Three kiwifruit cultivars of contrasting coloration were selected (Table 1). Sampling dates for each cultivar are reported as days after pollination (DAP). Samples were taken about every 20 days (Supplementary Table S1). At each sampling time, 20 fruits were collected at random from three vines of each of the three cultivars. Fruits were collected of each genotype at seven preharvest stages: on 25, 45, 65, 85, 105, 125, and 145 DAP. ‘Hongyang’ and ‘Hayward’ are commercial cultivars in China and their times of harvest were determined according to the normal industry criteria: ‘Hongyang,’ when the soluble solids content (SSC) was ≥7.0° Brix, and ‘Hayward,’ when the SSC was ≥6.5° Brix. Meanwhile, ‘Jinnong-2’ is a commercial cultivar in Shaanxi province but not widely grown outside China. We set its commercial harvest time as being when the SSC ≥ 10.0° Brix because this criterion is commonly used for screening breeding populations of A. chinensis. For a long time, the dry matter content of ≥15.0% has been a industry criterion for determining harvest time for kiwifruit. By 145 DAP, the fruits of all three cultivar had reached commercial harvest maturity (Table 1). Fruits were then harvested and stored at room temperature (c. 22°C). A total of 20 fruits of each genotype were examined after 10 days storage (155 DAP) and a further 20 fruits at full ripeness – fruit firmness was about 9.8 N on 170 DAP (‘Hongyang’ and ‘Jinnong-2’) or on 190 DAP (‘Hayward’).

Of the 20 fruits of each cultivar at each stage, 10 fruits were used to measure the various physiological parameters (i.e., SSC, dry matter content, firmness and color indices). After peeling (removing the skin to approximately 1 mm depth), the other 10 fruits were carefully separated into outer and inner pericarps (seeds were removed) and the separated tissues immediately frozen in liquid nitrogen. Petals, stems, leaves and ovaries were collected from vines of ‘Hongyang.’ Fruits of another three, red-fleshed cultivars (‘Qihong,’ ‘Donghong’ and ‘Purpurea’) and of three, green-fleshed cultivars (‘Xuxiang,’ ‘Cuixiang’ and ‘Jinkui’) were also collected at maturity. All samples were taken from the National Center of Kiwifruit Breeding, Mei County, Shannxi province, China. The tissues were frozen in liquid nitrogen and stored at -80°C pending analysis.

Fruit flesh firmness was determined by penetration at the two opposite cheeks of each fruit after removal peel (1 mm thick, 1 cm²) using the GUSS Fruit Texture Analyzer (GS-15, Strand, South Africa) with a 8 mm plunger (Minas et al., 2016). The same fruit was used to measure the SSC, expressed as ° Brix, by taking two juice samples from the equatorial part of each fruit and measuring with a hand-held refractometer (PAL-1, Atago, Japan). The 2 mm transversal slice came from the equator of each fruit was uesd to measure the dry matter content. The fruit slices after weighed their fresh weights (FW) were then dried at 105°C in a vacuum oven (DZF-6050, Shanghai, China) to constant weight (DW). The dry matter content is FW/DW ^∗ 100% (Famiani et al., 2012). Three repeated were carried out for these physiological parameters.

The colors of the outer and inner pericarp tissues at the various stages were measured using a chroma meter (CR-400, Konica Minolta, Japan) based on the CIE L^∗, a^∗, b^∗ mode (Hanbury and Serra, 2002). Five random measurements were carried out on the outer and inner pericarps of each fruit transverse section. The data are described as L^∗ (lightness), a^∗ (red-green sensation), b^∗ (yellow-blue sensation) and h° [hue angle, tan^-1(b^∗/a^∗)]. Three replicates were performed for each sample.

Chlorophyll and carotenoids were extracted from the fruits following the method of Montefiori et al. (2009b), and quantified using the method of Tian et al. (2014). Three replicates were performed for each sample.

The anthocyanins were extracted with 1% (v/v) hydrochloric acid/methanol and held for 24 h at 4°C in the dark. The supernatant was filtered through a 0.22 μm syringe. Extracts were analyzed as previously described (Bi et al., 2014). Three replicates were performed for each sample.

