The purple-fleshed sweet potato cv. YUZI263 was used to clone the gene and analyze gene expression and anthocyanin contents. Sweet potato YUZI263 were planted in the field and 5-month-old plants were harvested and frozen in liquid nitrogen for future using. Total RNAs and anthocyanins were extracted from storage roots (Φ 0.5 cm, Φ 3.0 cm), veins, petioles, leaves and vein tips. Nicotiana tabacum was used as plant materials for genetic transformation.

The total RNAs of tubers were used as initial material for gene cloning. The full-length cDNA of IbDHKR was obtained using rapid amplification of cDNA ends (RACE) technology. A pair of primers, fibdhkr-core (5′-CTGCTGGCTTTATCGGCTCC-3′) and ribdhkr-core (5′-GGCACTGTTGTGGCCTCTTG-3′), were used to amplified the core fragment of IbDHKR. The 3′ end of IbDHKR was amplified using a gene-specific primer ibdhkr3 (5′- GGAAGAAGGGGTTACTTCCCTA-3′) and NUP (5′-AAGCAGTGGTATCAACGCAGAGT-3′). The 5′ end of IbDHKR was amplified using a gene-specific primer ibdhkr5 (5′-GGATCCTTGGAATCGAAATCCATA-3′) combined with NUP. The coding sequence of the full-length IbDHKR was amplified using a pair of primers, fibdhkr (5′- ATGGTGGACGGTAATCATCCAA-3′) and ribdhkr (5′- TCAAGCTTTTAAGGGCACTACC -3′). Homology and sequence identity were confirmed using the BLAST program.

Anthocyanins were extracted according to the reported proce dures with minor modification. The solution containing anthocyanins was measured by UV Spectrophotometer U-1100 (HITACHI, Japan), the absorbance at 530 nm was measured three times and the average values were represented for the relative contents in different tissues.

Competent Escherichia coli cells of the strain BL21 (DE3) were transformed with the pET28a(+)-IbDHKR constructs. The E. coli cells were subcultured to a density of OD₆₀₀ = 0.5 in Luria–Bertani medium containing kanamycin (100 mg/l). IPTG was added to a final concentration of 0.5 mM for the induction of protein expression, and the bacteria were harvested following an additional 16 h of growth at 16°C and stored frozen at –70°C. The recombinant IbDHKR proteins were partially purified from crude bacterial extracts and monitored by SDS-PAGE and detection with the Anti-His antibody and visualized using WesternBreeze^TM Chemiluminescent kit (Invitrogen, Carlsbad, CA, USA), the Western blot analysis was preferred according to the kit manufacturer’s protocol. The 6× His tagged recombinant proteins were purified using Ni-NTA columns under native conditions. Substrate-feeding assay was conducted according to Xie et al. (2004) with some modifications. The initial standard assay conditions included incubation at 30°C for 30 min in 100 μl total volume containing 48 μl of 100 mM Tris HCl (pH7.0), 40 μl of 0.45 μg/μl IbDHKR purified protein, 10 μl of 10 mM NADPH, and 2 μl of 10 μg/μl substrate (DHK). Enzyme reactions were stopped by adding 200 μl of ethyl acetate and vortexing. Samples were centrifuged for 2 min at 10,000g and 150 μl of ethyl acetate extract was removed to a new tube and evaporated with a stream of nitrogen gas. Residues were dissolved in 30 μl of methanol and used directly for LC/MS analysis. An Agilent 6320 ion trap LC/MS system was used under default Smart Parameter settings. The analyzer and ion optics were adjusted to achieve optimal resolution using the ESI Tuning Mix. The mass spectrometer was scanned in the m/z range of 50–2200 at 8100 mass units/s with an expected peak width of ≤0.35 mass units. For automated MS/MS, the trap isolation width was four atomic mass units. The associated Agilent 1200 LC was fitted with a Zorbax 80A EXTEND-C18 column (150 mm × 2.1 mm, 5 μ particle size) maintained at 35°C. The binary solvent system consisted of 90:10 (v/v) water/acetonitrile containing 0.1% formic acid and 0.1% ammonium formate (solvent A) and 10:90 (v/v) water/acetonitrile containing 0.1% formic acid and 0.1% ammonium formate (solvent B). The separation gradient was 100% A to 100%B in 4.5 ml over 30 min.

