OjANR1 (Fig. 3C). These data suggest that OjDFR1 appears to fulfill crucial roles in the anthocyanin biosynthesis in O. japonica.

Heterologous expression in Escherichia coli and in vitro biochemical characterization

To provide direct evidence that the OjDFR1 encoded the DFR enzyme, in vitro enzyme activity assay was implemented. The cDNA of OjDFR1 was therefore subcloned into the pET-32a expression vector and soluble recombinant proteins were successfully purified followed by SDS-PAGE identification (Fig. 4A). Next, the purified recombinant proteins were incubated with DHQ, DHM and DHK in the presence of NADPH, and the respective reaction products were characterized by HPLC through comparing to the UV spectra and the relative retention time (Cheng et al., 2013). As shown in Fig. 4B and 4C, HPLC analyses detected at 280 nm revealed that new peak was observed from reactions using DHQ and DHM as substrate, and these two peaks were not observed from control (protein from E. coli carrying the pET-32a empty vector). However, reduction of DHK to the respective leucopelargonidin was not observed (Fig. 4D). In fact, the recombinant OjFR1 and OjANR1 were also tested by DHQ, DHM and DHK, but no formation of the respective leucoanthocyanidin was observed (Fig. S1).

Functional analysis of OjDFR1 in Arabidopsis tt3-1 mutants

To validate the functional identity of OjDFR1, its cDNA was subcloned into the pBI121 expression vector containing the CaMV 35S promoter. Via floral dip method, OjDFR1 was introduced into Arabidopsis tt3-1mutant which failed to accumulate brown proanthocyanidins in the seeds and anthocyanins in the cotyledons and hypocotyls. Transgenic plants were selected on MS medium with 50 mg/L kanamycin, and then seeds of wild-type, tt3-1 as well as T2 transgenic lines were grown for 7 days on medium supplemented with 3% sucrose. Phenotypic investigation showed that transgenic OjDFR1 plants successfully restored the coloration of their seeds and hypocotyls (Fig. 5A), while the empty vector control was still green (raw data). In order to confirm this phenotype was produced by ectopic expression of OjDFR1, RT-PCR was carried out. As expected, OjDFR1 was highly expressed in transgenic plants, and no amplicons were observed in the control (Fig. 5C).

In addition, 7-day-old seedlings were also used to quantitatively determine the amounts of anthocyanins, and the results indicated that anthocyanins level in transgenic Arabidopsis was higher than mutant, and similar to wild type (Fig. 5B). As results present in Fig. 6, anthocyanins were not detected in tt3-1, but these absent peaks were fully complemented in transgenic seedlings expressing OjDFR1, though the contents were lower. Totally, our anthocyanin analyses strongly prove that O. japonica DFR gene is functional for the biosynthesis of proanthocyanidins and anthocyanins in vivo.

OjDFR1 contributes to darker flower color in transgenic tobacco

In order to characterize the effects of OjDFR1 transgene on anthocyanin profiles in flowers, it was overexpressed in tobacco plants. Totally 12 independent transgenic lines that confirmed by gene-specific PCR analysis were obtained. Based on visual observations, transgenic tobacco plants possessed darker pink flowers than wild type (Fig. 7A). Then the presence of OjDFR1 was examined by RT-PCR (Fig. 7B). Additionally, anthocyanin levels in corollas were also measured through HPLC (Fig. S2). Compared with control, corollas of transgenic tobacoo overexpressing OjDFR1 contained significantly higher amount of anthocyanins on fresh weight basis (Fig. 7C), which implying OjDFR1 protein might interact with the endogenous enzymes of anthocyanin pathways in tobacco and lead to increased anthocyanin accumulation in transgenic flowers. For further investigate the effect of OjDFR1 on endogenous tobacco in anthocyanin biosynthetic genes, we performed qPCR analysis. Among these genes (NtCHS, NtCHI, NtF3’H, NtF3’5’H, NtDFR, NtANS, NtUFGT, NtAN1a, NtAN1b and NtAN2), the expressions of NtANS, NtUFGT and NtAN2 were higher in transgenic tobacco flowers, whereas the transcripts levels of NtCHS, NtCHI, NtAN1a and NtAN1b were slightly lower in transgenic flowers than in wild-type (Fig. 8).