Total RNA was extracted using the Plant RNA Kit (Omega Bio-tek, Norcross, GA, United States). The RNA concentration and quality were determined by UV spectrophotometry and by running on a 1.0% agar ethidium bromide-stained gel. Approximately 1 μg of total RNA was used for cDNA synthesis with the PrimeScript RT reagent kit (TaKaRa, Dalian, China). Quantitative real time PCR was carried out with SYBR Premix ExTaq II Kit (TaKaRa, Dalian, China), and amplification was monitored on an Icycler iQ5 (Bio-Rad, Berkeley, CA, United States) in a reaction volume of 20 μL.

The amplification program consisted of one cycle of 95°C for 40 s followed by 40 cycles of 95°C for 30 s and 59°C for 30 s. Melting curve analysis was carried out after 40 cycles to ensure the proper amplification of target fragments. Actin was used for normalization. All analyses were repeated three times using biological replicates. Primer sequences are listed in Supplementary Table S2.

Using the sequences of MYB, bHLH and WD40 which have been identified to promote anthocyanin synthesis in other plants, we blasted the genome of kiwifruit¹, resulting in five MYBs, five bHLH and two WD40 members being identified. The nucleotide sequences were used for phylogenetic analyses. Sequences were aligned using ClustalW and adjusted manually as necessary. The resulting data were analyzed by the Neighbor Joining method using the MEGA 6.0 program. They were named based on the results of the phylogenetic analyses.

Full-length cDNAs of AcMYBF110 were amplified from ‘Hongyang’ fruits by specific primers (given in Supplementary Table S2) based on the available sequences in the kiwifruit genome database¹. The amino acid sequences encoded by AcMYBF110 and anthocyanin-promoting MYBs from various plants were used for multiple alignment using DNAMAN software.

The coding region without stop codon of AcMYBF110 with C-terminal GFP fusion was inserted into the multiple cloning site of plant binary expression vector pVBG2307 (Ahmed et al., 2012) to form 35S:AcMYBF110-GFP. The primers are shown in Supplementary Table S2. Vector which inserted only with the GFP gene (35S:GFP) was used as positive control. Two constructs were transformed to Agrobacterium tumefaciens strain GV3101 using a freeze-thaw method. Then the Agrobacterium strains were infiltrated into 6-week-old Nicotiana benthamiana leaves. After incubating at 24°C with 16 h light for 48–72 h, the fluorescence was observed and DAPI was used to locate the fluorescent proteins in the nucleus.

Six-week-old N. tabacum leaves were used for infiltration. The above strains containing 35S:AcMYBF110-GFP and control (35S:GFP) were infiltrated into the abaxial leaf surface. Each infiltration was carried out on three leaves. The infiltrated plants were placed in the dark at 24°C overnight and then transferred to a growth chamber at 24°C with a 16 h light/8 h dark cycle under low light conditions. Plants were photographed and sampled 6 days after being infiltrated. All samples were frozen in liquid nitrogen and stored at -80°C, pending analyses.

Colors of all three kiwifruit cultivars were green at the beginning of development and with a similar hue angle – both the outer and inner percarps (Figures 1A–D). Obvious redness appeared in the inner percarp of ‘Hongyang’ (HY) at 85 DAP, while for its green outer percarp, as well as both outer and inner pericarps of ‘Hayward’ (HWD), remained green throughout, accelerated color change started after 155 DAP in both outer and inner pericarps of ‘Jinnong-2’ (JN) (Figures 1A–C). Hue angle decreased dramatically during development in inner pericarp of HY, followed by JN (both outer and inner pericarps), but less so in three green pericarps (Figure 1D). A PCA analysis also showed that all outer and inner pericarps of three cultivars were divided into red, yellow, and green groups (Figure 1E). Combining the colors of the three cultivars can be described satisfactorily using hue angle which is used in the following analyses.

Based on the observations of the phenotypes and on the analyses of hue angles in the fruits of the three kiwifruit cultivars, six stages could be identified for pigment analysis. These stages were S1, S2, S3, S4, and S5 referring to fruits at 45, 85, 125, 145, and 155 DAP, respectively and S6 when fully ripe.