The coding regions from IbDHKR was subcloned into the Bgl II and BstE II site of the binary vector pCAMBIA1304⁺ (Kai et al., 2009), replacing the mGFP5 and gusA genes. The pCAMBIA1304⁺-IbDHKR derivative and un-modified pCAMBIA1304⁺ were introduced into Agrobacterium tumefaciens EHA105 bacteria and used to transform tobacco leaf disks. Transformation and regeneration of tobacco leaf disks used the common methods. Agrobacterium was cleared from the explants using 250 mg/ml carbenicillin. The disks were cultured on the MS containing 0.1 mg/L NAA, 1.0 mg/L 6-BA, 250 mg/L carbenicillin (Cb), and 20 mg/L hygromycin (Hyg) at 25°C underlight for one month, then subcultured on MS with 250mg/L Cb and 20mg/L Hyg for 4 to 5 weeks to regenerate shoots. Subsequently, the regenerated shoots were excised from the original explants and cultured on MS (IBA 0.1 mg/L + Hyg 20 mg/L + Cb 250 mg/L) to grow to plants. Mature plants were transferred to pots and maintained under standard greenhouse conditions. The transgenic plants were checked by PCR method using the primers: F-gp (5′- ACTATCCTTCGCAAGACCCT-3′) and R-gp (5′- ATCACTCCAGCAGCTCTCATC-3′).

The full-length cDNA of IbDHKR (GenBank accession No. HQ441167) was 1392 bp in length with a 1185-bp coding sequence encoding a 394-amino-acid polypeptide with the calculated molecular weight of 44 kD. The BLAST searching results indicated that IbDHKR showed higher similarity with the DFRs from Ipomoea nil (94% similarity), Nicotiana alata (77% similarity), and Arabidopsis thaliana (58% similarity). The NADPH-binding motif (V10-Y32) (Shimada et al., 2005), as well as the substrate-binding domain (T133-K158) (Beld et al., 1989; Johnson et al., 2001), was found in IbDHKR according to the multiple alignments (Figure 2). The homology-based modeling showed that the structure of IbDHKR was very similar with the grape DFR protein, the NADPH/substrate-binding motif exposed on the surface (Petit et al., 2007). The phylogenetic analysis indicated that the plant DFRs were divided into eudicot type and monocot type because the DFRs from eudicots plants formed a clade separate from the clade composed of monocot DFRs (Figure 3). The bioinformatic analysis suggested that IbDHKR belonged to the DFR family.

Purple-fleshed sweet potato YUZI263 is the first cultivar authorized by China Ministry of Agriculture, which has high-yield starch and contains rich anthocyanins in tubers (Zeng et al., 2013). The tissue profiles of two anthocyanin-biosynthetic genes including flavanone 3-hydroxylase (IbF3H) (Figure 4A) and IbDHKR were analyzed in different organs by qPCR (Figure 4B). IbDHKR expression level was much higher in storage roots (Φ 3.0 cm or 0.5 cm) than in up-ground organs including veins, petioles, leaves and vein tips. For example, the expression level of IbDHKR was about 9.4-fold higher in storage roots (Φ 3.0 cm) than in petioles. Notable, bigger storage roots (Φ 3.0 cm) had significantly higher expression level of IbDHKR (3.1-folds) than slimmer storage roots (Φ 0.5 cm), suggesting that IbDHKR expression was associated with YUZI263 storage root development. For sweet potato IbF3H, it was also highly expressed in storage roots, just like the tissue profiles of IbDHKR. The previous research reported that anthocyanidin synthase (IbANS) of YUZI263 was expressed preferentially in storage roots in which the anthocyanin content was much higher than other organs (Liu et al., 2010). The root-preferential tissue profiles of anthocyanin-biosynthetic genes suggested that storage roots were the main organs to synthesize anthocyanins. The coordinate expression of the anthocyanin-biosynthetic genes facilitated anthocyanin biosynthesis in storage roots. Analysis of the anthocyanin contents in different organs indicated that storage roots also contained much higher level of anthocyanins than any other organ (Figure 4C). The anthocyanin content in storage roots (Φ 3.0 cm) was about 10–58 folds of that in up-ground organs. The storage roots with 3 cm diameter had higher level of anthocyanins than storage roots with 0.5 cm diameter, suggesting that anthocyanin accumulation was positively linked with development of storage roots. These results on analyzing gene expression and anthocyanin accumulation showed that the anthocyanin accumulation was strongly in accordance with the expression levels of anthocyanin-biosynthetic genes in purple-fleshed sweet potato YUZI263.