In the outer pericarp tissues of the three kiwifruit cultivars, the chlorophyll b content of JN was higher than that of HY and that of HWD (lowest) before fruits were harvested (Figure 2A). But at the last, it was only 1.16 μg⋅g^-1 FW in JN, significantly lower than that in HY and HWD. So did the contents of chlorophyll a (Figure 2B). In the inner pericarp tissues, the contents of both chlorophyll b and chlorophyll a reached peaks at S2 for all three kiwifruit cultivars, followed by a decrease until S6. At this stage the contents were significantly higher in HWD than that in HY and JN (Figures 2E,F).

The contents of lutein in the outer pericarp tissues generally decreased slightly but fluctuated somewhat over the development and ripening stages, showing no obvious pattern (Figure 2C). It is noteworthy that the content of beta-carotene in HY and HWD generally decreased, but in JN it increased strongly at post-harvest (Figure 2D). This change was well correlated with the visible phenotype, indicating beta-carotene seems to be associated with the formation of yellow color in the yellow-fleshed kiwifruit. Further, it was found that in the inner pericarp tissues of all three kiwifruit, the contents of both lutein and beta-carotene in HWD peaked at S2 and then gradually decreased (Figures 2G,H). In contrast for HY and JN, the content of lutein showed a slight decrease during development, while the content of beta-carotene increased strongly post-harvest as well as showing change in the outer pericarp of JN.

To better analyze the pigment differences among the three kiwifruit cultivars, anthocyanins were also detected by HPLC. Unlike carotenoids and chlorophylls detected in all samples, anthocyanins were only detected in the fruit of HY and cyanidin 3-O-xylogalactiside was the major component (Supplementary Figure S1). In the fruit of HY, only traces of anthocyanins (or none at all) were detected in the early growth stages in both the inner and outer pericarps of HY (Figures 3A,B). In the inner pericarp, it started to increase markedly from S2 until S6, while in the outer pericarp, it changed very slightly – very much lower than in the inner pericarp.

In summary, these results suggest the yellow color of kiwifruit is due not only to the degradation of the chlorophylls but also to the increase in beta-carotene. While the red color of the red-fleshed kiwifruit is due to anthocyanins accumulation.

To further investigate what controls the color diversity of green-, yellow-, and red-fleshed kiwifruit, the inner pericarps of the three cultivars were used to analysis the expressions of the key structural genes involved in pigment biosynthesis and degradation using qPCR.

The genes involved in chlorophyll biosynthesis and degradation were measured (Figures 4A,B). CAO and RBCS expressions varied considerably throughout fruit development in the three kiwifruit cultivars. In contrast, GLUTR expression showed higher levels for all stages in the green-fleshed kiwifruit than that in the red- or yellow-fleshed ones. The expression of CBR and PAO showed no obvious regularity for three kiwifruit cultivars. PPH expression in green-fleshed kiwifruit reached a peak, higher than in other two kiwifruit in S4 but except for this peak, overall PPH expression was higher in the yellow-fleshed kiwifruit in the later stages. SGR expression in red- and yellow-fleshed kiwifruit showed an obvious increase with a peak at S5 that was very much higher than in the green-fleshed kiwifruit.

Among six carotenoid biosynthesis genes, ZDS was not detectable in any samples (data not shown), the relative expression of PSY, PDS, CRTISO and LCY-𝜀 in the green-fleshed kiwifruit was higher at most stages than in the other two kiwifruits (Figure 5A). While expression of LCY-β showed an increase in all three cultivars, the higher levels were measured in the red- and yellow-fleshed kiwifruits than in the green-fleshed one. As for the carotenoid degradation genes CCD and NCED, both of them showed the highest expression level in the yellow-fleshed fruit at early stages. However, in the final stage, their expressions in yellow-fleshed kiwifruit was significantly lower than in the other two (Figure 5B).

The expression profiles of the anthocyanin biosynthetic pathway genes, including CHS, CHI, F3H, F3′H, DFR, ANS and UDP flavonoid glycosyltransferases (UFGT), were investigated in three kiwifruit during fruit development (Figure 6). Apart from UFGT, most of the others, were expressed in both the red and also in the non-red tissues. In contrast, UFGT was highly expressed in the red inner pericarps but not in the green or yellow inner pericarps.

To find significant statistical correlations between gene expression and flesh color, the Pearson’s correlation coefficient (r) test was used with the above qPCR data, pigment contents and hue angles. SGR showed a significant positive correlation with the degradation of chlorophylls (Supplementary Table S3), and LCY-β correlated well with the beta-carotene and hue angles (Supplementary Table S4). Also, UFGT showed a significant negative correlation with hue angle and a significant positive correlation with anthocyanin concentration (Supplementary Table S5). Apart from the genes identified above, none of the others showed significant correlations with flesh color.

Structural plant genes are usually regulated by transcription factors, but we found no transcription factors regulating the chlorophylls and carotenoids from the genome of kiwifruit² based on previous reports. While twelve anthocyanin transcription factors, including five MYBs, five bHLHs and two WD40s, were screened in kiwifruit and named according to the phylogenetic analysis (Supplementary Figure S2) (Hichri et al., 2010; An et al., 2012; Liu et al., 2013b; Schaart et al., 2013; Tuan et al., 2015; Lai et al., 2016). The expression profiles of these kiwifruit transcription factors were then examined by real-time qPCR (Figures 7A–C).

Except for the first stage, MYB5b was either not expressed at all or showed extremely low expression levels for all samples at all stages of fruit development (Figure 7A). In contrast, MYB5a, MYB5c and MYB3 were expressed in all samples and showed higher expressions in green and yellow fruits than in red fruits. Overall, the variations of these four MYBs were correlated neither with anthocyanin content nor with the expression of UFGT (Supplementary Table S6). However, the transcript level of MYBF110 in all three fruits showed significant correlations with both anthocyanin content (|r| = 0.550) and with the expression of UFGT (|r| = 0.609). Especially during the coloring stages (S2 to S4), its expression increased and reached maximum value at S4 in the red fruit, while for the green and yellow fruits, it decreased, showing significantly lower levels than in the red fruit (Figure 7A). These findings suggest an essential role for AcMYBF110 in the anthocyanin biosynthesis of red-fleshed kiwifruit.

Of the five bHLH genes, all showed much higher expressions in the green fruit than in the red and yellow ones at all stages of development (Figure 7B). For bHLH1, bHLH2 and bHLH3, it increased until S4, at which stage expression reached a maximum and then decreased in the later stages. For bHLH4 and bHLH5, expression fluctuated in the early stages but showed an overall decreasing trend during fruit development. All five genes showed no obvious correlation with anthocyanin biosynthesis (Supplementary Table S6).

Lastly, the expression of both WDR1 and WDR2 was significantly higher in green fruit than in red and yellow fruits at all stages (Figure 7C), being not consistent with anthocyanin content and UFGT expression (Supplementary Table S6), although it showed a tendency to increase in red fruit at the coloring stages.

The full-length cDNA of AcMYBF110 was cloned from ‘Hongyang’ fruit. The predicted AcMYBF110 protein contained a highly conserved N-terminal R2R3 repeat of a DNA-binding domain with a bHLH motif (Zimmermann et al., 2004). Second, the ANDV motif (Box A) and the [R/K]Px[P/A/R]xx[F/Y] motif (Box B) are present in AcMYBF110 (Figure 8A). Both of these motifs are well conserved in anthocyanin-promoting MYBs (Kranz et al., 1998; Lin-Wang et al., 2010). AcMYBF110 had a high degree of homology with other plant anthocyanin-promoting MYBs (Figure 8A). For example, the amino acids in the R2R3 DNA-binding domain share 95.15% identity to AcMYB110, 83.50% identity to LcMYB1, 82.52% identity to FaMYB10, and 81.30% identity to MrMYB1.