To unveil the stereospecific reduction, the recombinant protein of His-tagged IbDHKR was purified from engineered E. coli. When engineered E. coli was induced by IPTG, a specific 44 kD protein band was found, but this band was not found in engineered E. coli (pET28a(+)-IbDHKR) without IPTG treatment and in E. coli with the blank vector pET28a(+). With the extension of IPTG induction time, this protein band became thicker and thicker. In 4-hour induction of E. coli by IPTG, engineered E. coli produced the specific protein of which quantity was enough for analysis (Figure 5A). After harvesting ultrasonically broken E. coli using centrifuge, this protein was also found in supernatant of the broken debris of engineering E. coli (Figure 5A), facilitating protein purification. Then, the recombinant His-tagged IbDHKR was purified using Ni²⁺ column, of which molecular weight was about 44 kD (Figure 5B), consistent with the calculated molecular weight. The molecular weight of IbDHKR was similar with the reported DFRs from Camellia sinensis (L.) O. Kuntze (Singh et al., 2009), Medicago truncatula (Xie et al., 2004) and Vitis vinifera (L.) (Petit et al., 2007). Then, the recombinant His-tagged IbDHKR was detected using antibody against His-tags by western blot (Figure 5B). Finally, three different types of dihydroflavonol substrates including DHK, DHQ and DHM were, respectively, fed to the recombinant IbDHKR. The results indicated that IbDHKR catalyzed the reduction of DHK alone to produce leucopelargonidin that was confirmed by LC-MS (Figure 5C). When DHQ or DHM was fed to recombinant IbDHKR, no products were detected. These results suggested that this recombinant protein was a dihydroflavonols reductase specifically reducing DHK (dihydrokaempferol reductase) to produce leucopelargonidin.

In order to investigate the effect of overexpressing IbDHKR on anthocyanin biosynthesis, IbDHKR was used for genetic transformation of Nicotiana tubaccum. Five independent plants were obtained on screening media with hygromycin, but only three of them (Line 1, 7, and 9) had the hygromycin-resistant gene and IbDHKR (Figure 6), suggesting that they were the authentic transgenic plants. Then the three lines were, respectively, propagated and used for anthocyanin content analysis. The expression of IbDHKR was detected in Line 7 and 9, and was not detected in Line 1. Line 9 showed much higher expression level of IbDHKR than Line 7. The flowers of Line 7 showed no visible difference with the flowers of wild-type tobacco. This might be due to the low expression level of the transgene IbDHKR. Compared with the light pink corolla of wild-type tobacco, the corolla of Line 9 was more highly pigmented (Figure 6). In Line 9, the anthocyanin content was about 1.5-fold higher than the wild-type. Overexpression of cranberry DFR in tobacco increased the anthocyanin accumulation in tobacco corolla that kept more highly pigmented than control (Polashock et al., 2002). Because only one transgenic line (Line 9) showed very higher expression level of IbDHKR, the definitive conclusion of overexpressing IbDHKR on anthocyanin accumulation needed more independent transgenic lines. In fact, overexpression of DFR definitively promoted anthocyanin biosynthesis in many plants (Xie et al., 2004; Lee et al., 2008; Huang et al., 2012; Chu et al., 2014).