To examine whether the expression patterns of AcMYBF110 coincided spatially or inter-varietally with anthocyanin accumulation in kiwifruit, real-time qPCR was used to investigate its expression levels in different tissues of ‘Hongyang’ and in the other cultivars. The results showed significantly higher levels of expression were found in two red tissues and in the three red-fleshed cultivars (Figure 8B). These correlated well with the high anthocyanin contents of these samples. In contrast, lower levels were observed in the other green tissues in which anthocyanins were barely detectable. These findings further confirm that AcMYBF110 may be involved in regulating anthocyanin biosynthesis in kiwifruit.

To determine the subcellular localization of the putative protein encoded by AcMYBF110, the transient expression assay of this protein fused to the green fluorescent protein (GFP) was carried out in 6-week-old N. benthamiana leaves. As shown in Figure 8C, the AcMYBF110-GFP fusion protein signal was observed both in the nucleus and cytoplasm. This suggests the AcMYBF110 protein could act as a transcriptional regulator in plant cells.

To further confirm the regulation role of AcMYBF110 in anthocyanin biosynthesis, the transcriptional activity of AcMYBF110 was tested using a tobacco transient color assay. The above fused construct under the action of 35S (35S:AcMYBF110-GFP) was syringe-infiltrated into the abaxial surfaces of expanding N. tabacum leaves. The empty vector (35S:GFP) served as a control. An intense red pigmentation was observed at the infiltration sites 6 days after being transformed with 35S:AcMYBF110-GFP (Figure 8Di), while the infiltration sites remained green in leaves transformed with 35S:GFP. Anthocyanin extraction and HPLC analysis showed the leaves transformed with 35S:AcMYBF110-GFP contained anthocyanin but not those with 35S:GFP (Figure 8Dii). Moreover, AcMYBF110 was highly expressed in leaves transformed with 35S:AcMYBF110-GFP but no transcript was detected in leaves transformed with 35S:GFP (Figures 8Diii,iv). We then carried out qPCR analyses of the expression of key anthocyanin synthesis genes in tobacco leaves, including NtDFR, NtANS and NtUFGT. The expression levels of these were higher in the leaves transformed with 35S:AcMYBF110-GFP than in those transformed with 35S:GFP by 17.73-, 2185.20-, and 801.58-times, respectively.

Chlorophyll, carotenoid and anthocyanin are the three most important color pigments in plant tissues, including in fruit. We measured the concentrations of these three pigments in the outer and inner pericarps of three different-colored kiwifruits at various developmental stages. The concentrations of chlorophyll b and chlorophyll a showed similar variation trends in all kiwifruits (Figures 2A,B,E,F) but were significantly lower in the yellow fruits (JN) than that in the red (HY) or green (HWD) fruits when fully ripe. Meanwhile, the beta-carotene content decreased slightly in the early stages in all three cultivars. Subsequently, beta-carotene increased significantly until S6 in JN but remained low in HY and HWD (Figures 2C,D,G,H). In contrast, there was no obvious pattern for lutein content, which agrees with Ampomah-Dwamena et al. (2009). These results indicate that the formation of yellowness in kiwifruit is related not only to a degradation of chlorophylls but also to an increase in beta-carotene (Ampomah-Dwamena et al., 2009; Pilkington et al., 2012). Anthocyanin was detected only in HY fruits with very low or no detectable levels found in the green outer pericarps from S1 to S6 of and green inner pericarp at S1, while higher levels were detected in the red inner pericarps from S3 to S6 (Figure 3). This indicates anthocyanin is the pigment primarily responsible for the formation of the red color in the inner pericarp of HY (Li et al., 2015; Man et al., 2015).

Chlorophyll de-greening is a coordinated process. It seems to be involved with a down-regulation of chlorophyll biosynthesis and an up-regulation of chlorophyll degradation. SGR proteins have been shown previously to play a critical role in the initiation of chlorophyll degradation and senescence (Luo et al., 2013; Sakuraba et al., 2015). We measured the transcript levels of the chlorophyll biosynthesis and degradation genes and found the genes are expressed in all three cultivars (Figure 4). Based on correlation analysis between gene expression and hue angles and chlorophyll content, SGR showed a significant positive correlation with the degradation of chlorophylls (Supplementary Table S3). This is in agreement with results obtained by Pilkington et al. (2012), who showed SGR2 was a potential regulatory step of chlorophyll degradation. To date, the relationship between carotenoid gene expression and carotenoid accumulation has been investigated in a range of crop species. For example, PDS was found to contribute directly to beta-carotene accumulation in citrus (Fanciullino et al., 2008). Ha et al. (2007) suggest that high concentrations of carotenoids are associated with high transcription levels of PDS and PSY in pepper. Meanwhile, the expression level of PSY was found to be higher in high carotenoid tomatoes, compared with low carotenoid ones (Morris et al., 2004). In our study, only the transcription level of LCY-β was well correlated with beta-carotene contents and hue angles (Figure 5 and Supplementary Table S4). This indicates that LCY-β may be a key determinant of beta-carotene biosynthesis in kiwifruit, suggesting that LCY-β is involved in controlling and regulating the yellow color formation in kiwifruit (Ampomah-Dwamena et al., 2009). These results suggest carotenoid biosynthesis may be differently controlled in different species. All the above results suggest that yellow color formation may be controlled by SGR together with LCY-β. During fruit development and ripening, SGR activates the degradation of chlorophyll, meanwhile LCY-β catalyzes beta-carotene biosynthesis. Together these cause flesh color to change from green to yellow.

It has been shown UFGT enzyme is key to anthocyanin biosynthesis in peach (Tuan et al., 2015), pear (Wang et al., 2013), strawberry (Gonzalez et al., 2008). While for apple, most stuctural genes in the anthocyanin pathway function in concert to determine fruit anthocyanin level (El-Sharkawy et al., 2015). In the present study, we measured and compared the expression levels of CHS, CHI, F3H, F3′H, DFR, ANS and UFGT in the green-, yellow-, and red-inner pericarps of the three cultivars during development (Figure 6). Apart from UFGT, most of the others, were expressed in both the red and also in the non-red tissues, showing a weak correlation with anthocyanin content (Figure 6 and Supplementary Table S5). In contrast, UFGT showed the strongest correlation with anthocyanin content, being highly expressed in the red inner pericarps but not in the green or yellow inner pericarps. This agrees results obtained by Montefiori et al. (2011), who suggests the red inner pericarps of ‘HD22’ is the result of high expressions of AcF3GT1.

In plants, the structural genes are always regulated by transcription factors. There are, however, few studies of transcription factors regulating the genes of chlorophyll and carotenoid biosynthesis and degradation. Those studies that exist are mostly in model plants (Tang et al., 2012; Liu et al., 2013a). Hence, no transcription factors regulating chlorophylls and carotenoids have been screened in our study. More importantly, anthocyanin content was the most significant difference between the three kiwifruit cultivars in this study. Therefore, our study only analyzed the anthocyanin transcription factors in kiwifruit. A growing body of evidence suggests anthocyanin biosynthesis in plants is controlled by a transcription complex, composed of two transcription factors belonging to the R2R3-MYB and the bHLH-MYC protein families. A WD40 co-factor protein is also involved. These three proteins activate the expression of a downstream structural gene in the anthocyanin pathway by forming the MYB-bHLH-WD40 (MBW) complex (Schaart et al., 2013; Xu et al., 2014, 2015).

At this stage, the MYBs determining anthocyanin biosynthesis have been well characterized in many species (Chagne et al., 2013; Perez-Diaz et al., 2016; Zhai et al., 2016). In our study, we used the above anthocyanin regulation related MYBs to screen the genome of kiwifruit³. We found five MYBs and named these: AcMYBF110, AcMYB5a, AcMYB5b, AcMYB5c and AcMYB3 based on the names of similar genes (Supplementary Figure S2A). Expression and correlation analyses showed that transcript of AcMYBF110 was significantly correlated with both anthocyanin content and also with the expression of UFGT (Figure 7A and Supplementary Table S6). Similar analyses for the other MYBs showed much weaker correlations. This indicates AcMYBF110 may plays a crucial role in regulating anthocyanin biosynthesis in kiwifruit. Therefore, this was cloned from fruit of ‘Hongyang’ and its expression levels were found to be correlated with anthocyanin contents (Figures 8A,B). Moreover, the subcellular localization analysis showed that AcMYBF110 is located both in the nucleus and the cytoplasm. This suggests that the AcMYBF110 protein is the transcriptional regulator in these plant cells (Figure 8C). Further, transient color assays showed that, as with peach PpMYB10.1 (Zhou et al., 2015) and Actinidia chinensis AcMYB110 (Montefiori et al., 2015), AcMYBF110 can autonomously induce anthocyanin accumulation in N. tabacum leaves by activating transcription of NtDFR, NtANS and NtUFGT (Figure 8D). This differs from apple MdMYB10 and MdMYB110 which induce anthocyanin accumulation by necessarily interacting with bHLH3 (Chagne et al., 2013). These results confirm AcMYBF110 is a key R2R3 MYB transcription factor regulating anthocyanin biosynthesis in red-fleshed kiwifruit. Further studies are required to clarify which anthocyanin structural genes are regulated by AcMYBF110.

The bHLH proteins function as anthocyanin regulators and have been reported in model plants and fruit species including: maize (Ludwig et al., 1989), Arabidopsis (Zhang et al., 2003), grape (Hichri et al., 2010), peach (Rahim et al., 2014) and litchi (Lai et al., 2016). They usually serve as co-factors interacting with the R2R3 MYBs to induce anthocyanin biosynthesis. In this study, we found five bHLHs from the genome of ‘Hongyang,’ and their expression levels were measured by qPCR (Figure 7B). All five bHLHs were much more highly expressed in green fruit than in red or yellow fruit at all stages of development. Moreover, they showed no significant correlation with anthocyanin accumulation (Supplementary Table S6). This indicates that, as in apple (Espley et al., 2007), peach (Rahim et al., 2014) and litchi (Lai et al., 2016), bHLHs likely depends on co-expression with other regulatory factors (such as R2R3 anthocyanin-promoting MYBs) to induce anthocyanin biosynthesis in the red-fleshed kiwifruit. This instead of contributing directly to anthocyanin biosynthesis. However, in most plants more than one bHLH factor controls anthocyanin biosynthesis and other metabolic pathways. An increasing body of evidence supports the view that the different bHLHs have specific functions within each species. Hence, it is worth further study to ascertain which of the bHLH serves as a partner to AcMYBF110 so as to regulate coloration in the red-fleshed kiwifruit.

The WD40 group is members of the MBW transcription complex. This group is related to anthocyanin biosynthesis and has been isolated in a number of model plant species. Initially, it included AN11 which controls flower pigmentation in petunia (de Vetten et al., 1997) and AtTTG1 in Arabidopsis (Walker et al., 1999). More recently, WD40 members involved in anthocyanin biosynthesis have been isolated in a number of fruit species (Brueggemann et al., 2010; Matus et al., 2010; Schaart et al., 2013). Based on these WD40s, we found two WD40 members (WDR1 and WDR2) in kiwifruit (Supplementary Figure S2C). Analysis of qPCR showed the expression of WDR1 was significant higher in green fruit than in red or yellow fruits at all stages. It also showed a trend for increase in the red fruit in the coloring stages. Similar profiles were found for the expression of WDR2 (Figure 7C). Overall, the expressions of WDR1 and WDR2 were not consistent with anthocyanin content and UFGT expression in the three different colored kiwifruits examined here (Supplementary Table S6). However, if analyzed only in the red fruit, their expressions were highly correlated with both anthocyanin content and with UFGT expression (data not shown). These findings suggest these WDRs are likely to play different roles in different kiwifruit cultivars. For example, in red kiwifruit they are primarily involved in regulation of anthocyanin biosynthesis, while in the green and yellow kiwifruits, their involvement seems mainly to be the regulation of other physiological and biochemical processes. In other plants, they have been shown to be involved in defense and response to various biotic and abiotic stresses (Lee et al., 2010; Chen and Brandizzi, 2012).